A Critical Review of the Use of Bismuth Halide Perovskites for CO₂ Photoreduction: Stability Challenges and Strategies Implemented

Abstract

Inspired by natural photosynthesis, the photocatalytic CO₂ reduction reaction (CO2RR) stands as a viable strategy for the production of solar fuels to mitigate the high dependence on highly polluting fossil fuels, as well as to decrease the CO₂ concentration in the atmosphere. The design of photocatalytic materials is crucial to ensure high efficiency of the CO2RR process. So far, perovskite materials have shown high efficiency and selectivity in CO2RR to generate different solar fuels. Particularly, bismuth halide perovskites have gained much attention due to their higher absorption coefficients, their more efficient charge transfer (compared to oxide perovskites), and their required thermodynamic potential for CO2RR. Moreover, these materials represent a promising alternative to the highly polluting lead halide perovskites. However, despite all the remarkable advantages of bismuth halide perovskites, their use has been limited, owing to instability concerns. As a consequence, recent reports have offered solutions to obtain structures highly stable against oxygen, water, and light, promoting the formation of solar fuels with promising efficiency for CO2RR. Thus, this review analyzes the current state of the art in this field, particularly studies about stability strategies from intrinsic and extrinsic standpoints. Lastly, we discuss the challenges and opportunities in designing stable bismuth halide perovskites, which open new opportunities for scaling up the CO2RR.

1. Introduction

The continuous increase in the concentration of CO₂ in the atmosphere is driven by the high global electricity demand, estimated at approximately 1400 Terawatt-hours per year [1]. In 2021, the CO₂ emissions derived from the energy sector were 36.3 Gt, the highest annual quantity ever registered, caused by the sudden rebound from the global pandemic lockdown in 2020 [2]. Therefore, the pollution related to the increased CO₂ concentration in the atmosphere continues to be a matter of concern that requires action and global decision making.

Several solutions have been proposed to mitigate the environmental impact of the high CO₂ concentration in the atmosphere. The efforts have been diverse, and there are certain approaches that stand out. For instance, the promulgation of environmental regulations focuses on carrying out preventive actions to avoid the emission of pollutants into the atmosphere. Another option to reduce CO₂ emissions is to improve the energy efficiency of the already-established fossil fuels, since this would result in a better use of these energetic resources [3,4]. Likewise, it is possible to use renewable fuels whose combustion reaction generates products with a lower environmental impact, e.g., natural gas, methanol, biogas, or alternative energy sources that do not involve the use of fossil fuels. However, these fuels are commonly obtained through the use of nonrenewable and highly polluting fossil fuels. Alternatively, some technologies allow the use of CO₂ as a raw material in different processes in order to synthesize many industrial utility products (commodities) [5,6,7,8,9]. One of these processes is the production of short-chain hydrocarbons (solar fuels) through heterogeneous photocatalysis from CO₂ reduction [10,11,12,13,14,15]. This is one of the most promising technologies to help mitigate CO₂ pollution, since it can make use of sunlight to reduce CO₂, employing semiconductors and water. In this process, it is possible to capture and transform CO₂ into nonharmful products that, in some cases, can serve different purposes. However, the implementation of this technology at higher scales has been limited by the low efficiency of the photocatalysts. Nonetheless, materials with perovskite structure have been reported to exhibit an outstanding efficiency in CO2RR. Thus, in this work, we discuss recent advancements, challenges, and improvement strategies using a family of compounds which have been little-explored despite their interesting properties: bismuth halide perovskites (A₃Bi₂X₉; ABi₃X₁₀, etc.; where A = monovalent cations and X = halogens). This review aims to propose innovative strategies to design promising materials for photocatalytic CO2RR that could serve in helping close the gap in current efforts to reach global sustainability and reduce the excessive concentration of CO₂ in the atmosphere.

2. Photocatalytic CO₂ Reduction

The photocatalytic CO₂ reduction reaction (CO2RR) is an attractive option due to the simplicity of the process, which requires CO₂, H₂O, and a photocatalyst as raw materials to generate solar fuels. These materials are introduced to a batch or continuous photocatalytic reactor with CO₂, H₂O, and in some cases, additional sacrificial agents. Then, the reactor is irradiated with natural or artificial sunlight to promote the formation of the electron (e⁻) and hole (h⁺) pair that reacts with the adsorbed chemical species (e.g., CO₂) on the catalyst surface, carrying out its reduction to different solar fuels. This process can generate a variety of C1 and C2 compounds with value on the market, such as formic acid (HCOOH), carbon monoxide (CO), formaldehyde (HCHO), methanol (CH₃OH), acetic acid (CH₃COOH), methane (CH₄), ethanol (C₂H₅OH), or ethane (C₂H₆), as shown in Figure 1 [10,15].

Among the aforementioned products, one that requires a lower thermodynamic potential for its formation is CH₄, while the production of CO and HCOOH both demand the lowest number of electrons, which can be easily generated by kinetic energy. However, other factors can modulate the photocatalytic activity of material and the production of a specific solar fuel, among which the following stand out:

• The energy of photoexcitation. The energy and density of photons significantly affect the distribution of photoexcited electrons, resulting in different reaction mechanisms and influencing their C1 and C2 selectivity [16]. It has been shown that strong light reception of the material, higher photon energy, and high light intensity favor the generation of a higher number of electron–hole pairs during the photocatalytic CO₂ reduction.
• Band engineering. Engineering the band structure of photocatalysts allows for the modulation of reduction potentials in the semiconductor materials, which can influence the selectivity of a generated product. For this matter, doping is one of the simplest and most effective ways to modulate the band structure of photocatalysts, along with the use of heterojunctions, which harness the reduction potentials of various photocatalysts [17].
• Intermediate products. A weak interaction of the by-products with the photocatalyst will result in a faster desorption to the reaction medium. In this manner, the selectivity of the desired product could be modified through the nonintervention of by-products [18,19,20]. Previous work demonstrated that weak physical adsorption between CO and C promotes its desorption and improves the selectivity towards CO. On the other hand, a strong chemisorption between CO and N-containing groups on carbon can promote the photocatalytic CO₂ reduction to CH₄ [18].
• The separation of photogenerated charges. The density of photogenerated electrons on the surface of photocatalysts can influence surface reactions and affect the selectivity of a given generated product. This density mainly depends on the separation efficiency of photogenerated electrons and holes in the photocatalysts. This can be modulated with metal charging, heterojunctions, the adjustment of intrinsic charge carrier mobilities, and the construction of an internal electric field [21,22,23].
• CO₂ affinity. The CO₂ adsorption on the photocatalyst’s surface significantly affects the reaction efficiency and product selectivity. The CO₂ molecule in the gas phase is an inert linear molecule with an ∠OCO angle of 180° and a length of C–O bonds close to 1.1 Å. Upon adsorption on photocatalysts, CO₂ is activated into negatively charged CO₂^δ– species, and the ∠OCO angle is decreased to 140°, while the length of C–O bonds is stretched up to 1.3 Å. CO₂^δ– is a key intermediate for producing carbon-containing products from photocatalytic CO₂ reduction [24]. Thus, CO₂ adsorption is one of the most critical steps during the photocatalytic reaction. Recent works suggest that the type of CO₂ binding on the photocatalyst’s surface influenced the selectivity for the generation of CO, CH₄, CH₃OH, and HCOOH in continuous and batch reactors [25]. Moreover, studies show that the CO₂ affinity of the photocatalysts is promoted by favoring basic sites on their surface [26,27], N-functionalization [28], and oxygen vacancy engineering [25,29]. Likewise, it is possible to improve CO₂ adsorption by modifying the reaction conditions; for instance, lowering reaction temperature and increasing partial pressure [30].
• Surface area. Porous materials are superior candidates for enhancing the CO₂ adsorption on the photocatalysts’ surfaces for their photocatalytic conversion, which is normally achieved by loading active components onto porous supports [31,32,33]. This property can influence the selectivity of the photocatalytic CO₂ reduction to obtain products of higher molecular weight. For example, the selectivity of the reaction for CH₃OH and HCOOH generation can be increased by supporting Cu₂O nanoparticles on porous materials [31]. At the beginning of the reaction, unidentate CO₂ species are adsorbed in bare and supported Cu₂O nanoparticles, which can be converted to the carboxyl radical (COOH•) under irradiation. Then, the carboxyl radical reacts with H⁺ to form HCOOH, which is dehydrated to HCO•. This intermediate is easily adsorbed on the photocatalyst surface, and it could determine the selectivity of the reaction. Lower surface areas promote an easier desorption of HCO•, favoring the formation of HCOH; meanwhile, higher surface areas can maintain this compound hydrogenated until the formation of CH₃OH.

3. Photocatalysts with Outstanding Performances for CO2RR

Scaling CO2RR at pilot and industrial levels requires the design of photocatalyst materials with outstanding efficiencies and stabilities over extended periods. Other requirements may include chemical stability, low toxicity, high CO₂ conversion efficiency, high surface area, cost-effectiveness, accessibility, resistance to photocorrosion, and adequate redox potential to generate different solar fuels [8,10,34].

Inoue et al. were the first research group to report the photocatalytic CO2RR using different photocatalysts (e.g., TiO₂, WO₃, ZnO, CdS, GaP, and SiC) suspended under UV-Visible irradiation [35], generating HCOOH, HCOH, CH₃OH, and CH₄ as reaction products. Some of these photocatalysts are still used for CO₂ reduction; however, to obtain higher efficiencies, it is necessary to implement certain modifications. Some of these variations may include heterojunctions, anion functionalization, or encapsulation using porous materials [36,37,38,39,40].

Titanium dioxide (TiO₂) has been the most widely used photocatalyst since the earliest reports. It has many valuable properties: high photocatalytic activity, stability, inertness to chemical corrosion, and low toxicity. In addition, TiO₂ can perform CO2RR to generate value-added products with good efficiencies, such as formic acid (HCOOH; 2954 μmol g⁻¹ h⁻¹) and formaldehyde (HCHO; 16,537 μmol g⁻¹ h⁻¹), under UV irradiation; however, to achieve these high efficiencies it was necessary to use high pressures (7–20 bar), a high temperature (80 °C), and Na₂SO₃ as a sacrificial agent [41].

Alternatively, other photocatalytic materials have also been utilized for CO2RR with promising efficiencies. Some examples of these photocatalysts are scheelites AMO₄ (A = Ca, Sr, Ba) [42], AV₂O₆ (A = Ca, Sr, Ba) [43], rGO-CuO [44], Ti-MOF [45], Ni/BaTiO₃ [46], Cu/In₂O₃ [47], simple perovskites (ABO₃, A = Li, Na, K, B = Ta, Nb) [48,49], SrTiO₃ [50], LaCoO₃ [51], and recently, halide perovskites ABX₃ (A = formamidinium, FA, methylammonium, MA, or Cs; B = Pb, Bi, or Sn; and X= Cl, Br or I) [52]. The aforementioned materials have stood out because they exhibit remarkable properties for use in photocatalytic CO2RR, such as:

(i)

Ease of synthesis. These materials are distinguished by their low enthalpy of formation, which often grants them facile synthesis processes at normal temperatures (~25 °C) and pressures (~1 atm). Also, it allows for the formation of different morphologies with a high number of active sites [53,54].

(ii)

Structure modulation. It has been demonstrated that perovskites can modulate their crystalline structures and energy bands by incorporating anions or cations from different groups [55,56,57,58]. In addition, the polarization refinement structure can induce an intensive internal electric field, which facilitates charge migration to the surface, improving the efficiency of CO2RR [59]. On the other hand, the structure modulation can be carried out by variation of the morphology (e.g., bulk, nanoplates, nanorods, quantum dots, etc.) [60,61,62].

(iii)

Light harvesting. The modulation of the composition of a halide perovskite can significantly alter its band gap, which can make it an excellent light harvester compared with other semiconductors. At an atomic level, the most common strategies for light-harvesting enhancement include extending the absorption range via band gap engineering, applying the plasmonic resonance enhancement effect, and the arrangement of light absorption using tandem absorber devices [63]. Recently, Jian et al. have proposed the formation of CsPbBr₃ nanocrystals with O-defective WO₃ composites as a solution for stable and highly efficient materials that can harvest the broadest solar spectrum possible [64]. This strategy increases CO2RR efficiency up to 7-fold compared to pristine materials.

(iv)

Exciton generation. The efficient generation of electron–hole pairs under excitation can enhance the photoconversion efficiency in CO2RR and other applications, such as photovoltaics and photon detection [65]. This phenomenon is mainly seen in narrow-band perovskites, which can promote multiple exciton generation (MEG). Perovskite materials can exhibit a variety of exciton species under different excitation conditions, including biexciton and triple exciton [66]. The surface defects of perovskite are the main source of the production of charged excitons. Under weak light excitation, this is mostly generated by single exciton recombination; meanwhile, at medium excitation intensity, biexciton recombination is favored. On the other hand, by using strong light excitation, Auger recombination dominates the generation of charged excitons, and perovskite materials are mainly characterized by their recombination of these excitons. Lin et al. demonstrated that the decay time of single excitons and biexcitons in Mn–CsPbBr₃ perovskites can be modified by applying an external magnetic field (300 mT), suppressing the recombination of the photogenerated charges and subsequently improving the efficiency of CO and CH₄ generation from CO2RR [67].

(v)

Long carrier diffusion lengths. Long carrier lifetime makes halide perovskite materials high-performance photovoltaic materials. Chen et al. proposed that the band-edge carrier lifetime increases when the system transitions from a lower rotational entropy to another phase with higher entropy [68]. These results suggest that the recombination of the photogenerated pair is inhibited by screening, leading to the formation of polarons and thereby extending their lifetime.

(vi)

Band structure. Generally, perovskite materials have adequate redox potentials in their valence and conduction bands for CO2RR. Figure 2 shows the band structures of selected perovskite-structured materials. The bands were calculated by Equations (1) and (2), considering the band gap (Eg) of each material, the absolute electronegativity (X), and the energy of free electrons (E^e) on the hydrogen scale (4.5 eV). In general, halide perovskites exhibited less negative conduction potentials than the simple perovskites; however, they have the thermodynamical potential required to reduce the CO₂ to different solar fuels.

$$E_{VB}=X-E^e+0.5E_g$$

(1)

$$E_{CB}=X-E^e-0.5E_g$$

(2)

4. Metal Halide Perovskites for CO2RR

Perovskite materials exhibit several promising features that have recently brought attention to their field of study. Perovskite structures are isostructural configurations to calcium titanate (CaTiO₃; a mineral commonly found in nature) and other cation variations (e.g., Ta or Nb). The general formula of perovskite materials is ABX₃, where A stands for organic or inorganic cations (e.g., MA⁺, FA⁺, NH⁴⁺, K⁺, Sr²⁺, Ca²⁺, etc.) having larger radii than B (e.g., Ti⁴⁺, Bi³⁺, Pb²⁺, Ta⁴⁺, Nb⁴⁺, etc.) and X are anions, either oxides (O₂⁻) or halogens (I⁻, Br⁻, Cl⁻, or F⁻) [69,70,71]. In these structures, the central atom is not in direct contact with its neighboring coordination atoms, commonly giving these compounds the rare property of ferroelectricity [72] and other promising optoelectronic characteristics. Additionally, the use of organic ions on the A- and X-sites of perovskites leads to several geometric and structural degrees of freedom, which allows for detailed studies on the composition–structure–property relations of these materials [73]. Hydrogen bonds commonly assemble amino groups and halide ions in hybrid organic–inorganic perovskite structures; in which case, their ionic bonding facilitates their synthesis at low temperatures [74].

There are several factors that bolster the photovoltaic properties of metal halide perovskites, such as their high symmetry, the periodicity of their crystal structure (commonly referred to as 1D, 2D, or 3D perovskites), the arrangement of their polar organic cations, the ionic energy of their halides’ anions, and the electronic configuration of the B-site cation of their structure [75]. These characteristics also increase their likelihood to obtain high photocatalytic yields [76]. More specifically, lead halide perovskites (LHP) have shown outstanding efficiencies for CO₂ photoreduction [77], owing to their narrow band gaps and adequate band structure suitable for carrying out the photocatalytic CO2RR. In addition, these materials have good charge transport properties [78], which allows charges to be efficiently separated and transferred within LHP band structure [79].

4.1. Lead Halide Perovskites (LHP)

The general formula of LHPs is APbX₃, where A = organic and inorganic cations (e.g., Ca, K, Cs, Na, CH₃NH₃, or CH(NH₂)₂) and X = halides (e.g., I, Br, Cl, or F). Inorganic cations are preferable, as they are more stable [80]. The unit cell of the APbX₃ crystal lattice usually has a cubic geometry in which the Pb²⁺ cation is in the center of the octahedra, coordinating with six X anions to form [BX₆]⁴ ions, while the A-site cation is located at the vertices.

CsPbBr₃ perovskites have a particularly high photocatalytic performance for CO₂ photoreduction [77,81,82]. The highest yield registered to date for this metal halide perovskite was described by Chen et al., who reported the generation of CO/CH₄ with efficiencies up to 1724 μmol g⁻¹ under UV light [81]. Other authors have also reported the activity of this perovskite to produce CO from CO2RR in heterojunctions with graphitic carbon nitride (149 μmol g⁻¹ h⁻¹) [82], Bi₂WO₆ (CO+CH₄; 503 μmol g⁻¹) [77], and MXene (CH₄, 27 μmol g⁻¹ h⁻¹) [83]. Also, the CsPbBr₃ perovskite demonstrated high stability in 10 cycles of consecutive CO2RR evaluation for CO/CH₄ generation [84].

4.2. Lead-Free Halide Perovskites (LFHP)

Despite the promising activity of LHPs, the presence of lead in their structure could generate environmental toxicity problems. Therefore, various efforts have been made to replace this metal without sacrificing its superior performance, leading to the development of lead-free halide perovskites (LFHP). One of the most frequent ones is the alteration of the periodicity in the crystalline arrangement of LHP materials, which is achieved by modifying their layers to optimize their optical properties [85]. For this purpose, different synthesis methods have been developed using various reactants, solvents, and ligands that allow for the obtention of desired optical properties in these materials and a high environmental stability [86]. As a result, many elements have been used to form various perovskite structures.

Different families of chemical elements can form LFHP structures, such as some transition metals (Pd [87], Cd [88]), chalcogenides (Zr, Hf) [89], and La [90]. There is great potential in replacing Pb with group 14 metals; however, Sn²⁺ and Ge²⁺ cations tend to undergo oxidation due to the high-energy 5s orbitals, which promotes a highly unstable structure at ambient exposure [91]. There are also reports on the use of mixed or trivalent cations to form structures of the A₃B₂X₉ or A₂B⁺B³⁺X₆ types. Alternatively, Bi³⁺ and Sb³⁺ ions are promising candidates to replace Pb²⁺ and develop stable, lead-free perovskites. In some cases, these structures exhibit low-dimensional A₃M(III)₂X₉ structures whose properties have not been extensively studied, especially in photocatalytic CO2RR.

The perovskite structures possess the general formula of ABX3, as discussed above. The optical properties of the perovskite materials can be tuned by changing the A, B, or X ions present in the ABX3 structure. The size of A, B, and X should satisfy the Goldschmidt tolerance factor (t), as shown in Equation (3):

$$t=\frac{r_A+r_X}{\sqrt{2}\left(r_B+r_x\right)}$$

(3)

In this equation, variables r_A, r_B, and r_X are the ionic radii of each element in the ABX3 structure [92]. The perovskite materials show good stability when the tolerance factor is equal or close to 1. In the case of LFHPs, bismuth has a similar ionic radius to lead (as well as comparable properties), the tolerance factor rule is satisfied, and the stability of the Bi-based perovskite materials is enhanced when the substitution takes place. Moreover, it was found that Bi-based perovskite materials possess a higher absorption coefficient, which renders them an efficient light-absorbing material for solar cell applications and, recently, for photocatalytic applications [92,93,94,95,96,97]. This is supported by reports on the use of Cs₃Bi₂I₆Br₃ films as solar cells, for which the highest power conversion efficiency (PCE) reported so far was 1.15% [93]. In addition, bismuth halides have been demonstrated to be good candidates for oxidizing a variety of organic compounds, such as thiamazole [94], vanillyl alcohol [95], rhodamine B [96], phenol [97], etc.

5. Bismuth Halide Perovskites

Among the different options for replacing the Pb²⁺ cation, bismuth provides an attractive option due to its nontoxic nature and chemical stability [98]. Its trivalent (3+) oxidation state causes materials to form an A₃Bi₂X₉ structure, where A can be K⁺, Rb⁺, Cs⁺, or methylammonium MA⁺, and X can be I, Br, or Cl. Some of them have shown low toxicity, stability in air, and ease of synthesis [99]. In addition, as shown in Figure 2, this family has the required thermodynamic potentials to reduce CO₂ to different fuels, such as HCOOH, CO, CH₃OH, CH₄, C₂H₅OH, or C₂H₆, which positions bismuth halide perovskites as promising materials to carry out the photocatalytic CO2RR.

5.1. Crystal Structure of the A₃Bi₂X₉ Family

The A₃Bi₂X₉ family of materials crystallize in different systems, depending on the halide anion used. For example, Cs₃Bi₂Cl₉ has a monoclinic unit cell; meanwhile, Cs₃Bi₂Br₉ and Cs₃Bi₂I₉ have hexagonal structures (Figure 3a) [100]. Particularly, the structure of Cs₃Bi₂I₉ is composed of bioctahedral (Bi₂I₉)³⁻ clusters surrounded by Cs cations, which are situated between these clusters in such a way that a nearly ideal hexagonal-type close packing of the cesium and iodine atoms takes place. Each layer in the close packing has the same configuration, with three iodide atoms for each Cs atom.

Figure 3b shows the structure of MA₃Bi₂I₉ for reference purposes. This structure has a hexagonal symmetry (P63/mmc) with face-sharing BiI₆ dioctahedral (Bi₂I₉)³⁻ clusters isolated by MA⁺. One MA⁺ group is surrounded by six metal halide octahedra [101]. Figure 3c shows the unit cell of the Rb₃Bi₂I₉, which contains corrugated layers of corner-connected BiI₆ octahedra [102]. The crystal structure of Rb₃Bi₂I₉ can be described as a distorted defect variant of the perovskite type (P21/n) in which only two-thirds of the octahedral sites are occupied. An ordered cubic close packing of Rb and I also occurs, where the BiI₆ octahedra are coordinated by eight Rb atoms in a distorted cube, the unit cell of the perovskite type. On the other hand, another set of perovskite materials can be formed using K⁺ as the A cation. The K₃Bi₂I₉ perovskite crystallizes in a monoclinic structure with a P21/n space group [103]. This forms a 2D-layered structure with distorted BiI₆ corner-sharing octahedra. However, there are no reports about its use in photocatalytic applications.

5.2. Performance of A₃Bi₂X₉ Family in CO2RR

So far, most of the reports related to CO2RR using this family employ Cs⁺ as the A-site cation. Shen et al. reported the photocatalytic activity of quantum dots (QD) of the Cs₃Bi₂X₉ (X=Cl, Br, and I) family from the perspective of surface engineering. They mainly studied the effect of the type of halogen associated with surface regulation [104]. The results showed that Cs₃Bi₂Br₉ displayed the highest activity for CO production (135 μmol g⁻¹), with 98.7% selectivity after 5 h of artificial solar light; meanwhile, Cs₃Bi₂I₉ exhibited poor activity, possibly due to its intrinsic instability under irradiation (Figure 4a). The highest activity of the Cs₃Bi₂Br₉ perovskite was attributed to a more efficient charge transfer with longer lifetimes, which often grants more free charges for CO2RR, as shown in Figure 4b.

Since band structures influence the reactivity of a photocatalyst, a more negative potential in CB would indicate a higher capacity for photogenerated electrons to reduce CO₂ to different fuels. Figure 4c shows that both Cs₃Bi₂Br₉ and Cs₃Bi₂Cl₉ possess suitable bandgaps and adequate band-edge potentials to carry out the CO2RR to CO. Remarkably, the band-edge positions exhibit a VB upshift with halogen substitution from Cl to Br, while there was little change on CB. These results confirm that the presence of Br in Cs₃Bi₂X₉ promoted an enhanced solar energy utilization and photoreduction capability due to its profitable electronic band structures. Moreover, the density functional theory (DFT) method was used to clarify the internal electron transfer direction change in using Br or Cl in the Cs₃Bi₂X₉ perovskite. As is shown in Figure 4d, Bi atoms are depleted, and X atoms are charge-accumulated, which reflects thicker electron cloud density in the Cs₃Bi₂Br₉ perovskite (mainly in the Br-site). Additionally, the authors demonstrated a higher CO₂ affinity in Cs₃Bi₂Br₉ than in the rest of the samples, according to the b-CO₃²⁻ signals obtained by FTIR. These results confirm that the selection of Br as the anion in the Cs₃Bi₂X₉ structure is the best pathway to design a new and more efficient photocatalyst for CO2RR.

On the other hand, the effect of modifying the A-site with Rb⁺, Cs⁺, and CH₃NH₃⁺ (MA⁺) in the A₃Bi₂I₉ structure was investigated by Bhosale et al. [98]. The origin of the application of bismuth halide perovskites derives from the promising performance of Cs₃Bi₂I₉ and MA₃Bi₂I₉ in solar cells of 1.15% and 3.17%, respectively [93,105]. The A₃Bi₂I₉ (A = Rb, Cs, and MA) perovskites were synthesized by a top-down ultrasonic method that favored the formation of nanoparticles with diameters lower than 20 nm. Interestingly, Cs₃Bi₂I₉ and MA₃Bi₂I₉ perovskites crystallized in hexagonal structures of space group P63/mm; while the Rb₃Bi₂I₉ sample had a monoclinic system with space group P21/n. All the samples exhibited activity reducing the CO₂ to CO and CH₄. However, some differences were detected. For example, the highest efficiency for CH₄ production (17 μmol g⁻¹) was obtained with the Rb₃Bi₂I₉ perovskite; meanwhile, Cs₃Bi₂I₉ exhibited the highest selectivity for CO generation (77 µmol g⁻¹), which is one of the best results obtained in comparison with other perovskites under similar experimental conditions. The differences in the selectivity of both samples were attributed to the formation of monodentate and bridge carbonates in Rb₃Bi₂I₉ and Cs₃Bi₂I₉, respectively.

Electron Paramagnetic Resonance (EPR) demonstrated that in Rb₃Bi₂I₉ and Cs₃Bi₂I₉ samples, holes are efficiently stabilized with Bi⁴⁺ and oxygen ions; nonetheless, for MA₃Bi₂I₉, holes are additionally stabilized by ^•CH₂NH₃⁺. As a result, the transfer of h⁺ into water is less efficient in MA₃Bi₂I₉, resulting in a lower charge transfer efficiency. The crystal structure of the studied samples also influenced the efficiency of CO2RR. Rb₃Bi₂I₉ crystallizes in a distorted defect variant of the perovskite structure in which every third Bi layer within the face [001] is depleted. Therefore, since Bi constitutes the active site of the catalyst, its depletion in Rb₃Bi₂I₉ favored lower CO2RR efficiencies compared to Cs₃Bi₂I₉.

These works demonstrated the potential of the A₃Bi₂X₉ family to act as a photocatalyst in CO2RR, showing efficiencies higher than other traditional and LHP perovskites. Nevertheless, it is well-known that the low stability of these materials at ambient conditions hinders its commercialization and upscaling to the industrial level. For instance, the Cs₃Bi₂X₉ (X = Cl, Br) family showed apparent stability after 5 h of simulated solar light [104]; meanwhile, the A₃Bi₂I₉ (A = Rb, Cs, and MA) structure remained constant after 7 days of exposure to humidity (70%) and UV irradiation (305 nm) [98]. However, there is no evidence regarding the recycling of these materials in consecutives cycles of evaluation. Considering this, several authors have proposed different strategies to increase the stability of the halide perovskites. These are discussed in the next section.

5.3. Stability Strategies for Bismuth Halide Perovskites

Bismuth halide perovskites have emerged as promising candidates for CO2RR. The efficiency of these materials for CO and CH₄ production from CO2RR has increased through the optimization of their perovskite structure and the engineering of suitable synthesis methods. Despite these advancements, the instability of the A₃Bi₂X₉ family to environmental factors (e.g., humidity, light, oxygen, or temperature) is one of the most critical factors limiting their application and, thus, their commercialization. Under different environmental conditions, these materials can undergo morphological and structural changes and lower light absorption, negatively affecting their photocatalytic performance [106,107,108,109]. For example, in the presence of water, the halide perovskites tend to hydrolyze back to their precursors. Specifically, water can penetrate the perovskite lattice to form mono- and di-hydrated perovskite phases in which the A (Cs⁺ and Bi³⁺) cations are no longer strongly bonded to the octahedral; as a result, a phase segregation of AI₂ crystals takes place (Equation (4)). Then, the hydrolysis of AI₂ occurs to promote the formation of bismuth oxyhalides (BiOX, X = I, Br, Cl) (Equation (5)) [108]. Furthermore, water can protonate the excess halide to form an acid (HX), which is highly volatile at room temperature. Therefore, the continuous release of these compounds can drive the decomposition of the perovskite layer, decreasing the efficiency of photocatalytic reactions, such as CO2RR.

$${\mathrm{A}}_3\mathrm{B}{\mathrm{i}}_2{\mathrm{X}}_9\to 3\mathrm{A}\mathrm{X}+2\mathrm{B}\mathrm{i}{\mathrm{X}}_3$$

(4)

$$\mathrm{B}\mathrm{i}{\mathrm{X}}_3+{\mathrm{H}}_2\mathrm{O}\to \mathrm{B}\mathrm{i}\mathrm{O}\mathrm{X}+2\mathrm{H}\mathrm{X}$$

(5)

On the other hand, the combination of oxygen and light can also promote the decomposition of the bismuth halide perovskites through a photo-oxidation process [109]. This process occurs due to the photocatalytic properties of the material itself and starts with the absorption of light with energy equal to or higher than the band gap value of the material, which promotes an electron transfer from the CB to the VB. Then, the photogenerated electrons in the perovskite react with the adsorbed O₂ to form superoxide radicals (O₂^∙-), which can start the decomposition of this material. This process has been previously demonstrated in the MAPbI₃ perovskite [109,110], and its mechanism is schematized in Figure 5. The first step is the oxygen diffusion into the lattice (Figure 5a), then the light absorption in the perovskite leads to the generation of electrons and holes (Figure 5b), which react with the adsorbed oxygen to generate the superoxide radicals (Figure 5c). Finally, this, promotes the degradation of the perovskite to PbI₂, H₂O, I₂, and CH₃NH₂ (Figure 5d) [109].

Considering these scenarios and the degradation mechanisms, several strategies have been proposed to increase halide perovskites’ stability, which are summarized in Figure 6. These strategies can be classified into two different approaches:

(a)

Intrinsic. This approach is widely used to improve the stability of metal halide perovskites through structural modification, which involves the alteration of the composition of cations and anions that are inherent to the perovskite structure [111,112]. The formation of solid solutions and double perovskites are examples of this approach. Likewise, the obtention of low-dimensional-networked perovskites promotes the formation of more stable crystal structures by selecting appropriate manufacturing techniques. Stoichiometry modification then settles the network in a quasi-zero-dimensional (0D) configuration, where the metal halide octahedra are almost isolated [113]. Another successfully intrinsic strategy is passivation [114]. This strategy allows the removal of dangling bonds, which act as recombination centers associated with defects, promoting better stability and optical properties. Antisolvent engineering is another strategy that speeds up nucleation during perovskite synthesis through solvent extraction [115,116,117]. This strategy allows the obtention of dense, high-quality films of bismuth halide perovskites with enhanced stability.

(b)

Extrinsic. This approach is characterized by the modification of the perovskite’s external properties without altering its internal composition. Among these strategies, encapsulation is one of the most used. It consists of the coverage of the perovskite to prevent exposure to moisture, oxygen, UV light, and temperature, thus protecting the structure. The encapsulation can be done by glass–glass covering of the edges of the substrate with an adhesive, as well as the deposition of a polymer coating. The adhesives can be ethylene methacrylate (EMA), ethylene vinyl acetate (EVA), polyisobutylene (PIB), and epoxy resins activated with UV light [118]. Alternatively, encapsulation can be done by depositing polymer coatings, e.g., polyethylene oxide, polyvinyl pyridine. The construction of heterostructures or Z-schemes between perovskites and other semiconductors has contributed to the design of promising stabilities for photocatalytic reactions, since electrons are not available to react with the adsorbed oxygen. This avoids the degradation of the perovskite structure [119,120,121]. Another option is the dispersion of the perovskite nanoparticles in porous support for enhanced electron and hole separation, more accessible active sites, and close contact with the reaction species [54]. At the same time, adding adequate reactants to the medium can promote good stability of the perovskite structure. For example, it has been proved that the addition of HI during the evaluation of MA₃Bi₂I₉ promotes an excellent phase stability and enhanced photocatalytic activity for H₂ evolution [122].

Some of these strategies have been implemented in bismuth halide perovskites to achieve better stabilities and efficiencies in photocatalytic CO2RR, mainly for Cs₃Bi₂X₉ perovskites. The implementation of these strategies for other isostructural materials has not been reported so far. Thus, these applied approaches in Cs₃Bi₂X₉ will be discussed in the following section.

5.4. Implemented Strategies to Increase the Efficiency and Stability of the Cs₃Bi₂X₉ Family in CO2RR

As previously mentioned, the most widely studied materials in the A₃Bi₂X₉ family for photocatalytic CO2RR uses Cs⁺ as the A-site cation. So far, the stability of the Cs₃Bi₂X₉ perovskite has been boosted by implementing two intrinsic (solid solutions and double perovskites) and two extrinsic (the formation of heterojunctions with other semiconductors and the encapsulation of the perovskites in porous supports) approaches.

5.4.1. Intrinsic Approaches to Improve the Stability of Cs₃Bi₂X₉ in CO2RR

(a)

Solid solutions

The formation of solid solutions enables more stable structures than their unmixed counterparts because increasing the availability of configuration states also increases the configurational entropy of the system [123]. A recent report proposed the synthesis of Cs₃Bi₂X₉ (X = Cl, Cl_0.5Br_0.5, Br, Br_0.5I_0.5, I) by solution recrystallization and demonstrated that this set of materials exhibited suitable band gaps and energy band positions to achieve CO production from CO2RR [124]. The Cs₃Bi₂(Br_0.5I_0.5)₉ solution showed an optimal photoreduction performance for CO generation (54 μmol g⁻¹ in 3 h) with 100% selectivity upon visible light irradiation. This efficiency is almost five times higher compared to Cs₃Bi₂I₉, which is attributed to a suitable band gap structure, wide light absorption range, and higher photocurrent. Furthermore, there was no obvious decrease in the CO evolution rate during 10 h of irradiation, which indicated that the solid solution might have moderate stability under these conditions.

(b)

Double perovskites

Research on double perovskites has recently grown in importance due to their higher stability and nontoxicity compared to the widely used lead halide perovskites. Replacing two divalent (Pb²⁺) ions with heterovalent cations (e.g., one trivalent and one monovalent cation) leads to the formation of double perovskites.

In double perovskites (of A₂BB’X₆ formula), two toxic divalent lead ions are replaced by nontoxic monovalent (B) and trivalent metal ions (B’), which neutralize the overall charge in the conventional perovskites. Double perovskites owe their name to the fact that their unit cell is twice the cell of metal halide perovskites [125].

In the search for alternative photocatalytic materials with improved stability for CO2RR, some double perovskite materials have been proposed based on the incorporation of Ag as a monovalent cation and Bi as a trivalent cation [126,127,128,129,130,131]. The efficiency of the double perovskite Cs₂AgBiX₆ (X=I, Cl, Br) has been demonstrated in the generation of CO from CO2RR [126,129]. These works also evidenced that the Cs₂AgBiBr₆ nanoplates showed a more than 8-fold enhancement compared to nanocubes of the same perovskite (255 μmol g⁻¹ vs. 31 μmol g⁻¹). This improvement of photocatalytic performance was related to the anisotropically confined charge carriers and their in-plane long diffusion length for the nanoplates compared to nanocubes.

The stability of Cs₂AgBiBr₆ was recently demonstrated by Zhou et al. [128]. They investigated the stability of this double perovskite under different solvents after 21 days of exposure. The different solvents employed were polar (dimethyl formamide, acetone), protonic (isobutanol), mild polar (ethyl acetate, chloroform), and nonpolar (octane). The results are shown in Figure 7. The perovskite was decomposed by most of the solvents studied due to their strong polarity and the dynamic equilibrium between bound and free ligands, leading to the degradation of Cs₂AgBiBr₆. In contrast, the perovskite showed superior stability in chloroform and octane for more than 3 weeks. Furthermore, the use of ethyl acetate allowed for the stabilization of the perovskite for 5 days. Therefore, these results offered an opportunity for upcoming research on the CO2RR by using double halide perovskites. Under this approach, the production of CO and CH₄ was 14 and 9 μmol g⁻¹, respectively [128,130].

5.4.2. Extrinsic Approaches to Improve the Stability of Cs₃Bi₂X₉ in CO2RR

(a)

Heterojunctions of Cs₃Bi₂X₉ with other semiconductors

The use of heterostructured materials has resulted in outstanding efficiencies in different photocatalytic reactions [132]. Remarkably, their use in photocatalytic CO2RR has allowed for the control of selectivity [133], the enhancement of the charge transfer, and, thus, an increase in the production of different solar fuels [134,135,136]. Considering this, various research groups have recently proposed implementing this extrinsic approach to increase the stability of bismuth halide perovskite.

Liu et al. proposed an interesting heterojunction based on Cs₃Bi₂I₉/Bi₂WO₆ [137]. They suggested that cosharing the Bi atom enabled intimate contact and strong electron coupling between both perovskites, which eventually increased the interfacial charge transfer through the Z-scheme pathway. The band potentials of each semiconductor were obtained from Mott–Schottky plots, which served as the basis for the proposal of the band coupling of Cs₃Bi₂I₉/Bi₂WO₆. The Fermi energy level of the halide perovskite was higher than that of Bi₂WO₆, which suggested that the electrons in Bi₂WO₆ tend to migrate to Cs₃Bi₂I₉ to balance the Fermi energy level, promoting an electron cloud coupling between both materials, as shown in Figure 8a. Additionally, the authors proposed that Bi₂WO₆ favors H₂O oxidation; meanwhile, Cs₃Bi₂I₉ favors CO₂ reduction. This phenomenon accords with the low CO production observed in the Bi₂WO₆ reference, whereas the heterojunction promoted the highest CO generation (66 μmol g⁻¹), which is over four times higher than its reference Cs₃Bi₂I₉ (Figure 8b). On the other hand, the stability of the heterojunction was confirmed after three cycling tests. The results indicated that the Cs₃Bi₂I₉–Bi₂WO₆ heterojunction retains its activity after different evaluations (Figure 8c). The X-ray diffraction results evidenced that the crystalline phases of both materials remained, demonstrating its good stability in the gas–solid reaction system under simulated sunlight irradiation.

Cerium oxide (CeO₂) is another semiconductor that was proposed to form a heterostructure with Cs₃Bi₂I₉. This oxide is known to possess a good CO₂ affinity [138], which is a prerequisite to designing alternative photocatalyst materials to assure high efficiencies in CO2RR. Therefore, Feng et al. synthesized the Cs₃Bi₂I₉–CeO₂ Z-scheme heterojunction by a self-template-oriented method using BiOI–Bi₂O_2.7 nanosheets as a template and Bi precursor [139]. The authors also proposed a charge transfer mechanism similar to the one previously discussed by Liu et al. In this mechanism, the electrons in CeO₂ are transferred to the perovskite to reduce the CO₂ molecule to CO and CH₄. As a result, the total yield of the products (CO/CH₄) reached 238 μmol g⁻¹, and the electron consumption yield was calculated as 877 μmol g⁻¹, which is 7 and 15 times higher than the reference materials, respectively. It is important to highlight that this is the best efficiency obtained so far among the reported bismuth-based perovskite (Table 1). Also, the authors investigated the stability of Cs₃Bi₂I₉–CeO₂ after five consecutive cycles and found that in the first three cycles (12 h), the activity remains; however, the activity partially decreased after the fourth (71%) and fifth cycles (55%), suggesting that the catalyst showed good stability in the first 12 h of reaction.

Another option to design heterojunctions with bismuth halide perovskites is by using sulfurs. Zhang et al. proposed an S-scheme based on the coupling of Cs₃Bi₂Br₉ with In₄SnS₈ [140]. The authors demonstrated that the coupling of both semiconductors boosted the charge separation by forming an internal electric field by in situ irradiated X-ray photoelectron spectroscopy, electron spin resonance, femtosecond time-resolved absorption spectroscopy, and density functional theory calculations. The Cs₃Bi₂Br₉–In₄SnS₈ heterojunction exhibits activity towards CO generation (9.5 μmol g⁻¹ h⁻¹) from CO2RR with high selectivity (92.9%), with results up to 3.8 times higher than the pristine materials. The high selectivity was associated with a decreased energy barrier in the CO₂ reduction to CO through an adsorbed ^⁕COOH intermediate.

In addition, the combination of two approaches has been proposed to increase the stability of the bismuth halide perovskites. Zhang et al. synthesized the double perovskite Cs₂AgBiBr₆ on the surface of MXene nanosheets [131]. Here, Mxene (M_n+1X_nT_x) reduces the exciton binding energy of the double perovskite and simultaneously promotes a more efficient formation of the charge carriers. The combination of these strategies promoted the formation of CO (11.1 μmol g⁻¹), CH₄ (1.3 μmol g⁻¹), and H₂ (8.9 μmol g⁻¹) under visible irradiation. The apparent quantum yield showed values of 0.083%, 0.071%, and 0.005% at 485, 535, and 595 nm, respectively.

(b)

Encapsulation in porous supports

The encapsulation of bismuth halide perovskites is another option to provide stability through their support on porous materials. Previous studies suggest that mesoporous encapsulation enhances the stability of halide perovskites and prevents halide ion exchange [141]. However, it is still unclear whether the mesoporous structure provides stability against moisture, since H₂O molecules can penetrate into the pores to degrade the perovskite. Even so, a recent report demonstrated good stability in the Cs₃Bi₂Br₉ perovskite encapsulated in the porous structure of the MCM-41 molecular sieve, as is shown in Figure 9a. This strategy allows the stable production of CO (17 μmol g⁻¹ h⁻¹) after 8 consecutive cycles (Figure 9b) [142].

Similarly, the implementation of extrinsic and intrinsic strategies allowed for the obtention of the double perovskite Cs₂AgBiBr₆ grown on mesoporous TiO₂ nanoparticles that provide remarkably high stability for CH₄/CO generation, which was recently proposed by Sun et al. [127]. This work reports the best selectivity (88.7%) for CH₄ generation compared to other lead and bismuth halide perovskites. This result was associated with the high CO₂ uptake of the mesoporous support (3.4 cm³ g⁻¹) and the Bi-active sites, which mediate the hydrogenation of CO (^*HCO) under light irradiation. Also, the presence of TiO₂ suggests a possible heterostructure formation with Cs₂AgBiBr₆, enhancing the charge transfer.

Table 1. Summary of the solar fuels obtained from CO2RR by the implementation of different stability strategies in the Cs₃Bi₂X₉ perovskites.

Material	Morphology	Product (μmol g−1)	Irradiation	Reaction Medium	Stability	Ref.
Cs3Bi2X9 (X = Cl, Br, I)	Quantum dots	CO (135)	AM 1.5G	H2Ovapor	5 h	[104]
A3Bi2I9 (A= Rb+, Cs+ or CH3NH3+, MA+)	Nanoparticles	CO (77) CH4 (5)	UV lamp (32 W, 305 nm)	Trichloromethane	12 h	[98]
Cs3Bi2X9 (X = Cl, Cl0.5Br0.5, Br, Br0.5I0.5, I)	Nanocrystals	CO (54)	Xenon lamp (300 W, 420 nm)	H2Ovapor	10 h	[124]
Cs2AgBiX6 (X = Cl, Br, I)	Nanoplatelets	CO (40) CH4 (25)	Laser (405 nm)	Ethyl acetate	6 h	[126]
Cs2AgBiBr6	Cubic	CO (14) CH4 (9)	AM 1.5G (100 W)	Ethyl acetate	21 days	[128]
Cs2AgBiI6	Quasi-spherical	CO (19)	Xenon lamp (300 W, 420 nm)	Toluene	3 h	[129]
Cs2AgBiBr6 /TiO2	Spheres	CO (20) CH4 (151)	Xenon lamp (300 W, 70 mW/cm2)	Isopropyl alcohol	5 h (5 cycles)	[127]
Cs2AgBiBr6 /g-C3N4	Semi-spheres	CO (1.8) CH4 (0.2)	AM 1.5G	Ethyl acetate and methanol	12 h (4 cycles)	[130]
Cs2AgBiBr6 /MXene	Nanocrystals	CO (11) CH4 (1) H2 (9)	Xenon lamp (400 nm)	H2Ovapor	5 h (3 cycles)	[131]
Cs3Bi2I9 /Bi2WO6	Nanosheets	CO (65)	Xenon lamp (300 W, 100 mW/cm2)	H2Ovapor	5 h (3 cycles)	[137]
Cs3Bi2I9 /CeO2	Nanosheets	CO (170) CH4 (65)	Xenon lamp (300 W)	H2Ovapor	12 h (5 cycles)	[139]
Cs3Bi2Br9 /In4SnS8	Quantum dots	CO (60)	Xenon lamp (300 W)	H2Ovapor (80 kPa)	8 h (3 cycles)	[140]
Cs3Bi2Br9 /MCM-41	Nanoparticles	CO (34)	Xenon lamp (300 W, 420 nm)	H2Ovapor	2 h (8 cycles)	[142]

Table 1. Summary of the solar fuels obtained from CO2RR by the implementation of different stability strategies in the Cs₃Bi₂X₉ perovskites.

Material	Morphology	Product (μmol g−1)	Irradiation	Reaction Medium	Stability	Ref.
Cs3Bi2X9 (X = Cl, Br, I)	Quantum dots	CO (135)	AM 1.5G	H2Ovapor	5 h	[104]
A3Bi2I9 (A= Rb+, Cs+ or CH3NH3+, MA+)	Nanoparticles	CO (77) CH4 (5)	UV lamp (32 W, 305 nm)	Trichloromethane	12 h	[98]
Cs3Bi2X9 (X = Cl, Cl0.5Br0.5, Br, Br0.5I0.5, I)	Nanocrystals	CO (54)	Xenon lamp (300 W, 420 nm)	H2Ovapor	10 h	[124]
Cs2AgBiX6 (X = Cl, Br, I)	Nanoplatelets	CO (40) CH4 (25)	Laser (405 nm)	Ethyl acetate	6 h	[126]
Cs2AgBiBr6	Cubic	CO (14) CH4 (9)	AM 1.5G (100 W)	Ethyl acetate	21 days	[128]
Cs2AgBiI6	Quasi-spherical	CO (19)	Xenon lamp (300 W, 420 nm)	Toluene	3 h	[129]
Cs2AgBiBr6 /TiO2	Spheres	CO (20) CH4 (151)	Xenon lamp (300 W, 70 mW/cm2)	Isopropyl alcohol	5 h (5 cycles)	[127]
Cs2AgBiBr6 /g-C3N4	Semi-spheres	CO (1.8) CH4 (0.2)	AM 1.5G	Ethyl acetate and methanol	12 h (4 cycles)	[130]
Cs2AgBiBr6 /MXene	Nanocrystals	CO (11) CH4 (1) H2 (9)	Xenon lamp (400 nm)	H2Ovapor	5 h (3 cycles)	[131]
Cs3Bi2I9 /Bi2WO6	Nanosheets	CO (65)	Xenon lamp (300 W, 100 mW/cm2)	H2Ovapor	5 h (3 cycles)	[137]
Cs3Bi2I9 /CeO2	Nanosheets	CO (170) CH4 (65)	Xenon lamp (300 W)	H2Ovapor	12 h (5 cycles)	[139]
Cs3Bi2Br9 /In4SnS8	Quantum dots	CO (60)	Xenon lamp (300 W)	H2Ovapor (80 kPa)	8 h (3 cycles)	[140]
Cs3Bi2Br9 /MCM-41	Nanoparticles	CO (34)	Xenon lamp (300 W, 420 nm)	H2Ovapor	2 h (8 cycles)	[142]

5.5. General Remarks

Table 1 lists the reports of bismuth halide perovskites for CO2RR and the stability reached within each system. As is shown, the stability of the bismuth halide perovskites in gas–solid reactions varies from 5 h to 21 days. In general, to assure good efficiencies, the literature suggests the addition of organic solvents (e.g., ethyl acetate, trichloromethane, toluene, or isopropyl alcohol). However, this strategy makes it difficult to differentiate the actual capacity of the perovskites to produce solar fuels from CO2RR. Also, the use of double perovskites shows promising results to obtain materials with remarkably high selectivity for products such as CH₄, which can be directly used for the generation of clean and sustainable energy.

So far, the best result for CO generation was obtained with the Cs₃Bi₂I₉–CeO₂ heterostructure (135 μmol g⁻¹), whereas the implementation of three stability strategies such as double perovskites, encapsulation, and heterostructures allowed for the highest efficiency in producing CH₄ (151 μmol g⁻¹). Both efficiencies (with good stability during the evaluation) demonstrated the feasibility of using CO2RR to generate clean and renewable solar fuels.

6. Next Steps to Design and Evaluate Bismuth Halide Perovskites in CO2RR

Thus far, there is a limited number of reports of using bismuth halide perovskites for CO2RR despite their interesting and remarkable properties, ideal for this reaction. This opens a window of opportunity to implement stability strategies to achieve better efficiencies in this process for the larger scale.

According to the literature reviewed, some phenomena remain to be further investigated in this field. First, it seems that the halide anion (X) and A cation selection in the A₃Bi₂X₉ structure influences the photocatalytic performance of the perovskite in CO2RR. Particularly, there have been better results for CO2RR with the use of Br as X and Cs as A cation for CO generation, as well as Rh as A cation for CH₄ production. Nevertheless, other cations of smaller size, such as K, have not been explored in the study of the A₃Bi₂X₉ family.

What is more, there are no reports about the possible formation of a heterostructure employing bismuth oxyhalides (BiOX) and bismuth halide perovskites, mostly because of their interaction with H₂O (equations (4) and (5)). This would be interesting, since the cosharing Bi and X atoms could promote better interaction with the components. This type of passivation on perovskites opens the possibility of their evaluation in batch reactors in the liquid medium to produce liquids products (e.g., HCOOH, HCOH, CH₃OH, CH₃CH₂OH), which offers great advantages related to their intrinsic chemical energy content and the eases of implementation, along with storage and transportation. Furthermore, the encapsulation of bismuth halide perovskites in polymeric films has not been explored.

In addition, it was demonstrated that combining extrinsic and intrinsic approaches is critical to assure good efficiencies and remarkable stabilities for bismuth halide perovskites in CO2RR. Therefore, when scientists begin designing bismuth halide perovskites, it is vital to implement at least one stability strategy to ensure good efficiencies and allow the community to move forward with scaling the CO2RR process.

Another critical factor is the need for a standardization process for reactor design, operational parameters (e.g., temperature, pressure, or flow mass), light source and filters, as well as solvents to assure a fair comparison of the efficiencies obtained, which will eventually promote the scaling of CO2RR at the industrial level to the continuous operation of solar refineries that use the CO₂ emissions, H₂O, and solar light as raw materials. All these critical steps to design new and more efficient bismuth halide perovskites are summarized in Figure 10.

7. Conclusions and Perspectives

This review aims to summarize the stability strategies recently proposed for bismuth halide perovskite to assure high efficiencies in CO2RR. This topic is highly interesting for the scientific community since bismuth halide perovskites have enormous potential to be used in CO2RR to generate clean and renewable solar fuels. However, although significant progress has been achieved in bismuth halide perovskites, there are still many challenges regarding their stability, selectivity, and efficiency for continuous solar fuel generation.

The stability strategies of the bismuth halide perovskites have been classified into two approaches: intrinsic and extrinsic. The first approach is related to the modification of the bismuth halide’s crystal structure by forming solid solutions or by adding a second cation to improve the stability of the materials against O₂, H₂O, and light. On the other hand, the second approach is associated with incorporating additional components to the perovskite to maintain its stability (e.g., heterojunctions, encapsulation on porous supports, and modifying the reaction parameters). The reports so far evidence that it is necessary to include at least one stability strategy of each approach to obtain highly stable structures with remarkable efficiencies and selectivity to produce solar fuels.

This study tries to provide a brief guide for scientists in designing new and advanced bismuth halide perovskites with high stability using different strategies that boost their practical applications in artificial photosynthesis.