Complex Catalytic Materials Based on the Perovskite-Type Structure for Energy and Environmental Applications

This review paper focuses on perovskite-type materials as (photo)catalysts for energy and environmental applications. After a short introduction and the description of the structure of inorganic and hybrid organic-inorganic perovskites, the methods of preparation of inorganic perovskites both as powders via chemical routes and as thin films via laser-based techniques are tackled with, for the first, an analysis of the influence of the preparation method on the specific surface area of the material obtained. Then, the (photo)catalytic applications of the perovskites in energy production either in the form of hydrogen via water photodecomposition or by methane combustion, and in the removal of organic pollutants from waste waters, are reviewed.


Introduction
The 21st century has brought new challenges to scientists in solving the issues generated, on the one hand, by the increased demand for energy and, on the other hand, by the more stringent exigencies in environmental protection. In recent years we witnessed a continuous search to find more economically viable processes to generate cleaner energy starting from well known as well as from new renewable resources using more performant catalysts than conventional ones. In this respect, researchers have been focused on developing new catalysts and to perfect the existing ones in order to enhance the sustainability of energy-generating processes. An example illustrating the development of the research trends concerns the utilization of perovskite-type materials as catalysts for different processes involving energy production as well as removal of pollutants by oxidation processes. Complex oxides having an ABO 3 perovskite-type structure are possible candidates for catalytic combustion, their potential as oxidation catalysts being studied for the first time approximately 50 years ago [1,2]. These oxide materials are very active, highly thermally stable and present low volatility. Due to these properties, they were considered as potential candidates to replace the expensive oxide-supported noble metals (especially platinum and palladium) which are currently the combustion catalysts used on a large scale.
for this aim allows surpassing the issues posed by the formation of noble metals volatile oxides or by their sintering at high temperatures [3][4][5].
Photocatalysis is considered the key of solving energy generation and environmental pollution problems since it can use sunlight, which is one of the cleanest energy sources. Perovskites are intensively studied as photocatalysts, for several reactions such as: generation of hydrogen by photodecomposition of water and total photo-oxidation of different organic pollutants [6,7].
The purpose of this literature review is to highlight the state of the art concerning the influence of the physico-chemical properties of perovskite-type materials on their catalytic and photocatalytic efficiencies, focusing mainly on photoelectrochemical water splitting, the photodecomposition of organic pollutants and the catalytic combustion of methane reactions. To this aim, the scientific information from more than 300 references published during the last 20 years has been screened.

Inorganic Perovskites
Most of the chemical compounds having the general formula ABO3 with r(A n+ ) ≈ 2r(B m+ ) (where r is the ionic radius of A and B cations with different valences forming compounds such as: A I B V O3, A II B IV O3 and A III B III O3) show an inorganic perovskite-type structure. Perovskites' name derives from the first found mineral CaTiO3 which has the same atomic arrangement [8]. The unit cell of CaTiO3 is represented as a cube, where the corners are occupied by calcium ions, the titanium ions are positioned at the body's center, and oxygen ions at the faces center ( Figure 1). The name and the simple cubic structure were preserved even though in 1946 Megaw determined that CaTiO3 has an orthorhombic structure [9]. In the perovskite structure, the larger cation A is 12-fold coordinated with oxygen ions and responsible for the stability properties, having minor effects on the catalytic properties. Generally, the cations A are elements with inert d 0 and f 0 electronic structure, such as alkaline or rare earth cations. The catalytic activity is influenced by the smaller B cations with an octahedral coordination, which are usually 3d, 4d or 5d transition metals acting as active sites, because they have the ability to perform redox cycles without structure destruction [8,10]. In 1926, Goldschmidt [11] has defined a tolerance factor: t = (rA + rO)/2 1/2 (rB + rO) (1) where rA, rB and rO are the ionic radii of A and B cations and O anions, respectively, which is correlated to the thermodynamic and structural stability. The ideal perovskite cubic structure is stable only if 0.8 < t < 0.9, this range being slightly larger in the case of distorted perovskite structures. At room temperature, a distortion in the structure may exist, but at high temperatures, the cubic structure is formed. Besides the ideal cubic structure, other well-known symmetries for perovskites are: In 1926, Goldschmidt [11] has defined a tolerance factor: t = (r A + r O )/2 1/2 (r B + r O ) (1) where r A , r B and r O are the ionic radii of A and B cations and O anions, respectively, which is correlated to the thermodynamic and structural stability. The ideal perovskite cubic structure is stable only if 0.8 < t < 0.9, this range being slightly larger in the case of distorted perovskite structures. At room temperature, a distortion in the structure may exist, but at high temperatures, the cubic structure is formed. Besides the ideal cubic structure, other well-known symmetries for perovskites are: orthorhombic, rhombohedral, monoclinic and triclinic [12]. In catalysis, the perovskite stability during the catalytic cycles is an essential factor and it depends on the stability of the structural lattice, the valence stability of the transition metals in the reaction environment and, last but not least, the capability of defects' formation [8]. Perovskites have the ability to incorporate mixed valance cations in their structure by either isostructural substitution of cations in a mother structure or by formation of structural anion vacancies, affecting in this way the structural and catalytic properties [8].
The most used substitutions involve the replacing of half of B cations with other cations having different charge (AB x+ 0.5 B y+ 0.5 O 3 ). This type of replacement determines a shift of the oxygen ions present in the ordered structure toward the cations with higher valence. Moreover, the A and B cations in perovskites can both be easily substituted leading to doped compounds with the formula A 1-x A' x B 1-y B' y O 3 [12]. Perovskite materials also show excellent properties for several applications, such as: dielectric, piezoelectric, semiconductors, electro-optic or superconductors [13]. Their utilisation as catalysts for energy production and environmental protection is detailed in Section 5 of this review.

Hybrid Organic-Inorganic Perovskites
Hybrid organic-inorganic perovskites, denoted HOIPs, are a class of materials derived from inorganic perovskites with general formula ABX 3 . Compared to the pure inorganic perovskite, in HOIPs the A site and/or X site are substituted by cations of organic amines and different inorganic/organic anions X, respectively. In this way, the rigidity and the compactness of the perovskite structure are diminished, and the organic part offers more functionalities and enhanced flexibility [14][15][16]. Several examples of A and X organic ions that can be integrated in a hybrid organic-inorganic perovskite structure are presented in Figure 2.

Co-Precipitation Method
One of the simplest and utilized synthesis methods for perovskite catalysts is co-precipitation. This method implies the precipitation of the metal precursors (oxides, alkoxides, nitrates or other inorganic salts), when the solubility limit decreases as an effect of adding a chemical reagent. Nitric acid can be used for dissolving the metal precursors having a low solubility [18,19], while an In 1978, the German researcher Dieter Weber reported for the first time the cubic phase of a hybrid perovskite (MA)PbX 3 , where MA is metylammonium and X is a halogen anion (Cl − , Br − or I − ) [16]. This structure shows an octahedral coordination around Pb 2+ cations with halogen anions at the center of the faces of the cube, and the metylammonium cations occupy the A sites at the cube corners. Similar to the inorganic perovskites, the structure of HOIPs is defined by a Goldschmidt tolerance factor (t). The replacement of A and X sites in HOIPs with organic linkers leads to an adaption of this factor generating a formula with higher complexity, as follows: t = (r Aeff + r Xeff )/2 1/2 (r B + h Xeff /2)  3 Mechanochemical 9.5 [51] In Table 2 the influence of the preparation method on the specific surface area of lanthanide cobaltate-based perovskites is resumed. Generally, the co-precipitation method is used for both pure and doped cobaltates having rare-earth elements in the A site. The specific surface area of perovskites prepared by this method ranges from 1.6 to 8 m 2 /g [3,56,57]. The calcination temperature of the cake (precipitate), the calcination time and the heating rate are experimental factors controlling the specific surface area. Similar values of SSA were obtained by using the same preparation method for PrCoO 3 , NdCoO 3 and GdCoO 3 , respectively [56]. The citrate method is a suitable technique for cobaltate perovskites with higher specific surface area. It is frequently used for both pure and doped complex perovskites. The specific surface areas are extended over a larger range starting from 6 m 2 /g for pure LaCoO 3 and reaching a value of 18 m 2 /g for Ce-doped LaCoO 3 [50,58,59]. Solid state reaction offers a smaller range of specific surface areas with a maximum value of 5.1 m 2 /g for Ba-doped LaCoO 3 (20% of A site). Moreover, a slight increase of SSA with the increase of the ionic radii of dopants (Ca 2+ , Sr 2+ , Ba 2+ ) is observed [45]. Several microstructured powders of different perovskite-type materials based on the La 1-x Sr x Co 1-y B' y O 3 formula prepared by freeze-drying method were reported to show SSA ranging from 10.4 to 22.7 m 2 /g. Yamazoe et al. reported that the freeze-drying method allows the formation of higher surface area materials compared to other thermal evaporation techniques due to its great control of evaporation processes and lower decomposition temperatures [60]. A much higher value of SSA was reported for La 0.9 Ce 0.1 CoO 3 prepared by flame pyrolysis (62 m 2 /g) compared to the same perovskite prepared by the citrate method (10 m 2 /g) [58,61]. This method implies the combustion in the flame of a solution containing the metallic precursors. It is a versatile technique which ensures an excellent control of the material crystallinity and particle size through its experimental parameters [62]. Manganates represent a class of perovskite materials having manganese in B site positions (AMnO 3 ). The effect of the synthesis method on the SSA of lanthanide manganate-based perovskites is presented in Table 3. Compared to cobaltates, the SSA of manganates prepared by co-precipitation are higher with a maximum value of ca. 15 m 2 /g for pure LaMnO 3 [3]. For manganese-based perovskites, one of the highest values of the SSA (68 m 2 /g) was obtained by the citrate method, which consisted of adding citric acid and ethylene glycol to a solution containing lanthanum and manganese nitrates. It has been observed that the specific surface area increases with the citric acid etching time [66]. A detailed study concerning both the neat and Mn-doped LaAlO 3 perovskites reported that SSA increases with the Mn content reaching a maximum value of 33 m 2 /g for LaAl 0.2 Mn 0.8 O 3 . When aluminum cations are completely replaced with manganese cations, SSA decreases to 22 m 2 /g [55]. Pure LaMnO 3 prepared by flame-pyrolysis shows an SSA of 56 m 2 /g. By substituting (10 at.%) of La cations with Sr 2+ , a slight decrease of the SSA can be observed (51 m 2 /g). By increasing the dopant concentration (20 at.%), the SSA increases up to 70 m 2 /g. In contrast to Ce-doped cobaltates prepared by flame-pyrolysis (62 m 2 /g), La 0.9 Ce 0.1 MnO 3 shows a higher SSA of about 84 m 2 /g, with a particle size of 30-45 nm [61]. Yu et al. reported an accurate study concerning the effect of the preparation method on the SSA of Pd-doped LaMnO 3 [67] showing that an inappropriate ratio between the metallic precursors and the organic fuel in the combustion synthesis leads to a very low SSA (1 m 2 /g), while higher SSA can be obtained by using amorphous citrate (12 m 2 /g) and flame pyrolysis (32 m 2 /g) methods. The ultrasonic spray combustion method generates Pd-doped LaMnO 3 perovskite material showing the highest SSA (39 m 2 /g). In contrast to the classical combustion synthesis, this method uses an ultrasonic spray gun to initiate the reaction [68]. Notably, the crystal structure of the perovskite depends on the preparation method as well. Indeed, a rhombohedral crystal structure was obtained via ultrasonic spray combustion and citrate methods, while combustion and flame pyrolysis methods generated an orthorombic perovskite structure [67]. Ciambelli et al. reported the preparation of Ce-and Y-doped rare earth manganates by the mechanochemical method. Starting from oxide and carbonate precursors, the resulting materials showed SSAs of 14 and 19 m 2 /g, respectively [51].  Table 4 shows the effect of the synthesis method on the SSA for lanthanide ferrite-based perovskites. The co-precipitation method leads to LaFeO 3 (21 m 2 /g) solids having even higher SSA than manganates (15 m 2 /g) and cobaltates (8 m 2 /g) [3,57,76]. The citrate method is one of the most used preparation techniques for pure LaFeO 3 and doped LaFeO 3 perovskites. Pure LaFeO 3 showing the SSA of 2.9 m 2 /g was prepared by this method. It was observed that the SSA increases with the concentration of Mg 2+ used as B-site dopant. This tendency is maintained for a maximum Mg 2+ concentration of 40 at.%. A further increase of the dopant content leads to the reduction of SSA [77]. A similar value of the SSA for neat LaFeO 3 prepared via citrate method was reported elsewhere [78]. No changes of the SSA were observed for La 0.7 Ca 0.3 FeO 3 compared to simple LaFeO 3 . However, when increasing the Ca 2+ content to 50 at.%, the SSA dramatically decreases down to 0.7 m 2 /g. Other publications report values of SSA of ca. 20 m 2 /g for undoped LaFeO 3 prepared by the citrate method. Also, it increases to ca. 38 m 2 /g for Ca-doped LaFeO 3 [76,79]. The substantial differences between the SSA of the same perovskite materials can be correlated to the calcination temperature used during the preparation procedure. Materials with smaller SSA were calcined at 800 • C for 5 h, while the others were obtained at 700 • C in 6 h [76,79]. Pd-doped LaFeO 3 shows similar values of SSA as Pd-doped LaMnO 3 , when prepared via citrate and combustion methods. However, the SSA decreases for Pd-doped LaFeO 3 prepared by flame pyrolysis and ultrasonic spray combustion methods. In contrast to Pd-doped LaMnO 3 , the formation of the orthorombic crystal structure of Pd LaFeO 3 is independent on the preparation method [67]. Also, the mechanochemical method is used for the preparation of doped ferrites. The resulting materials possess SSA smaller than 10 m 2 /g [51]. The effect of the preparation method on the SSA of supported perovkites is resumed in Table 5. The wet impregnation technique was used to stabilize commercial γ-Al 2 O 3 (having initial SSA of 200 m 2 /g) with 5 wt.% La 2 O 3 . Lanthanum nitrate, manganese acetate and urea were used to load the previously prepared support with 30 % LaMnO 3 by using the deposition technique. This method is based on the precipitation of the active phase, which in this case is LaMnO 3 perovskite, to the support surface. Finally, the sample was dried at 120 • C and then calcined in air for 3 h at 800 • C. The resulting material shows an SSA of 88 m 2 /g. Yet, the calcination temperature is not enough to obtain a pure perovskite, tiny amounts of La 2 O 3 and La(OH) 3 side-phases being observed by X-ray diffraction. MgO-supported LaMnO 3 with an SSA of 25 m 2 /g was also prepared by the same method [70]. La 0.8 Sr 0.2 MnO 3 was successfully dispersed on different MAl 2 O 4 spinels supports (M = Mg, Ni, Co) via wet impregnation technique. The perovskite loading was 20 wt.% and the SSA varies between 18 and 34 m 2 /g [73]. Similar SSA were obtained for Ag-doped LaMnO 3 loading metallic foils made of Fe-Cr-Al using the same impregnation technique. Before the impregnation, the metallic foil was washcoated with 84.8% Al 2 O 3 , 14.4% TiO 2 , 0.8% La 2 O 3 through the citrate method [80]. Neat LaFeO 3 prepared via citrate method was dispersed on similar supports, leading to materials with SSA of 7.7 m 2 /g [81]. The highest specific surface area was obtained for ZrO 2 -supported LaMnO 3 (132.5 m 2 /g) [82]. Pulsed laser deposition (PLD) is a technique belonging to the physical vapor deposition (PVD) class which takes place in a vacuum chamber. The PLD setup is schematically presented in Figure 3. Pulsed laser deposition (PLD) is a technique belonging to the physical vapor deposition (PVD) class which takes place in a vacuum chamber. The PLD setup is schematically presented in Figure 3. A pulsed laser beam is focused onto the surface of a target consisting of the desired material to be deposited. When the laser pulses have high enough energy density, a plasma plume forms at the target surface as an effect of vaporizing or ablating small parts from it. The flux of material necessary for the film growth is provided by the ablation plume and it is collected on a substrate. In other words, the deposition process using laser ablation is based on the vaporization of the target material, followed by the deposition of the vaporized material onto a collecting substrate. The substrate is situated at a well-known distance from the target and it is placed parallel to the target, as can be seen in Figure 4 [85,86].  A pulsed laser beam is focused onto the surface of a target consisting of the desired material to be deposited. When the laser pulses have high enough energy density, a plasma plume forms at the target surface as an effect of vaporizing or ablating small parts from it. The flux of material necessary for the film growth is provided by the ablation plume and it is collected on a substrate. In other words, the deposition process using laser ablation is based on the vaporization of the target material, followed by the deposition of the vaporized material onto a collecting substrate. The substrate is situated at a well-known distance from the target and it is placed parallel to the target, as can be seen in Figure 4 [85,86]. Pulsed laser deposition (PLD) is a technique belonging to the physical vapor deposition (PVD) class which takes place in a vacuum chamber. The PLD setup is schematically presented in Figure 3. A pulsed laser beam is focused onto the surface of a target consisting of the desired material to be deposited. When the laser pulses have high enough energy density, a plasma plume forms at the target surface as an effect of vaporizing or ablating small parts from it. The flux of material necessary for the film growth is provided by the ablation plume and it is collected on a substrate. In other words, the deposition process using laser ablation is based on the vaporization of the target material, followed by the deposition of the vaporized material onto a collecting substrate. The substrate is situated at a well-known distance from the target and it is placed parallel to the target, as can be seen in Figure 4 [85,86]. The fundamental processes taking place during the laser ablation are as follows: the heating of the irradiated material; the melting, evaporation or sublimation of the heated material; the formation The fundamental processes taking place during the laser ablation are as follows: the heating of the irradiated material; the melting, evaporation or sublimation of the heated material; the formation of the plume and, finally, the plume expansion. One of the most important ablation parameters is the energy density of the laser pulse or its fluence (J/cm 2 ). The preparation of thin films by using PLD can be performed both in a vacuum and in a gas atmosphere, which influences the deposition process. Figure 5 displays a photograph of a TiO 2 target [85,86]. of the plume and, finally, the plume expansion. One of the most important ablation parameters is the energy density of the laser pulse or its fluence (J/cm 2 ). The preparation of thin films by using PLD can be performed both in a vacuum and in a gas atmosphere, which influences the deposition process. Figure 5 displays a photograph of a TiO2 target [85,86]. Pulsed laser deposition is a non-conventional technique, relatively simple and suitable for the production of thin films that can be successfully applied for many classes of materials when other techniques fail. It presents the following advantages: • the laser radiation can be well focused on very small spot sizes at the target surface, increasing in this way the efficiency, the control and the flexibility of the process; • the deposition chamber can be considered a "clean reactor" because the energy source (laser) is external, being independent of the deposition medium; also, the laser parameters (energy density and wavelength) can be easily adjusted to ensure the reproducibility of the sample preparation; • it is a simple and versatile technique from the point of view of experimental achievement, offering the possibility to obtain all kind of materials (complex stoichiometry, organo-metallic compounds); • the properties of the obtained thin films (thickness, crystalline structure, stoichiometry and composition) can be rigorously controlled, because they depend on the laser parameters (wavelength, laser fluence, the spot area, the duration of pulse, the repetition rate etc.) which are easily controlled from the outside of the deposition chamber; • it ensures large deposition rates (1-5 Å/pulse).
However, the pulsed laser deposition technique has also a number of disadvantages, such as: • the possibility to cover only substrates having small area (~1 cm 2 ); • the appearance of material droplets or clusters on the surface of the thin films leading to an increased roughness, which can affect the crystallinity, the optical, electrical and magnetic properties of the manufactured thin films [85,86].
The latter disadvantage can be diminished or even eliminated by optimizing the PLD system and deposition conditions as follows: • the selection of a suitable target material: a target made by dense and very small particles, ensures uniform conditions during the ablation process. A material presenting a lot of defects or Pulsed laser deposition is a non-conventional technique, relatively simple and suitable for the production of thin films that can be successfully applied for many classes of materials when other techniques fail. It presents the following advantages: • the laser radiation can be well focused on very small spot sizes at the target surface, increasing in this way the efficiency, the control and the flexibility of the process; • the deposition chamber can be considered a "clean reactor" because the energy source (laser) is external, being independent of the deposition medium; also, the laser parameters (energy density and wavelength) can be easily adjusted to ensure the reproducibility of the sample preparation; • it is a simple and versatile technique from the point of view of experimental achievement, offering the possibility to obtain all kind of materials (complex stoichiometry, organo-metallic compounds); • the properties of the obtained thin films (thickness, crystalline structure, stoichiometry and composition) can be rigorously controlled, because they depend on the laser parameters (wavelength, laser fluence, the spot area, the duration of pulse, the repetition rate etc.) which are easily controlled from the outside of the deposition chamber; • it ensures large deposition rates (1-5 Å/pulse).
However, the pulsed laser deposition technique has also a number of disadvantages, such as: • the possibility to cover only substrates having small area (~1 cm 2 ); • the appearance of material droplets or clusters on the surface of the thin films leading to an increased roughness, which can affect the crystallinity, the optical, electrical and magnetic properties of the manufactured thin films [85,86].
The latter disadvantage can be diminished or even eliminated by optimizing the PLD system and deposition conditions as follows: • the selection of a suitable target material: a target made by dense and very small particles, ensures uniform conditions during the ablation process. A material presenting a lot of defects or different structural mechanical strains, which can appear during the processing procedure, affects the deposition process. Moreover, the target material has to present a high absorption coefficient at the used laser wavelength; • the rotating and the translation of the target material toward the laser beam during the deposition process; • the optimizing of the deposition parameters (the laser fluence, the laser spot area, the repetition rate); • the utilization of a supplementary laser beam parallel to the substrate surface which can split the material clusters.
Additionally, the PLD can be coupled with a radiofrequency (RF) plasma source for better performance. Compared to the standard PLD setup, in this case a RF plasma source directed to the substrate is added. The function of the RF source is to ensure a supplementary control of the anionic composition of the manufactured thin films by using gases from the RF plasma source [87]. Figure 6 shows the experimental setup of the RF-assisted PLD vacuum chamber. • the rotating and the translation of the target material toward the laser beam during the deposition process; • the optimizing of the deposition parameters (the laser fluence, the laser spot area, the repetition rate); • the utilization of a supplementary laser beam parallel to the substrate surface which can split the material clusters.
Additionally, the PLD can be coupled with a radiofrequency (RF) plasma source for better performance. Compared to the standard PLD setup, in this case a RF plasma source directed to the substrate is added. The function of the RF source is to ensure a supplementary control of the anionic composition of the manufactured thin films by using gases from the RF plasma source [87]. Figure 6 shows the experimental setup of the RF-assisted PLD vacuum chamber. In Table 6 the experimental conditions for different types of perovskite materials grown by PLD and PLD-RF techniques are presented.

Matrix-Assisted Pulsed Laser Evaporation (MAPLE)
A major disadvantage of the manufacturing of thin films by laser ablation arises from the deposition process. The plasma formation and the condensation of elements on a substrate are not suitable for soft organic and polymeric materials, because their structure can be decomposed easily or even completely destroyed as an effect of interaction with laser beam. In order to depose this kind of materials, a modification of the experimental setup is required, this modification being related to the deposition target. Therefore, in the matrix-assisted pulsed laser evaporation (MAPLE) technique the target is prepared by dissolving the polymeric (or organic) material into a volatile solvent and the resulting mixture is frozen in liquid nitrogen. A photograph of a MAPLE target can be seen in Figure  7. The laser wavelength is selected in such a way that only the solvent reacts when the laser beam hits the frozen target [143,144]. In Table 6 the experimental conditions for different types of perovskite materials grown by PLD and PLD-RF techniques are presented.   [142] a P O2 -the oxygen pressure during the deposition process; b T-the substrate temperature during the deposition process; c RF-prepared via PLD-RF; d n.s.-not specified; e RT-room temperature.

Matrix-Assisted Pulsed Laser Evaporation (MAPLE)
A major disadvantage of the manufacturing of thin films by laser ablation arises from the deposition process. The plasma formation and the condensation of elements on a substrate are not suitable for soft organic and polymeric materials, because their structure can be decomposed easily or even completely destroyed as an effect of interaction with laser beam. In order to depose this kind of materials, a modification of the experimental setup is required, this modification being related to the deposition target. Therefore, in the matrix-assisted pulsed laser evaporation (MAPLE) technique the target is prepared by dissolving the polymeric (or organic) material into a volatile solvent and the resulting mixture is frozen in liquid nitrogen. A photograph of a MAPLE target can be seen in Figure 7. The laser wavelength is selected in such a way that only the solvent reacts when the laser beam hits the frozen target [143,144]. The difference between this technique and PLD lays in the target preparation, which generates other laser-material interaction mechanisms [145]. The setup used for MAPLE deposition is presented in Figure 8.  The difference between this technique and PLD lays in the target preparation, which generates other laser-material interaction mechanisms [145]. The setup used for MAPLE deposition is presented in Figure 8. The difference between this technique and PLD lays in the target preparation, which generates other laser-material interaction mechanisms [145]. The setup used for MAPLE deposition is presented in Figure 8. In this case, two main processes occur when the laser beam falls onto the target surface: the evaporation of the frozen target and the ejection of the organic material. Generally, the concentration of polymeric (or organic) substance is very low (ca. 1-5%), the evaporation of the solvent and of the desired substance taking place simultaneously. The energy of the photons absorbed by the solvents In this case, two main processes occur when the laser beam falls onto the target surface: the evaporation of the frozen target and the ejection of the organic material. Generally, the concentration of polymeric (or organic) substance is very low (ca. 1-5%), the evaporation of the solvent and of the desired substance taking place simultaneously. The energy of the photons absorbed by the solvents is converted into thermal energy which induces the heating and the evaporation of the target. The polymeric chains have sufficient kinetic energy to cross the distance between the target and the substrate, while the solvent molecules are eliminated from the reaction chamber by the vacuum pumps. When the experimental conditions are optimized, the polymeric (or organic) substances can be transferred from the target to the substrate without structural damage [146,147]. This method has also been used to obtain some perovskite thin films (Table 7). Abbreviations: D T-S -the distance between the material target and the substrate; rGO-reduced graphene oxide; MAI-methylammonium iodide; ITO-indium tin oxide; IR-infrared radiation; DMSO-dimethyl sulfoxide; MEG-ethylene glycol; FTO-fluorine-doped tin oxide.

Energy Production
Nowadays, one of the major problems facing humanity is to fully cover the global energy demands which have severely increased in the last few years. Most energy is generated by using fossil fuels. The high energy demand on the market accelerates the consumption of these exhaustible resources based on carbon. Moreover, the intense utilization of fossil fuels causes the greenhouse effect and generates huge amounts of pollutants affecting the environmental safety and human's health [151]. Hence, a new alternative for energy production must be urgently and efficiently implemented. One of the most promising ways to solve the current energy problem is to use solar energy. The energy provided by the sun is abundant, among the cleanest energy resources, which does not make the global warming status worse. Furthermore, solar energy is an ecological and renewable source, and these features are propelling it as a suitable candidate for the global supply. Solar energy can be converted to both electrical and chemical energies by using photovoltaic and photocatalytic concepts. However, until technologies based on these processes are suitable for industrial implementation, new alternatives in order to increase fossil fuels' efficiency and to minimize the pollution are required. Among others, catalytic combustion of methane is one of the promising ways to increase the efficiency and to minimize pollution [152].

The Production of Energy in the Form of Hydrogen via Water Photodecomposition
Hydrogen is considered to be one of the most suitable options to replace the carbon-based fuels. It can be generated using different methods from various renewables (hydro, solar) and non-renewables (coal, natural gas, nuclear) sources. Among these, hydrogen can be produced from water through different processes, such as high-temperature decomposition (thermochemical water splitting) [153][154][155], electrolysis [156], photocatalysis [157] and photoelectrochemical water splitting [158]. Today, hydrogen is mainly synthesized by steam reforming of hydrocarbons (especially methane). It finds uses generally in petroleum refining [159], the production of ammonia [160] and in the metal refining industry [161]. For the future, it is intended to use hydrogen in fuel cells for high-efficiency power production systems which can directly produce electricity at low temperatures, with no emission of toxic by-products. In practice, inside a fuel cell, the hydrogen or a hydrogen-rich fuel is reacting with pure oxygen or oxygen from air, the resulting product being water [162]. As already discussed in Section 4.1, the solar energy is considered the most suitable alternative as global energy supplier and thus the water photodecomposition reaction is one of the most studied methods for chemical energy (hydrogen) production.
There are two similar processes for the photodecomposition of water: photocatalysis, which is based on a particulate system, and photoelectrocatalysis (also known as photoelectrochemistry) based on photoelectrochemical (PEC) cells. A schematic representation of the mechanisms involved in these two processes can be found in Figure 9. Generally, in photocatalysis, the photocatalyst powders are freely suspended in a solution (as a slurry) or are fixed in a reactor bed. This is considered one of the most simple and cheapest methods for water photodecompositon which does not require transparent electrodes or directional illumination. As can be observed from Figure 9, the semiconductor photocatalyst particles are irradiated and the photons having energies higher than the material band gap are absorbed. The photons' absorption is followed by the charge carriers formation, electrons (e − ) are excited to the conduction band (CB), positive holes (h + ) being simultaneously created in the valence band (VB). Once formed, the charge carriers migrate to the photocatalyst surface. When they arrive to the solid photocatalyst/liquid electrolyte interface, electrons are reducing water to molecular H2, while positive holes are oxidizing it to O2. Unfortunately, only a small portion of the photogenerated charge carriers participate in the redox reaction, and most of them suffer bulk or surface recombination, as can be seen in Figure 10. In fact, the charge recombination is one of the biggest challenges of this type of process, because a high loss of the excited charge carriers leads to a decrease in the reaction yield [164,165]. Generally, in photocatalysis, the photocatalyst powders are freely suspended in a solution (as a slurry) or are fixed in a reactor bed. This is considered one of the most simple and cheapest methods for water photodecompositon which does not require transparent electrodes or directional illumination. As can be observed from Figure 9, the semiconductor photocatalyst particles are irradiated and the photons having energies higher than the material band gap are absorbed. The photons' absorption is followed by the charge carriers formation, electrons (e − ) are excited to the conduction band (CB), positive holes (h + ) being simultaneously created in the valence band (VB). Once formed, the charge carriers migrate to the photocatalyst surface. When they arrive to the solid photocatalyst/liquid electrolyte interface, electrons are reducing water to molecular H 2 , while positive holes are oxidizing it to O 2 . Unfortunately, only a small portion of the photogenerated charge carriers participate in the redox reaction, and most of them suffer bulk or surface recombination, as can be seen in Figure 10. In fact, the charge recombination is one of the biggest challenges of this type of process, because a high loss of the excited charge carriers leads to a decrease in the reaction yield [164,165].
holes are oxidizing it to O2. Unfortunately, only a small portion of the photogenerated charge carriers participate in the redox reaction, and most of them suffer bulk or surface recombination, as can be seen in Figure 10. In fact, the charge recombination is one of the biggest challenges of this type of process, because a high loss of the excited charge carriers leads to a decrease in the reaction yield [164,165]. In order to evaluate the performance of a photocatalyst, the concept of the apparent quantum yield (QE %) was introduced and it is described by the following formulae [166,167]: In addition to the high recombination of the photogenerated charge carriers, the photocatalytic system has other major disadvantages. The photogenerated hydrogen and oxygen gas are formed on the same photocatalyst surface and, hence, they are predisposed to form water via the so-called "surface back-reaction" (SBR). In order to avoid this reaction, the photocatalytic systems require the utilization of additional sacrificial reagents which are electron acceptors or donors. Moreover, the final separation of the gas produced requires additional energy consumption. Also, because the solid particles are suspended in the liquid electrolyte, a part of irradiation light will be absorbed by the liquid medium, decreasing the energy of incident light which interacts with the photocatalyst [167].
On the other hand, in a PEC system the standard configuration is based on a photoelectrode (called working electrode) and a counter electrode, as presented on Figure 9. The working electrode consists of a semiconductor material (of either n-type or p-type) deposited on a conductive substrate. When the semiconductor material is of n-type, the working electrode acts as photoanode (Figure 11a), while when it is of p-type, it operates as photocathode (Figure 11b). Moreover, the PEC system can also work as a tandem system with both photoanode and photocathode, as shown in Figure 11c. also work as a tandem system with both photoanode and photocathode, as shown in Figure 11c. For example, in a standard PEC configuration, when photons are absorbed by a n-type semiconductor, the positive holes from the valance band are migrating to the photoelectrode surface, where water is oxidized to O2. The electrons are collected by the conductive substrate and they are sent through an external circuit to the counter electrode (Pt), where water molecules are reduced to H2. Generally, in order to ensure a better separation of the charge carriers, an external electrical or chemical (pH difference) bias between electrodes is applied. When the working electrode is made of p-type semiconductor, O2 is formed at the counter electrode surface, while H2 is generated on the ptype photoanode surface. Also, the same principle is applied for tandem PEC configuration, water Figure 11. Photoelectrochemical water-splitting systems with: (a) n-type semiconductor photoanode; (b) p-type semiconductor photocathode; (c) tandem system. Adapted from reference [163].
For example, in a standard PEC configuration, when photons are absorbed by a n-type semiconductor, the positive holes from the valance band are migrating to the photoelectrode surface, where water is oxidized to O 2 . The electrons are collected by the conductive substrate and they are sent through an external circuit to the counter electrode (Pt), where water molecules are reduced to H 2 . Generally, in order to ensure a better separation of the charge carriers, an external electrical or chemical (pH difference) bias between electrodes is applied. When the working electrode is made of p-type semiconductor, O 2 is formed at the counter electrode surface, while H 2 is generated on the p-type photoanode surface. Also, the same principle is applied for tandem PEC configuration, water molecules are reduced to H 2 at the p-type semiconductor surface while being oxidized to O 2 at the surface of the n-type semiconductor [163].
The PEC system shows excellent advantages compared to the particulate system, the oxygen is generated at the photoanode, while the hydrogen is formed at the photocathode. In this way, the resulting gas can be independently collected with no further costly gas separation methods or efficiency lowering because of SBR. Furthermore, appropriate and superior quality contact between the conductive substrate and semiconductor ensures excellent charge transfer leading to great photoelectrochemical efficiencies [168].
The PEC water-splitting processes can be divided into two electrochemical redox reactions: the hydrogen evolution reaction (HER) and the oxygen evolution reaction (OER). The HER can occur by the Volmer-Heyrowsky or Volmer-Tafel mechanisms following a two-electron transfer process (2H 2 O + 2e − H 2 + 2HO − ) [169]. On the other hand, the OER is a more complicated redox process implying the transfer of four electrons (2H 2 O 4e − + O 2 + 4H + ) [170,171]. Depending on the experimental conditions, as well as, the nature and the physico-chemical properties of the photoelectrodes, the backward reactions can affect the global water splitting efficiency. The hydrogen oxidation reaction (HOR) is based on the oxidation at the cathode of H 2 to H + by the same mechanism as HER. By contrast, at the anode, the oxygen reduction reaction (ORR) can directly or indirectly occur by the same four-electron transfer process. The direct reduction of O 2 leads to H 2 O formation, and it can be associative or dissociative. On the other hand, the indirect process leads to water formation from intermediate compounds such as hydrogen peroxide [172,173]. The dissociation of H 2 O molecule into H 2 and 1 2 O 2 is a thermodynamically non-spontaneous reaction and, hence, in order to occur, it requires an excess of energy corresponding to a Gibbs free energy of +237 kJ/mol H 2 (1.23 eV/electron transfer). In other words, the water photo-dissociation can theoretically occur only if the absorbed photons have energies higher than 1.23 eV, which correspond to 1100 nm. As can be seen, the infrared light is the most suitable for water photodecompositon reaction, at least from a theoretical point of view. Nevertheless, in practice a higher energy is required because of the energy losses (electron/hole transfer, the recombination of electron/hole, kinetic losses). Generally, it is recommended to use photocatalysts/photoelectrodes having a band gap value of 2.0-2.2 eV to overcome losses. Another important factor for the overall efficiency of water splitting reaction is related to the energy position of the bands extrema. In order to make possible the HER the bottom of the conduction band (CB) of the photocatalyst must be located at a potential which is more negative than the reduction potential of H + /H 2 (0 V vs. NHE-normal hydrogen electrode) and the top of the valence band (VB) has to be located at more positive potential than the oxidation potential of H 2 O/O 2 (1.23 V vs. NHE) [174]. A schematic representation is presented in Figure 12.
resulting gas can be independently collected with no further costly gas separation methods or efficiency lowering because of SBR. Furthermore, appropriate and superior quality contact between the conductive substrate and semiconductor ensures excellent charge transfer leading to great photoelectrochemical efficiencies [168].
The PEC water-splitting processes can be divided into two electrochemical redox reactions: the hydrogen evolution reaction (HER) and the oxygen evolution reaction (OER). The HER can occur by the Volmer-Heyrowsky or Volmer-Tafel mechanisms following a two-electron transfer process (2H2O + 2e − ⇄ H2 + 2HO − ) [169]. On the other hand, the OER is a more complicated redox process implying the transfer of four electrons (2H2O ⇄ 4e − + O2 + 4H + ) [170,171]. Depending on the experimental conditions, as well as, the nature and the physico-chemical properties of the photoelectrodes, the backward reactions can affect the global water splitting efficiency. The hydrogen oxidation reaction (HOR) is based on the oxidation at the cathode of H2 to H + by the same mechanism as HER. By contrast, at the anode, the oxygen reduction reaction (ORR) can directly or indirectly occur by the same four-electron transfer process. The direct reduction of O2 leads to H2O formation, and it can be associative or dissociative. On the other hand, the indirect process leads to water formation from intermediate compounds such as hydrogen peroxide [172,173]. The dissociation of H2O molecule into H2 and ½O2 is a thermodynamically non-spontaneous reaction and, hence, in order to occur, it requires an excess of energy corresponding to a Gibbs free energy of +237 kJ/mol H2 (1.23 eV/electron transfer). In other words, the water photo-dissociation can theoretically occur only if the absorbed photons have energies higher than 1.23 eV, which correspond to 1100 nm. As can be seen, the infrared light is the most suitable for water photodecompositon reaction, at least from a theoretical point of view. Nevertheless, in practice a higher energy is required because of the energy losses (electron/hole transfer, the recombination of electron/hole, kinetic losses). Generally, it is recommended to use photocatalysts/photoelectrodes having a band gap value of 2.0-2.2 eV to overcome losses. Another important factor for the overall efficiency of water splitting reaction is related to the energy position of the bands extrema. In order to make possible the HER the bottom of the conduction band (CB) of the photocatalyst must be located at a potential which is more negative than the reduction potential of H + /H2 (0 V vs. NHE-normal hydrogen electrode) and the top of the valence band (VB) has to be located at more positive potential than the oxidation potential of H2O/O2 (1.23 V vs. NHE) [174]. A schematic representation is presented in Figure 12.  In order to evaluate the performance and efficiency of materials and devices used in PEC systems, it is required to use benchmark metrics of assessment. One of the most important methods for the performance quantification of PEC systems is a photocurrent density-voltage (J-V) curve. This technique implies the measurement of the photocurrent density (mA/cm 2 ) as a function of the applied voltage (V) under chopped or continuous irradiation. The presence of the anodic photocurrent is correlated to the O 2 generation, while the cathodic photocurrent with H 2 generation, respectively. Figure 13 displays the J-V curves for a photoanode in dark with/without catalyst and under illumination with/without catalyst [175]. The onset potential is defined as the applied potential where the current starts to increase; a lower onset potential is desired from an economical energy consumption point of view.
Materials 2020, 13, x FOR PEER REVIEW 16 of 19 In order to evaluate the performance and efficiency of materials and devices used in PEC systems, it is required to use benchmark metrics of assessment. One of the most important methods for the performance quantification of PEC systems is a photocurrent density-voltage (J-V) curve. This technique implies the measurement of the photocurrent density (mA/cm 2 ) as a function of the applied voltage (V) under chopped or continuous irradiation. The presence of the anodic photocurrent is correlated to the O2 generation, while the cathodic photocurrent with H2 generation, respectively. Figure 13 displays the J-V curves for a photoanode in dark with/without catalyst and under illumination with/without catalyst [175]. The onset potential is defined as the applied potential where the current starts to increase; a lower onset potential is desired from an economical energy consumption point of view. Alternatively, a direct correlation between solar energy and hydrogen evolution is necessary to calculate the solar to hydrogen (STH) efficiency. In order to determine this efficiency, it is necessary to quantitatively measure the produced H2. STH efficiency gives the overall reaction efficiency of a PEC system used for water splitting reaction under simulated solar illumination and with no external voltage applied. Also, it can be calculated only if the electrodes are immersed in the same electrolyte without other sacrificial reagents [176]. STH efficiency can be calculated using the following formula: where, ν H 2 is the number of mmol of photogenerated H2, 237,000 mJ/mmol represent the Gibbs free energy, Plight is the power density of the incident light and A is the illuminated area of the electrode (area of the spot light). AM 1.5G (air mass 1.5 G) represents the standard reference solar spectra used to calculate the conversion efficiency of solar energy to electrical or chemical energy. It defines the effect produced by the atmosphere of the Earth on the solar radiation [177]. Also, STH efficiency can be directly calculated from the photocurrent density when the measurement of the H2 amount is not possible, using the formula [178]: Alternatively, a direct correlation between solar energy and hydrogen evolution is necessary to calculate the solar to hydrogen (STH) efficiency. In order to determine this efficiency, it is necessary to quantitatively measure the produced H 2 . STH efficiency gives the overall reaction efficiency of a PEC system used for water splitting reaction under simulated solar illumination and with no external voltage applied. Also, it can be calculated only if the electrodes are immersed in the same electrolyte without other sacrificial reagents [176]. STH efficiency can be calculated using the following formula: where, ν H2 is the number of mmol of photogenerated H 2 , 237,000 mJ/mmol represent the Gibbs free energy, P light is the power density of the incident light and A is the illuminated area of the electrode (area of the spot light). AM 1.5G (air mass 1.5 G) represents the standard reference solar spectra used to calculate the conversion efficiency of solar energy to electrical or chemical energy. It defines the effect produced by the atmosphere of the Earth on the solar radiation [177]. Also, STH efficiency can be directly calculated from the photocurrent density when the measurement of the H 2 amount is not possible, using the formula [178]: where, J ph is the photogenerated current density, V redox is the redox potential of interest (1.23 V) and η f is the faradaic efficiency of the hydrogen evolution reaction. For the determination of the faradaic efficiency of hydrogen, it is crucial to know the active surface area of the catalyst (photoelectrode) and the current density passing between the two electrodes [178].
For the system where an external bias is applied, other tools are required for the calculation of the efficiency. One of them is the applied bias photon-to-current conversion efficiency (ABPE) calculated with the following formula: where V applied is the potential applied between the working and counter electrodes. This formula can be applied for PEC devices working in a three-electrode system which implies the addition of a reference electrode. As in the case of STH efficiency, for the calculation of ABPE efficiency the utilization of sacrificial reagents and chemical bias should be avoided [176].
In addition, the incident photon-to-current efficiency (IPCE) is one of the most important tools for the evaluation of PEC performances. It offers the possibility to determine the practical efficiency limits of a material that can be used in PEC systems. It can be determined using the following formula [176]: where 1239.8 (V × nm) represents a multiplication of h (Planck's constant) and c (the speed of light), P mono is the power density of the monochromated light used for photoelectrode irradiation and λ is the wavelength of the light.
In 1972, Fujishima et al. reported the first photoelectrochemical water splitting reaction with solar hydrogen production opening new opportunities in the PEC field [179]. The authors used n-type TiO 2 as photoanode and an irradiation source emitting in the ultraviolet (UV) region. Since Fujishima and Honda discovered this effect, TiO 2 has been extensively used as photoelectrode because it is cheap, non-toxic and possesses high photostability. However, it has a wide band gap (ca. 3.2 eV; λ ≈ 396 nm) and it can absorb only the UV light which represents a small part of solar light (3-5%) [180,181].
Taking into account the excellent efficiencies of HOIPs in the photovoltaic applications, they seem to be possible candidates for integration in PEC devices. In a PEC system, the photoelectrodes must be in direct contact with an aqueous solution of electrolyte in order for the chemical reaction to take place. As already mentioned, the organic perovskites are very unstable in aqueous media, hence the presence of moisture has crucially affected their application in photoelectrochemistry [182,183]. However, many researchers have tried to overcome this drawback in order to generate a device showing excellent stability and high efficiency for PEC water splitting applications. The passivation of CH 3 NH 3 PbI 3 -based photoanode with a very thin Ni layer for both waterproof coating and the holes transfer has been reported. The device activity was almost entirely lost in an electrolyte solution 0.1 M Na 2 S (pH 12.8) after 20 min [184]. The stability was increased up to 30 min when a carbon nanotube/polymer composite layer was used as waterproof coverage layer [185]. Also, Wang et al. reported an innovative functionalization of CH 3 NH 3 PbI 3 -based photoanode with Ni as protective layer against humidity showing 30 min stability and 2.08 mA/cm 2 value of the photogenerated current [186]. All these photoelectrodes are based on the planar perovskite solar cells, including expensive Au electrodes and hole transporting materials, making the final device very costly and irrelevant due to its unsatisfactory stability [187].
Inorganic perovskites having the ABO 3 formula are of great interest for H 2 and O 2 production due to their exceptional stability and high photocatalytic activity. In Table 8 a series of different perovskites used as photoelectrodes for H 2 and O 2 generation via a photoelectrochemical water splitting reaction are presented.
Titanates were tested in photoelectrochemical water splitting in both acidic and basic environments. For example, BaTiO 3 with a large band gap value (3.11 eV) shows a photocurrent density of 0.007 mA/cm 2 at 0.5 V vs. SCE (saturated calomel electrode). The SCE electrode is a reference electrode based on mercury and mercury chloride [188]. The addition of PbTiO 3 to TiO 2 leads to a device with enhanced photoelectrochemical response (0.3 mA/cm 2 ) compared to pure TiO 2 (almost 0 mA/cm 2 ) under visible light irradiation. The observed difference in the photocurrent density indicates that PbTiO 3 acts as a visible light absorber [189]. Excellent photoelectrochemical activities are reported for SrTiO 3 , CaTiO 3 , LaNiO 3 and NaTaO 3 under UV radiation, but the utilization of visible light is restricted by their wide band gap values [190][191][192][193][194][195][196]. Therefore, perovskites having smaller band gap values are of interest in the photocatalytic field for enhanced water-splitting efficiencies. Ferrite perovskites are a class of materials which generally show the band gap energies in the visible region. The rhodamine B was successfully reduced under visible irradiation on YFeO 3 that showed a much higher photocatalytic activity compared to TiO 2 which possess a low absorption under visible light [197], and it was successfully tested in the form of both nanoparticles and compact films in photoelectrochemical production of H 2 [198].
In particular, BiFeO 3 (BFO) and LaFeO 3 (LFO) were extensively utilized as photoelectrodes in photoelectrochemical applications, mostly for water photolysis and dye photodegradation. The dyes photodecomposition process will be detailed in the next section of this review. BFO is one of the few multiferroic materials at room temperature with a band gap value of ca. 2.7 eV and an excellent photocatalytic activity [199]. The ferroelectric properties of BFO are able to improve the electron-hole separation and, hence, the photocatalytic activity. The effect of spontaneous polarization on the charge carriers separation is extensively detailed in literature [200]. Both undoped and doped BFO prepared by PLD were tested in photoelectrochemical applications as shown in Table 8. In 0.2 M Na 2 SO 4 electrolyte under sunlight illumination, the BFO photoanode generates a photocurrent density of 0.12 µA/cm 2 at 0.8 V vs. Ag/AgCl. After the material was heated in a reductive atmosphere of hydrogen (5%), the photocurrent increases ca. 6 times. The difference is due to smaller recombination and more efficient separation of free carriers on the hydrogen-treated sample [201]. A value of 17% for incident photon-to-current efficiency was reported for BFO photoelectrodes prepared by chemical vapour deposition (CVD). The modification of BFO surface with Ni improves the reaction kinetics and decreases the overpotential for water oxidation. The photocurrent density increases from 0.17 to 0.72 mA/cm 2 with Ni addition [202]. Nanowires as well as nanocubes of BFO generate only O 2 when irradiated with UV light. A mixture of BFO and SrTiO 3 (STO) can lead to H 2 formation under visible light irradiation [203,204]. Also, in the case of polycrystalline BFO films, it was reported that by increasing the film crystallinity, the PEC response is enhanced. Moreover, the PEC response is strongly dependent on the spontaneous polarisation of BFO polycrystalline films [205,206]. Similarly, Cho et al. have reported epitaxial thin films of BFO deposited by PLD on STO showing excellent PEC properties [207]. Ti-doped BFO was reported to generate 0.04 mA/cm 2 at 1.23 V vs. RHE (reversible hydrogen electrode) in a basic solution of NaOH under UV irradiation [208]. The RHE is a reference electrode with the potential independent on the pH changes. The onset potential was ca. 0.81 V vs. RHE [208]. A detailed study on Y-doped BFO used as photoanode with high PEC efficiency for water splitting reaction (J ph = 0.8 mA/cm 2 at 1.4 V vs. RHE) was recently reported in literature [209]. Moreover, complex BFO-heterostructures were used as photoelectrodes for water splitting reaction. For example, nanofibers of BiFeO 3 /Bi 2 Fe 4 O 9 with enhanced stability show higher efficiency for H 2 production compared to pure BFO and Bi 2 Fe 4 O 9 . This improved efficiency was attributed to the formation of well-defined heterojunctions between the component materials, which helps the separation of charge carriers and avoids their recombination [210]. A significant increase in the photocurrent density compared to pure BFO was also observed for BiFeO 3 /Fe 2 O 3 . The heterostructure having a concentration of 9% Fe 2 O 3 shows a J ph value of ca. 0.19 mA/cm 2 (at 0.6 V vs. Ag/AgCl), using visible irradiation light, which is about three times higher than that obtained for pure BFO (0.055 mA/cm 2 ) [211]. Recently, a complex device based on WO 3 /BiVO 4 /BiFeO 3 /FTO (fluorine-doped tin oxide) was reported to show a huge value of the photocurrent density under solar irradiation in 0.5 M Na 2 SO 4 electrolyte (46.9 mA/cm 2 at 2.53 V vs. RHE) [212]. The great response is correlated to the great separation of photogenerated charge carriers due to the similar band alignment of the three component materials deposited on the FTO substrate. The presence of an internal electric field at the BiVO 3 /BFO p-n junction interface together with BFO polarization, which improve the electron-hole pairs separation, have also been noticed [212].
LaFeO 3 is a perovskite with a smaller band gap value than BiFeO 3 , more specifically ca. 2.07 eV, offering the possibility for a better absorption in the visible range [180,213]. Several years ago, Celorrio et al. studied the photoelectrochemical properties of nanostructured LFO photocathode manufactured by the screen-printed technique. They reported higher cathodic than anodic photocurrent values [214]. In screen-printed technology, the perovskite material is combined to an organic compound which is used to improve the adhesion to the substrate. The mixture is printed to the substrate using a heater, followed by calcination [215]. Also, in another study, the simultaneous generation of O 2 and H 2 using LFO powders was noted [216]. The values of the photogenerated current are around 0.8 mA/cm 2 at potentials higher than 1 V vs. Ag/AgCl for utilisation of LFO as photoanode under simulated solar light [217]. Yu and coll. have deposited p-LFO and p-LFO/n-Fe 2 O 3 on ITO (indium tin oxide) conductive substrates by PLD. The fabricated photocathodes show enhanced stability in alkaline NaOH electrolyte solution under visible irradiation even after 120 h [218]. May et al. [219] have reported an accurate study on the effect of thickness of LaFeO 3 /Nb:SrTiO 3 (LFO/STON) ultrathin films prepared by PLD on the photoelectrochemical properties. The cathodic photocurrent was shown to be strongly dependent on the film thickness, being observed only for samples with 10-25 nm thickness. The anodic photocurrent is observed for both thinner and thicker than 10 nm LFO films [219]. A photoanode based on Cu-doped LFO was reported to show a photocurrent density of 2 mA/cm 2 at 1.1 V vs. Ag/AgCl under visible irradiation [220]. Notably, no reports on utilization of devices based on heterostructures of LFO/BFO perovskites for water splitting reaction have been published to date.  The utilization of co-catalysts can further improve the global efficiency of water splitting reaction. Generally, noble metals (Pt, Rh, Pd) and metal oxides (NiO, RuO 2 ) are extensively used as co-catalysts. When a photoelectrode is loaded with a co-catalyst, the photogenerated electrons which arrive at the surface are entrapped by the co-catalyst. In this way, the utilisation of co-catalysts, stimulates the surface chemical reaction and constrains the backward reaction, leading to an increased overall water splitting efficiency [90]. A schematic representation of a photocatalyst and a photoelectrode loaded by co-catalysts is presented in Figure 14.
Also, the working principle of the reaction in the presence of photocatalysts (a) and photoelectrodes (b) loaded with an oxidation co-catalyst for both dark and illuminated processes is illustrated in Figure 15. Figure 15. The working principle of the reaction in the presence of photocatalysts (A) and photocurrent density vs. the applied potential curves of photoelectrodes (B) loaded with an oxidation co-catalyst for both dark and illuminated processes, where OC-oxidation cocatalyst, Euncatactivation energy without OC and Ecat-activation energy with OC. Adapted from references [168,228].
Polycrystalline films of BFO synthesized via the citrate method, decorated with nanoparticles of Ag which act as co-catalysts, were reported. This heterostructure shows almost two times higher photocurrent density compared to pure BFO [229]. Also, a device based on reduced graphene oxide (rGO)/BFO was reported to ensure an excellent charge carriers separation, leading in this way to a higher IPCE [230]. Moreover, thin films of BiFeO3 show water-splitting activity without any external bias applied [200,221]. Also, the working principle of the reaction in the presence of photocatalysts (a) and photoelectrodes (b) loaded with an oxidation co-catalyst for both dark and illuminated processes is illustrated in Figure 15. Also, the working principle of the reaction in the presence of photocatalysts (a) and photoelectrodes (b) loaded with an oxidation co-catalyst for both dark and illuminated processes is illustrated in Figure 15. Polycrystalline films of BFO synthesized via the citrate method, decorated with nanoparticles of Ag which act as co-catalysts, were reported. This heterostructure shows almost two times higher photocurrent density compared to pure BFO [229]. Also, a device based on reduced graphene oxide (rGO)/BFO was reported to ensure an excellent charge carriers separation, leading in this way to a higher IPCE [230]. Moreover, thin films of BiFeO3 show water-splitting activity without any external bias applied [200,221]. Figure 15. The working principle of the reaction in the presence of photocatalysts (A) and photocurrent density vs. the applied potential curves of photoelectrodes (B) loaded with an oxidation co-catalyst for both dark and illuminated processes, where OC-oxidation cocatalyst, E uncat -activation energy without OC and E cat -activation energy with OC. Adapted from references [168,228].
Polycrystalline films of BFO synthesized via the citrate method, decorated with nanoparticles of Ag which act as co-catalysts, were reported. This heterostructure shows almost two times higher photocurrent density compared to pure BFO [229]. Also, a device based on reduced graphene oxide (rGO)/BFO was reported to ensure an excellent charge carriers separation, leading in this way to a higher IPCE [230]. Moreover, thin films of BiFeO 3 show water-splitting activity without any external bias applied [200,221].

Catalytic Combustion of Methane
Natural gas (a hydrocarbon gas mixture consisting mainly of methane) has a major impact on our lives, being one of the most used energy sources. Thermal combustion of light hydrocarbons meets many challenges because of its low efficiency and high pollution degree. This conventional combustion process takes place at high temperatures (above 1300 • C) generating high amounts of nitrogen oxides (NO x ), which severely affect the environment and, furthermore, human health. In the last few years, the market request has shown a significant increase of fuel consumption, leading, at the same time, to an advanced pollution risk. The catalytic combustion is preferred instead of the flame combustion due to its major advantages, such as enhanced fuel efficiency and lower emission of pollutants in the exhaust gas [3,5]. Important quantities of methane (CH 4 ) are annually consumed in the fuel industry, it being considered one of the cleanest sources of fossil energy [231]. The simplified chemical reaction of methane combustion is: However, beside the production of CO 2 and H 2 O, during the methane combustion process, many other reactions can take place, such as partial combustion of methane with CO formation, steam reforming and others [232]:  (14) In contrast to the classical combustion, the heterogeneous catalytic combustion of methane operates at lower temperatures (<800 • C), increases the conversion efficiency of methane into energy, and severely decreases the atmospheric pollutants emission [233]. It is worth noting that catalytic combustion of methane is important not only for power generation but, also, for air pollution abatement [234]. The ideal catalyst for methane combustion is required to possess high thermal and chemical stability, but at the same time it should present catalytic oxidation activity in the low-temperature range [4]. The most promising low-temperature catalysts for this reaction are the supported noble metals. Excellent catalytic activities for methane combustion are reported on supported platinum [235], rhodium [236] and palladium [237]. Unfortunately, they are expensive and in a strong oxidative atmosphere they are susceptible to form volatile oxides leading to a decrease of the catalytic activity. Also, at high temperatures, the noble metal particles tend to accumulate and to form a compact material (the so-called sintering effect). This effect decreases the specific surface area (SSA) of the catalyst and, implicitly, its catalytic activity [4]. These disadvantages made researchers look for a possible replacement of this kind of catalysts. A good alternative is to use active catalysts for hydrocarbons combustion based on single or binary oxides of transition metals. A few single oxides active for hydrocarbons combustion are Co 3 O 4 , CuO, NiO, MnO 2 . The combination of oxides generally gives a greater thermal stability and higher combustion activity compared to the single oxides [45].
Inorganic perovskites ABO 3 have been found to be adequate catalysts for total oxidation since they combine the high catalytic activity with the low volatility [3]. For the perovskite-type oxide catalysts, two types of oxygen species having different bonding strength are present: (i) adsorbed on the surface and (ii) lattice oxygen. It is believed that the adsorbed oxygen is more active, reacting with methane at lower temperatures than the lattice oxygen. At higher temperatures, the coverage of the adsorbed oxygen decreases, while the lattice oxygen becomes highly active. Voorhoeve et al. have proposed two reaction types for catalytic combustion of methane over perovskite materials: suprafacial and intrafacial. The suprafacial reaction is mainly present at low temperature, while with increasing the temperature, the intrafacial process occurs [45]. Table 9 presents the performances of different undoped perovskites used as catalysts for methane combustion. Lanthanum cobaltate (LaCoO 3 ), having a SSA smaller than 15 m 2 /g. shows T 50 (temperature corresponding to 50% methane conversion) values ranging between 525 and 709 • C [45,50,[56][57][58][59]63,64] depending on the reaction conditions, i.e., methane and oxygen concentrations in the feed gas and the magnitude of the gas flow rate. With the increase of the SSA, T 50 decreases, reaching a minimum temperature of 449 • C for LaCoO 3 with an SSA of 43 m 2 /g [65]. This temperature can be further lowered by adding noble metals (e.g., Pt, Pd) to the LaCoO 3 perovskite structure [61]. The replacement of La 3+ with Pr 3+ leads to a major increase of T 50 and of the activation energy (E a ), which is, in fact, the required energy for a chemical reaction to take place. This behavior is correlated to the redox properties of Pr 4+ /Pr 3+ couple which can affect the valence state of cobalt ions during methane combustion reaction [56]. Lanthanum manganate (LaMnO 3 ) shows excellent catalytic activity for the methane combustion with T 50 in the 500-600 • C interval [50,55,57,64,70,72,75,80,82,238]. However, the value of T 50 can be further decreased by changing the synthesis method of LaMnO 3 perovskites. For example, LaMnO 3 prepared by the flame-hydrolysis method shows a T 50 of 489 • C, while that prepared by the citric acid and flame pyrolysis methods show T 50 values of 446 and 435 • C, respectively [61,65,66]. No major changes in the catalytic activity were observed when La 3+ was replaced by Pr 3+ or Gd 3+ [69]. However, it severely decreases for NdMnO 3 and SmMnO 3 because of their easier reductibility [72]. It is worth noting that, due to its good thermal stability and high activity, which are further improved after dispersion on a lanthanum-stabilized alumina-coated monolith [70], LaMnO 3 is an active phase of choice in high-pressure structured catalytic combustors [239][240][241][242][243]. Lanthanum ferrite (LaFeO 3 ), which shows the highest stability compared to LaCoO 3 , LaNiO 3 and LaMnO 3 , displays a catalytic activity similar to LaMnO 3 [3,45,57,61,[63][64][65]69,[76][77][78]244]. Ciambeli et al. have reported a detailed study on ferrite perovskites (AFeO 3 ) containing different rare earth cations (La 3+ , Sm 3+ , Nd 3+ ) in A sites. It was demonstrated that LaFeO 3 catalyst shows the best catalytic activity for methane combustion, with total selectivity to CO 2 , having the T 50 = 528 • C and the temperature corresponding to 90% methane conversion (T 90 ) of ca. 615 • C [77]. Spinicci et al. have revealed that La 0.9 FeO 2.85 shows better activity compared to stoichiometric LaFeO 3 . The further decrease of lanthanum and oxygen contents leads to the loss of the catalytic activity [245]. The catalytic activity can be limited by the low specific surface area of the perovskite catalysts [45]. One of the most used methods to increase the SSA of catalysts based on perovskites is to disperse them on appropriate supports [246]. Indeed, it has been shown that the SSA of the unsupported BaTiO 3 and PbTiO 3 increased from 0.4 and 0.5 m 2 /g, respectively, to 193 and 175 m 2 /g for their alumina-supported counterparts, respectively. The increase of the SSA led to an enhanced catalytic activity for methane combustion [44,247].  The catalytic activity of the perovskites is strongly dependent on B-site cations, while A-site cations are responsible for structure stability, as already mentioned in Section 2.1. Therefore, by substituting small portions of A and/or B cations with other transition metals, both the catalytic activity and thermal/chemical stability of the final catalyst can be improved. Generally, the most used A-site dopants for perovskites are alkaline earth metals (Sr, Ca and Ba) [3,45,53,58,61,71,73,74,76,78,79] and lanthanides (Ce, Eu) [45,51,58,61,71], as can be observed in Table 10 where the performance of A-site doped perovskites in methane combustion is resumed. Arai et al. [45] reported that La 0.6 Sr 0.4 MnO 3 perovskite prepared by solid state reaction exhibits similar catalytic activity to Pt/Al 2 O 3 at low temperatures (350-550 • C). At temperatures higher than 600 • C it becomes less active than the noble metal catalyst. However, La 0.6 Sr 0.4 MnO 3 was shown to be the most active catalyst compared to Ca-doped LaMnO 3 , Sr-doped LaFeO 3 , Ba-doped LaCoO 3 , Ce-doped LaCoO 3 and Ca-doped LaCoO 3 [45]. The high activity of La 0.6 Sr 0.4 MnO 3 for methane combustion at low temperatures was confirmed in other studies [71,74]. This behavior was correlated with its high capability to adsorb oxygen on the surface, most probably due to the fact that the presence of the bivalent dopant determines the formation of Mn 4+ species, [45,71,74]. When the temperature increases, the amount of adsorbed oxygen decreases and, hence, its catalytic activity. Calcium is the most used element for LaFeO 3 doping, as can be seen in Table 10. Pecchi et al. [76] reported a detailed study concerning La 1-x Ca x FeO 3 perovskites prepared by citrate and co-precipitation methods. For the samples prepared by co-precipitation, there are almost no changes in T 50 with the calcium content, while for those prepared via the citrate method the addition of calcium minimizes T 50 [76]. Similar results on Ca-doped LaFeO 3 prepared by the citrate method were also reported elsewhere [79]. In Table 11 the performances of B-doped perovskites in the catalytic combustion of methane are presented. Saracco et al. studied LaCr 1-x Mg x O 3 perovskite prepared by the citrate method and observed that its catalytic activity increases with the Mg content [47]. The opposite behaviour was observed for LaAl 1-x Mn x O 3 , whose catalytic activity decreases with the Al content. LaAl 0.2 Mn 0.8 O 3 shows the best activity, while the activity of LaAl 0.4 Mn 0.6 O 3 is similar to the pure LaMnO 3 [55]. A high fraction of Fe 4+ is observed in LaFe 1-x Mg x O 3 perovskite and it increases with the addition of Mg content leading to lower catalytic activity at low temperatures [77]. Taguchi et al. [248] reported that Ca(Mn 0. 6 [78,248,249]. Moreover, even a lower T 50 value was reported for Al-doped LaMnO 3 . Thus, for 10% Al 3+ content, the T 50 is ca. 520 • C and increases with the Al content. Notably, the T 50 value decreases with the increase of the SSA, the highest SSA (22 m 2 /g) corresponding to the La(Mn 0.9 Al 0.1 )O 3 system [250]. Recently, Miao et al. [251] reported a catalyst based on La(Mn,Fe)O 3 with an excellent stability at 550 • C in the catalytic combustion of methane. The methane conversion was almost entirely preserved after eight combustion cycles [251].

Applications of Perovskite-Type Materials in the Removal of Pollutants from Waste Waters
In the last few years, the high extent of pollution has become one of the biggest problems facing humanity because it can severely affect human life [252]. The residual dyes emerging from different industries (pharmaceutical, textile, paper and others) are considered to be the most common water pollutants. For example, in the textile industry, every kilogram of final product generates 50-100 L of waste water [253][254][255]. Numerous techniques for water treatment have already been developed, such as adsorption [256], chemical coagulation [257], electrocoagulation [258], advanced oxidation processes [259], photocatalysis [260] and photoelectrocatalysis [261]. The adsorption method used in waste water treatments involves the adhesion of pollutants (organic dyes, in this case) to the surface of a solid material, called adsorbent. The adsorbent properties of solid materials can be explained by the force fields which govern the surface properties. Depending on the type of forces established between the adsorbent and the adsorbed molecules, the adsorption can be: physical adsorption (for weak forces) and chemisorption (chemical bonds). The main characteristics controlling the adsorption efficiency of an adsorbent are the specific surface area and porosity. The adsorption method has the advantage that it can be used for a large amount of water with low pollutants content using simple and relatively cheap operating systems [262]. The most commonly used adsorbents for the elimination of organic dyes from water are: activated carbon [263,264], alumina [265,266], bentonite [267,268] and zeolites [269]. The main drawbacks of these techniques are correlated to the adsorbents' capacity which decreases in time and the high costs of their regeneration. Moreover, depending on the pollutant's nature, the adsorbent could be irreversibly blocked [262]. Chemical coagulation is an extensively used method for the waste water treatment. It is based on the utilization of coagulants which have the ability to form precipitates. The pollutants are trapped in the formed precipitate which is settled, offering the possibility to easily separate the supernatant (water) from sediments. The process is strongly dependent on the pH, coagulant concentration and mixing. Depending on the nature of the pollutant, the coagulant can be either inorganic (AlCl 3 ; FeCl 3 ) or organic (poly diallyldimethyl-ammonium chloride; polyacrylamide). This method is able to remove high amounts of organic dye from water, but it leads to an increased cost of the process due to two drawbacks: (i) it requires high quantities of chemical coagulant, and (ii) it produces large amounts of sludge [270,271]. The electrocoagulation is similar to the chemical coagulation, in this case the coagulants being electrochemically generated. It uses low-voltage and two sacrificial iron (or aluminum) electrodes. Under an applied current voltage, at anode are generated Fe 3+ (or Al 3+ ) species, while at cathode water is reduced to H 2 and hydroxides. The ions formed at the anode surface react with hydroxide groups, leading to the coagulants' production. The resulting quantity of sludge is much lower compared to the chemical coagulation [270,272]. However, it requires often the replacement of the sacrificial electrodes to preserve a high efficiency. The operating cost of this technique increases significantly when applied in non-electrified areas, this being a common case for waste water found in nature [273]. One of the most used advanced oxidation methods for the degradation of dyes from waste water is ozonation. Ozone has a very high oxidation potential being capable to oxidize the organic dyes [259]. Unfortunately, ozone is very toxic for the human body, its strong oxidizing character can cause various diseases [274]. The photocatalytic and photoelectrochemical processes are considered to be the key solution to remove environmental pollution because they are clean methods which use renewable solar energy, their advantages have been already discussed in this review.
Rhodamine B is an important cationic xanthene dye being one of the mostly used model organic dye in photodegradation studies. Important quantities of this organic dye are coming from the textile industry contributing to environmental pollution [275][276][277][278]. Equally, methyl orange (anionic dye) and methylene blue (cationic dye) are considered harmful dyes released from textiles and printing industries and it is desired to minimise as much as possible their concentration in the environment [279][280][281][282][283][284]. Besides these, other common organic pollutants targeted in photodegradation applications are congo red (highly toxic and carcinogenic pollutant) [285], neutral red [286], phenol red [287], 4-methylphenol [288] and tetracycline [289], their chemical structures being presented in Table 12.

TC [295]
TC [295] The photooxidation reaction of organic dyes is similar to the processes involved in the water-splitting reaction. When the semiconductor is irradiated, the photons having energies higher than its band gap energy are absorbed. The photogenerated charge carriers migrate to the photocatalyst surface. Once there, electrons react with adsorbed oxygen molecules (O 2(ads) ) generating initially superoxide radicals (O 2 • ) and, then, hydroperoxide radicals (HOO • ). On the other hand, the positive holes interact with the surface-adsorbed water molecules (H 2 O (ads) ) producing free hydroxyl radicals (OH • ). When the organic compounds (dyes) are adsorbed on the photocatalyst surface, they are rapidly oxidized by the highly reactive radicals to CO 2 and H 2 O [296]. The photodegradation of organic compounds can be schematically presented by the following series of reactions, where P = photocatalyst: As in the case of hydrogen production from water, TiO 2 is the most utilized photocatalyst for water depollution. It was used as photocatalyst for organic dye photooxidation in both powder and immobilized form, with good photocatalytic activity [297][298][299]. Its activity was further improved when an external bias was applied to the system. However, pure TiO 2 has high rate of the charge carrier's recombination and wide band gap value, which are the main drawbacks of this semiconductor in photocatalysis [164,300].
Structural and compositional properties of oxide perovskites are suitable for the photodegradation of organic molecules, their performances being summarised in Table 13. SrTiO 3 shows excellent photocatalytic activity under UV irradiation for the degradation of rhodamine B [301,302] and methyl orange [303]. The rhodamine B was completely degraded after 1.3 h, while 3 h were necessary to completely eliminate methyl orange dye. Under visible irradiation, the degradation efficiency of SrTiO 3 is lower than 50%, because of its large band gap [304].
The doping of SrTiO 3 with different elements (Fe, Nb) leads to an improvement of the photocatalytic efficiency under visible light irradiation [289,304,305]. LaCoO 3 shows even higher photocatalytic activity under UV irradiation for the degradation of rhodamine B, with a complete photooxidation of the organic dye after 0.8 h [305]. Its activity considerably decreases under visible irradiation [306].
The degradation of methyl orange was performed with conversions higher than 90% over BiFeO 3 photocatlaysts under both UV and visible light [307]. Moreover, nanoparticles of BiFeO 3 and Gd-doped BiFeO 3 (Bi 1-x Gd x O 3 , where x = 0, 0.05, 0.1, 0.15) prepared by classical citrate method were tested in the photodegradation of MB. It was observed that while increasing the Gd content, the photocatalytic activity increases as well, reaching the highest conversion of the organic dye (94%) with Bi 0.9 Gd 0.1 FeO 3 . The increased ferromagnetic nature of Bi 0.9 Gd 0.1 FeO 3 is considered responsible for its different photocatalytic behavior in MB degradation [308]. A similar degradation efficiency (94%) was observed for rhodamine B over Gd-doped BiFeO 3 [309]. Congo red dye was almost completely oxidized on Ca-doped BiFeO 3 ultrafine nanofibres (Ca x Bi 1-x FeO 3 , where x = 0, 0.05, 0.1, 0.15) and the photocatalytic conversion of the dye increased along with the Ca content [310]. La x Bi 1-x FeO 3 (x = 0, 0.05, 0.1, 0.15) perovskites were used as photocatalysts for the photooxidation of the same dye, almost 80% of it being oxidized on La 0.1 Bi 0.9 FeO 3 after 3 h. Further increasing the La content decreases the conversion to ca. 50% [311].
A heterostructure based on BiFeO 3 /CuWO 4 was tested for the degradation of methyl orange (MO) and rhodamine B (Rh B). The dyes' concentrations were spectrometrically analyzed in time, and the absorption peaks of MO and Rh B disappeared almost entirely after 120 and 75 min, respectively. The higher photocatalytic activity observed on BiFeO 3 /CuWO 4 compared to single components was correlated with the heterojunction formation between the two oxides. The most beneficial BiFeO 3 :CuWO 4 ratio for the organic dyes photodegradation was found to be 1:1 [312]. Methyl violet (MV) was photooxidized up to 93% on a heterostructure based on p-n heterojunction BiFeO 3 /TiO 2 under visible irradiation as an effect of the poor recombination of charge carriers and strong absorption properties of the device [313].
Great efficiencies were reported for pure LaFeO 3 used as photocatalyst under visible irradiation for the degradation of 4-methylphenol [314], rhodamine B [315,316], and methyl orange [317]. Nanocubes, nanospheres and nanorods of LaFeO 3 were used as photocatalysts for Rh B oxidation under visible irradiation. No remarkable differences were observed between the nanorod and nanosphere morphologies, the organic dye conversions being similar. On the other hand, for the nanocube morphology the conversion efficiency was much lower. Irrespective of their morphology, all LFO samples showed better activities than Degusa P25 TiO 2 .
The higher efficiency of LFO is explained by its strong absorption properties in the visible region. Furthermore, the presence of an enhanced density of adsorbed oxygen or oxygen from surface hydroxyl groups on LFO photocatlaysts, which can act as efficient oxidizing agent, was confirmed by X-ray photoelectron spectroscopy (XPS) [318]. Wei et al. [319] reported that the doping of LaFeO 3 with Mn ions leads to a perovskite structure of LaFe 0.5 Mn 0.5 O 3 having higher oxygen vacancies and excellent absorption properties for visible light. All these, together with the variable valency of Mn cations contribute to the generation of a photocatalyst with improved activity for photodegradation of MO. It was also noted that the reaction mechanism is pH-dependent: the global yield decreased with the increasing of the pH value [319]. Recently, numerous scientific papers based on a heterostructure of nanosheets of graphitic carbon nitride (g-C 3 N 4 )/LFO nanoparticles were reported. High values of the photogenerated current were obtained on 15% g-C 3 N 4 /LFO (~9 µA/cm 2 ) and 20% g-C 3 N 4 /LFO (~17 µA/cm 2 ) compared to pristine LFO (0.04 µA/cm 2 ). Rh B was almost completely oxidized on g-C 3 N 4 /LFO after ca. 2 h under visible light irradiation. Moreover, the photocatalytic device showed excellent stability, its activity being entirely maintained after 4 catalytic cycles [320,321]. This photocatalyst was also used with good results for Brilliant Blue (BB) degradation [322].

Conclusions and Perspectives
The properties and (photo)catalytic behavior of perovskite materials in water splitting, catalytic combustion of methane and photo-degradation of pollutants from water have been presented and discussed based on more than 300 relevant papers, leading to the following conclusions: There are various synthesis methods for both powder and thin films, which determine their physicochemical properties. The specific surface area of the perovskite powders, which is a key characteristic of a solid catalyst, is strongly influenced by the preparation method used, but remains low. Indeed, the highest surface areas, mainly obtained by citrate and flame-pyrolysis methods, do not exceed several tens of m 2 /g. On the other hand, pulsed laser deposition is one of the most suitable preparation methods for inorganic perovskite thin films, due to its high material transfer efficiency, precise control and the great flexibility of the process. Depending on the experimental conditions, the stoichiometry of the material, as well as the thickness and the crystallinity of the films can be controlled. The most commonly used lasers for the preparation of perovskite films are those emitting in UV spectrum (193 nm, 248 nm and 355 nm). The films' thickness starts from less than 1 nm and rises up to ca. 600 nm. Oxide ferroelectric perovskites show excellent efficiency for the conversion of solar energy into chemical energy (H 2 ) via water splitting. Both photocatalytic and photoelectrochemical systems are extensively studied in this application domain. The high spontaneous polarization of BiFeO 3 is beneficial for a very efficient electron-hole separation. LaFeO 3 presents strong absorption properties of visible light, which represents ca. 42% of the entire solar spectrum. The photoelectrodes are tested for a wide range of pH values, starting from semi-acidic to strong alkaline media. The highest photocurrent density (46.9 mA/cm 2 at 2.53 V RHE ) is obtained for a complex heterostructure based on WO 3 /BiBO 4 /BiFeO 3 . The best stability (more than 120 h) was reported in 1M NaOH for p-LaFeO 3 /n-Fe 2 O 3 . Due to their good thermal stability, perovskite materials were successfully used in the catalytic combustion of methane for both power generation and methane emission abatement. Although the performance of pure perovskites is limited by their small specific surface area, their efficiency can be improved either by dispersion onto support materials possessing high surface area and thermal stability or by doping with other transition metals. Indeed, substitution in A and B sites of the perovskite structure with small amounts of other cations can improve both the stability and activity of the catalyst. Improved activity and stability can also be obtained by coating of the supported perovskite either on ceramic or metallic monoliths. The most used A-site dopants for perovskites are alkaline earth metals (Sr, Ca and Ba) and lanthanides (Ce, Eu), while for B-sites metals from the 3 and 4 periods (Mg, Al, Mn and Cu) in particular are preferred. The most active perovskites for the low-pressure methane combustion is La 0.6 Sr 0.4 MnO 3 with a value of T 50% of 360 • C. The high activity of this catalyst is due to its enhanced ability to adsorb oxygen on the surface. The photodegradation of organic dyes on inorganic semiconducting perovskites showed excellent results. Their high stability under extreme chemical conditions, strong absorption properties and efficient charge separation lead to high photocatalytic activity even after several reaction cycles. Catalytic systems containing BiFeO 3 perovskites as such or modified with different dopants exhibited an exceptionally high activity in the photocatalytic degradation of both anionic and cationic organic dyes.
Due to their high chemical and thermal stabilities together with their large compositional flexibility, perovskites remain (photo)catalytic materials of choice for applications in energy production and environmental protection. Obviously, new preparation procedures and the improvement of existing ones will allow in the future not only higher surface areas to be obtained but also new morphologies with consequences for their (photo)catalytic behavior.