Prospects of Halide Perovskites for Solar-to-Hydrogen Production

1. Introduction

Photocatalytic water splitting is one of the efficient and green energy technologies to produce hydrogen fuel with less to no toxic emissions. Ever since the first demonstration of the photoelectrochemical water splitting by Honda and Fujishima, solar energy driven hydrogen generation by employing semiconductors as an active material created a buzz in the research community. Several materials including TiO₂, ZrO₂, hetero-phase-mixed oxides, other white titanates and zirconates and oxide perovskite semiconductors have been under constant investigation for hydrogen generation; however, most of these semiconductors are active in UV region and there had been an ongoing demand for investigates materials with absorption in the visible region or competent with a broader absorption spectrum to utilize a substantial amount of solar radiation. While doping of the existing active materials to tune the band gap and achieve absorption in the visible region is one approach, alternatively the materials which are already proven to be satisfactory for other photovoltaic applications (like LEDs and solar cells) intrigued the researchers.

Materials with AMX₃ (A being the large cation M being the small metal cation and X being halide anion) are often referred to as Halide-perovskites (HPs), while the term perovskite pertains to the materials crystal structure. These materials are composed by two cations: A the larger cation, B the smaller cations, usually metal ion and X the halide anion. The investigation of Pb-HP CsPbX₃ is dated back to 1958 by Moller.[1] The following decade Pb- HPs fascinated the researchers with their intriguing optoelectronic and structural properties. Pb-HPs are known for their high light absorption coefficient and high carrier mobility with low excitonic binding energy, which make them potential candidates for photon-induced activities. In 2009, with Miyasaka’s report of replacing dye with organic-inorganic halide perovskite CH₃NH₃PbI₃ as a sensitizer in solar cell, HPs made a dynamic entry to the photovoltaic applications.[2] These materials have been extensively studied and employed for solar cells, LEDs, photodetectors, gas sensors, switching devices, FETs, metal-ion batteries etc., In 2016, the most promising Pb-HP perovskite made its debut into solar driven hydrogen generation.[3] However, realizing the inevitable stability and toxicity issues associated with lead HP, there have been constant efforts to develop sustainable and stable HP perovskites for renewable hydrogen generation.

In the current review, we summarize the Pb-free HPs for solar driven hydrogen generation (/evolution) from the aqueous solutions. As a part of this discussion, a general introduction to the solar driven hydrogen generation systems, status and issues associated with Pb-HPs in hydrogen generation are elaborated. A detailed discussion is provided on the novel Pb-free HP compositions developed for hydrogen generation, strategies to improve the performance well as stability of these materials.

2. Solar-Driven Hydrogen Generation Systems

Figure 1 shows the HP-based water splitting process and systems. Conventionally solar-driven hydrogen generation systems are classified into three systems based on the operating mechanisms- photocatalytic, photoelectrochemical (PEC) and photovoltaic-powered electrolysis (PV-E) hydrogen generation

Figure 1a summarizes the basic principle of photocatalytic water splitting. To split water into its constituent H₂ and O₂ molecules via photocatalytic process requires a visible-light-responsive photocatalyst with a sufficient potential to overcome the positive Gibbs free energy of splitting reaction of the water. Therefore, the photocatalyst should have a suitable a sufficiently narrow band gap to harvest visible photons and its band gap must encompass both the water reduction and oxidation potentials. These potentials stand at +0 V and +1.23 V respectively, referenced to the normal hydrogen electrode (NHE). Generally, the photocatalysis mechanism involves three major steps: (1) photogeneration of charge carriers (electron-hole pairs) which can be enhanced by bandgap tunability and high absorption coefficient (2) Diffusion of photogenerated charge carriers towards the redox-active sites on the surface. Charge separation and collection efficiency, can be enhanced with high carrier lifetime, defect and surface passivation hetero/homo-junctions ensuring that more electrons and holes successfully participate in the desired redox reactions. These systems are characterized by two-step excitation and/or charge separation process and mimic the natural photosynthesis process, harnessing wide range of visible light for splitting because a change in Gibbs free energy (∆G₀= 237.18 kJ·mol⁻¹) required to drive each photocatalyst can be reduced as compared to the one-step water splitting system and that the separation of evolved H₂ and O₂ is possible. (3) The final step is the photocatalytic oxidation and reduction of the water molecules for evolution of hydrogen and oxygen. As the photocatalyst is dispersed in aqueous solutions, the stability of these materials plays a crucial role with limited choices.

Figure 1c shows the schematic of the PEC water splitting system consisting of a photocathode and photoanode immersed in an aqueous solution, where the half-cell reactions constitute oxidation at anode and reduction at cathode – resulting in water splitting into H₂ and O₂. The performance of the PEC water splitting system is quantified by Solar-to-hydrogen (STH) efficiency (η_STH), usually measured under standard solar irradiance of AM 1.5G with no applied bias. η_STH is calculated from the photocurrent density generated J_sc (in mA.cm^-2) and Faradaic efficiency (η_F) contribution to the hydrogen evolution reaction (HER) and 1.23 V represents the standard potential required for the electrolysis of water. Alternatively, H₂ production/evolution can be measured by gas chromatography or mass spectrometry.

$${\mathsf{\eta }}_{STH}={\displaystyle \frac{(J_{sc}\times 1.23\times {\eta }_F)}{P_{in}}}\times 100$$

Figure 1d shows the schematic of a PV-E based water splitting system wherein a PV device provides the required potential to drive oxidation-reduction reactions for H₂ evolution.

3. Lead-Halide Perovskites for Hydrogen Generation

In 2016, Park et al.[3] pioneered the use of HPs for hydrogen generation through hydroiodic (HI) acid splitting, by leveraging the dynamic equilibrium of ions in methylammonium lead iodide (MAPI) perovskite in aqueous HI and HBr reaction mediums as shown in Figure 2a. The study demonstrated continuous photocatalytic activity of MAPI for approximately 6.5 days, showcasing robust phase stability and potential for hydrogen generation. Dissolution of the PbI₃^- ions in aqueous solution requires controlled accessibility of I^- ions and with an excess of I^- ions in the reaction medium leads to the formation of PbI₃^- and PbI₄^-2 ions. The availability of H⁺ ions is to be monitored to restrict the formation of hydrated phases and induce forced precipitation of MAPI crystals. Therefore, MAPI, in aqueous solution, forms hydrated phases or disassociates into it constitute metallic salt and requires a specific range of concentration of H⁺ and I^- ions for precipitation and stabilization. This understanding led Park et al. to employ aqueous HI and HBr reaction mediums to produce hydrogen gas by photocatalytic activity of MAPI. The study highlights a continuous photocatalytic activity of MAPI under illumination for about 6 and a half days. MAPI exhibited a robust phase stability due to the dynamic equilibrium of the material and the saturated reaction system. A series of experiments conducted under variable conditions, including variable intensities and wavelengths of visible light, and in presence of the Pt co-catalyst and H₃PO₂ reducing agent, indicated the potential of employing MAPI in aqueous environment for hydrogen generation by its photocatalytic activity.

Following this, a series of mixed halide perovskites, hydrogen generation from aqueous hydrohalic acids. Wu et.al significantly enhanced the photocatalytic hydrogen evolution activity by incorporating Br ion into MAPI.[4] They prepared powdered samples of MAPbBr_3–xI_x with a gradient of iodide concentration, gradually increasing from the interior to the surface of each particle to a facile light-assisted halide exchange, resulting in a band gap funnel structure as shown in Figure 2b(i). The distribution of Cl and I ions within the perovskite crystal is as shown in Figure 2b(ii). In this unique architecture, photoluminescence (PL) experiments revealed swift charge carrier transport from the bromide-rich region to the iodide-rich region. The iodide-rich sites acted as the primary locations for charge recombination, while the bromide-rich sites served as the absorbing and carrier generation sites. This funnel structure effectively minimized carrier recombination, leading to improved H₂ photogeneration. To demonstrate this, the experiments were conducted under aqueous mixed HI/HBr acid-saturated solutions, along with H₃PO₂ selectively reducing I^– ions that could interfere with light absorption. Under visible light irradiation (λ ≥ 420 nm, 100 mW cm^–2), MAPbBr_3–xI_x exhibited H₂ evolution of 1021.20 μmol h^–1g^–1 (compared to 11.20 μmol h^–1g^–1 for pristine MAPbBr₃). This efficiency was further enhanced to 2604.80 μmol h^–1g^–1 by incorporating a 3 wt % Pt loading to separate electron–hole pairs. The solar-to-light chemical efficiency at the start of the reaction without H₃PO₂ was found to be 1.05%, and the apparent quantum efficiency of the MAPbBr_3–xI_x/Pt system under 450 nm irradiation reached 8.10%. Remarkably, this system demonstrated excellent stability over 30 hours of continuous experiments, repeated in 6 cycles of 5 hours each.[4] The authors extended their band gap funnel approach to explore all-inorganic mixed perovskite, specifically the CsPbBr_3–xI_x system, using the same preparation method as before.[5] The band gap alignment and photographic images of the synthesized perovskites are shown in Figure 2c(i,ii). In this case, CsPbBr_3–xI_x was also loaded with Pt co-catalyst. Under visible light irradiation (λ ≥ 420 nm, 120 mW cm^–2) in an aqueous HBr solution saturated with CsPbBr₃, the CsPbBr_3–xI_x system demonstrated remarkable H₂ photogeneration, reaching 896.0 μmol h^–1 g^–1, while maintaining high stability for up to 50 hours. The apparent quantum efficiency for this system was found to be 2.15% under 450 nm irradiation.[5]

Employing composites, a traditional strategy to improve photocatalytic activity of semiconductors, is adapted by research in case halide perovskites as well. Wu and his team developed a composite of MAPI with reduced graphene oxide (rGO) by dispersing GO powders in an aqueous HI acid solution saturated with MAPbI₃.[6] Upon exposure to visible light (λ≥420 nm) during the photoreaction, GO was reduced to rGO, as confirmed by Raman spectroscopy. Further analyses using Transmission Electron Microscopy (TEM) and Fourier transform IR (FTIR) validated the attachment of r-GO to MAPbI₃ and the formation of new chemical bonds between them. Figure 2d shows the SEM image of the MAPI/rGO; inset shows the schematic of photocatalytic activity involved. Photocatalytic H₂ evolution reactions were carried out on powder samples of the MAPbI₃/r-GO composite in an aqueous HI solution saturated with MAPbI₃ (along with H₃PO₂, in order to reduce I₃⁻). Compared to pristine MAPbI₃, the composite exhibited a remarkable 67-fold improvement in photoactivity, reaching a rate of 939.90 μmol h⁻¹g⁻¹. Interestingly, the authors found that rGO performed even better than Pt as cocatalysts, as measured by hydrogen photogeneration in Pt-loaded MAPI. Moreover, the composite displayed excellent stability over 20 cycles of repeated experiments, with each cycle lasting 10 hours. Photoluminescence (PL) measurements confirmed that the composite facilitated more rapid charge transfer and effective electron−hole separation, leading to an increased H₂ evolution rate.[6]

Wang and colleagues employed a similar strategy to enhance HI splitting by MAPbI₃, using Pt/TiO₂ nanoparticles as nanoscale electron-transporting channels to efficiently extract electrons when in contact with MAPI. To improve the interaction between the metal halide perovskite (MHP) and Pt/TiO₂ nanoparticles, they heated the solution containing both semiconductors to promote MAPI dissolution and subsequent reprecipitation during cooling. This process strengthened the dynamic attachment of MAPI to Pt/TiO₂, facilitating their interaction. Remarkable enhancement in hydrogen evolution was achieved with a 50% loading of Pt (0.75 wt %)/TiO₂, resulting in an 89-fold improvement compared to pristine Pt/MAPI and leading to the evolution of 436.6 μmol of hydrogen in 6 hours. The composite’s stability was confirmed over four successive photocatalytic cycles. Based on photoluminescence (PL) measurements, the authors proposed a mechanism in which MAPI particles dispersed in a saturated solution acted as light absorbers, generating electrons and holes upon light irradiation. The proper band alignment between Pt/TiO₂ and MAPbI₃ facilitated the transfer of electrons generated on MAPI to Pt/TiO₂, where they reduced protons to form H₂. Simultaneously, the holes on MAPI oxidized I⁻ to I₃⁻, enhancing HI splitting efficiency. The authors also experimented with a composite using commercial Ta₂O₅ and Nb₂O₅ nanoparticles loaded with Pt. While Ta₂O₅ showed some effectiveness in promoting H₂ evolution, its potentially unfavourable band alignment resulted in negligible catalytic activity compared to TiO₂. However, it is important to note that these findings were preliminary, and further investigations could be pursued to explore the potential of such composites.[7]

While the traditional PV devices suffer from low open circuit voltages and does not overcome the thermodynamic driving forces and overpotentials of the water splitting reactions, HP solar cells possess sufficiently high V_ocs are advantages for water splitting. [3,8,9,10] However, integration of PV systems with the electrochemical cell is a challenging task. Designing multijunction/tandem junctions is an effective strategy to improve power conversion efficiency of the solar cells. Combining wide band gap HP with a narrow-band gap crystalline Si for a tandem solar cells recently surpassed Shockley Queisser limit, researchers from KAUST and Helmholtz Zentrum Berlin’s (HZB) demonstrated perovskite-Si tandem solar cells with efficiencies beyond 32%. With the PV-assisted water splitting which require high operating potential, these devices with high open circuit voltage (V_oc > 1.23 V) are of special interest. For the first time Dutta et.al integrated electrochemical cell and a monolithic lead HP- Si tandem solar cell to achieve solar-to-hydrogen (STH) efficiency > 21%.[9] With the operation of electrochemical cell operating at high photocurrent of 18 mA and operating potential over 1.4 V which is close to the maximum power point of the employed solar cell, the designed water splitting system is conveniently without light concentrating techniques.[9] K. Fehr et.al developed a durable HP photochemical cell fabricated with the conductive adhesive-barrier (CAB). [10] The CAB protected the HP from the corrosion and degradation of the material from the water splitting reaction and the stability and the Solar-to-hydrogen (STH) efficiencies are demonstrated for two electrochemical systems- one with a co-planar photoelectrodes and the other with a HP-Si photoanode. For a Co-planar photoelectrodes system, the researchers achieved a solar-to-hydrogen (STH) efficiency of 13.4% and a lifetime of 16.3 hours to reach t60 (60% of initial photocurrent). Monolithic silicon-perovskite tandem photoanode outperformed its counterpart with a remarkable STH efficiency of 20.8% and a lifetime of 102 hours to reach t₆₀. [10]

4. Issues with Lead HPs for Water Splitting

4.1. Toxicity

Halide perovskite energy conversion systems, solar cells as well as water splitting systems are dominated exhibit superior power conversion efficiency when conventional efficiency and figure of merit. However, it is worth noting that the most efficient PSCs are constructed using lead halide salts (PbI₂ and PbBr₂), which renders them potentially hazardous due to their toxic nature. Figure 3 shows the possible ways of lead into ground water, followed by entering the food chain and eventually into human intake.[11] Based on energy market projections for 2050, we can estimate the required amount of lead to produce these solar cells. Ram et al. conducted calculations and predicted that the electricity energy demand from PVs in 2050 would be 12,210 TWh/year, accounting for 60% of the total electricity demand. Assuming that PSCs will contribute approximately 30% of the PV energy share, the demand for lead for PSCs would amount to 1000 tons per year for the European Union (EU) alone. [12]

Despite the amount of lead used in PSCs is comparably low (as compared to other applications), there is a concern arises from the relatively high-water solubility of lead halide salts, higher than other heavy metal compounds commonly employed in the PV industry, such as CdS, PbS, and CdTe. For instance, the water solubility of PbI₂ is approximately one billion times higher than that of PbS and PbSe.[13] This solubility poses a problem as water/ moisture (as well as oxygen) can cause damage to solar cell, leading to the decomposition of perovskite into PbI₂, hydroiodic acid, and methylamine.[14] The interaction between MA⁺ and H₂O forms a strong hydrogen bond, weakening the original bond between MA⁺ and the Pb-I octahedral. This accelerates the deprotonation of the organic cation, and the proton can transfer to I⁻ via H₂O, resulting in the production of volatile species like CH₃NH₂ and HI leaving behind PbI₂ on the device.[15] Particularly failure of encapsulation of photoanode leads to the leaching out of lead into ground water ingress and is associated with elevated bioavailability, means a mere 10% increase in lead contamination from perovskite in soil could result in a staggering over 100% increase in lead contamination within plants. Consequently, plants grown in soils contaminated with lead, even at concentrations within stringent international limits for agricultural soils, exhibit evident lead intoxication.[13,16]

4.2. Stability

Photo-energy conversion systems used in practical applications must be designed to withstand diverse weather conditions, with temperature and humidity being critical factors. The International Electrotechnical Commission (IEC) standard includes three stability tests: thermal cycling, damp heat, and humidity freeze, which closely relate to these two factors. Ensuring stability against temperature and humidity can be assessed by considering various aspects, including the perovskite material, functional layers, interfaces, and back electrode. These elements collectively contribute to the overall performance and durability of solar cells under challenging environmental conditions. However, popular lead perovskite halides, include MA-based and FA-based ones face severe thermal instability.[17] Thermogravimetric analysis reveals that organic lead halide perovskites exhibit a high decomposition temperature beyond 85 °C, questioning their application for terrestrial photovoltaics. Reports indicate that decomposition can occur at 85 °C due to the volatilization and loss of various organic species in an open space.[18] Using less volatile alternatives like FA⁺ and inorganic Cs⁺ instead of MA⁺ is a common approach. However, the non-photoactive yellow δ-phase of FAPbI₃ and CsPbI₃, rather than the black phase, is thermodynamically stable at room temperature.[19]

4.2.1. Moisture induced Degradation

Despite their popularity, phase transitions, instability to the thermal stress and exposure to the ambient air and moisture have been constantly viewed as potential barriers for the organic-inorganic lead halide perovskite to be adopted of for technology. Earlier in 2014, researchers identified the degradation of CH₃NH₃PbI₂ into aqueous HI and Methyl amine and followed by the interference of O₂, the HI decomposes into I₂ and H₂O.[20,21,22,23,24] Figure 4a shows the proposed decomposition mechanism of MAPI in presence of water molecules. Further Snaith et.al established the roots of degradation of the perovskite to the dangling hydrogen bond and escaping of the MAI from the lattice structure, leaving behind PbI₂. [25] While Yang et.al investigated the degradation of MAPI using in-Situ GIXRD and absorption and found out the first step to degradation is formation of perovskite dihydrate containing isolated PbI₆^4– octahedra,[26] Frost et al. suggested reversible degradation mechanism that involves the phase transition of the perovskite molecule into its mono hydrate and successive breaking down into organic salt and the and PbI₂.[20]

4.2.2. Oxygen and Photo- Induced Degradation

In 2015, Aristidou et.al accounted for the degradation of CH₃NH₃I₃ in moisture-free condition to the oxygen present in the atmosphere.[27] Despite careful encapsulation, the material tends to decompose in the presence of oxygen and light; the photo-exited perovskite results in generation of the superoxide by transferring an electron to oxygen molecule, in return this superoxide attacks perovskite and decomposes into the organic and inorganic salts with water as by-product. The decomposition can be summarized as follows:

Degradation mechanism of CH₃NH₃PbI₃ in presence of oxygen

$${MAPbI}_3\stackrel{\begin{array}{c}Light\\ NoMloisture\end{array}}{\to }{{MAPbI}_3}^{*}$$

$$O_2\stackrel{{MAPbI}_3^{*}}{\to }{O}_2^{.-}$$

$${MAPbI}_3+O_2^{.-}\stackrel{Deprotonation}{\to }{MeNH}_2+{PbI}_2+{\displaystyle \frac{1}{2}}{I}_2+H_{2}O$$

* photo-excited spices of MAPI perovskite

The following year, the same group identified that contrary to reports suggesting adverse effects of the moisture induced degradation, the rate of light and oxygen induced degradation is higher. Figure 4b shows the degradation of MAPI under various conditions and clearly the degradation of the material in dry air and open air under light condition doesn’t exhibit much difference, indicating moisture not being the major culprit alone.[21]

Tang et.al studied the photo degradation of the perovskite thin film under various conditions (in nitrogen, vacuum, and air) and unveiled the photoinduced degradation of the CH₃NH₃PbI₃ under various experimental conditions (N₂, vacuum and atmospheric conditions under light and dark environments). Figure 4(c) shows the XRD patterns of MAPI films exposed to various conditions. The light-exposed sample under vacuum shows no sign of absorption, ascertaining the detrimental effect of photons on the perovskite. This investigation identified the similarity of the photo-degradation of CH₃NH₃PbI₃ and photolysis of PbI₂. While earlier discussed reported suggests that the degradation of MAPI under vacuum conditions primarily involves the formation of non-ionic molecular (MAI) defects, rather than ionic defects, implies that the loss of the CH₃NH₃⁺ cation is closely linked to the sequestration of iodide from the anion octahedron lattice (PbI₆)^4-, thereby creating iodine vacancies within the MAPI films. This study highlights that the initiation of photoinduced degradation in MAPI within a vacuum arises from the generation of iodine vacancies due to MAI loss.[22]

Sun et.al studied the photo-induced degradation of perovskite processed using two different precursors- Pb.Ac₂ (Ac being acetate) and solution-engineered (sol-eng) processed perovskite films and co-related the diffusion of O₂ and variation in the microstructure of the material. Interestingly sol-eng perovskite showed faster degradation and as compared to the Pb.Ac₂ perovskite and this attributed to the small gains of the sol-eng perovskite with high defect density at the boundaries and thereby making the material more susceptible to the oxygen diffusion (Figure 4d). On the other hand Pb.Ac₂ perovskite are characterized by large gains and O₂ attacks the grain boundaries followed the propagation of the defects throughout the material. Figure 4d shows the SEM images of the Pb.Ac₂ perovskite and diffusion for O2 across the grains and the relative degradation mechanism.[23]

4.3. Band Gap of Pb-HPs and Need of Pb-Free HPs in PECs

Conventional lead perovskite with iodide halide has a narrow band gap of 1.54 eV and is known to deliver high saturation current. However, the material is not an ideal choice for water splitting due to the photoelectric voltage which barely supports the water oxidation and charge injection potentials. On the other hand, chlorine-based perovskites are limited by the low absorption spectrum. Alternatively, bromide perovskites exhibit suitable energy band gaps and are apt for designing of low-onset potential photoanodes. Particularly, cesium lead bromide (CsPbBr₃) has the larger band gap at 2.4 eV, yet the lowest theoretical saturation current, which required combating strategies to compensate the saturation current. Moreover, while all inorganic perovskites require high temperature synthesis, methylammonium lead bromide (MAPbBr₃) have poor thermal and moisture stability. This opened doors for formamidinium lead bromide (FAPbBr₃) as a candidate material for photoanodes for water splitting. Recently, Yang et.al investigated monolithic FAPbBr₃ photoanode for water splitting. The photoanode delivered high photon-to-current efficiency of 8.5% with an open circuit voltage of 1.4 V. Interestingly, the energy band off-set allowed the coupling of the photoanode with carbon/graphite protective layer, improving the stability of the of the system to operational 100 hours.[28] On the other hand, 2D perovskites inherently have higher band gaps, which ensure sufficiently high potential to facilitate water oxidation/ splitting. For example (HDA)₂PbI₄ perovskite, a lead compound with hexadecylammonium (HDA) organic cations incorporated between the metal halide nanosheets, exhibit a bandgap of 2.4 eV and offer high stability in water environment because of the hydrophobicity of the long chain organic molecule.[29] Similarly, long chain alkyl amine perovskites (C_nH_2n+1NH₃)₂PbI₄ (n = 14, 16 and 18) have absorption edge extended to ~ 540 nm and demonstrated excellent stability up to 96 hours when dispersed in water.[30] However, lead-based 2D perovskites for water splitting are not significantly unexplored, probably due to the limited charrier mobility governed by structural low dimensionality.

While instability of conventional (organic) lead halide perovskite is a significant problem that needs to be addressed, some progress has been made in moving towards low –dimensional perovskites as well as all-inorganic perovskites. Proposing the development of 2D/3D heterojunction perovskites that could provide a trade-off between band gap, stability, and charge mobility emerges as a potential solution.[31] However, the toxicity concern primarily centers around lead, and efforts have been made to identify alternative materials that can substitute it.

5. Lead-Free Perovskites in Water Splitting/Hydrogen Generation

5.1. Classification of HP Perovskites (Based on Their Structure)

Based on the ionic radius of the cations and anions, perovskite materials adapt either of the following structural dimensionalities- 0-D or 2-D or 1-D or 3-D structures. Figure 5 shows the structure of different dimensional perovskites. Considering a 3-D perovskite with a general formula of AMX₃, a material has continuous corner-sharing metal halide [MX₆]⁴⁻ octahedra and a 3-D perovskite can be transformed into 2-D perovskite by incorporating a spacer cation which separate the metal-halide octahedra. Whereas 1-D perovskites consist of either face-sharing or corner sharing octahedra and in case of the 0-D perovskites, the metal-halide octahedral clusters exist individually and are surrounded by the A type inorganic or organic cations. Considering the strictly periodical spatial arrangement the metal-halide octahedra and wrapping/ surrounding of the A-type cation species of the octahedra, the lower dimensional (2-D, 1-D, and 0-D) perovskites can therefore be treated as assemblies of 2D quantum wells or 1D quantum wires or 0D molecules/cluster. Thus, lower dimensional perovskites are structurally unique and usually differ from morphological forms of 3-D perovskites like nanosheets/nanoplatelets, nanowires/nanorods, and nanoparticles/quantum dots based on 3D AMX₃.[32] For example, the 1-D perovskites morphologically resemble nanorods or wires with a strong interaction between the metal-halide species. These interactions lead to the formation of the electronic band formation, yet with the material being limited in length, favoring the quantum confinement effect.[33] Contrary to this, in a molecular level 1-D perovskites anionic metal halide species are surrounded by organic cations and are isolated from one another. This framework results in bulk assemblies of quantum wire-like structures, that is macroscopic crystals that exhibit properties of nanomaterials. However, lead-based 1-D perovskites did not progress greatly and therewith Pb-free 1-D perovskites have not been investigated extensively.[34,35] For instance Chenkun et.al attempted to synthesize 1D (C₄N₂H₁₄)SnBr₄ alongside 0D (C₄N₂H₁₄Br)₄SnBr₆ and observed the photoinduced degradation into 0-D perovskite. [36]

Furthermore, the dimensionality of the perovskite plays a crucial role in the stability, electrical and optoelectronic properties of the material. Typically, 3-D perovskites with small (alkyl chain organic) cations are susceptible to moisture, temperature and/or photo–induced degradation. Incorporating a bulky/long chain organic cation into 3-D perovskite matrix not only transforms it into a lower dimensional 2-D perovskite, but also improves the air/ environmental stability of the materials. Higher formation energy and hydrophobicity of the A site cations of the lower dimensional perovskites, particularly 2-D perovskites offer greater stability over its 3-D counter parts. On the other has 0-D perovskites, owing to their isolated and undisturbed metal-halide cores, offer highest stability; in fact, unstable tin-based perovskites on subjected to photodegradation or oxidation of Sn end up stabilizing into stable 0-D perovskites.[36,37] To understand the electronic and optoelectronic properties, it is essential to acknowledge the origin of the electronic band formation. Generally, with reduction in the dimensionality the bandgap and excitonic binding energy increases, which hinders their application in solar cells. However, with the quantum confinement(-like) effect due to electronic bands generated by the alternative A-type cation spices and metal-halide species, lower dimensional (particularly 2-D) perovskites are prominent in LED and photodetector applications. For a perovskite, the conduction band and valency band are formed from the hybridized orbitals of the metal cation and halide anion, with a little to no contribution from the A-type (organic and/or smaller) cation; with the reduction in the dimensionality the connection between the metal-halide framework is weakened, which is the basic reason for increase in the bandgap as well as reduction in the electronic dimensionality, otherwise low charge mobility (higher effect mass of the charge) of material. Despite this, the electronic dimensionality and effective mass of the charges of the perovskite or even the band gap can be improved by wisely choosing/replacing the inactive A-type cation otherwise orienting/tilting the metal-halide octahedra.[38,39] This viability 2-D as well as 0-D perovskites is greatly exploited in tuning these materials for specific applications.[40,41]

Apart from classification of perovskite in terms of the structural dimensionality perspective, replacement of the Pb in by different metal cations leads to the following perovskite/ perovskite-inspired materials. When Pb⁺² is replaced with either trivalent metal cations (like Bi⁺³ or Sb⁺³) leading to the formation of perovskite- inspired materials, which are often structurally 0-D and have a general formula of A₃B₂X₉, these 0-D perovskites be structurally tuned to from a layered 2-D perovskite as well; Lee et.al introduced polyethylene glycols (PEGs) ligands to a 0-D Cs₃-nBi₂X₉ to synthesize a 2-D [(PEG₆^-NH₃⁺)_nCs_3-nBi₂X₉, perovskite.[42] On the other hand, when Pb⁺² is replaced with a pair of metal cations, M⁺¹ and M⁺³, forms the vacancy order double perovskites with a general formula of A₂M(I)M(III)X₆. Also, other class of the vacancy order perovskites consist of metal cation with tetravalent metal cation and have a general formula of A₂MX₆; Tin- based (Sn-) HPs, Cs₂SnX₆, are one such popular vacancy order perovskite known for their air-stability and explored for solar cells as well as photocatalysis.

5.2. Structure and Bandgap in (Lead-Free) Perovskite

Consider a 3-D perovskite AMX₃ such as CsSnX₃ (X = I, Br, Cl), Huang et.al studied the electronic band structure of the material using the quasiparticle self-consistent GW method (QSGW). For a cubic phased perovskite, the VB maxima is comprised of the mostly the hybridized X’s ns states and a band which is antibonding between Sn 5s and X’s np orbitals and the CB minima is mainly comprised of the Sn’s 5p states. Furthermore, conduction band minima (CBM) exhibits threefold degeneration without SO coupling and splits into a doublet and a quadruplet (including spin) when SO is included.[43] Similarly for a 3-D MASnX₃ and 2-D perovskite the contribution from the larger (/organic) cation is insignificant. In fact, the hybridized orbitals of the Sn-s to X−p form strong band dispersion at the valence band maxima (VBM) and lead to the higher hole mobility, thus the material was first reported as solid hole transport layer for the dye sensitized solar cells.[44] However, absence of the spin-orbital and band splitting at the CBM, the electron mobility of is not in par with the conventional Pb-perovskite. Moreover, the stabilization of tetravalent ions of the metal ions increases from Pb to Sn to Ge to Si over their divalent forms, which argues the synthesis and applications of the 3-D Ge and Si perovskites.

Conventional method to tune the band gap by partial to full replacement of halide anion and the A-type cation in the traditional lead-based perovskite is equally effective for the 3-D lead-free perovskites. Typically, for a 3-D tin-based perovskite MASnCl₃ the bandgap is reported to be 1.1 eV and replacement of MA with FA or Cs led to increasing band gap to 1.4 and 1.36 eV respectively.[45] Although literature suggest the non-involvement of the A-type cation in the CB and VB formation, the A-type cation is responsible for the titling of the metal-halide octahedra, thereby affecting the band gap.[46] Other arguments validated the signature of the C- and N- orbitals of the organic cation in the formation of the VB.[47] Replacement of MA with a larger cation like dimethylammonium not only increased the band gap (2.05eV to 2.9) but also transformed the crystal structure from pure cubic to orthorhombic.[48]

The exploration of group-VA cations (M³⁺=Bi³⁺ and Sb³⁺) as potential replacements for Pb²⁺ aims to maintain the chemistry of the lone-pair ns² state, which is known for its benefits in achieving high photovoltaic performance. However, due to their higher +3 oxidation state, the normal AMX₃ perovskite structure, consisting of corner-sharing [MX₆] octahedra, cannot form. Instead, A₃M₂X₉ emerges as the stable stoichiometry in this context. Generally, researchers have experimentally reported two main phases of A₃M₂X₉: the hexagonal phase, featuring zero-dimensional (0D) bi-octahedral face-sharing clusters [M₂I₉]³⁻ (also known as the dimer phase, and the phase composed of two-dimensional (2D) corrugated layers with partially corner-sharing MX6 octahedra (referred to as the layered phase,). The 0D dimer phase can be easily synthesized through low-temperature synthesis methods.

Perovskites with two metal cations with a general formula A₂M(I)M(III)X₆ or A₂M(IV)X₆, (otherwise called as vacancy order perovskites) and A cation typically being a large monovalent ion, often Cs⁺ cation, M(I) and M(III) cation often being transition metal ions, such as Ag+ (K+ or Tl+ ) combined with Bi³⁺ (Sb³⁺) or it would be a simple tetravalent metal ion like Sn⁴⁺ (Te⁴⁺) in case vacancy order perovskite. The crystal structure of double perovskites can be visualized as a framework of corner-sharing octahedra. The larger A-site cations occupy the voids between the octahedra. The smaller metal cations are located at the center of the octahedra. The octahedra share their corners, forming a three-dimensional network. This arrangement provides structural stability to the material. The double perovskite structure is highly flexible, allowing for the incorporation of different combinations of cations at the M(I) and M(III) sites. This flexibility gives rise to a wide range of properties and functionalities in double perovskite materials. Furthermore, in A₂M(IV)X₆, the octahedron is no longer connected by the halide anion resulting from the alternative arrangement of vacancies and M(IV) cations. Generally, considering their valence atomic orbitals, the possible substituting for metal cations can be metals- 1. with the semi-core d states, 2. with the lone pair s state, and 3. without the d or s states. And the presence of lone pair due to the ns² electrons from the metal cation determines the nature of the bandgap. For suppose, the predominant Ag based double perovskite systems Cs₂AgBiX₆ exhibit indirect bandgaps with mismatch of the CBM and VBM at L point and X point respectively and angular momentum mismatch of the orbitals that comprise the CBM and VBM. Moreover, With M(I) being monovalent and M(III) being a trivalent metal cation, the band edges are determined by the [M(III)X₆] octahedra rather than M(I). The CB and VB arise from the np orbital of the M(III) and antibonding orbitals of the M(III)’s ns and X’s np orbitals. This leads to the isolation of the [B(I)X₆] octahedra and is responsible for the electronically 0-D structure. Thus, the indirect and large bandgap alongside the reduced electronic dimensionality leading to the relatively large carrier effective masses and reduced mobility makes the Cs₂AgBiX₆ not an ideal choice for solar cells. However, the material has significant importance in LEDs and in recent times explored for photocatalysis.

In case of double perovskite A₂M(I)M(III)X₆, VBM of formed is by bonding orbitals of MIII (ns), MI (nd), and X (np), whereas the conduction band minimum (CBM) is formed by antibonding orbitals MIII (np) and X (np). This electronic band structure is similar to that of a 3-D perovskite as well as GaAs or CdSe. [49] On the other hand, absence of the antibonding orbitals at either of the band edges, leads to poorer defect-tolerant materials. To understand the electronic band of a 3-D double perovskite, let us consider novel Bi-Ag perovskite system, say Cs₂AgBiX₆. The outer shell electronic configuration of Bi³⁺ is 6s²6p⁰ and is same as Pb⁺². The VB is predominated by the X[np] and the Ag [4d] while the CB is formed by X[np], Ag [5s] and Bi [6p] states. Further, the mismatch in the localization of the CBM and VBM gives rise to the indirect bandgap in these materials. An ordered Cs₂AgBiBr₆ exhibits a indirect bandgap of 2.04 eV and easily be switched to direct bandgap disordered Cs₂AgBiBr₆. It is known the direct bandgap materials have a higher mobility (lesser effective mass of electrons) as compared to that of the than their counter part with of the same bandgap. Thus, disordered Cs₂AgBiBr₆ has suitable direct band gap of 1.59 eV with higher mobility and is apt for photovoltaic and photocatalytic applications. [50]

Stability of the A₂M(I)M(III)X₆ is mostly dependent on the cations A and M(I) and the halide X. The theoretical calculations suggest a perovskite system with larger A⁺ cations (e.g., Cs⁺ is preferred over Li⁺) and smaller halides (e.g., F^– is preferred over I^–) has higher stability. Most investigated double perovskites and contain Cs cation and either Cl or Br anions, and to the best of our knowledge, the bandgap of these systems is suitable for PV and PEC applications. For example, Cs₂AgFeCl₆ exhibited a bandgap of ∼1.55 eV and replacing the Fe with Ga lead to reduction in the bandgap to ∼1.37 eV. [51] Also, stability is influenced by the size of the metal cation M(I) and often perovskite systems with Ag+ have higher stability compared to Cu⁺. Introducing larger cation like iodine into the perovskite system for tuning the bandgap could be an unfavorable approach and can lead to collapsing of perovskite structure due to crystal instability. Interestingly, alloying the trivalent metal cations can be adopted for finely tuned by bandgap and switching between direct and indirect bandgap material.

Furthermore, in case of double perovskites, a rational substitution of the metal cations allow tuning from indirect to direct bandgap materials. The key to choose M(I) /M(III) cations in A₂M(I) M(III)X₆ with lone-pair in ns orbital, resulting in strong s–p coupling between the ns orbital and X p orbitals, or induce a disordered metal alloy at the respective metal ion sites. Alternatively, double perovskites with metal cations fully occupied pseudo-closed s² shells have similar band structure as lead perovskite and pose direct bandgap, high mobility and high absorption coefficient. However, unlike Bi/Sb- Ag based double perovskite, not many of these perovskites are realized. Zhao et.al reported design of the double perovskites for photovoltaic application, which reveals that Cs₂InBiCl₆ and Cs₂InSbCl₆ exhibits a direct bandgap, with the CBM formed from antibonding states of metal cations’ np orbitals and Cl 3p orbitals, and VBM from the antibonding states of metal cations’ ns orbitals and Cl 3p orbitals.[52] Theoretical studies identified that (MA)₂TlBiBr₆ is shows the similar electronic structure as conventional lead perovskite, and experimental results confirmed the direct band of the material to be 2.16 eV. However, the presence of Tl, which is again a toxic element, is the major setback for this material.[53]

Perovskites crystallizing in 2-dimensional units separated by interlayer species are termed layered perovskites. Depending on the spacer cations the layered perovskites can fall in on the two categories – Ruddlesden-Popper (RP) perovskite (with a chemical composition A_n−1M_nX_3n+1) and Dion-Jacobson (DJ) perovskite (with a chemical composition F′A_n−1M_nX_3n+1 where F and F′ are monovalent (+1) and divalent (+2) organic spacer cations, respectively. To simplify the understanding, RP perovskites have spacer molecules (generally significantly larger organic cations) interlayered between the edge-sharing inorganic Octahedra to create a 3-D network. whereas DJ perovskites still have organic spacer layers but connected in a zig-zag fashion and corner-sharing inorganic Octahedra resulting in 2-D framework. Interestingly, the structural stability of the RP perovskite is vulnerable due to variable thickness, on the other hand DJ perovskites are limited by tolerance factor due to restricted movement of corner sharing species. Apart from these structural differences, in both the layered perovskites share the similar traits- Metal-halide framework is mainly responsible for the optoelectronic properties, the spacer cations play role in physical properties like exciton binding energy and charge mobility and exhibit wider bandgap. better stability and moisture resistance compared to their 3-D counter parts. It’s worth noting that in 1994, D.B. Mitzi in his Communications about synthesis of the Sn layered perovskite, commented about the tuning of electronic and magnetic properties in layered perovskites without disturbing the core active layers.[54] Thanks to the interlayer arrangement of the organic and inorganic layers in the layered/ 2-D perovskites, the material exhibits better overall stability because of the following factors- 1. Restriction of the movement of the ions 2. Structural confinement and maintained integrity of the material 3. Restriction of the moisture ingress by the organic spacer layers with tunable length 4. Tolerance to mechanical strain and thermal cycles. In fact, incorporating/ introducing a small amount (0.08 M) of layered (Sn) perovskite in the 3-D FASnI₃ perovskite improved the crystallinity of material and reduced the trap states, which are otherwise caused by the oxidation of the Sn⁺² to Sn⁺⁴ in 3-D tin perovskites. This 2D/3D hybrid tin perovskite devices exhibited 9 % PCE (Vs. 6% PCE for the 3-D Sn perovskite device) with enhanced light and environmental stability. [55]

For a (2-D Ruddlesden–Popper) layered perovskites, the inorganic metal-halide [MX₆]⁴⁻ layers are separated by insulating organic (/separator layers), and these separator layers are connected via van der Waals force. Provided the thickness of the inorganic sheet is comparable to the de Broglie wavelength of carriers, 2-D materials experience a strong quantum confinement and quantum size effects.[56] This results in unique quantum well like electronic band structures in with a formation of strongly bound exciton and increased photon absorption.[57] Thus, 2D perovskites exhibit different optical properties compared to 3D ones. Evidently, the size and shape of the organic and the spacer cations determine the tilting and relative positioning of the [MX₆]⁴⁻, interfering with the optoelectronic properties of the material. Madal et.al studied the influence of the positioning of the fluorine substituted PEA in the phenyl ring of (PEA)₂SnI₄. However, the study revealed that the bandgap variation is insignificant.[58]

By understanding of 3-D, 2-D and 0-D perovskite structures and structure-bandgap correlation and stability, researchers identified various potential halide-perovskite compositions for hydrogen generation. Based on the conduction and valence band potential levels with respect to the NHE, these lead-free perovskites are employed either in photocatalytic or photo-electro chemical systems. The following discussion provides a summary of Pb-free compositions used for hydrogen generation.

5.3. Photocatalytic Water Splitting Using Pb-Free Halide Perovskites

Figure 6 shows the band gap alignment of various Pb-free HP with rest to NHE that exhibited promising results in the solar-driven hydrogen generation from the water and/or aqueous solution.

5.3.1. Bismuth and Antimony HPs

Stable oxidation state of bismuth and antimony (Bi⁺³ and Sb⁺³) is isoelectronic with Pb²⁺ with comparable ionic radius and similar valence band configuration (6s²), making bismuth a potential replacement of Pb⁺² in conventional perovskite. However, owing to the hetero-valent metal replacement, the perovskite adapts a 0-D structure with A₃Bi(/Sb)₂X₉ formula. Thus 0-D perovskites are often difficult blend into a continuous uniform film and possess a indirect band gap. These factors restrict their application in photovoltaic applications. However, they are promising lead-free halide perovskites for photocatalytic activities. Table. 1 lists out the 0-D Bi and Sb HPs employed for the photocatalytic water splitting.

Table 1. 0-D bismuth and antimony HP photocatalytic systems discussed in Section 5.3.

No	Material	Reaction Medium	Illumination	HER/ Photocurrent	Stability	Ref
1	MA3Bi2I9/ Pt	Aqueous HI/H3PO2	300 W Xe-lamp with a 400 nm cutoff filter	169.21 µmol g−1 h−1	10 hrs of 7 cycles	59
2	Cs3Bi2I9	Aqueous HI/H3PO2	Visible Light	22.5 μmol h-1 11.7 H2 molecules per second	5 hrs of 3 cycles	60
3	Cs3Bi2I9	HI in n ethyl acetate	100 mW cm-2	1504.5 μmol h-1 g-1	2 hrs of 4 cycles	61
4	Cs3Bi2I9/Pt	Aqueous HI/H3PO2 Aqueous Me.OH	100 mW cm-2 (λ > 420 nm)	2304 μmol g-1 35.5 μmol g-1	4 hrs	62
5	MA3Sb2I9/ Pt	Aqueous HI/H3PO2	100 mW cm-2 (λ > 400 nm)	883	3 hrs of 4 cycles	63
6	Cs3Sb2I9/ Pt	Aqueous HI/H3PO2	100 mW cm-2 (λ ≥ 400 nm)	804.54 μmol g−1	50 h	64
7	2-AMPSbI5-1	Sodium sulfate: H2O	150 W xenon lamp	106.7	4 cycles	65
8	2-AMPSbI5-2	Sodium sulfate: H2O	150 W xenon lamp	96.3	4 cycles	65
9	PtSA/Cs2SnI6	Aqueous HI	100 mW cm−2 (λ ≥ 420 nm,	430 μmol h−1 g−1	180 h	66

Table 1. 0-D bismuth and antimony HP photocatalytic systems discussed in Section 5.3.

No	Material	Reaction Medium	Illumination	HER/ Photocurrent	Stability	Ref
1	MA3Bi2I9/ Pt	Aqueous HI/H3PO2	300 W Xe-lamp with a 400 nm cutoff filter	169.21 µmol g−1 h−1	10 hrs of 7 cycles	59
2	Cs3Bi2I9	Aqueous HI/H3PO2	Visible Light	22.5 μmol h-1 11.7 H2 molecules per second	5 hrs of 3 cycles	60
3	Cs3Bi2I9	HI in n ethyl acetate	100 mW cm-2	1504.5 μmol h-1 g-1	2 hrs of 4 cycles	61
4	Cs3Bi2I9/Pt	Aqueous HI/H3PO2 Aqueous Me.OH	100 mW cm-2 (λ > 420 nm)	2304 μmol g-1 35.5 μmol g-1	4 hrs	62
5	MA3Sb2I9/ Pt	Aqueous HI/H3PO2	100 mW cm-2 (λ > 400 nm)	883	3 hrs of 4 cycles	63
6	Cs3Sb2I9/ Pt	Aqueous HI/H3PO2	100 mW cm-2 (λ ≥ 400 nm)	804.54 μmol g−1	50 h	64
7	2-AMPSbI5-1	Sodium sulfate: H2O	150 W xenon lamp	106.7	4 cycles	65
8	2-AMPSbI5-2	Sodium sulfate: H2O	150 W xenon lamp	96.3	4 cycles	65
9	PtSA/Cs2SnI6	Aqueous HI	100 mW cm−2 (λ ≥ 420 nm,	430 μmol h−1 g−1	180 h	66

Guo et al. (2019) successfully synthesized bright-red colored bismuth perovskite MA₃Bi₂I₉ using a simple solvothermal method for the photocatalytic hydrogen generation from aqueous HI solution (hypophosphorous acid as a selective reducing agent). The perovskite powder has a VBM of −5.76 eV and a CBM of −3.78 eV, which are suitable for photocatalytic HI splitting. The researchers demonstrated the exceptional phase stability of MA₃Bi₂I₉ in hydriodic acid (HI) aqueous solutions with varying concentrations under visible light irradiation. The photocatalytic activity of the material with the addition of Pt as a cocatalyst is significantly enhanced the photocatalytic rate for H₂ evolution, reaching approximately 169.21 µmol g⁻¹ h⁻¹, a 14-fold improvement compared to the bare perovskite. The solar chemical conversion efficiency was measured at 0.48%. The synthesized perovskite exhibited excellent phase and photocatalytic stability, even after 70 hours of repeated H₂ evolution the material remained robust in the reaction solution. The enhanced stability of MA₃Bi₂I₉ was attributed to the oxidation state of bismuth (III).[59]

As photocatalytic activity takes place on the surface of the photocatalyst, nanostructure and morphology- which influence the surface area of the catalyst, and anisotropic properties-which vary based on the orientation of the catalyst, play a crucial role in the photocatalytic activity. Choudary et.al., (2022) study the Cs₃Bi₂I₉ nanodiscs (NDs) for photocatalytic aqueous HI splitting as well the photoelectrochemical conversion. The authors were successful in stabilizing the NDs at various HI by maintaining a dynamic equilibrium between the saturated perovskite solution and the material precipitated. The perovskite NDs exhibited exceptional stability at a diluted concentration of HI for about 8 hours. The synthesized nanostructures exhibited a remarkable H₂ evolution rate of 22.5 μmol h^-1 under visible light in an aqueous HI with apparent quantum efficiency of 1.3%. Furthermore, these NDs demonstrate a low overpotential of 533 mV at -100 mA per square centimeter for electrocatalytic H₂ evolution in the same 6.34 M HI solution, with a turnover frequency of 11.7 H₂ molecules per second. Notably, the NDs display exceptional recyclability and durability for both photocatalytic and electrocatalytic HI splitting processes.[60] Li et.al., (2022) studied the anisotropic charge transfer of in the bismuth perovskite and indicated the favorable facets for the effective charge separation and enhanced photocatalytic hydrogen generation. While (100) and (006) exhibited higher photocatalytic activity, the researchers developed a simple solvothermal process to achieve defined morphologies and facet-specific Cs₃Bi₂I₉ perovskites to capture this advantageous anisotropic properties. Perovskite hexagonal prisms showed superior photocatalytic H₂ evolution performance, splitting of HI in n ethyl acetate (EA) reached 1504.5 μmol h^-1g^-1 of H₂ evolution rate, which is 22.1 times that of a disordered Cs₃Bi₂I₉ photocatalyst compared to their counterpart samples with disordered structures and low (100)/(006) ratios. The stability and morphology of the perovskite were maintained after 8 h of photocatalytic reaction, with trace iodine particles formed on the (100) facets because of iodide photo-oxidation.[61] Miodynska et.al., (2023) attempted to understand the of effect of the larger cation in Bi- based 0-D perovskite in terms of morphology, stability and band edge variation. As series of Bi-perovskite, A₃Bi₂I₉ (A = Cs, Rb, MA, FA) are investigated for this purpose and identified that although the crystal structure and band gap are not affected, morphology and greatly influenced by the A type cation. While Cs forming hexagonal prisms, Rb forming irregular aggregates, and MA and FA forming irregular structures of aggregated particles with a band gap varying between 1.82-1.88. Owing to the higher degree of crystallinity and defined morphology, Cs₃Bi₂I₉ exhibited superior photocatalytic performance in both methanolic and acidic electrolytes: The Cs₃Bi₂I₉ photocatalyst yielded approximately 35.5 μmol g^-1 and 2304 μmol g^-1. in the reaction when employed with 10% MeOH and HI/H₃PO₂ electrolytes, respectively under 4 hours of illumination.[62]

Recently, Sb-based 0-D perovskite turned out to be an attractive option for photovoltaic and photocatalytic applications owing to its stability and non-toxic nature. Particularly, DFT studies of MA₃Sb₂I₉ perovskite reveal high charge mobility, direct band gap and good defect tolerance. This motivated researchers to exploit the material properties for hydrogen generation. Ahmad et.al (2023) synthesized stable MA₃Sb₂I₉ via solvothermal reaction for the hydrogen generation from aqueous HI using H₃PO₂ as a scavenger. The material produces 300 μmol g^-1 h^-1 of H₂ under visible light irradiation. The addition of Pt as a co-catalyst enhances the H₂ production rate up to 883 μmol g^-1 h^-1. The material exhibits a high stability and maintains its activity after five cycles.[63] The same group (2023), reported Cs₃Sb₂I₉ with a comparable performance as compared to the MA-Sb based perovskite. Cs₃Sb₂I₉ showed hydrogen generation of 95.54 μ mol g⁻¹. Platinum was used as bi-catalyst and obtained results for Cs₃Sb₂I₉/Pt (2 mg) showed excellent hydrogen generation of 804.54 μ mol g⁻¹. As anticipated the material exhibited higher stability as compared to MA-based counterpart owing to its all-inorganic composition.[64] Recently Rokesh et.al., (2023) reported 2-(aminomethyl pyridine)SbI₅ perovskite inspired material for photocatalytic and PEC hydrogen generation. Two photocatalysts, AMPS-1 and AMPS-2 are essentially synthesized from two Sb- precursors, iodide and oxide and their hydrogen evolution rates are compared. The band of AMPS-2 is lightly larger than AMPS-1, which could possibly due to the presence of oxide impurities. However, these impurities have not significantly affected the photocatalytic performance of the material: while AMPS-1 produced about 106.7 μmol h⁻¹ g⁻¹ hydrogen, AMPS-2 yielded 96.3 μmol h⁻¹ g⁻¹.[65]

5.3.2. Tin and Germanium HPs

While tin-halide perovskites established their prominence in replacing the lead perovskites in solar cell devices, they phase instability and low band gaps remain detrimental for their potential utilization in photocatalytic applications. On the other hand, Ge, being from the same group as Pb, Ge is looked up as prospective replacement for Pb in conventional perovskites. However, as far as we know very little literature is available for conventional 3-D tin and Ge perovskites used in hydrogen generation. Fortunately, like that of lead perovskite, the electronic structure of tin perovskites is strongly influenced by the organic cation and halogen, which can be exploited to tune their electronic energy levels and photo reactivity. Based on this strategy, Ricciarelli et.al., (2022) suggested modifying conversional 3-D tin perovskite by replacing MA cation with DMA, resulting in a stable perovskite with a suitable band gap and energy level alignment for hydrogen generation.[67] Furthermore, PEA₂SnBr₄, DMASnBr₃ and PhBz₂GeBr₄ showed remarkable stability in water and are employed in conjunction with g-C₃N₄ as heterojunction photocatalysts for water splitting.[68,69,70] These are discussed in heterojunction photocatalysts with g-C₃N₄ section.

5.3.2. Vacancy-Order HPs

Another class of perovskites, vacancy-order (double) perovskites have recently been investigated in various optoelectronic, electronic as well as energy conversion systems owing to their robust crystal structure and optoelectronic properties in par with conventional lead perovskites. Particularly, Sn- based perovskite Cs₂SnX₆, Bi-Ag based perovskites Cs₂AgBiX₆ have proven to be suitable substitutes for the Pb-HPs for hydrogen generation.

Zhou et al. (2021) synthesized all-inorganic Cs₂SnI₆ perovskite material that can anchor single-atom platinum (Pt) sites with a unique Pt-I₃ configuration on its surface (referred to as PtSA/Cs₂SnI₆) enhancing the photocatalytic activity and stability of hydrogen evolution reaction. The Pt single-atom sites on the perovskite surface decorated by using hydrothermal treatment, impregnation, and activation techniques. The PtSA/Cs₂SnI₆ catalyst has exceptional photocatalytic performance in hydrogen production from aqueous HI solution, achieving a remarkable turnover frequency of 70.6 h⁻¹ per Pt. This value is approximately 176.5 times higher than that seen for Pt nanoparticles supported on Cs₂SnI₆ perovskite. Here is a summary of the photocatalytic test results. This superior performance of PtSA/Cs₂SnI₆ is attributed to the unique Pt–I₃ configuration and the strong metal-support interaction effect, which enhanced the charge separation and transfer, and reduced the energy barrier for hydrogen production.[66]

5.3. Pb-Free HPs for Water Splitting via PEC Systems

Table 2 summaries the Pb-free HP based photoelectrode systems used in PEC for hydrogen generation. Often 3-D perovskites (ABX₃) suffer with instability issues in aqueous and are usually employed with multiple protection layers to avoid direct contact with electrolytic medium, particularly in PCE water splitting systems. Usually employed protection layers are InBiSn alloy, metal foils including titanium and nickel, carbon materials and atomic layer deposited oxides. On the other hand, vacancy-order Pt, Ag-Bi, Re based-perovskites exhibited resistance to degradation in harsh aqueous mediums which enabled the use of perovskite electrodes without protection layers, opening new doors to PEC water splitting in variable pH mediums.

Cs₂PtX₆ vacancy order perovskite is known for its exceptional stability over a wide range of pH, variable thermal and prolonged ambient exposure conditions. Hamdan et.al., (2020), studied the electrochemistry of Cs₂PtI₆ in electrolytic solution of pH varying from 1 to 14. Cyclic voltammetry (CV) is conducted to investigate the redox processes and catalytic activity of the materials in aqueous solutions of different pH values. The key finding was that Cs₂PtI₆ showed reversible oxidation/reduction peaks corresponding to I^-/I^3- and Pt⁴⁺/Pt²⁺, and that it was stable in acidic, neutral, and basic media. Interestingly, Cs₂PtI₆ had mixed valency of Pt⁴⁺ and Pt²⁺, and that the redox processes involved the formation and reduction of triiodide (I^3-). Furthermore, Cs₂PtI₆ exhibited a photocurrent of 0.8 mA cm^-2 at 1.23 V(vs. RHE) and over 12 hours of PEC stability without loss of performance. However, this value is much lower than the reported values for lead halide perovskites which is possibly due to low charge carrier mobility and lifetime of Cs₂PtI₆. The Cs₂PtI₆ material faces multiple challenges: its high production cost due to platinum usage limits its commercial viability; its 1.4 eV band gap isn’t optimal for single-junction photovoltaics, but this could be improved through halide/chalcogen substitution in the X anion; and its poor contact with conducting glass leads to high series resistance and a low fill factor, future work on optimizing deposition methods and charge transport layers.[71] Sikarwar et.al (2023) et.al explored stable Ag-M (M= In, Bi, Sb) vacancy-order perovskite photoelectrodes for PEC water splitting. The synthesized materials exhibit remarkable resistance to oxidation retaining their structural stability over 100 cycles of electrochemical cycling. This allows for the utilization of these substances in the process of photoelectrochemical (PEC) water oxidation in both CH₃CN and H₂O, with and without the presence of an IrOx cocatalyst.[74]

Nandigana et.al., (2023) studied two dimensional variants of bismuth perovskite vacancy-order 2-D Cs₂AgBiCl₆ and 0-D Cs₃Bi₂Cl₉. The hydrothermally synthesized double perovskite Cs₂AgBiCl₆ exhibits multifaceted crystal structures, featuring uniform di-pyramidal units, while Cs₃Bi₂Cl₉ particles lack the defined structural formation. Interestingly both the perovskites exhibit high thermal stability, the TGA studies reveal that both the materials are stable upto 470 ^oC. The perovskite photoanodes are tested for the photoelectrochemical performance using three electrode electrochemical workstation in the 1 M KOH (pH ∼13.7) alkaline electrolyte with Pt counter electrode and Ag/AgCl (Sat. KCl) as the counter electrode. The dark and light current densities of the Cs₂AgBiCl₆ are 3.30 mA cm^-2 and 3.85mA cm^-2 and that of Cs₃Bi₂Cl₉ are 1.78mA cm^-2 and 2.18mA cm^-2. The photo-response of both the photoanodes is instantaneous and persisted for about 300 sec. The PCE for the photoelectrochemical system is determined by the following equation and PCE of Cs₃Bi₂Cl₉ and Cs₂AgBiCl₆ is achieved to be 0.09% and 0.13% at the 0.93 V vs Ag/AgCl. The materials exhibited steady performance over 10 hours. The low charge transfer resistance and high charge transfer density of Cs₂AgBiCl₆ as compared to the Cs₃Bi₂Cl₉ resulted in its superior performance, which is understandably from the high crystallinity and defined morphology of the nanocrystals.[73]

Chandra et.al., (2023) the first time investigated for Re-based vacancy-ordered perovskite, Cs₂ReX₆ (X= Cl and Br) for PEC generation hydrogen. The hydrothermally synthesized Re-perovskites have similar structure to Cs₂PtX₆ compounds and exhibited great thermal stability 600 °C. Furthermore, the Cs₂ReX₆ materials show panchromatic light absorption covering the entire visible region hinting high photon harnessing ability and with their valence band positions are below the water oxidation potential, they are suitable for photoanodes for solar water oxidation. Cs₂ReX₆ materials displayed their impressive electrochemical stability in a pH 11 solution, with no oxidation currents even at high potentials (up to 1 V vs. Ag/AgCl). The materials consistently yield photocurrent densities of 0.15–0.20 mA cm⁻² at 0.4 V vs. Ag/AgCl under AM1.5G illumination, remaining stable under both interrupted and continuous illumination. Also, Re-Br perovskite, due to its higher and suable flat-band potential and lower charge transfer resistance generated better photocurrent than Re-I perovskite.[75]

Recently, Liu et.al., (2023) studies the Cu-Ag based perovskite inspired material photoelectrodes for PEC water splitting. The synthesized Cu_1.4Ag_0.6BiI₅ nanocrystals exhibited a direct band gap of 2.19 eV and band structures and well aligned with the redox potentials of water splitting reaction. To demonstrate the PEC from the material the photoanode is fabricated without an protective layer, and impressively the photoanode showed –0.05 V_RHE (RHE = reversible hydrogen electrode) onset potential and a photocurrent density of 4.62 mA cm^–2 at 1.23 VRHE, as well as an applied bias photon-to-current efficiency (ABPE) of 2.94%. The charge recombination dynamics and transient absorption studies reveal the n-type nature of the material. Furthermore, the material displayed good stability in both solvent and water, which is attributed to the hydrophobic capping ligands involved during the synthesis process. The NCs-based photoanode for photoelectrochemical (PEC) water splitting exhibits a T50% lifetime of ~310 min, which is one of the longest reported lifetimes for lead-free perovskite-inspired materials.[72]

6. Enhancing Photocatalytic Performance of Pb-Free HPs

Improving the photocatalytic performance and/or photocurrent of the electrode is crucial for ensuring sustainability of the lead-perovskite materials for hydrogen generation. The photocatalytic activity is highly dependent on 1. Ability to generate electron-hole pairs by harnessing maximum solar energy 2. Effective separation of the generated electron-hole pairs. 3. Supporting redox reactions (on the surface in case of photocatalysis or photoelectrode immersed in the reaction solution. Approaches including the 1. band gap engineering by adapting suitable composition, 2. Forming heterojunctions are essentially employed for improving the HER from the system.

Compositional engineering, on the other hand, is a promising tool to design stable perovskites. Careful choice of larger cation has a significant effect on the moisture induced degradation of the perovskite. Molecules like DMA and PEA exhibit strong hydrophobicity, which protects the refrain the penetration of the water molecules into the perovskite lattice/structure. Unlike the usual Sn-perovskites, DMASnI₃ exhibited exceptional stability. The material remained unaffected even after continuous exposure to the water for several hours.

6.1. Band Gap Tuning by Compositional Engineering

While the optical band gap, absorption co-efficient and the nature of the band gap are responsible for absorbing the photons and generating the charge carriers, tuning these inherent properties in the (lead-) halide perovskites is widely explored by developing mixed halide perovskite compositions and doping metal ions. In general, adjusting the band gaps by varying the halide composition of the perovskite is the most explored method and interestingly mixed halide perovskites have superior optoelectronic properties- while bromide and chloride-based perovskites have long charge-carrier lifetimes, exceeding milisecond timescale, while iodide-based HPs have short charge-carrier lifetimes, not exceeding ns timescale however the absorption range is in the inverse order. Therefore, designing a mixed halide perovskites to balance out the properties based on the application is essential. For example, in lead HPs, creating funneling band gap by varying the halide composition, creates a favorable energy band gradient for the effective charge transfer across the perovskite material.[76] Similarly, attempts are made to adjust the band gap energy of CsPbX₃ by anion exchange method resulting in the interfacial charge transfer and lifetime between the perovskite-TiO₂ heterojunction. In addition to the band gap tuning the studies revealed that chloride substitution reduces the trapping states, while iodide substitution leads to slower charge-carrier relaxation and indirect excitons with longer lifetime at the HPs/TiO₂ interface.[77] This approach of tuning the band gap and/or introducing the funneling band gap to lead-free perovskites to improve the photocatalytic activity has been studied by few researchers.

Tang et.al (2022) created a bandgap funneling effect for charge carriers within the bismuth-based double halide perovskites MA₃Bi₂Cl_9–xI_x by synthesizing the perovskite material via facile solvothermal anion-exchange technique. The synthesized double halide perovskite exhibited cubic crystal structure with lattice expansion, attributed to the I^- substitution. Furthermore, the EDS mapping showed the gradient distribution of I^- from the surface to the interior of the perovskites, indicating a bandgap funnel structure. Figure 7a shows the schematic representation of band gap funneling effect in the MA₃Bi₂Cl_9–xI_x. The charge carried dynamics suggest a lower recombination rate and longer carrier lifetime. These eventually lead to higher photocurrent in mixed halide perovskites as compared to that of MA₃Bi₂Cl₉. The photocatalytic performance of the perovskite samples was assessed for H₂ production in a saturated HCl/HI solution (with H₃PO₂ as a reducing agent) under visible light irradiation (λ ≥ 420 nm, 100 mW cm⁻²), MA₃Bi₂Cl_8.8I_0.2 perovskite demonstrated highly efficient H₂ evolution, yielding a rate of ≈214 µmol h⁻¹g⁻¹. To further enhance H₂ generation and inhibit the reverse reaction, 3.0 wt.% Pt cocatalyst was incorporated and the system exhibited enhanced photocatalytic activity, achieving a hydrogen evolution rate of ≈341 ± 61.7 µmol h⁻¹g⁻¹.[78]

Chen et al. (2020) developed a novel lead-free perovskite material, Cs₃Bi_2xSb_2–2xI₉ (CBSI-x), with potential applications as a photocatalyst for generating hydrogen from aqueous HI solutions. The researchers investigated the effects of Sb doping on the crystal structure, electronic structure, optical absorption, charge transfer, and defect properties of Cs₃Bi_2xSb_2–2xI₉ and the photocatalytic activity of the synthesized perovskite is compared to the conventional MAPI’s performance. The series of perovskites with variable Sb- dopant concentrations were synthesized through a co-precipitation method. The perovskite exhibited a hexagonal structure with space group P63/mmc with the peaks shifting to higher 2θ values with the increase of Sb in the perovskite. Figure 7b shows the absorption spectra of the Cs₃Bi_2xSb_2–2xI₉ doped with variable concentration of Sb. While the doping of Sb in Cs₃Bi_2xSb_2–2xI₉ effectively reduces the contribution of Bi³⁺ on the conduction band, this resulted in smaller bandgaps of the Bi-SB perovskites with stronger optical absorption than individual metal perovskites, particularly CBSI-0.3 posed smallest bandgap (≈1.63 eV) and the smallest absorption tail. Moreover, the incorporation of Sb effectively reduced mid gap states, thereby leading to reduced non-radiative defects and improving charge transfer and separation mechanisms. Consequently, this material exhibited superior photocatalytic activity with enhanced photocurrent (when compared to the conventional lead-based perovskite).[79]

Yin et.al. (2021), synthesized a series of stable and multifunctional Pt doped Sn-based perovskites Cs₂Pt_xSn_1-xCl₆ (0<x<1) via a simple hydrothermal method. The perovskites demonstrated the switchable function for optoelectronics and photocatalysis applications. Figure 7c(i) and (ii) shows the Crystal structure and transformation relationship of Cs₂Pt_xSn_1−xCl₆ and its charge-carrier dynamics model for x= 0.25 and 0.75. The perovskites have a face-centered cubic crystal structure with Fm3¯m space group and form a series of continuous Cs₂Pt_xSn_1-xCl₆ solid solutions with isolated [SnCl₆]^2- and [PtCl₆]^2- octahedra. The study reveals that the Pt substitution not only enhances the thermodynamic stability and the thermal stability of Cs₂Pt_xSn_1-xCl₆ but also positively impacts the charge-carrier dynamics and the radiative recombination process of Cs₂Pt_xSn_1-xCl₆. With the Pt incorporation, the electron density surrounding chlorine improved the Pt-Cl bonding improving the stability of the crystal lattice. Additionally, the sub-band gap states control the photoluminescence and photocatalytic functions of Cs₂Pt_xSn_1-xCl₆. The samples with low Pt content have higher sub-band gap state-density and longer charge-carrier lifetime than those with high Pt content. In other words, as the Pt content increases, efficient self-trapping occurs, resulting in an enhanced radiative transition process. The substitution of Pt proves to be an effective method for adjusting the radiative recombination process and extended lifetime of photogenerated charge carriers. This has a direct correlation with the ultimate improvement of photocatalytic activity. The photocatalytic hydrogen evolution activity of Cs₂Pt_xSn_1-xCl₆ with triethanolamine (TEOA) as a sacrificial reagent under simulated solar light illumination. The hydrogen production rate followed an opposite trend of the photoluminescence intensity and quantum yield, indicating that different types of photogenerated charge carriers were involved in the two processes. This concludes that the hydrogen evolution activity to the sub-band gap states in Cs₂Pt_xSn_1-xCl₆, which could capture or respond to the photoexcited carriers and facilitate the electron transfer process. The highest rate was 16.11 mmol g^-1h^-1 for Cs₂Pt_0.05Sn_0.95Cl₆.[80]

6.2. Pb-Free HP Heterojunctions

Effective charge separation by suppressing the recombination losses is one of the key factors to improve the photocatalytic (photoelectrochemical) activity of the hydrogen generation/production system. The electron-hole recombination is inevitable phenomenon due to the Coulombic force between oppositely polar charges, moreover this swift process, with time limited to nanoseconds, is challenging to regulate. However, this can be minimized by engineering the photocatalyst and the optimizing the system. Introducing heterojunctions is a classic and promising strategy to address the (surface) recombination in photocatalysts and is well explored in the traditional solar-driven hydrogen-generation systems. Band alignment plays a crucial role in developing an appropriate heterojunction that ensures rapid charge separation, thereby improving the overall photocatalytic performance of the system. Lead-perovskite- heterojunctions with TiO₂, rGO, g-C₃N₄. etc.., have been investigated and employed for photocatalytic activities including hydrogen generation, CO reduction and organic contaminants degradation.[81,82,83,84,85,86,87] These studies have inspired researchers to adapt lead-free perovskite-based heterojunction architectures for hydrogen generation.

6.2.1. Semiconductor/Pb-Free Heterojunctions

Table 3 summarizes the semiconductor/Pb-free HP heterojunctions employed for the H₂ generation. Jayaraman et.al (2021) investigated heterojunction of hydrothermally synthesized vacancy-order Pt-perovskites A₂PtI₆ (A= Cs⁺, Rb⁺, or K⁺) with BiVO₄ for photoelectrochemical oxidation of water. While synthesized Pt-perovskites exhibited an impressive stability for about 2 weeks under ambient conditions, Cs₂PtI₆ showed higher thermal stability with decomposition temperature at ~364 ^oC. Furthermore, Cs₂PtI₆ exhibited remarkable stability in both acidic and alkaline environments, while Rb₂PtI₆ and K₂PtI₆ demonstrated instability. The heterojunction between BiVO₄ and Cs₂PtI₆ led to enhanced photocurrent in comparison to BiVO₄ alone, primarily attributed to improved charge separation and enhanced light absorption. Figure 8(a) shows the photocurrent response of BiVO₄ and BiVO₄/Cs₂PtI₆ heterojunction (Inset- Band alignment of the photoanode as a function of the wavelength). Furthermore, an IrO_x cocatalyst further boosted the photocurrent, achieving a level of 2 mA cm⁻² at 1.23 V (vs RHE). However, the study is confined to the oxidation of water and can be potentially exploited for hydrogen generation for either acidic or basic reaction solutions.[88]

Tang et.al (2020) demonstrated a novel perovskite/perovskite heterojunction catalyst by establishing interface between methylammonium bismuth iodide (MA₃Bi₂I₉) and tri(dimethylammonium) hexa-iodobismuthate (DMA₃BiI₆) via a facile solvent engineering technique; a series of heterojunction perovskites (BBP-0, BBP-1, BBO-5 and BBP10) by reacting MAI, HI and Bi(NO₃) in IPA and variable amounts of DMF. The synthesis process and the formation of the MA₃Bi₂I₉ and DMA₃BiI₆ is schematically represented as shown in Figure 8(b) The VBM and CBM positions of MA₃Bi₂I₉ and DMA₃BiI₆ showed a well-matched type-II heterostructure, which can facilitate the interfacial charge transfer between the two phases. It is suggested that the photoexcitation results in electron transportation from the conduction band of DMA₃BiI₆ to that of MA₃Bi₂I₉, while the holes on the valence bands migrate in the opposite direction. Thus, this perovskite-perovskite heterojunction enhances the charge separation and thereby improves the photocatalytic active. The synthesized photocatalysts are employed for the H₂ evolution from aqueous HI solution and identified that BBP-5 exhibited superior performance- BBP-5’s H₂ production rate is 198.4 µmol h⁻¹g⁻¹, which is 15- and 8-fold increases relative to BBP-0 and BBP-10, respectively.[89]

Jiang et.al (2021) attempted to synthesize Cs₂AgBiBr₆ supported on nitrogen-doped carbon (N-C) for efficient photocatalytic hydrogen evolution from aqueous HBr solution under visible light irradiation. Conventional one-pot synthesis is employed for the synthesis of the Cs₂AgBiBr₆/N-C) composited, schematic representation of the synthesis process is shown in Figure 8c(i). The crystal structure of the perovskite is determined to be cubic and is not effected by the N-C interaction. The band alignment of the various heterojunction photocatalysts are shown in Figure 8c(ii)- the CBM of Cs₂AgBiBr₆ is higher than the Fermi level of N-C, which enables photoexcited electron transfer from Cs₂AgBiBr₆ to N-C for subsequent H₂ generation. The energy level difference is the largest for Cs₂AgBiBr₆/N-C-140, resulting in the strongest driving force of charge transfer. Furthermore, BET analysis reveals that N-C-140 has the highest surface area and nitrogen content, which facilitate electron diffusion and interfacial charge separation implicating that Cs₂AgBiBr₆/N-C-140 is expected to outperform. The photocatalytic activity of the photocatalyst is measured by the HER reaction from aqueous HBr solution. Cs₂AgBiBr₆/N-C-140 achieves a high hydrogen evolution rate of 380 μmol g⁻¹ h⁻¹, which is about 19 times faster than that of pure Cs₂AgBiBr₆ and 4 times faster than that of a physical mixture of Cs₂AgBiBr₆ and N-C-140. And the material exhibited a decent stability for 6 cycles with 3 hours of hydrogen reaction per cycle.[90]

Lin et.al (2023) studied a new 0-D perovskite halide, Cs₃Rh₂I₉ (with dimer unit [Rh₂I₉]³⁻ separated by Cs ions) functionalized by nitrogen-doped carbon (NC) for hydrogen evolution for the chlorine-alkali electrolyte. The Rb perovskite is prepared via solid state reaction using CsI as the flux followed by dissolution-precipitation method to form Cs₃Rh₂I₉ nanoclusters on a polar nitrogen-doped carbon (NC) as shown in Figure 8d(i). Cs₃Rh₂I₉, due to their 0-D, experience downsizing to smaller clusters creating active sites of the NC due to repeated dissolution and precipitation process in aqueous DMF solution. The perovskite material undergoes self-electrochemical self-reduction and a bottom-up transformation, resulting in the formation of distinctive Rh nanoparticles. The fully reconstructed Cs₃Rh₂I₉/NC-R catalyst significantly lowers the barrier for water dissociation in alkaline HER. The process is schematically represented in Figure 8d ii and iii. Consequently, Cs₃Rh₂I₉/NC-R exhibits an impressive mass activity of 772.1 mA mg⁻¹Rh in a chlor-alkali electrolyte. This figure is approximately 2.5 times greater than that of liquid-reduced Rh/NC catalysts with similar particle sizes and 35.5 times higher than electrochemically reduced Cs₃Rh₂I₉-R catalysts with larger particle sizes.[91]

6.2.2. g-C₃N₄/Pb-Free HP Heterojunction

Carbon based semiconductors like reduced graphene or g-C₃N₄ have played a promising role in photocatalytic activities. Developing g-C₃N₄ based composite/ heterojunction photocatalysts by anchoring the halide perovskite onto the g-C₃N₄ have an advantage of energy band alignment at the interfaces for enhancing the charge separation. Also, with surface functionalization of the g-C₃N₄ with amino or carboxyl groups, stronger chemical bond forms with the perovskite material rather than a mere physical adsorption, which further improves the charge injection at the interface. Moreover, g-C₃N₄ being a synthetic polymer, enables surface passivation of the perovskites and partially addresses the stability issues. In recent years, several researchers have reported several g-C₃N₄/Pb-free HP heterojunctions for hydrogen generation as listed in Table 4.

Researchers have exploited the stability of 2-D perovskites to develop superior photocatalyst for hydrogen generation. However, 2-D perovskites are known for high band gap and are to couple with the suitable co-catalysts to facilitate the HER and in this case g-C₃N₄ based heterojunction is used to broaden the absorption spectra as well as support the HER. Romani et.al (2020) demonstrated PEA₂SnBr₄, a water-resistant 2D Pb-free HP, coupled with g-C₃N₄ co-catalyst system is employed for hydrogen photogeneration and organic dye degradation under visible light. The synthesized material retains its crystal structure and 2.67 eV band gap even in contact with water for 4 hours. The band alignment of the PES₂SnBr₄/ g-C₃N₄ heterojunction is shown in Figure 10a. The hydrogen evolution rates (HERs) showed that the PEA₂SnBr₄/g-C₃N₄ composites had a remarkable enhancement of photocatalytic hydrogen production compared to pure g-C₃N₄ or PEA₂SnBr₄, with an optimal loading of 5 wt% perovskite. The composites also showed high HERs with glucose as a sacrificial agent. PEA₂SnBr₄/g-C₃N₄ composites exhibit synergistic effects, achieving high hydrogen evolution rates (1600 mmol g^-1 h^-1). This novel, water-resistant perovskite opens new possibilities for catalytic applications.[69] The group (2023) also introduced 2-D Ge- perovskites PhBz₂GeBr₄ (with band gap 3.64 eV) and PhBz₂GeI₄ (with band gap 3.32 eV) for the H₂ production from aqueous triethanolamine (TEOA). The HER showed synergistic enhancement with heterojunction and Pt co-catalyst when compared to the pure perovskites. For both PhBz₂GeBr₄ and PhBz₂GeI₄, the optimal perovskite loading was found to be 2.5 wt%, resulting in impressive hydrogen evolution rates of 81 mmol g^-1 h^-1 and 1,200 mmol g⁻¹ h⁻¹, respectively. The iodine-based perovskite exhibited superior performance owing to its lower band and better band alignment with g-C₃N₄ which lead to effective charge transfer and exhibited stability upto 4 cycles.[70]

Bresolina et.al (2020) synthesized Cs₃Bi₂I₉/ g-C₃N₄ binary photocatalyst for hydrogen generation as well as degradation of organic compounds. A simple processing by ultrasonication of Cs₃Bi₂I₉ and g- C₃N₄ nanosheets resulted in nitrogen-iodine chemical interactions between the two semiconductors The estimated band gaps of g- C₃N₄ and Cs₃Bi₂I₉ were about 2.68 eV and 1.88 eV, respectively. The band alignment of Cs₃Bi₂I₉/ g- C₃N₄ is shown in Figure 9b. The pristine g-C₃N₄ sample exhibits solar-light photocatalytic activity through the hydrogen evolution reaction (HER) at a rate of 496.85 µmol g^-1 h^-1, attributed to the favorable bandgap and distinctive electronic structure of g-C₃N₄. Upon integration with metal halide perovskite, the photocatalytic performance of the heterojunction demonstrates an enhancement of approximately 46%, yielding a HER rate of approximately 920.76 µmol g^-1 h^-1.[92]

Wang et.al (2020) combined Cs₂AgBiBr₆ (CABB) with rGO by synthesizing the perovskite solid-state reaction followed by photo-reduction. The SEM image of the synthesized photocatalyst is shown in Figure 9c(i)-inset. The schematic representation of the mechanism involved in the photocatalytic H₂ evolution from aqueous HBr via CABB/2.5%RGO is shown in Figure 9c(i). In the saturated solution, CABB perovskite on the surface of RGO, undergoes dynamic precipitation-dissolution equilibrium process and RGO could act as an attachment for CABB particles. Under illumination, CABB generates charge carries and the generated electrons are transferred to conductive RGO through the M–O–C bonds and reduce H⁺ to produce H₂ at the active sites of rGO. Meanwhile Br⁻ was oxidized to Br³⁻ by the holes on CABB particles. For CABB/2.5% rGO, the H₂ evolution was observed to be 80 times greater than that achieved with pure CABB when both were immersed in a saturated HBr:H₃PO₂ solution uninterruptedly for 120-hour.[93]

Romani et.al (2020) introduced water stable DMASnBr₃ as a co-catalyst for g-C₃N₄ for hydrogen evolution from 10% v/v aqueous triethanolamine (TEOA) reaction medium with the glucose sacrificial agent. Figure 9d shows the band alignment of the synthesized photocatalyst and the crystal structures of the DMASnBr₃ and g-C₃N₄. Owing to the hydrophobic nature of the dimethylammonium cation, heterojunction showed an impressive stability of 4 hour when dispersed in water. The composite showed an impressive HER of >1700 μmol g^-1 h^-1 in 10% (v/v) aqueous triethanolamine (TEOA) solution without any metal co-catalyst, under simulated solar light irradiation. This is a nearly 10-fold improvement compared to pure g-C₃N₄ and a 100-fold improvement compared to pure DMASnBr₃. The composite also showed a high HER of 300 μmol g^-1 h^-1 in 0.1 M aqueous glucose solution, with 3 wt% Pt as co-catalyst.[68] The same group adapted a similar scheme for the synthesis of stable Bi- based perovskite and fabrication of its heterojunctions with g-C₃H₄ and demonstrated their application in photocatalytic hydrogen generation. [70,93,94] Medina-Llamas et.al (2023) employed simple and scalable methods such as wet ball milling and thermal exfoliation to produce nanocrystalline Cs₃Bi₂Br₉ and g-C₃N₄ nanosheets, respectively. The composites showed significant improvement of H₂ production compared to the pure components. The highest HER is achieved at low perovskite loading (0.02 wt.%), while higher loadings result in a detrimental effect due to self-trapping phenomena in Cs₃Bi₂Br₉. Also, this work highlights the importance of the use of photocatalysts in the nanocrystals and nanosheets, which enhanced the photocatalytic by improving the surface area.[94]

Song et al., (2022) employed an in situ self-assembly method to yield CABB/g-C₃N₄ composite, which resulted in a type II heterojunction structure. The band alignment of the synthesized composited is shown in Figure 9e. The optimal CABB/g-C₃N₄ composite exhibited an impressive hydrogen evolution rate of 60 μmol g⁻¹ h⁻¹, a notable 2.5-fold increase compared to pristine Cs₂AgBiBr₆.[95]

Figure 9. g-C₃N₄/Pb-free HP Heterojunctions: (a) Band alignment of the PEA₂SnBr₄ and g-C₃H₄ relative to NHE potential. Reproduced with permission: Copyright 2020, Royal Society of Chemistry.[69] (b) Band alignment of the Cs₃Bi₂I₉ and g-C₃H₄ relative to NHE potential. Reproduced with permission: Copyright 2020, Elsevier B.V.[92] (c) i. Schematic representation of H₂ evolution mechanism by Cs₂AgBiBr₆-rGO under visible light irradiation (inset: SEM image of the photocatalyst). Reproduced with permission: Copyright 2019, Elsevier B.V.[93] (d) Band alignment of g-C₃N₄ and DMASnBr₃ aligned with respect to water, H⁺/H₂ and TEOA/TEOA⁺ redox level. Reproduced with permission: Copyright 2020, Wiley-VCH.[68] (e) Band alignment of the Cs₂AgBiBr₆ and g-C₃H₄ relative to NHE potential. Reproduced with permission: Copyright 2020, Springer Nature.[95]^.

Figure 9. g-C₃N₄/Pb-free HP Heterojunctions: (a) Band alignment of the PEA₂SnBr₄ and g-C₃H₄ relative to NHE potential. Reproduced with permission: Copyright 2020, Royal Society of Chemistry.[69] (b) Band alignment of the Cs₃Bi₂I₉ and g-C₃H₄ relative to NHE potential. Reproduced with permission: Copyright 2020, Elsevier B.V.[92] (c) i. Schematic representation of H₂ evolution mechanism by Cs₂AgBiBr₆-rGO under visible light irradiation (inset: SEM image of the photocatalyst). Reproduced with permission: Copyright 2019, Elsevier B.V.[93] (d) Band alignment of g-C₃N₄ and DMASnBr₃ aligned with respect to water, H⁺/H₂ and TEOA/TEOA⁺ redox level. Reproduced with permission: Copyright 2020, Wiley-VCH.[68] (e) Band alignment of the Cs₂AgBiBr₆ and g-C₃H₄ relative to NHE potential. Reproduced with permission: Copyright 2020, Springer Nature.[95]^.

7. Prospectives and Conclusions

Lead-free perovskites turned out to be attractive active materials for solar-driven hydrogen generation owing to their excellent and tunable optoelectrical properties, ease of synthesis and thin-film fabrication. In this review, the status of the lead-free perovskite for hydrogen generation using photocatalytic and photoelectrochemical systems and state-of-art strategies to improve the performance of the active material in-terms of increasing the HER are discussed. Based on the literature available we take the liberty to conclude and comment on the following observations:

(1).

Dimensionality and bandgap: In general, most of the non-lead metals (excluding Sn and Ge) tend to crystallize in lower-dimensional perovskite structures. These 0-D and 2-D perovskites inherently exhibit higher band gaps, making them suitable for water splitting applications.

(2).

Stability: Unlike conventional lead perovskite with MA cation, lead-free perovskites with Cs cation are structurally stable and allow crystallization of materials into its low dimensional perovskite phases. All inorganic 0-D Bi (/Sb) perovskite and vacancy order Sn, Ag-Bi etc.., exhibited excellent stability in water medium for several hours, proving their potential application in photocatalytic systems. Alternatively, polymer encapsulation, hydrophobic ligand assisted nanoparticle stabilization and core-shell perovskites also enhance stability and can be exploited for photocatalytic water splitting.

(3).

Co-catalysts: Loading Pt co-catalyst to improve the HER has become trivial, however this modification in the system improved hydrogen evolution drastically. Considering the total system cost, it is essential to explore alternatives to Pt. While halide-perovskite/Pt based photocatalytic systems are extensively studied, serval other metal (Ni, Cu, Mn) and oxide (-perovskite) co-catalysts in conjunction with lead-free perovskites can be explored.

(4).

Lead-free perovskites are usually employed in photocatalytic systems rather than in photoelectrochemical water splitting. Most possible reason would be the challenges in formation of the uniform thin films for the fabrication of photoanode due to their low dimensionality.

(5).

Owing to their wide gap, several lead-free perovskite compositions can be excellent choices for coupling with Si cells to develop tandem photoanodes for photoelectrochemical water splitting.

Overall, lead-free halide perovskites water splitting systems need further developments to be considered effectively as sustainable technology for hydrogen generation. With the ongoing dynamic and persistent research on the lead-free perovskite material study, innovative approaches to address the stability and improve the performance, these materials have a great scope in water splitting or solar driven energy conversion applications altogether.