A Review on Cs-Based Pb-Free Double Halide Perovskites: From Theoretical and Experimental Studies to Doping and Applications

Despite the progressive enhancement in the flexibility of Pb-based perovskites for optoelectronic applications, regrettably, they are facing two main challenges; (1) instability, which originates from using organic components in the perovskite structure, and (2) toxicity due to Pb. Therefore, new, stable non-toxic perovskite materials are demanded to overcome these drawbacks. The research community has been working on a wide variety of Pb-free perovskites with different molecular formulas and dimensionality. A variety of Pb-free halide double perovskites have been widely explored by different research groups in search for stable, non-toxic double perovskite material. Especially, Cs-based Pb-free halide double perovskite has been in focus recently. Herein, we present a review of theoretical and experimental research on Cs-based Pb-free double halide perovskites of structural formulas Cs2M+M3+X6 (M+ = Ag+, Na+, In+ etc.; M3+= Bi3+, In3+, Sb3+; X = Cl−, Br−, I¯) and Cs2M4+X6 (M4+ = Ti4+, Sn4+, Au4+ etc.). We also present the challenges faced by these perovskite compounds and their current applications especially in photovoltaics alongside the effect of metal dopants on their performance.


Introduction
During the last ten years, Pb-based perovskites as a promising subcategory of solar cell materials have attracted intense attention of the research community due to their unique advantages. In 2009, for the first time, Miyasaka et al. published a paper demonstrating the use of methylammonium lead triiodide (CH 3 NH 3 PbI 3 ) as a light absorber instead of dye in dye sensitized solar cells (DSSCs) with an efficiency of 3.8% [1]. A few years after, a solid hole transporting layer named Spiro-MeOTAD replaced the liquid electrolyte in these structures which helped to increase the efficiency of perovskite solar cells (PSCs) to 9.7% [2]. Since then, the number of research articles in this field exponentially increased, and considerable advances in power conversion efficiency exceeding 25.2% are reported for PSCs [3].
For the sake of completeness, we have provided a list of Cs/Bi 3+ -based halide double perovskites along with their theoretical and experimental bandgap values and other information including morphology and applied synthetic methods in Table 1. In order to get a good overview, we present the results from theoretical and experimental studies separately.  . Reproduced with permission from [54]. Copyright 2017, American Chemical Society.

Cs/Bi 3+ -Based Double Halide Perovskites
For many decades, Bi (bismuth) has been applied as an eco-friendly and non-toxic metal with interesting properties in diverse applications. Bi has been introduced as a suitable replacement for Pb because of comparable density and similar electronic configuration [55,56]. Due to the significant features, Bi has been one of the first examined candidates for developing Pb-free double halide perovskite materials for PV applications [48][49][50]. This idea was developed by three independent research groups which simultaneously, published their successful investigations by replacing Pb 2+ with heterovalent substitution of Bi 3+ and Ag + to form a lead-free double halide perovskite in 2016 [48][49][50].
For the sake of completeness, we have provided a list of Cs/Bi 3+ -based halide double perovskites along with their theoretical and experimental bandgap values and other information including morphology and applied synthetic methods in Table 1. In order to get a good overview, we present the results from theoretical and experimental studies separately. McClure et al. [49] studied Cs 2 AgBiBr 6 and Cs 2 AgBiCl 6 as the two first potential double halide perovskite compounds, by employing DFT analysis on VASP simulation software. The indirect bandgap of 2.06 eV for Cs 2 AgBiBr 6 and 2.62 eV for Cs 2 AgBiCl 6 were obtained as presented in Table 1. According to the atomic partial density of states, the indirect bandgap resulted from the combination of Ag 4d orbitals with halogens 3p/4p orbitals which led to sufficient changes in the valence band as shown in Figure 2a,b. The study has also shown that hole effective masses were lighter (Cs 2 AgBiCl 6 = 0.15 m h ; Cs 2 AgBiBr 6 = 0.14 m h ) than their CsPbX 3 analogs (CsPbCl 3 = 0.35 m h ; CsPbBr 3 = 0.37 m h ) and the electron effective masses were also comparable (Cs 2 AgBiCl 6 = 0.53 m e ; Cs 2 AgBiBr 6 = 0.37 m e ) with CsPbX 3 compounds (CsPbCl 3 = 0.41 m e ; CsPbBr 3 = 0.34 m e ).
Molecules 2021, 26, x FOR PEER REVIEW 7 of 42 character, respectively. This means that by moving up in halogen (from I − to Cl − ) and pnictogen group (from Bi 3+ to Sb 3+ ) in the periodic table, their p-state energy will be decreased, and consequently both VBM and CBM will be lowered. Another notable examined parameter was the smaller carrier effective masses in these compounds (between 0.1 and 0.4) compared to relevant MAPbI3. Filip et al. [57] focused on the electronic properties of Cs2AgBiX6 (X = Cl¯and Br − ) compounds through theoretical calculations. To obtain accurate bandgap values, they used the experimental crystal structures data within the DFT-LDA calculation in the quantum espresso suit with and without spin-orbit coupling. Based on obtained molecular orbitals diagram through the atom-projected density of states, they indicated that because of smaller energy difference in Ag-d and Br-p states in comparison with Ag-d and Cl-p, valence band width was reduced. Moreover, the delocalized nature of Br-4p orbitals resulted in more overlaps with Bi-6p, and thereby the width of conduction band increased as illustrated in Figure 3. Study of quasiparticle calculation for determination of the bandgap (1.8 eV for Cs2AgBiBr6 and 2.4 eV for Cs2AgBiCl6) showed a good agreement with their experimental findings (1.9 eV for Cs2AgBiBr6 and 2,2 eV for Cs2AgBiCl6) [57].
Xiao et al. [69] investigated the thermodynamic stability of Cs2AgBiBr6 by DFT calculation, and claimed that Ag vacancies were shallow accepters that resulted in intrinsic ptype conductivity in Cs2AgBiBr6. On the other hand, the existence of some dominant deep defects such as Bi vacancies (VBi) and AgBi antisites give rise to poor photovoltaic performance. Accordingly, Xiao et al. [69], in order to reduce the formation of deep defects, suggested that the synthesis of material preferably should be done under Bi-poor/Bi-rich growth conditions. Xiao et al. [67] studied Cs2In + M 3+ X6 (M = Bi 3+ , and X = halogens) double halide perovskites, both theoretically and experimentally. Their results indicated that high-energy-laying in 5s 2 state of In + is substantially responsible for promising photovoltaic performance. However, due to the oxidation of In + to In 3+ , and the reduction of Bi 3+ to its metal form, the whole perovskite structure becomes unstable.
Since optimizing the bandgap is an important factor for enhancing the efficiency of PSCs, Yang et al. [58] proposed by changing the atomic arrangement in Cs2AgBiBr6 crystal structure, the bandgap can be narrowed. Using Monte Carlo and DFT calculations performed in VASP code by employing HSE06 with SOC, they showed that increasing the temperature up to 1200 K would increase the energy and consequently, phase transitions would occur. In this condition, Ag + and Bi 3+ ions randomly occupy the M-site in Volonakis et al. [50] selected a different range of metals comprised of pnictogen (Bi 3+ , Sb 3+ (group VA of periodic table)) and noble metals (Ag + , Au + , Cu + ) for computational and experimental studies. By using density functional theory in the local density approximation (DFT-LDA), they found all these compounds to have indirect bandgaps of less than 2.7 eV. By changing both halides from I − to Cl − and the pnictogen from Bi 3+ to Sb 3+ , the bandgap was seen to be increasing. However, changing of the noble metals did not have any effect on this trend, since CBM and VBM mostly have pnictogen-p and halogen-p character, respectively. This means that by moving up in halogen (from I − to Cl − ) and pnictogen group (from Bi 3+ to Sb 3+ ) in the periodic table, their p-state energy will be decreased, and consequently both VBM and CBM will be lowered. Another notable examined parameter was the smaller carrier effective masses in these compounds (between 0.1 and 0.4) compared to relevant MAPbI 3 .
Filip et al. [57] focused on the electronic properties of Cs 2 AgBiX 6 (X = Cl and Br − ) compounds through theoretical calculations. To obtain accurate bandgap values, they used the experimental crystal structures data within the DFT-LDA calculation in the quantum espresso suit with and without spin-orbit coupling. Based on obtained molecular orbitals diagram through the atom-projected density of states, they indicated that because of smaller energy difference in Ag-d and Br-p states in comparison with Ag-d and Cl-p, valence band width was reduced. Moreover, the delocalized nature of Br-4p orbitals resulted in more overlaps with Bi-6p, and thereby the width of conduction band increased as illustrated in Figure 3. Study of quasiparticle calculation for determination of the bandgap (1.8 eV for Cs 2 AgBiBr 6 and 2.4 eV for Cs 2 AgBiCl 6 ) showed a good agreement with their experimental findings (1.9 eV for Cs 2 AgBiBr 6 and 2.2 eV for Cs 2 AgBiCl 6 ) [57]. In this section, we provide detailed synthesis methods of Cs/Bi 3+ -based double halide perovskites and their characterization results. This section based on obtained morphologies is divided into two categories, namely, crystalline and thin-film structures in 1.2.1 and 1.2.2, respectively. This detailed information of the experimental studies is also summarized in Table 1.


Cs2M + Bi 3+ X6: Single-Crystals, Polycrystalline and Nanocrystals-Based Perovskites Slaveney et al. [48] synthesized double halide perovskite Cs2AgBiBr6 with photoluminescence (PL) lifetime of approximately 660 ns with an indirect bandgap of 1.95 eV. Compared to MAPbI3, this compound showed higher stability against moisture and heat and had long room-temperature PL lifetime (736 ns to 1 μs for MAPbI3 films) which is a good characteristic for photovoltaic performance. Interestingly, the obtained PXRD (Powder X-ray diffraction) pattern did not show any evidence of material decompaction as shown in Figure 4a. Thermogravimetric analysis (TGA) showed that the material was stable up to 430° which could be due to the replacement of unstable organic cation MA + with Cs + ions as depicted in Figure 4b.
McClure et al. [49] prepared polycrystalline Cs2AgBiBr6 and Cs2AgBiCl6 by applying solid state and solution-based methods. The PXRD analysis confirmed the 3D structure of these compounds in corner connected octahedra with alternating Ag + and Bi 3+ in a rock salt ordering. However, according to PXRD and Rietveld refinements, due to large number of displacement parameters (Beq) and low bond valence sum of Cesium ion, phase transition could be expected below room temperature. UV-Vis diffuse reflectance spectra showed significant similarities between Cs2AgBiX6 and Pb-based analogs (CH3NH3PbX3) Figure 3. (a) Comparison between the molecular orbitals diagram of Cs 2 AgBrCl 6 and Cs 2 AgBrBr 6 and; (b) Square modulus of the wave functions for states marked from 1 to 3 on the molecular orbital diagram of Cs 2 BiAgCl 6 . 1 represents the Bi-6p 1/2 /halide-p antibonding orbitals at the bottom of the conduction band (at Γpoint). 2 is the antibonding Ag-5s/halide-p at the L-point of the conduction band), while 3 corresponds to the Ag-4d/halide-p antibonding orbitals found at the top of the valence band (at X-point). Reproduced with permission from [57]. Copyright 2016, American Chemical Society.
Xiao et al. [69] investigated the thermodynamic stability of Cs 2 AgBiBr 6 by DFT calculation, and claimed that Ag vacancies were shallow accepters that resulted in intrinsic p-type conductivity in Cs 2 AgBiBr 6 . On the other hand, the existence of some dominant deep defects such as Bi vacancies (V Bi ) and AgBi antisites give rise to poor photovoltaic performance. Accordingly, Xiao et al. [69], in order to reduce the formation of deep defects, suggested that the synthesis of material preferably should be done under Bi-poor/Bi-rich growth conditions. Xiao et al. [67] studied Cs 2 In + M 3+ X 6 (M = Bi 3+ , and X = halogens) double halide perovskites, both theoretically and experimentally. Their results indicated that high-energylaying in 5s 2 state of In + is substantially responsible for promising photovoltaic performance. However, due to the oxidation of In + to In 3+ , and the reduction of Bi 3+ to its metal form, the whole perovskite structure becomes unstable.
Since optimizing the bandgap is an important factor for enhancing the efficiency of PSCs, Yang et al. [58] proposed by changing the atomic arrangement in Cs 2 AgBiBr 6 crystal structure, the bandgap can be narrowed. Using Monte Carlo and DFT calculations performed in VASP code by employing HSE06 with SOC, they showed that increasing the temperature up to 1200 K would increase the energy and consequently, phase transitions would occur. In this condition, Ag + and Bi 3+ ions randomly occupy the M-site in A 2 M + M 3+ X 6 structure, and this leads to reduction of bandgap from 1.93 eV to pseudo-direct bandgap of 0.44 eV.

Cs 2 M + Bi 3+ X 6 : Experimental Studies
In this section, we provide detailed synthesis methods of Cs/Bi 3+ -based double halide perovskites and their characterization results. This section based on obtained morphologies is divided into two categories, namely, crystalline and thin-film structures in 1.2.1 and 1.2.2, respectively. This detailed information of the experimental studies is also summarized in Table 1.

•
Cs 2 M + Bi 3+ X 6 : Single-Crystals, Polycrystalline and Nanocrystals-Based Perovskites Slaveney et al. [48] synthesized double halide perovskite Cs 2 AgBiBr 6 with photoluminescence (PL) lifetime of approximately 660 ns with an indirect bandgap of 1.95 eV. Compared to MAPbI 3 , this compound showed higher stability against moisture and heat and had long room-temperature PL lifetime (736 ns to 1 µs for MAPbI 3 films) which is a good characteristic for photovoltaic performance. Interestingly, the obtained PXRD (Powder X-ray diffraction) pattern did not show any evidence of material decompaction as shown in Figure 4a. Thermogravimetric analysis (TGA) showed that the material was stable up to 430 • which could be due to the replacement of unstable organic cation MA + with Cs + ions as depicted in Figure 4b. although there were some minor differences at photon energies above the absorption onset as shown in Figure 5a,b. Using Kubelka-Munk equation and Tauc plot, the optical indirect bandgap of 2.19 eV and 2.77 eV were obtained for Cs2AgBiBr6 and Cs2AgBiCl6, respectively which were close to their theoretically calculated values (2.06 eV for Cs2AgBiBr6 and 2.62 eV for Cs2AgBiCl6) [46]. In order to investigate the light and chemical stability, samples were placed in both dark and light conditions under ambient environment for one month. The obtained UV-Vis diffuse reflectance spectra analysis of the samples after 14 and 28 days exhibited that Cs2AgBiBr6 was less light stable than Cs2AgBiCl6 and the conclusion was that encapsulation of the structures might be needed for PV application. Volonakis et al. [50] successfully prepared single-phase Cs2AgBiCl6 samples through the solid-state reaction in a sealed fused silica ampoule. The structural and optical measurements by using single-crystal X-ray diffraction (SCXRD), UV-Vis spectroscopy and photoluminescence showed that the Cs2AgBiCl6 perovskite belongs to Fm3̅ m space group while BiCl6 and AgCl6 octahedra are alternately placed in a rock-salt face-centered cubic structure with an indirect bandgap of 2.2 eV.
It is noticeable that the three abovementioned articles [48][49][50] presented different experimental bandgap values for Cs2AgBiCl6 (2.2 eV to 2.7 eV) and Cs2AgBiBr6 (1.83 eV to 2.19 eV) which is mainly due to the different synthesis methods (solid state preparation and solution processing) as well as utilizing different measurement techniques and methods such as UV-Vis spectroscopy and UV-Vis diffuse reflectance, Tauc plot, etc. for calculating bandgaps [57]. To overcome these discrepancies, Filip et al. [57] by employing the same solid state and solution-based synthesis methods prepared stable Cs2AgBiBr6 and Cs2AgBiCl6 single crystals under ambient environment. Through the optical measurements including UV-Vis spectroscopy and PL, they obtained an indirect bandgap of 1.9 eV and 2.2 eV for Cs2AgBiBr6 and Cs2AgBiCl6, respectively.
Xiao et al. [67] used In + instead of Ag + and synthesized Cs2In + Bi 3+ X6 (X = halogens) double halide perovskites using solid-state reaction method. Their experimental investigation using SCXRD (Single Crystal X-ray diffraction) and PXRD exhibited that In + -based halide double perovskites spontaneously decomposes into In 3+ -based ternary materials such as CsInI4, Cs3In2Br9 and Cs3In2Cl9 (Figure 6a). In fact, In 1+ is an unstable and unusual oxidation state in inorganic solids and there are only a few reported double perovskite compounds based on In + [70,71]. McClure et al. [49] prepared polycrystalline Cs 2 AgBiBr 6 and Cs 2 AgBiCl 6 by applying solid state and solution-based methods. The PXRD analysis confirmed the 3D structure of these compounds in corner connected octahedra with alternating Ag + and Bi 3+ in a rock salt ordering. However, according to PXRD and Rietveld refinements, due to large number of displacement parameters (Beq) and low bond valence sum of Cesium ion, phase transition could be expected below room temperature. UV-Vis diffuse reflectance spectra showed significant similarities between Cs 2 AgBiX 6 and Pb-based analogs (CH 3 NH 3 PbX 3 ) although there were some minor differences at photon energies above the absorption onset as shown in Figure 5a,b. Using Kubelka-Munk equation and Tauc plot, the optical indirect bandgap of 2.19 eV and 2.77 eV were obtained for Cs 2 AgBiBr 6 and Cs 2 AgBiCl 6 , respectively which were close to their theoretically calculated values (2.06 eV for Cs 2 AgBiBr 6 and 2.62 eV for Cs 2 AgBiCl 6 ) [46]. In order to investigate the light and chemical stability, samples were placed in both dark and light conditions under ambient environment for one month. The obtained UV-Vis diffuse reflectance spectra analysis of the samples after 14 and 28 days exhibited that Cs 2 AgBiBr 6 was less light stable than Cs 2 AgBiCl 6 and the conclusion was that encapsulation of the structures might be needed for PV application. Li et al. [56] by applying the high-pressure method, which previously had been used for size-dependent phase transformation [72] indicated that the bandgap of Cs2AgBiBr6 through high-pressure method was reduced by 22.3% from 2.2 eV to 1.7 eV. Using UVvis spectroscopy as optical measurement, they observed by increasing the pressure to ~15 GPa, the color of the product changed to brown black and absorption peak exhibited a continues redshift. Interestingly, even after releasing the pressure, the bandgap value was observed to be around ≈2.0 eV (Figure 6b).
The angle dispersive X-ray diffraction (ADXRD) analysis of the sample illustrated that by employing a wide range of pressure, a minimized octahedral tilting through ab plane occurs, and this decreased the Bi-Br-Ag bond angle from 180° to 166.4°. It is believed that this is one of the major factors of narrowing the bandgap of Cs2AgBiBr6.
In 2018, Hoye et al. [60] explored the carrier lifetime and recombination mechanism in the Cs2AgBiBr6 thin film in more detail. Both single crystals and thin film of Cs2AgBiBr6 were prepared through the slow precipitation method of saturated solution at low temperature. Based on optical measurements, the fundamental carrier lifetime for Cs2AgBiBr6 thin film was shown to be 1.4 μs. The carrier density of Cs2AgBiBr6 was calculated to be 2.2 × 1016 cm −3 under 1 sun, which was larger than MAPbI3 steady-state carrier density (5.2 × 1015 cm −3 under 1 sun). This clearly indicates the benefit of utilizing Cs2AgBiBr6 in solar cells. Volonakis et al. [50] successfully prepared single-phase Cs 2 AgBiCl 6 samples through the solid-state reaction in a sealed fused silica ampoule. The structural and optical measurements by using single-crystal X-ray diffraction (SCXRD), UV-Vis spectroscopy and photoluminescence showed that the Cs 2 AgBiCl 6 perovskite belongs to Fm3m space group while BiCl 6 and AgCl 6 octahedra are alternately placed in a rock-salt face-centered cubic structure with an indirect bandgap of 2.2 eV.
It is noticeable that the three abovementioned articles [48][49][50] presented different experimental bandgap values for Cs 2 AgBiCl 6 (2.2 eV to 2.7 eV) and Cs 2 AgBiBr 6 (1.83 eV to 2.19 eV) which is mainly due to the different synthesis methods (solid state preparation and solution processing) as well as utilizing different measurement techniques and methods such as UV-Vis spectroscopy and UV-Vis diffuse reflectance, Tauc plot, etc. for calculating bandgaps [57]. To overcome these discrepancies, Filip et al. [57] by employing the same solid state and solution-based synthesis methods prepared stable Cs 2 AgBiBr 6 and Cs 2 AgBiCl 6 single crystals under ambient environment. Through the optical measurements including UV-Vis spectroscopy and PL, they obtained an indirect bandgap of 1.9 eV and 2.2 eV for Cs 2 AgBiBr 6 and Cs 2 AgBiCl 6 , respectively.
Xiao et al. [67] used In + instead of Ag + and synthesized Cs 2 In + Bi 3+ X 6 (X = halogens) double halide perovskites using solid-state reaction method. Their experimental investigation using SCXRD (Single Crystal X-ray diffraction) and PXRD exhibited that In + -based halide double perovskites spontaneously decomposes into In 3+ -based ternary materials such as CsInI 4 , Cs 3 In 2 Br 9 and Cs 3 In 2 Cl 9 ( Figure 6a). In fact, In 1+ is an unstable and unusual oxidation state in inorganic solids and there are only a few reported double perovskite compounds based on In + [70,71]. Bekenstein et al. [59] selected hot-injection method for preparation of Cs2AgBiBr6 nanocrystals which previously was employed for synthesis of CsPbX3 and CsPbCl3 NCs by Protesescu et al. [73], and Xu et al. [74], respectively. The XRD and HRTEM analysis indicated that the crystals had the cubic   Li et al. [56] by applying the high-pressure method, which previously had been used for size-dependent phase transformation [72] indicated that the bandgap of Cs 2 AgBiBr 6 through high-pressure method was reduced by 22.3% from 2.2 eV to 1.7 eV. Using UV-vis spectroscopy as optical measurement, they observed by increasing the pressure to~15 GPa, the color of the product changed to brown black and absorption peak exhibited a continues redshift. Interestingly, even after releasing the pressure, the bandgap value was observed to be around ≈2.0 eV (Figure 6b).
The angle dispersive X-ray diffraction (ADXRD) analysis of the sample illustrated that by employing a wide range of pressure, a minimized octahedral tilting through ab plane occurs, and this decreased the Bi-Br-Ag bond angle from 180 • to 166.4 • . It is believed that this is one of the major factors of narrowing the bandgap of Cs 2 AgBiBr 6 .
In 2018, Hoye et al. [60] explored the carrier lifetime and recombination mechanism in the Cs 2 AgBiBr 6 thin film in more detail. Both single crystals and thin film of Cs 2 AgBiBr 6 were prepared through the slow precipitation method of saturated solution at low temperature. Based on optical measurements, the fundamental carrier lifetime for Cs 2 AgBiBr 6 thin film was shown to be 1.4 µs. The carrier density of Cs 2 AgBiBr 6 was calculated to be 2.2 × 1016 cm −3 under 1 sun, which was larger than MAPbI 3 steady-state carrier density (5.2 × 1015 cm −3 under 1 sun). This clearly indicates the benefit of utilizing Cs 2 AgBiBr 6 in solar cells.
Bekenstein et al. [59] selected hot-injection method for preparation of Cs 2 AgBiBr 6 nanocrystals which previously was employed for synthesis of CsPbX 3 and CsPbCl 3 NCs by Protesescu et al. [73], and Xu et al. [74], respectively. The XRD and HRTEM analysis indicated that the crystals had the cubic Fm3m structure with the side length of 8-15 nm as shown in Figure 7a  Bekenstein et al. [59] selected hot-injection method for preparation of Cs2AgBiBr6 nanocrystals which previously was employed for synthesis of CsPbX3 and CsPbCl3 NCs by Protesescu et al. [73], and Xu et al. [74], respectively. The XRD and HRTEM analysis indicated that the crystals had the cubic   The color of the sample was distinctly yellow, which was different from the previously prepared samples' color based on solid-state and solution-processed methods (orangered) [48][49][50]. The stability of the NCs was examined through the changing of solvents and changing of ligands. Solvents with higher polarity caused less stable NCs and by introducing the excess amount of primary and tertiary amines, absorption and emission features of the sample disappeared, which suggested that Cs 2 AgBiBr 6 and CsPbI 3 had similar surface chemistry. Furthermore, it was demonstrated [59], due to the Ag diffusion, reduction and coalescence, Cs 2 AgBiBr 6 NCs in solution were structurally unstable and degraded to Cs 3 Bi 2 Br 9 and Cs 3 BiBr 6 byproduct NCs. However, by controlling the solvents evaporation rate, less degradation was occurred on NCs superlattices assembled into the ordered NC solids including strong ligand-ligand interaction.
Zhang et al. [75], as presented in Table 2, replaced Ag + in Bi-based halide double perovskite with Na + ion for fabricating the solar cell, and Cs 2 NaBiI 6 perovskite was synthesized by facile one-step hydrothermal process. The material showed tolerance factor of 0.849 and octahedral factor of 0.466 with bandgap of 1.66 eV. The SEM images showed by increasing the concentration of solvent (HI), up to 6 M, the crystal growth got improved as shown in Figure 8a nated by the strong electron-phonon coupling due to the relatively large Huang-Rhys factor (S = 11.7). It was also revealed that the excitation resonant of the color center in ionic crystals of Cs2AgBiBr6 was more responsible for the PL emission rather than band to band transition. It was also proven that the color centers of ionic crystals were coupled with crystal lattice vibrations [76]. Local lattice deformation would occur if any of these centers were occupied. This led to energy offset in color center between ground state (unoccupied) and excited states (occupied) [61]. Figure 8. (a) multiple shuttle-like crystals from high acidity synthesis and; (b) clusters of nano-plates from non-acid synthesis. Reproduced with permission from [75]. Copyright 2018, Royal Society of Chemistry.


Cs2M + Bi 3+ X6: Film-Based Perovskites In following section, we present the studies in which Cs2M + Bi 3+ X6 perovskites prepared as thin film for PV application and the results are presented in Table 2. In 2019, Zelewski et al. [61] by combination of photoluminescence excitation (PLE) and Raman spectroscopy demonstrated that the PL emission in Cs 2 AgBiBr 6 , was dominated by the strong electron-phonon coupling due to the relatively large Huang-Rhys factor (S = 11.7). It was also revealed that the excitation resonant of the color center in ionic crystals of Cs 2 AgBiBr 6 was more responsible for the PL emission rather than band to band transition. It was also proven that the color centers of ionic crystals were coupled with crystal lattice vibrations [76]. Local lattice deformation would occur if any of these centers were occupied. This led to energy offset in color center between ground state (unoccupied) and excited states (occupied) [61].
• Cs 2 M + Bi 3+ X 6 : Film-Based Perovskites In following section, we present the studies in which Cs 2 M + Bi 3+ X 6 perovskites prepared as thin film for PV application and the results are presented in Table 2.
Although other research groups were employing halide acids as solvent, in 2017, Greul et al. [77] for the first time used dimethyl sulfoxide (DMSO) as the solvent in the synthesis of Cs 2 AgBiBr 6 film. By annealing at temperatures higher than 250 • all the side phases such as AgBr and Cs 3 Bi 2 Br 9 were removed. They reported that adopting a preheating step at 75 • right before the spin coating of Cs 2 AgBiBr 6 film led to enhanced surface coverage and increased optical absorption. However, the Cs 2 AgBiBr 6 film exhibited faster photoluminescence decay time (220 ns) in comparison with polycrystalline powders (568 ns) due to their large concentration of trap states ( Figure 9a). Also, SEM images of the film indicated some agglomeration on the surface of spin coated film which contributed to the fast crystallization process (Figure 9b,c). indicated some agglomeration on the surface of spin coated film which contributed to the fast crystallization process (Figure 9b,c). Wu et al. [78] employed low-pressure assisted (LPA) method for preparation of Cs2AgBiBr6 film. In this method, as-synthesized Cs2AgBiBr6 powder [48] was dissolved in DMSO and then spin-coated on an ITO/glass substrate. The prepared film then was quickly moved to a low-pressure chamber (20 Pa) and finally was annealed thermally (Figure 10a,b). Using the LPA method led to enhanced crystallinity and grain size. Gao et al. [81] inspired by the LPA synthesis method [88], adopted an anti-solvent dropping methodology for synthesis of Cs2AgBiBr6 film as depicted in Figure 11a,b. Combination of this method with post-annealing at high temperature resulted in smooth morphology, micro-sized grains and high crystallinity (Figure 11c). Wu et al. [78] employed low-pressure assisted (LPA) method for preparation of Cs 2 AgBiBr 6 film. In this method, as-synthesized Cs 2 AgBiBr 6 powder [48] was dissolved in DMSO and then spin-coated on an ITO/glass substrate. The prepared film then was quickly moved to a low-pressure chamber (20 Pa) and finally was annealed thermally (Figure 10a,b). Using the LPA method led to enhanced crystallinity and grain size.  Wu et al. [78] employed low-pressure assisted (LPA) method for preparation of Cs2AgBiBr6 film. In this method, as-synthesized Cs2AgBiBr6 powder [48] was dissolved in DMSO and then spin-coated on an ITO/glass substrate. The prepared film then was quickly moved to a low-pressure chamber (20 Pa) and finally was annealed thermally (Figure 10a,b). Using the LPA method led to enhanced crystallinity and grain size. Gao et al. [81] inspired by the LPA synthesis method [88], adopted an anti-solvent dropping methodology for synthesis of Cs2AgBiBr6 film as depicted in Figure 11a,b. Combination of this method with post-annealing at high temperature resulted in smooth morphology, micro-sized grains and high crystallinity (Figure 11c). Gao et al. [81] inspired by the LPA synthesis method [88], adopted an anti-solvent dropping methodology for synthesis of Cs 2 AgBiBr 6 film as depicted in Figure 11a,b. Combination of this method with post-annealing at high temperature resulted in smooth morphology, micro-sized grains and high crystallinity (Figure 11c).
Pantaler et al. [80] used chlorobenzene as anti-solvent before the annealing step during the film formation to obtain the smooth layer without any pinholes (Figure 12a). The SEM images of the film showed crystal grain size of 80 nm and 200 nm thickness (Figure 12b).  Pantaler et al. [80] used chlorobenzene as anti-solvent before the annealing step during the film formation to obtain the smooth layer without any pinholes (Figure 12a). The SEM images of the film showed crystal grain size of 80 nm and 200 nm thickness ( Figure  12b). Wang et al. [82] employed the sequential-vapor-deposition procedure for the synthesis of Cs2AgBiBr6 film as depicted in Figure 13. To obtain an optimized film, annealing Wang et al. [82] employed the sequential-vapor-deposition procedure for the synthesis of Cs 2 AgBiBr 6 film as depicted in Figure 13. To obtain an optimized film, annealing was carried out in two different steps. The film was first annealed at 200 • for 5 h, which was then followed by annealing at 240 • to form the double halide perovskite phase. was carried out in two different steps. The film was first annealed at 200° for 5 h, which was then followed by annealing at 240° to form the double halide perovskite phase. Figure 13. Scheme of sequential vapor deposition processing. Reproduced with permission from [82]. Copyright 2018, Wiley-VCH.
In 2019, Igbari et al. [83] selected two different synthesis methods of solution processing and vacuum processing for preparation of Cs2AgBiBr6 film as shown in Figure 14. The photophysical properties, electronic and crystalline structure as well as photovoltaic properties of the films were compared. The results of various characterizations demonstrated that Cs2AgBiBr6 film prepared by solution processing possessed higher crystallinity, longer charge carrier lifetime, narrower electronic bandgap of 1.98 eV (2.08 eV for vacuum-processed film) and higher mobility compared with vacuum-processed Cs2Ag-BiBr6 film. Doping/alloying process is an effective method for tuning the electronic properties of the semiconducting materials [89]. For perovskite materials the partial of original constituent elements are replaced by targeted ions which leads to diverse advantages including, enhanced optoelectronic performance, stabilized crystal structure, improvement in photoluminescence properties, etc. [90]. A variety of doped Cs2M + Bi 3+ X6 structures are reviewed and listed in Table 3. In 2017, Slavney et al. [91] engineered Cs2AgBiBr6 double halide perovskite's bandgap by dilute alloying Tl + (Cs2(Ag1−aBi1−b)TlxBr6; 0.003 ˂ x = a + b ˂ 0.075). DFT-PBE+SOC calculations showed that when Ag + was substituted by Tl + (x = 0.06), Figure 13. Scheme of sequential vapor deposition processing. Reproduced with permission from [82]. Copyright 2018, Wiley-VCH.
In 2019, Igbari et al. [83] selected two different synthesis methods of solution processing and vacuum processing for preparation of Cs 2 AgBiBr 6 film as shown in Figure 14. The photophysical properties, electronic and crystalline structure as well as photovoltaic properties of the films were compared. The results of various characterizations demonstrated that Cs 2 AgBiBr 6 film prepared by solution processing possessed higher crystallinity, longer charge carrier lifetime, narrower electronic bandgap of 1.98 eV (2.08 eV for vacuum-processed film) and higher mobility compared with vacuum-processed Cs 2 AgBiBr 6 film. was carried out in two different steps. The film was first annealed at 200° for 5 h, which was then followed by annealing at 240° to form the double halide perovskite phase. Figure 13. Scheme of sequential vapor deposition processing. Reproduced with permission from [82]. Copyright 2018, Wiley-VCH.
In 2019, Igbari et al. [83] selected two different synthesis methods of solution processing and vacuum processing for preparation of Cs2AgBiBr6 film as shown in Figure 14. The photophysical properties, electronic and crystalline structure as well as photovoltaic properties of the films were compared. The results of various characterizations demonstrated that Cs2AgBiBr6 film prepared by solution processing possessed higher crystallinity, longer charge carrier lifetime, narrower electronic bandgap of 1.98 eV (2.08 eV for vacuum-processed film) and higher mobility compared with vacuum-processed Cs2Ag-BiBr6 film. Doping/alloying process is an effective method for tuning the electronic properties of the semiconducting materials [89]. For perovskite materials the partial of original constituent elements are replaced by targeted ions which leads to diverse advantages including, enhanced optoelectronic performance, stabilized crystal structure, improvement in photoluminescence properties, etc. [90]. A variety of doped Cs2M + Bi 3+ X6 structures are reviewed and listed in Table 3. In 2017, Slavney et al. [91] engineered Cs2AgBiBr6 double halide perovskite's bandgap by dilute alloying Tl + (Cs2(Ag1−aBi1−b)TlxBr6; 0.003 ˂ x = a + b ˂ 0.075). DFT-PBE+SOC calculations showed that when Ag + was substituted by Tl + (x = 0.06), Doping/alloying process is an effective method for tuning the electronic properties of the semiconducting materials [89]. For perovskite materials the partial of original constituent elements are replaced by targeted ions which leads to diverse advantages including, enhanced optoelectronic performance, stabilized crystal structure, improvement in photoluminescence properties, etc. [90]. A variety of doped Cs 2 M + Bi 3+ X 6 structures are reviewed and listed in Table 3. In 2017, Slavney et al. [91] engineered Cs 2 AgBiBr 6 double halide perovskite's bandgap by dilute alloying Tl + (Cs 2 (Ag 1−a Bi 1−b )Tl x Br 6 ; 0.003 < x = a + b < 0.075). DFT-PBE+SOC calculations showed that when Ag + was substituted by Tl + (x = 0.06), the bandgap would decrease by 0.1 eV with direct transition, while replacing the Bi 3+ with Tl 3+ (x = 0.06) would lower the bandgap by 0.8 eV and transition would be indirect. Also, doping of 0.075 Tl in (Cs 2 (Ag 1−a Bi 1−b ) Tl x Br 6 structure decreased the bandgap to 1.40 eV (indirect) or to 1.57 eV (direct). Importantly, the long-lived carrier lifetime in microsecond obtained by time-resolved microwave photoconductivity (TRMC), suggested the efficient extraction of carriers in a solar cell. Both bandgap and carrier lifetime of Cs 2 (Ag 1−a Bi 1−b ) Tl x Br 6 structure were comparable to MAPbI 3 values, but the real setback is that as known Tl + is toxic.
Du et al. [92] through alloying of Sb 3+ /In 3+ in Cs 2 AgBiBr 6 perovskite structure, engineered the bandgap and demonstrated long carrier recombination lifetime. Based on UV-vis diffuse spectra, it was shown that by replacing Bi 3+ with In 3+ , when the level of alloyed metal (x) reach to 0.75 (Cs 2 AgBi 1−x In x Br 6 ), the indirect bandgap would increase to 2.27 eV. On the contrary, substitution of Bi 3+ with Sb 3+ up to 0.375 (Cs 2 AgBi 1−x Sb x Br 6 ; x = 0.375), resulted in narrowing the bandgap from 2.12 eV to 1.86 eV. The opposite bandgap shift direction was related to the different atomic configuration for these metals. The weak PL intensity of In-alloyed sample with high content (x = 0.75) was associated to deeper defect state and symmetry-forbidden transition from valence band to conduction band. Similarly, deep defect state of Sb 3+ -alloyed sample was responsible for the quickly suppressed emission intensity (Figure 15a-d).
the bandgap would decrease by 0.1 eV with direct transition, while replacing the Bi 3+ with Tl 3+ (x = 0.06) would lower the bandgap by 0.8 eV and transition would be indirect. Also, doping of 0.075 Tl in (Cs2(Ag1−aBi1−b) TlxBr6 structure decreased the bandgap to 1.40 eV (indirect) or to 1.57 eV (direct). Importantly, the long-lived carrier lifetime in microsecond obtained by time-resolved microwave photoconductivity (TRMC), suggested the efficient extraction of carriers in a solar cell. Both bandgap and carrier lifetime of Cs2(Ag1−aBi1−b) TlxBr6 structure were comparable to MAPbI3 values, but the real setback is that as known Tl + is toxic.
Du et al. [92] through alloying of Sb 3+ /In 3+ in Cs2AgBiBr6 perovskite structure, engineered the bandgap and demonstrated long carrier recombination lifetime. Based on UVvis diffuse spectra, it was shown that by replacing Bi 3+ with In 3+ , when the level of alloyed metal (x) reach to 0.75 (Cs2AgBi1−xInxBr6), the indirect bandgap would increase to 2.27 eV. On the contrary, substitution of Bi 3+ with Sb 3+ up to 0.375 (Cs2AgBi1−xSbxBr6; x = 0.375), resulted in narrowing the bandgap from 2.12 eV to 1.86 eV. The opposite bandgap shift direction was related to the different atomic configuration for these metals. The weak PL intensity of In-alloyed sample with high content (x = 0.75) was associated to deeper defect state and symmetry-forbidden transition from valence band to conduction band. Similarly, deep defect state of Sb 3+ -alloyed sample was responsible for the quickly suppressed emission intensity (Figure 15a  In 2019, Mn 2+ doped Cs2NaBiCl6 polycrystalline as a promising orange-red phosphor system was reported by Majher et al. [94]. The absorbed near-UV light by Bi 3+ ions in the host lattice transferred to Mn 2+ activators and through the spin-forbidden 4 T1 → 6 A1 transition, light emitting from 525 to 700 nm occurred. Also, they [94] demonstrated partial substitution of Cl − ions by Br − resulted in redshift of exciton spectra as well as an optical ab-  Cs 2 AgBi 1−x In x Cl 6 (x = 0, 0.25, 0.5, 0.75 and 0.9) NCs were prepared by Yang et al. [93], through anti solvent recrystallization. In this work, Cs 2 AgBi 1−x In x Cl 6 (x = 0.25, 0.5) with indirect bandgap were tuned to direct band gap by increasing the In 3+ content to 0.75 and 0.9. DFT calculation along with steady-state absorption, PL and transient absorption spectra measurements demonstrated dual color emission of violet (395 nm) (band to band transition) and bright orange (570 nm) (forbidden transition) in NCs.
In 2019, Mn 2+ doped Cs 2 NaBiCl 6 polycrystalline as a promising orange-red phosphor system was reported by Majher et al. [94]. The absorbed near-UV light by Bi 3+ ions in the host lattice transferred to Mn 2+ activators and through the spin-forbidden 4 T 1 → 6 A 1 transition, light emitting from 525 to 700 nm occurred. Also, they [94] demonstrated partial substitution of Cl − ions by Br − resulted in redshift of exciton spectra as well as an optical absorption peak. Zhang et al. [84], through doping of different amount of Rb + ion by replacing Cs + (stochiometric ratio of Cs + /Rb + : 100/0, 95/5, 90/10, 85/15) prepared different (Cs 1-x Rb x ) 2 AgBiBr 6 perovskite compounds. Optical measurements showed with doping ratio of Cs + /Rb + : 90:10, the intensity of PL spectra increased due to the reduction of defects. Moreover, based on IPEC spectrum, it was shown that doping of Rb + ion leads in enhancement of absorption at longer wavelength as depicted in Figure 16a  In 2020, Yao et al. [95] reported studies on lead-free double halide perovskite Cs2NaBiCl6 NCs as host, doped with Ag + , Mn 2+ , Eu 3+ ions through hot-injection approach in order to improve the optical properties of the material. The femtosecond time-resolved transient absorption technique was utilized to investigate the PL enhancement mechanism of ion doped NCs. The excitonic absorption energy of Ag-doped sample exhibited red-shift from 3.82 eV to 3.48 eV which led to significant increase of PLQY from 1.7% to 20%. The Mn 2+ -doped Cs2NaBiCl6 NCs showed a peak centered in 585 nm along with broad red-orange emission owing to 4 T1 → 6 A1 transition of coordinated Mn 2+ ions. Also, Eu 3+ -doped NCs, owing to the 5 D0 → 7 FJ (J = 1,2,3,4) transition, demonstrated four sharp emission lines at PL spectra corresponding to 591, 615, 652 and 700 nm, respectively.  In 2020, Yao et al. [95] reported studies on lead-free double halide perovskite Cs 2 NaBiCl 6 NCs as host, doped with Ag + , Mn 2+ , Eu 3+ ions through hot-injection approach in order to improve the optical properties of the material. The femtosecond time-resolved transient absorption technique was utilized to investigate the PL enhancement mechanism of ion doped NCs. The excitonic absorption energy of Ag-doped sample exhibited red-shift from 3.82 eV to 3.48 eV which led to significant increase of PLQY from 1.7% to 20%. The Mn 2+ -doped Cs 2 NaBiCl 6 NCs showed a peak centered in 585 nm along with broad red-orange emission owing to 4  As presented in Table 2, in 2017, Greul et al. [77] fabricated for the first time the Cs 2 AgBiBr 6 -based PSC, which after different parameter measurements exhibited power conversion efficiency (PECs) of 2.43% with V oc exceeding 1 Volt as well as higher stability under constant illumination in ambient environment compared to MAPBI 3 . In another study, Wu et al. [78] fabricated a PSC with Cs 2 AgBiBr 6 film as active absorbing layer with a PCE of 1.44% and V oc of 1.04 V, J sc of 1.78 mA cm −2 and FF of 78 under AM1.5 (100 mW cm −2 ) illumination. The low PCE of the device was attributed to larger exciton binding energy of Cs 2 AgBiBr 6 than MAPbI 3 . However, the Hole Transporting Layer (HTL)-free device showed higher stability in 4 months compared to MAPbI 3 based PSCs.
In 2018, Ning et al. [79] demonstrated the Cs 2 AgBiBr 6 solar cell using one-step spin coating process from single-crystal Cs 2 AgBiBr 6 solution. The device showed a PCE with maximum value of 1.22% and long photoexcited carrier diffusion length close to 110 nm. The V oc of 1.06 V, J sc of 1.55 mA cm −2 and FF of 74 were also achieved for the device. The low PCE value was attributed to the low efficiency of charge extraction by TiO 2 as the electron transporting layer (ETL) as well as presence of interfacial barrier due to the surface defects. They predicated by increasing the film thickness while preserving the quality of the absorber film and replacing the TiO 2 by other suitable ETL materials the efficiency of Bi-based halide double perovskite solar cells could be enhanced. Gao et al. [81] adopted anti-solvent dropping methodology for synthesis of Cs 2 AgBiBr 6 film as already shown in Figure 11a In 2018, Zhang et al. [75] replaced Ag + in Cs 2 AgBiBr 6 double halide perovskite with Na + ion in order to fabricate the solar device. Cs 2 NaBiI 6 perovskite was prepared by facile one-step hydrothermal process. However, the fabricated device demonstrated low J sc due to two main reasons, namely that Cs 2 NaBiI 6 perovskite had (1) low hole transport ability, and (2) was not able to efficiently convert the excitons to current. Nevertheless, the fabricated devices showed high stability for 5 months. Testing of a batch of 20 different devices showed the PCE of 0.42% with V oc of 0.47 V, J sc of 1.99 mA cm −2 and FF of 44. Wang et al. [82] fabricated a PSC by employing sequential-vapor-deposition procedure for synthesis of active layer Cs 2 AgBiBr 6 film as already showed in Figure 13. The device showed lower defect density compered to Cs 2 AgBiBr 6 film prepared by solution process as well as good stability after 240 h under ambient environment. The PCE of 1.37% and V oc of 1.12 V were reported. By employing solution-based processing and vacuum sublimation method, Igbari et al. [83] could achieve an optimized PCE of 2.51% and 1.41%, respectively for Cs 2 AgBiBr 6 -based PSCs.
In 2020, Wang et al. [64] studied the performance of Zinc chlorophyll (Zn-Chl) as the HTL in Cs 2 AgBiBr 6 -based PSC. It was shown that by employing the Zn-Chl not only enhanced the photovoltaic performance was achieved but also the light absorbing abilities of the Cs 2 AgBiBr 6 through the sensitizing of the perovskite material was improved. The PSCs based on Zn-Chl showed PCE of 2.79% with V oc of 0.99 V, J sc of 3.83 mA cm −2 and FF of 73.6. In another study in 2020, Yang et al. [65] demonstrated that by employing di-tetrabutylammonium cis-bis(isothiocyanato) bis (2,2 -bipyridyl-4,4 dicarboxylato) ruthenium (II) (N719) dye as an interlayer on the surface of Cs 2 AgBiBr 6 film, the efficiency and stability of Cs 2 AgBiBr 6 -based solar device were boosted. Applying the N719, also led to broadening the light absorption spectrum, reducing the charge carrier recombination, reducing the Cs 2 AgBiBr 6 film surface defects and accelerating the hole extraction. The optimized solar device showed PCE of 2.84% with V oc of 1.06 V, J sc of 5.13 mA cm −2 and FF of 52.4.
In 2021, Wang et al. [66] following their previous work [64], by employing carboxychlorophyll derivative (C-Chl) in the mesoporous TiO 2 film, improved the efficiency of Cs 2 AgBiBr 6 -based solar cells to more than 3%, which is the highest reported efficiency for Cs 2 AgBiBr 6 PSC. The fabricated PSC based on C-Chl-sensitized mesoporous TiO 2 film showed improved PCE of 3.11% with V oc of 1.04 V, J sc of 4.09 mA cm −2 , and FF of 73.
As summarized in this section, Cs 2 AgBiBr 6 -based solar cells have showed low power conversion efficiency compared to lead-based PSCs. However, the results obtained show clearly that Cs 2 AgBiBr 6 is a promising material for PV application even though the path to enhance the efficiency might be long. The aforementioned studies [77][78][79]82,84] show that employing of a variety of coating engineering such as using Low-pressure assisted method (LPA), anti-solvent method, vapor deposition method and introducing metal ion dopants resulted in enhanced Cs 2 AgBiBr 6 film morphology improving the efficiency. Furthermore, it was shown that modifying and optimizing ETL and HTL layers also have a significant impact on improvement of efficiency of Cs 2 AgBiBr 6 -based solar cells [64,66].
To increase the fill factor (FF) and the open source voltage (V oc ), the recombination has to be controlled. It is needed to obtain uniform and high-crystalline perovskite films; meaning that the defect density of the perovskite layer should be reduced [109]. In another words, the higher density traps of the perovskite film cause in more Shockley-Read-Hall (SRH) recombination and lower FF and consequently lead to poor PCE. Wang et al. [66] Used a concept based on ideality factor (N) in order to describe the SRH recombination due to defect density. N is defined by: where e is electron charge, K B is Boltzmann constant, I is different light intensity, T is the temperature and N shows the charge carrier recombination process. For ideal solar cells N must approach unity. When N approaches 2, the performance of the device dominated by Shockley-Read-Hall (SRH) recombination which is assisted by defect density. Therefore, lower N value shows suppression of SRH and reduced trap densities which results in higher fill factor and higher PCE [66].

Non-Photovoltaic Applications
Volonakis et al. [110] adopted first-principle calculation to determine the level of surfaces and surface termination energy of Cs 2 AgBiCl 6 , Cs 2 AgBiBr 6 , Cs 2 AgSbCl 6 and Cs 2 AgInCl 6 double halide perovskites. Their investigation demonstrated that according to ionization potential and electron affinity, amongst all these four materials, Cs 2 AgBiCl 6 and Cs 2 AgBiBr 6 were the most promising photocatalysts for solar-driven water splitting. Their study also indicated, by increasing the size of halogens in double perovskites, the electron affinity would increase as well, which was due to the shallower energies of the halogen p-states. In 2017, Pan et al. [111] for the first time, reported the application of Cs 2 AgBiBr 6 single crystals as X-ray detectors. Using of thermal annealing and surface treatment resulted in elimination of disordered Ag + /Bi 3+ and consequently, the resistivity of the crystals improved. The optimized device showed high sensitivity of 105 µC Gy air −1 cm −2 , low detection limit of 105 nC Gy air −1 s −1 under the external bias of 5 V as well as long-term operational stability which all are essential for X-ray detectors in order to medical diagnostics. The single crystals of Cs 2 AgBiBr 6 as suitable semiconductor directly converted X-rays into electrical signals due to its high average atomic number which results in higher X-ray absorption coefficient (α∝Z 4 /E 3 ), adequate µτ product (µ = carrier mobility; τ = carrier lifetime), low ionization energy and high resistivity. Yuan et al. [112] by introducing PEABr (phenylethylamine bromide) into the Cs 2 AgBiBr 6 perovskite precursors solution, obtained single crystals of Cs 2 AgBiBr 6 with enhanced ordering degree of [BiX 6 ] 3− and [AgX 6 ] 5− in octahedra arrangement. The improved order degree gave rise to lower defect density, tunable bandgap, decreased self-trapped exciton formation and increased carrier mobility. The X-ray detector displayed higher current response of 13 vs. 3190 µs, higher sensitivity of 288.8 µC Gy air −1 cm −2 under a bias of 50 µC Gy air −1 cm −2 , higher photoconductive gain and longer carrier drift distance. Li et al. [113] synthesized composites films comprised of Cs 2 AgBiBr 6 perovskite embedded in a polymer matrix by a simple drop-casting process. Hydroxyl functional groups of polymers significantly increased the dispersity of Cs 2 AgBiBr 6 in the composite films which led in large area dense films. The fabricated X-ray detector obtained by the composite films demonstrated a sensitivity of 40 µC Gy air −1 cm −2 under the external bias of 400 V, and due to the maximum tolerance of 5% tensile/compressive strain, bending/flexing of detectors did not have any degrading effect on photocurrent.
Lei et al. [114] adopted a one-step spin-coating synthesis method for preparation of Cs 2 AgBiBr 6 film as photodetector. The device showed high responsivity of 7.01 A/W, On/Off photocurrent ratio of 2.16 × 10 4 , specific detectivity of 5.66 × 10 11 Jones, EQE of 2164%, fast response speed of 956/995 µs. The other remarkable feature of unencapsulated photodetector was the high stability under ambient environment against water and oxygen degradation (36 h continuous operation) without no change in photodetection ability. Wu et al. [115] designed a HTL-free planer heterojunction device including ITO/ SnO 2 / Cs 2 AgBiBr 6 /Au as ultraviolet (UV) photodetector as shown in Figure 17a,b. The selfpowered devices exhibited two responsivity peaks at 350 nm and 435 nm which was associated with ultraviolet-A (320-400 nm). The mechanism explained by separation of photogenerated carriers at the of interface of Cs 2 AgBiBr 6 /SnO 2 heterojunction by its built-in field. A high responsivity of 0.11 A W −1 at 350 nm and the quick response of less than 3 ms was comparable with other semiconductor oxide heterojunction-based UV detectors. The unencapsulated UV detector also showed remarkable stability under ambient environment for more than 6 months without any noticeable degradation in photocurrent. tals improved. The optimized device showed high sensitivity of 105 μC Gyair −1 cm −2 , low detection limit of 105 nC Gyair −1 s −1 under the external bias of 5 V as well as long-term operational stability which all are essential for X-ray detectors in order to medical diagnostics. The single crystals of Cs2AgBiBr6 as suitable semiconductor directly converted Xrays into electrical signals due to its high average atomic number which results in higher X-ray absorption coefficient (α∝Z 4 /E 3 ), adequate μτ product (μ = carrier mobility; τ = carrier lifetime), low ionization energy and high resistivity. Yuan et al. [112] by introducing PEABr (phenylethylamine bromide) into the Cs2AgBiBr6 perovskite precursors solution, obtained single crystals of Cs2AgBiBr6 with enhanced ordering degree of [BiX6] 3− and [AgX6] 5− in octahedra arrangement. The improved order degree gave rise to lower defect density, tunable bandgap, decreased self-trapped exciton formation and increased carrier mobility. The X-ray detector displayed higher current response of 13 vs 3190 μs, higher sensitivity of 288.8 μC Gyair −1 cm −2 under a bias of 50 μC Gyair −1 cm −2 , higher photoconductive gain and longer carrier drift distance. Li et al. [113] synthesized composites films comprised of Cs2AgBiBr6 perovskite embedded in a polymer matrix by a simple drop-casting process. Hydroxyl functional groups of polymers significantly increased the dispersity of Cs2AgBiBr6 in the composite films which led in large area dense films. The fabricated Xray detector obtained by the composite films demonstrated a sensitivity of 40 μC Gyair −1 cm −2 under the external bias of 400 V, and due to the maximum tolerance of 5% tensile/compressive strain, bending/flexing of detectors did not have any degrading effect on photocurrent.
Lei et al. [114] adopted a one-step spin-coating synthesis method for preparation of Cs2AgBiBr6 film as photodetector. The device showed high responsivity of 7.01 A/W, On/Off photocurrent ratio of 2.16 × 10 4 , specific detectivity of 5.66 × 10 11 Jones, EQE of 2164%, fast response speed of 956/995 μs. The other remarkable feature of unencapsulated photodetector was the high stability under ambient environment against water and oxygen degradation (36 h continuous operation) without no change in photodetection ability. Wu et al. [115] designed a HTL-free planer heterojunction device including ITO/ SnO2/ Cs2AgBiBr6/Au as ultraviolet (UV) photodetector as shown in Figure 17a,b. The self-powered devices exhibited two responsivity peaks at 350 nm and 435 nm which was associated with ultraviolet-A (320-400 nm). The mechanism explained by separation of photogenerated carriers at the of interface of Cs2AgBiBr6/SnO2 heterojunction by its built-in field. A high responsivity of 0.11 A W −1 at 350 nm and the quick response of less than 3 ms was comparable with other semiconductor oxide heterojunction-based UV detectors. The unencapsulated UV detector also showed remarkable stability under ambient environment for more than 6 months without any noticeable degradation in photocurrent. Zhou et al. [116] reported the fabrication of Cs 2 AgBiBr 6 NCs by hot injection method for CO 2 photocatalytic reduction. The prepared double halide perovskite demonstrated significant stability against moisture, light and temperature. It also showed a total electron consumption of 105 µmol g −1 under simulated solar light (AM 1.5G) for 6 h. Zhang et al. [117] developed an alcohol-based photocatalytic system for dye degradation by applying Cs 2 AgBiBr 6 double halide perovskite under visible light irradiation. During the photocatalytic process, Cs 2 AgBiBr 6 kept its high chemical stability in ethanol, and due to the photocatalytic feature of Cs 2 AgBiBr 6 surface, pseudo-zeroth-order kinetics was obtained, and the reaction rate was also comparable to well-known CdS photocatalyst semiconductor. In 2017, Volonakis et al. [118] by using first-principle calculations, identified the direct bandgap of Cs 2 AgInX 6 (X = Cl, Br) halide double perovskites. According to DFT/LDA calculation, In-based perovskites showed smaller lattice constant compared to Bi-Based analogous due to the smaller size of In 3+ . Preliminary assessment of octahedral and tolerance factor exhibited that because of smaller ionic radii of In 3+ (0.8 Å), the coordination between In 3+ and I − ions would be impossible, and therefore, the synthesis of Cs 2 AgInX 6 (X = Cl, Br and Cl/Br) double halide perovskite may be amenable. Nominal bandgap for the Cs 2 AgInCl 6 based on theoretical calculation (HSE and PBE0 hybrid functionals) was reported to be of 2.07 eV with a bias of 0.6 eV (Figure 18a,b). They also demonstrated the VBM was mainly comprised of Cl-3p and In-4d/Ag-4d states while the CBM was occupied by Cl-3p and In-5s/Ag-5s states. The electron and hole effective masses were reported to be 0.20 m e and 0.28 m h , respectively. duced with permission from [115]. Copyright 2018, Wiley-VCH.

Cs/In
Zhou et al. [116] reported the fabrication of Cs2AgBiBr6 NCs by hot injection method for CO2 photocatalytic reduction. The prepared double halide perovskite demonstrated significant stability against moisture, light and temperature. It also showed a total electron consumption of 105 μmol g −1 under simulated solar light (AM 1.5G) for 6 h. Zhang et al. [117] developed an alcohol-based photocatalytic system for dye degradation by applying Cs2AgBiBr6 double halide perovskite under visible light irradiation. During the photocatalytic process, Cs2AgBiBr6 kept its high chemical stability in ethanol, and due to the photocatalytic feature of Cs2AgBiBr6 surface, pseudo-zeroth-order kinetics was obtained, and the reaction rate was also comparable to well-known CdS photocatalyst semiconductor. In 2017, Volonakis et al. [118] by using first-principle calculations, identified the direct bandgap of Cs2AgInX6 (X = Cl, Br) halide double perovskites. According to DFT/LDA calculation, In-based perovskites showed smaller lattice constant compared to Bi-Based analogous due to the smaller size of In 3+ . Preliminary assessment of octahedral and tolerance factor exhibited that because of smaller ionic radii of In 3+ (0.8 Å), the coordination between In 3+ and I − ions would be impossible, and therefore, the synthesis of Cs2AgInX6 (X = Cl, Br and Cl/Br) double halide perovskite may be amenable. Nominal bandgap for the Cs2AgInCl6 based on theoretical calculation (HSE and PBE0 hybrid functionals) was reported to be of 2.07 eV with a bias of 0.6 eV (Figure 18a,b). They also demonstrated the VBM was mainly comprised of Cl-3p and In-4d/ Ag-4d states while the CBM was occupied by Cl-3p and In-5s/Ag-5s states. The electron and hole effective masses were reported to be 0.20 me and 0.28 mh, respectively.  Zhao et al. [119] as provided in Table 4, inspired by Cu[In,Ga]Se 2 (CIGS) chalcopyrite solar absorbers [120][121][122], designed 36 different candidates of double halide perovskites where A = Cs, Rb, K; M + = Cu, Ag; M 3+ = Ga, In and X = Cl, Br, I through first-principles calculations. Although all the compounds showed direct bandgaps, three of them exhibited remarkable bandgap of 1.36 eV (Rb 2 CuInCl 6 ), 1.46 eV (Rb 2 AgInBr 6 ) and 1.50 eV (Cs 2 AgInBr 6 ). Generally, Cu-based compounds showed smaller bandgap than Ag-based perovskites. The screening thermodynamic stability of compounds exhibited positive ∆H dec that led to suppressed decomposition of compounds. Their study [119] showed that by increasing the film thickness up to 2 µm, spectroscopic limited maximum efficiency (SLME) increased to approximately 28%. This value was attributed to small direct bandgaps of Rb 2 CuInCl 6 , Rb 2 AgInBr 6 and Cs 2 AgInBr 6 materials which were close to optimal bandgap of 1.34 eV calculated for Shockley-Queisser limit. This SLME value (28%) was comparable to the obtained SLME values for CuInSe 2 (31.5%) and CH 3 NH 3 PbI 3 (30%).

Cs/In
The intrinsic defects of In-based double halide perovskite were investigated by Xu et al. [123] through the first-principle calculation. The theoretical study of band structure showed hybridization of Ag(d) and ionic X(p) orbitals along with negligible coupling with In (s) orbital resulted in direct bandgap where both VBM and CBM were placed at Γ point of Cs 2 AgInCl 6 . It was expected that Cs 2 AgInCl 6 shows more point defects as a quaternary compound, which makes the growth of high-quality film challenging. It was shown that in order to avoid of deep-level defects and unwanted secondary phases, the synthesis of film should be done in Ag-rich growth condition. It was also suggested that depending on chemical growth condition, the conductivity of Cs 2 AgInCl 6 could change from good n-type/poor n-type to intrinsic semiconducting.

•
Cs 2 M + In 3+ X 6 : Single-Crystals, Polycrystalline and Nanocrystals-Based Perovskites In order to verify the accuracy of theoretical studies on stability, direct bandgap and balanced effective masses of Cs 2 AgInX 6 (X = Cl, Br) Volonakis et al. [118] synthesized the materials through precipitation from the acid solution. The experimental measurements were in a good agreement with their theoretical results. The optical bandgap obtained by experimental measurements was 3.3 eV as presented in Table 4.
Zhou et al. [124] synthesized Cs 2 AgInCl 6 perovskite crystals by using hydrothermal method. The size of the crystals varied between 5 to 15 µm which was due to the different reaction time. Rietveld analysis for XRD measurement of Cs 2 AgInCl 6 powder identified the cubic unit cell with the space group of Fm-3m [AgCl 6 ] and [InCl 3 ] octahedra in 3D framework. Time-resolved emission spectra showed two different decay times of 16.3 and 100 µs which was attributed to surface/defect states and fundamental nonradiative recombination. The optical bandgap measured by UV-vis reflectance spectroscopy was 3.23 eV while this value determined to be 3.33 eV by using band structure and optical absorption calculations. However, the compound showed to maintain high light, moisture and heat stability.

Cs 2 M + In 3+ X 6 : Doping
In 2018, Nag et al. [96] prepared bulk Mn-doped Cs 2 AgInCl 6 perovskite employing the same method used by Volonakis et al. [118] in order to study the photoluminescence properties. The optical measurements of samples showed a weak intensity of undoped sample while by increasing of Mn 2+ concentration as dopant, the intensity of emissions raised up significantly. This was due to the de-excitation of Mn 2+ d-electrons from 4 T 1 → 6 A 1 state. Locandi and co-workers [97] prepared Cs 2 AgInCl 6 and Mn-doped Cs 2 AgInCl 6 NCs, using colloidal hot-injection method. Synthesis resulted in highly pure NCs without any undesired secondary phases for both Cs 2 AgInCl 6 and Mn-doped Cs 2 AgInCl 6 as well as high thermal stability up to 500 • . The experimentally obtained optical bandgap was larger compared to previously reported works (4.38 eV for Cs 2 AgInCl 6 NCs and 4.36 eV for Mn-doped Cs 2 AgInCl 6 NCs). However, the Mn-doped Cs 2 AgInCl 6 NCs exhibited bright PL emissions with a PLQY of 16 ± 4% which is comparable to Cs 2 AgInCl 6 NCs value (1.6 ± 1%). This result shows that doping with Mn 2+ would make the In 3+ -based double halide perovskites a good candidate for different applications such as LEDs.
Tran et al. [125] through a solid-state technique synthesized Cs 2 AgInCl 6 while the Cs 2 AgSb x In 1−x Cl 6 (x = 0.5, 0.4 and 0.2) solid solutions were prepared by combining the single crystals of hydrothermally synthesized of Cs 2 AgInCl 6 and Cs 2 AgSbCl 6 , in a stoichiometric ratio. UV-Vis diffuse reflectance measurement along with Tauc plot demonstrated, by increasing the Sb composition in Cs 2 AgSb x In 1−x Cl 6 (x = 0.5, 0.4 and 0.2 and 0), the bandgap would shift from direct to indirect while the value decreased from 3.53 eV to 2.54 eV in Cs 2 AgSbCl 6 (Figure 19a,b).  In 2019, Hu et al. [98] prepared colloidal Cs2Ag1-xNaxIn1-yBiyCl6 (x = 0-1; y = 0.03-0.16) NCs by a room temperature recrystallization process. The mean size of NCs were reported to be of 3 nm with diameter of 1.8-4.0 nm which could potentially confine the excitons. The remarkable blueshift photoluminescence was observed for the NCs which was comparable for bulk materials. Incorporation of partial amount of Na + and Bi 3+ resulted in bright near-white light emission with tunable color temperatures from 9759.7 to 4429.2 K. Furthermore, the introduction of Bi 3+ ions along with OA (ligand) passivation in Liu et al. [99] by utilizing a facile hot injection method, prepared Cs 2 AgInCl 6 and Bidoped Cs 2 AgInCl 6 NCs. It was indicated that by controlling the reaction time both ligands and HCl concentration, the purer NCs without any impurity phases with desirable size and shapes would be produced. By increasing the reaction temperature from 180 • to 280 • , the PLQY curve also rise and reached to highest PLQY of 11.4%. TEM and HRTEM analysis of NCs formed at 180 • exhibited of some spherical particles which were ascribed to Ag 2 O. Due to metal-ion-induced oxidation process the silver ions turn to silver nanoparticles. The size of undoped and doped (with 1% Bi 3+ ions) NCs were reported 9.79 nm and 10.59 nm, respectively. The optical measurements of NCs revealed a broad orange peak at 580 nm for Bi-doped Cs 2 AgInCl 6 and a blue emission peak at 470 nm for Cs 2 AgInCl 6 . It was also indicated that by doping Bi 3+ in Cs 2 AgInCl 6 NCs, the number of defects would decrease, and radiative localization would be promoted. The long-life time and broadened emission of Bi-doped Cs 2 AgInCl 6 NCs were attributed to self-trapped excitation (STEs) stemming from the Jahn-Teller distortion of [AgCl 6 ] octahedron in the excited state as well as the trivial sub-bandgap defect state transition.
In 2019, Hu et al. [98] prepared colloidal Cs 2 Ag 1−x Na x In 1−y Bi y Cl 6 (x = 0-1; y = 0.03-0.16) NCs by a room temperature recrystallization process. The mean size of NCs were reported to be of 3 nm with diameter of 1.8-4.0 nm which could potentially confine the excitons. The remarkable blueshift photoluminescence was observed for the NCs which was comparable for bulk materials. Incorporation of partial amount of Na + and Bi 3+ resulted in bright near-white light emission with tunable color temperatures from 9759.7 to 4429.2 K. Furthermore, the introduction of Bi 3+ ions along with OA (ligand) passivation in Cs 2 Ag 0.17 Na 0.83 In 0.88 Bi 0.12 Cl 6 nanocrystals led in PLQY of 64% which was the highest reported value for lead-free NCs to date.
In order to investigate the geometric, electric and photoluminescence properties of Mn 2+ -doped Cs 2 AgInCl 6 double halide perovskite, Wu et al. [126] employed firstprinciple calculations. Their study showed that the presence of Mn as a dopant resulted in defect complexes by replacement of Ag + with Mn 2+ atom and causing Ag vacancy (Mn Ag V Ag ), which was due to the charge balance and weak distortion of the metal octahedra. Subsequently, this defect configuration introduced two defect bands in the forbidden gap which was associated to 3d orbitals of the Mn 2+ ions. Therefore, the transition of electron from the first excited state to the ground state led to lower PL spectrum compared to bandgap which would make it beneficial for LEDs.
In 2019, in order to tune the bandgap of Cs 2 AgInCl 6 , microcrystal and colloidal nanocrystals of Yb 3+ doped Cs 2 AgInCl 6 were prepared by Mahor et al. [100] through the solution-process and hot-injection methods, respectively. The concentration of Yb 3+ content, analyzed by Inductively coupled plasma mass spectrometry (ICP-MS) was reported to have 0.1-1.6% for microcrystals and 6.2% for NCs. The Yb-doped samples showed an intense NIR emission centered at~994 nm. Optical measurements of the samples indicated the light at first was absorbed by the light and then non-radiatively transferred to excite the Yb 3+ ions and resulted in the de-excitation of (f) electrons in Yb 3+ ions ( 2 F 5/2 → 2 F 7/2 ). However, the PL decay was different for microcrystals compared to nanocrystals doped perovskites. The NCs exhibited a biexponential decay of 3 ms and 749 µs, and the microcrystals showed a single-exponential decay of Yb-emission with a lifetime of 2.7 ms. The samples also showed high stability under ambient environment.
In 2020, the Pb-free double halide perovskite Cs 2 NaInCl 6 :Sb 3+ was prepared by Gray et al. [101] through precipitation from an HCl solution with the aim of photoluminescent properties investigation. The PXRD analysis of doped and undoped compounds indicated Fm-3m crystal symmetry, a = 10.553344(4) Å with rock salt fully ordering of In 3+ and Na 3+ ions sites. The compound showed a large bandgap of~5.1 eV. However, the substitution of In 3+ with Sb 3+ resulted in strong absorption in the UV because of 5s 2 → 5s 1 5p 1 transitions of [SbCl 6 ] 3− . Through the transition from 3 P 1 → 1 S 0 with radiative relaxation back to 5s 2 ground state, strong blue luminescence at 445 nm with a PLQY of 79% was observed. Nevertheless, with increasing of Sb 3+ content in Cs 2 NaInCl 6 , more than 3% the PL intensity decreased. Furthermore, Cs 2 NaInCl 6 :Sb 3+ showed smaller Stocks shift (0.94 and 1.38 eV) compared to vacancy ordered double perovskite Cs 2 SnCl 6 , which was due to the change of coordination number from 6 in Cs 2 NaInCl 6 to 5 in Cs 2 SnCl 6 .
In a similar study, Zeng et al. [102] showed that doping of 10% Sb 3+ in Cs 2 NaInCl 6 perovskite would break the parity forbidden transition as well as modulating of density of state (DOS) population which led to an optimized blue PLQY of 78.9% assigned to be self-trapped excitons (STEs).

Cs 2 M + In 3+ X 6 : Applications
• Non-Photovoltaic Applications In 2017, Luo et al. [127] fabricated single crystals of Cs 2 AgInCl 6 by thermodynamic synthesis method for UV photodetection. The prepared single crystals showed light yellow color on surface with colorless interior. This phenomenon was explained by different compositions of sample which means oxygen or oxygen containing functional groups change the surface compositions, consequently the optical properties of sample would change. Furthermore, the optical measurements of sample verified the existence of parity-forbidden transition in Cs 2 AgInCl 6 , which had been already demonstrated by Yan et al. [53] through theoretical calculation. It was suggested that the large difference between experimentally obtained optical bandgap (3.2 eV) and photoluminescence emission energy (2.1 eV) was caused by parity-forbidden transitions. However, the fabricated UV photodetection device exhibited ultralow trap-density of 8.6 ± 1.9 × 10 8 cm −3 which was comparable with Pbbased perovskites value (1.80 ± 1.07 × 10 9 cm −3 ), as well as high ON-OFF ratio of around 500, fast photo response of 1 ms, low dark current of 10 pA at 5 V bias and high detectivity of 10 12 Jones.

Cs 2 M + Sb 3+ X 6 : Theoretical Results
In 2017, Tran et al. [125], as presented in Table 5, exhibited a new design for engineering the convergence of direct and indirect bandgap in double halide perovskites based on chemical adjustment of (s) and (p) orbitals character in CBM. Because of differences in orbital overlaps, the relative crystals momenta of VBM and CBM determine whether a bandgap is direct or indirect. This means that bands derived by s-orbitals will increase in energy from Γ to X in a cubic Brillion zone while the bands derived by p-orbitals reduce in energy. Therefore, if the conduction band can be adjusted from s orbitals to p orbitals with a negligible change in the valance band consequently, the difference would result in shift from direct to indirect bandgap. By means of this theory, they examined the feasibility of their design strategy with the experimental preparation of Cs 2 AgInCl 6 , Cs 2 AgSbCl 6 and Cs 2 Sb x In 1−x Cl 6 (x = 0.5, 0.4 and 0.2). Their experimental study successfully demonstrated that Cs 2 AgInCl 6 and Cs 2 AgSbCl 6 showed direct and indirect bandgaps, respectively. By employing the optimized amount of 60%: 40% for In 3+ : Sb 3+ in Cs 2 Sb x In 1−x Cl 6 , it was shown that the bandgap changes from indirect to direct. Zhou et al. [130] investigated the color-tuning phenomenon of Cs 2 AgSbCl 6 crystals which occurred during the synthesis procedure by PBE approaches through first-principle calculation while anti-site defect was established as a model. It was shown that with exchanging site-equal Ag and Sb ions, the two allotropes of nearest neighbor (NN) and second nearest neighbor (2NN) were stable with only 7-12 meV per atom larger than balanced structure (E balanced > E NN > E 2NN ), and the relatively small lattice expansion resulted in different bandgaps.
In 2019, Wei et al. [62] investigated the crystallinity and symmetry of Cs 2 AgSbBr 6 double halide perovskites by employing DFT calculations. It was shown that even though there was a similarity between band structure of Cs 2 AgSbBr 6 and Cs 2 AgBiBr 6 , the 5s 5p orbitals in Sb lowered the CBM significantly, which led to smaller bandgap in Cs 2 AgSbBr 6 compared to Cs 2 AgBiBr 6 . Lin et al. [68] presented a strategy for developing quadruple perovskite halides. Through DFT calculations and symmetry analysis, Cs 4 CdSb 2 Cl 12 and Cs 4 CdBi 2 Cl 12 were identified as two stable perovskites with vacancy ordered 3D crystal structure along with 3D electronic dimensionality with direct forbidden bandgaps.

•
Cs 2 M + Sb 3+ X 6 : Single-Crystals, Polycrystalline and Nanocrystals-Based Perovskites Zhou et al. [130] hydrothermally synthesized Cs 2 AgSbCl 6 crystals and demonstrated by increasing the amount of HCl as a solvent from 0.5 to 1.5 mL, the color of prepared powders changed from yellow to near black, which directly influenced on bandgap. While the darker samples showed lower bandgaps and the lighter color samples indicated higher bandgap as depicted in Figure 20a- In 2019, Wei et al. [62] investigated the crystallinity and symmetry of Cs2AgSbBr6 double halide perovskites by employing DFT calculations. It was shown that even though there was a similarity between band structure of Cs2AgSbBr6 and Cs2AgBiBr6, the 5s 5p orbitals in Sb lowered the CBM significantly, which led to smaller bandgap in Cs2AgSbBr6 compared to Cs2AgBiBr6. Lin et al. [68] presented a strategy for developing quadruple perovskite halides. Through DFT calculations and symmetry analysis, Cs4CdSb2Cl12 and Cs4CdBi2Cl12 were identified as two stable perovskites with vacancy ordered 3D crystal structure along with 3D electronic dimensionality with direct forbidden bandgaps.


Cs2M + Sb 3+ X6: Single-Crystals, Polycrystalline and Nanocrystals-Based Perovskites Zhou et al. [130] hydrothermally synthesized Cs2AgSbCl6 crystals and demonstrated by increasing the amount of HCl as a solvent from 0.5 to 1.5 mL, the color of prepared powders changed from yellow to near black, which directly influenced on bandgap. While the darker samples showed lower bandgaps and the lighter color samples indicated higher bandgap as depicted in Figure 20a-c. Vargas et al. [129] by incorporation of Cu 2+ and Sb 3+ through a solution method, prepared a 2D layered double halide perovskite, Cs4Sb2CuCl12, which exhibited a promising direct bandgap of around ~1 eV, due to the unpaired electron in the 3d orbital of Cu 2+ . The crystalline structure of the perovskite comprised of [CuCl6] 4− octahedra which was placed between [SbCl6] 3− layers and corner-shared to [SbCl6] 3− octahedra. The prepared perovskite showed high stability against moisture, light and temperature. Vargas et al. [129] by incorporation of Cu 2+ and Sb 3+ through a solution method, prepared a 2D layered double halide perovskite, Cs 4 Sb 2 CuCl 12 , which exhibited a promising direct bandgap of around~1 eV, due to the unpaired electron in the 3d orbital of Cu 2+ . The crystalline structure of the perovskite comprised of [CuCl6] 4− octahedra which was placed between [SbCl6] 3− layers and corner-shared to [SbCl6] 3− octahedra. The prepared perovskite showed high stability against moisture, light and temperature.
In 2019, Wei et al. [62] prepared a bulk form of Cs 2 AgSbBr 6 double halide perovskite through the hydrothermal method. The prepared samples showed a low indirect bandgap of 1.64 eV for single crystals. The color of as-prepared samples changed from black to brown by increasing the temperature to 480 K for 5 min, which could be ascribed to charge transfer between Sb 3+ and Sb 5+ in a black phase.
Dahl et al. [85] synthesized Cs 2 AgSbCl 6 and Cs 2 AgInCl 6 nanocrystals by using a modified hot injection method. Instead of adding cesium oleate which is regularly used in the synthesis of double halide perovskites acyl halides was added into the solution of metal acetate precursors under ambient environment at mild temperatures. It was found that the concentration and type of acyl halide had a deep effect on synthesized nanocrystals. The prepared crystalline nano-cubes showed an edge length of 10 nm terminated with (200) facets as well as small silver nano-crystallites decorating the cubes. By developing a titration essay, in order to test the stability of prepared NCs, Dahl et al. [85] observed that Cs 2 AgSbCl 6 dissolved in the presence of minimum concentration (0.01-0.1 mM) of octylamine and released more than twice decomposition energy compared to Cs 2 AgInCl 6 and CsPbCl 3 as illustrated in Figure 21a brown by increasing the temperature to 480 K for 5 min, which could be ascribed to charge transfer between Sb 3+ and Sb 5+ in a black phase. Dahl et al. [85] synthesized Cs2AgSbCl6 and Cs2AgInCl6 nanocrystals by using a modified hot injection method. Instead of adding cesium oleate which is regularly used in the synthesis of double halide perovskites acyl halides was added into the solution of metal acetate precursors under ambient environment at mild temperatures. It was found that the concentration and type of acyl halide had a deep effect on synthesized nanocrystals. The prepared crystalline nano-cubes showed an edge length of 10 nm terminated with (200) facets as well as small silver nano-crystallites decorating the cubes. By developing a titration essay, in order to test the stability of prepared NCs, Dahl et al. [85] observed that Cs2AgSbCl6 dissolved in the presence of minimum concentration (0.01-0.1 mM) of octylamine and released more than twice decomposition energy compared to Cs2AgInCl6 and CsPbCl3 as illustrated in Figure 21a,b. A solution of metal acetates (cesium, silver and indium, antimony or bismuth) in xylene with oleic acid and oleylamine is heated to 100 °C, then an acyl chloride precursor is injected to form NCs of the corresponding Cs2AgMCl6 (M = In, Sb, Bi) double perovskite. The reactivity of the acyl chloride precursor is an important factor for tuning nanocrystal formation; (b) analysis of stability for Cs2AgSbCl6, Cs2AgInCl6 and Cs2AgBiCl6 and CsPbCl3 NCs. Individual points represent the fraction of initial absorption remaining after 4 h as a function of the different concentrations of octylamine that the nanocrystal solutions were exposed to. Reproduced with permission from [85]. Copyright 2019, American Chemical Society.
Lin et al. [68] synthesized Cs4CdSb2Cl12 and Cs4CdBi2Cl12 3D perovskites by solvothermal method in order to support their theoretical results. The steady-state PL exhibited warm orange emission, while the transient PL showed carrier recombination lifetime of microseconds at low temperature.
Garcia-Espejo et al. [63] by employing a mechanochemical approach, prepared polycrystalline inorganic Cs2AgSbBr6 double perovskites. Bromide derivative salts were added to a high energy ball mill with different molar ratio under atmospheric condition. Figure 21. (a) reaction sketch for the preparation of double halide Cs 2 AgSbCl 6 , Cs 2 AgInCl 6 and Cs 2 AgBiCl 6 perovskite NCs. A solution of metal acetates (cesium, silver and indium, antimony or bismuth) in xylene with oleic acid and oleylamine is heated to 100 • C, then an acyl chloride precursor is injected to form NCs of the corresponding Cs 2 AgMCl 6 (M = In, Sb, Bi) double perovskite. The reactivity of the acyl chloride precursor is an important factor for tuning nanocrystal formation; (b) analysis of stability for Cs 2 AgSbCl 6 , Cs 2 AgInCl 6 and Cs 2 AgBiCl 6 and CsPbCl 3 NCs. Individual points represent the fraction of initial absorption remaining after 4 h as a function of the different concentrations of octylamine that the nanocrystal solutions were exposed to. Reproduced with permission from [85]. Copyright 2019, American Chemical Society.
Lin et al. [68] synthesized Cs 4 CdSb 2 Cl 12 and Cs 4 CdBi 2 Cl 12 3D perovskites by solvothermal method in order to support their theoretical results. The steady-state PL exhibited warm orange emission, while the transient PL showed carrier recombination lifetime of microseconds at low temperature.
Garcia-Espejo et al. [63] by employing a mechanochemical approach, prepared polycrystalline inorganic Cs 2 AgSbBr 6 double perovskites. Bromide derivative salts were added to a high energy ball mill with different molar ratio under atmospheric condition. Due to the thermodynamic instability of Cs 2 AgSbBr 6 , XRD for Cs 2 AgSbBr 6 showed varied X-ray diffraction peaks showing the formation of side phases such as AgBr, CsAgBr 2 and Cs 3 Sb 2 Br 9 . It was demonstrated that 2D layered of Cs 3 Sb 2 Br 9 is more stable than 3D Cs 2 AgSbBr 6 double perovskite [131]. The Cs 2 AgSbBr 6 bandgap from Tauc plot was estimated to be 1.93 eV.
Deng et al. [132] integrated Cs 2 AgSbCl 6 powder and Cs 2 AgSbCl 6 /TiO 2 heterojunction nanoparticles through solution state method and hydrothermal process, respectively. Optical bandgaps of 2.60 eV from optical absorption curve for Cs 2 AgSbCl 6 were obtained. However, the optical absorption was seen improved for Cs 2 AgSbCl 6 /TiO 2 heterojunction sample in the visible region as a result of interface states formation and lowered bandgap showing the facilitation of the photo-induced optical transitions. The comparison of charge transfers of Ag 2 Sb 2 Cl 8 /TiO 2 and Cs 4 Cl 4 /TiO 2 indicated that photo-induced carrier separation was more efficient at Cs 4 Cl 4 /TiO 2 interface.

Cs 2 M + Sb 3+ X 6 : Doping
Tran et al. [125] through the solid-state technique synthesized Cs 2 AgSbCl 6 perovskite and Cs 2 AgSb x In 1−x Cl 6 (x = 0.5, 0.4 and 0.2) were prepared by combining the single crystals of hydrothermally synthesized Cs 2 AgInCl 6 and Cs 2 AgSbCl 6 in a stoichiometric ratio. UV-Vis diffuse reflectance measurement along with Tauc plot demonstrated that by increasing the Sb composition in Cs 2 AgSb x In 1−x Cl 6 (x = 0.5, 0.4 and 0.2 and 0), the bandgap would shift from direct to indirect while the value decreased from 3.53 eV to 2.54 eV in Cs 2 AgSbCl 6 .
Karmakar et al. [103] by investigating of Cs 2 AgSbCl 6 and Cu 2+ -doped Cs 2 AgSbCl 6 double halide perovskites reported of a well-ordered crystal structure with integration of Cu 2+ ions into the lattice where Ag + ions were replaced by Cu 2+ . The results of optical measurements showed that Cu 2+ ions had a direct effect on reduction of bandgap from 2.56 eV (Cs 2 AgSbCl 6 ) to 1.02 eV for Cu-doped Cs 2 AgSbCl 6 (x = 0.1 Cu 2+ ). The prepared Cu-doped Cs 2 AgSbCl 6 perovskite exhibited significant stability up to 365 days. The DFT calculation demonstrated small carrier effective masses (>0.4 m e ). Furthermore, it was found that using Cu 2+ ion as dopant increased the conductivity of semiconductor. Kshirsagar et al. [104] prepared Cs 2 AgSb 1−x Bi x Cl 6 alloy NCs with 0 ≤ x ≤ 1 by employing the hot-injection method. Incorporation of Bi 3 in Cs 2 AgSb 1-x Bi x Cl 6 NCs (x = 0.36) raised the PL emission intensity to the maximum level of 2.74 eV. Also, a broadened red-shift emission was observed at 2.17 eV which was ascribed to the carrier-phonon coupling leading intrinsic self-traps. The average length of 10 nm was characterized by TEM images, which proved that even with incorporation of Bi, the size and shape of NCs remained unchanged. The absorption spectra of Cs 2 AgSbCl 6 exhibited an absorption peak at 3.45 and 4.08 eV. Since the edge length of synthesized NCs were larger than Bohr radius (1.02 nm), and due to the absence of quantum confinement effect in NCs, no modifications in absorption spectrum were observed. However, the addition of Bi in Cs 2 AgSb 1−x Bi x Cl 6 (x = 1), decreased the bandgap from 3.45 eV to 3.39 eV as a result of larger spin-orbit coupling strength of Bi as well as anti-site disorders as shown in Figure 22a,b.
incorporation of Bi, the size and shape of NCs remained unchanged. The absorption spectra of Cs2AgSbCl6 exhibited an absorption peak at 3.45 and 4.08 eV. Since the edge length of synthesized NCs were larger than Bohr radius (1.02 nm), and due to the absence of quantum confinement effect in NCs, no modifications in absorption spectrum were observed. However, the addition of Bi in Cs2AgSb1−xBixCl6 (x = 1), decreased the bandgap from 3.45 eV to 3.39 eV as a result of larger spin-orbit coupling strength of Bi as well as anti-site disorders as shown in Figure 22a,b. To the best of our knowledge, there have not been any significant studies on Sb 3+based double perovskite for PV application. The only fabricated PSC based on Cs 2 AgSbBr 6 thin film was reported by Wei et al. [62] in 2019 with a very low photovoltaic efficiency (0.01%), which was attributed to the presence of secondary phases with large bandgaps.

Cs-Based Vacancy-Ordered Double Halide Perovskites
Vacancy-ordered double perovskites are another form of double perovskite in which one B-site cation is replaced by a vacancy and the other B-site cation is in a oxidation state of B 4+ (A 2 M 1+ M 3+ X 6 → A 2 M 4+ X 6 → A 2 M 4+ X 6 ; where indicates a vacancy). This category of perovskite materials also has the close-packed anionic lattice like ABX 3 . While in ABX 3 structure, the stability of the perovskite is predicted by Goldschmidt tolerance factor, in A 2 BX 6 structures the radius ratio is calculated by A-site cation radius to the 12-coordinate void [133][134][135][136] as shown in Figure 23a To the best of our knowledge, there have not been any significant studies on Sb 3+based double perovskite for PV application. The only fabricated PSC based on Cs2AgSbBr6 thin film was reported by Wei et al. [62] in 2019 with a very low photovoltaic efficiency (0.01%), which was attributed to the presence of secondary phases with large bandgaps.

Cs-Based Vacancy-Ordered Double Halide Perovskites
Vacancy-ordered double perovskites are another form of double perovskite in which one B-site cation is replaced by a vacancy and the other B-site cation is in a oxidation state of B 4+ (A2M 1+ M 3+ X6 → A2□M 4+ X6 → A2M 4+ X6; where □ indicates a vacancy). This category of perovskite materials also has the close-packed anionic lattice like ABX3. While in ABX3 structure, the stability of the perovskite is predicted by Goldschmidt tolerance factor, in A2BX6 structures the radius ratio is calculated by A-site cation radius to the 12-coordinate void [133][134][135][136] as shown in Figure 23a

Cs 2 M 4+ X 6 : Theoretical Results
In 2014, Lee et al. [137], as provided in Table 6, investigated the band structure of perovskite by DFT calculations and it was shown that Cs 2 SnI 6 possessed a direct bandgap of 1.3 eV at the Γ point comprised of filled I-5p and empty I-6p/Sn-5p orbitals which contributed to VBM and CBM, respectively. Both VBM and CBM were dispersed in energy resulting in high charge carrier mobility of Cs 2 SnI 6 .
Maughan et al. [106] through DFT calculations indicated that the phonon interaction was stronger in Rb 2 SnI 6 compared to Cs 2 SnI 6 . This was caused by the larger number of nondegenerate lower frequency phonons which contributed to the lattice dielectric response and decreased charge carrier mobilities. Debbichi et al. [138] theoretically studied the atomic and electronic band structures of single crystal and polycrystal Cs 2 Au 2 I 6 . It was shown that Au demonstrated mixed-valence of +1 and +3 in the B-site, which made the Cs 2 Au 2 I 6 to exhibit the same electronic structure as double perovskite materials resulting in an optimal bandgap of 1.31 eV. An in-depth optical simulation done by Debbichi et al. [138] suggested that polycrystalline Cs 2 Au 2 I 6 would be a good candidate for PV application because of its remarkable optical absorption and electrical performance such as small effective masses and long diffusion length. It was predicted by employing polycrystalline Cs 2 Au 2 I 6 as an active layer in PSC, the short-circuit-current of 30 mA cm −2 and photoconversion efficiency of 20% could be obtained. They used full-wave electromagnetic simulations based on the finite element methods (FEM) to do their calculation.
In 2017, Ju et al. [139] introduced a new class of vacancy-ordered perovskite based on Ti 4+ ion. Their experimental and theoretical results indicated that both Cs 2 TiI 2 Br 4 and Cs 2 TiBr 6 possessed the ideal bandgap values of 1.38 eV for single junction and 1.78 eV for tandem solar cells. Moreover, both materials exhibited good environmental stability. Sakai et al. [140], by employing DFT/GCA calculations showed that the electron and hole effective masses of Cs 2 PdBr 6 were to be 0.53 and 0.85 m e , respectively, which resulted in n-type semiconductivity in Cs 2 PdBr 6 . Lee et al. [137] prepared Cs 2 SnI 6 perovskite as a stable molecular iodosalt serving as a hole transporting layer in solid-state DSSCs. This iodosalt compound which Sn is on its +4-oxidation state showed high stability against moisture and air compared with CsSnI 3 and MASnI 3 perovskites.
In 2019, Kong et al. [141] adopted a solution-based process for synthesis of Cs 2 TiX 6 (x = Cl and Br) at room temperature. The highly uniform and thermally stable crystals and thin film of Cs 2 TiBr x Cl 1−x (0 < x < 1) were prepared. The obtained materials showed a quasi-direct bandgap of 1.7 eV for Cs 2 TiBr 6 , 1.95 eV for Cs 2 TiBr 2 Cl 4 and 2.5 eV for Cs 2 TiCl 6 . Furthermore, steady-state PL exhibited an emission peak centered at~535 nm for Cs 2 TiCl 6 ,~635 nm for Cs 2 TiBr 2 Cl 4 and~670 nm for Cs 2 TiBr 6 . The FWHM values for all prepared samples were larger than 100 nm, which is comparable with Pb-based perovskites. However, in order to obtain more efficient thin films and for engineering the bandgaps, they suggested different synthesis approaches and alloying with other metallics or halide elements were required. Sakai et al. [140] by employing the solution process technique, synthesized Cs 2 PdBr 6 halide perovskite. During the synthesis process, the oxidation state of Pd 2+ changed to Pd 4+ which was generated in situ. The obtained compound crystallized in a cubic structure with a space group of fm-3m. The optical measurements indicated that compound had an indirect bandgap of 1.6 eV. In order to study the sample photoconductivity a sandwich structure was fabricated including ITO/Cs 2 PdBr 6 /Ag. The carried-out experiment confirmed the feasibility of Cs 2 PdBr 6 perovskite for different optoelectronic applications such as LEDs, PV and photon-sensors.

•
Cs 2 M 4+ X 6 : Film-Based Chen et al. [87], as already presented in Table 2, by employing the facile low-temperature vapor deposition methodology, prepared a high-quality thin film of Cs 2 TiBr 6 halide perovskite. The synthesized thin film demonstrated an optimal bandgap of 1.78 eV with a carrier diffusion length of more than 100 nm.

Cs 2 M 4+ X 6 : Doping
In 2016, the solid-solution of Cs 2 Sn 1−x Te x I 6 was prepared by Maughan et al. [136] in order to investigate of its structure-property relationship. The substitution of Sn by Te causing the increase of electronic dispersion due to the closer contact distances of I-I bonding resulted in reduction of conductivity, carrier mobility and carrier concentration. DFT calculation revealed the hindered formation of intrinsic iodine vacancy donor defects resulting in insulating characteristics of Cs 2 TeI 6 . However, the theoretical calculation of Cs 2 SnI 6 native defects showed a low enthalpy of iodine vacancies formation and a level of defect energy that was a shallow donor to the conduction band. This makes the material tolerant to defect states. In Te-doped compound, the covalency of Te-I bonding suppresses the formation of iodine vacancy state causing the reduction in conductivity.
Maughan et al. [106] studied the electrical and structural changes created by substitution of Cs + with Rb + . Both Cs 2 SnI 6 and Rb 2 SnI 6 showed a native n-type semiconductivity. However, the replacement of Cs + with Rb + decreased the charge carrier mobility bỹ 50 times compared to Cs 2 SnI 6 that makes Cs 2 SnI 6 as an interesting material for optoelectronic applications.
In 2018, Tan et al. [107] selected vacancy-ordered Cs 2 SnCl 6 perovskite as a host and Bi 3+ as the luminescence dopant. The prepared Cs 2 SnCl 6 : Bi perovskite showed an intense rise of PLQY (78.9%) with an emission peak at 445 nm. Furthermore, the doping of Bi resulted in narrowing of bandgap from 3.9 eV in Cs 2 SnCl 6 to 3.0 eV in Cs 2 SnCl 6 : Bi due to the generating of defect bands [Bi Sn + V Cl ]. The formation of BiOCl layer as well as stable oxidation state of Sn 4+ enhanced the thermal and water stability. Also, the combination of Cs 2 SnCl 6 : Bi with commercial yellow phosphors along with commercial UV-LED chips led to high warm light emission with a correlated color temperature of 4486 K and a commission Internationale de I'Eclairage (CIE) coordinate of (036; 0.37).
In 2019, Ma et al. [108] studied the effect of Ge 4+ substitution on CsSn 1−x Ge x I 6 (x = 0.25, 0.5, 0.75, 1) properties. First-principle calculation demonstrated that the concentration of Ge 4+ and the value of the bandgap had linear relationship. The crystal cells of traditional ABX 3 perovskites were easily influenced by doping small cations which caused tilting and contraction of the cell [142]. In this study, doping of Ge did not result in tilting up the BI 6 octahedra, but the contraction of crystal cell occurred, which gave rise to the reduction of bandgap. Employing DOS and PDOS calculations, it was shown that doping of Ge 4+ led to the change in composition of CBM and the bandgap. In 2017, Qui et al. [86], as already provided in Table 2, by developing a two-step sequential deposition method, prepared B-γ-CsSnI 3 thin film for PV application. However, under the ambient environment the oxidation of Sn changed from 2+ to 4+, which resulted in vacancy-ordered air-stable Cs 2 SnI 6 perovskite with a bandgap of 1.48 eV. The best power conversion efficiency of 0.96% was measured for PSC with perovskite film of 300 nm thickness, exhibiting the V oc of 0.51 V, J sc of 5.41 mA cm −2 and FF of 35. The inefficient electron/hole extraction towards the electrodes was explained by the mismatched energybarrier between TiO 2 /Perovskite/HTL layers resulting in poor efficiency of~1% as shown in Figure 24a-c. oxidation state of Sn 4+ enhanced the thermal and water stability. Also, the combination of Cs2SnCl6: Bi with commercial yellow phosphors along with commercial UV-LED chips led to high warm light emission with a correlated color temperature of 4486 K and a commission Internationale de I'Eclairage (CIE) coordinate of (036; 0.37). In 2019, Ma et al. [108] studied the effect of Ge 4+ substitution on CsSn1−xGexI6 (x = 0.25, 0.5, 0.75, 1) properties. First-principle calculation demonstrated that the concentration of Ge 4+ and the value of the bandgap had linear relationship. The crystal cells of traditional ABX3 perovskites were easily influenced by doping small cations which caused tilting and contraction of the cell [142]. In this study, doping of Ge did not result in tilting up the BI6 octahedra, but the contraction of crystal cell occurred, which gave rise to the reduction of bandgap. Employing DOS and PDOS calculations, it was shown that doping of Ge 4+ led to the change in composition of CBM and the bandgap. In 2017, Qui et al. [86], as already provided in Table 2, by developing a two-step sequential deposition method, prepared B-γ-CsSnI3 thin film for PV application. However, under the ambient environment the oxidation of Sn changed from 2+ to 4+, which resulted in vacancy-ordered air-stable Cs2SnI6 perovskite with a bandgap of 1.48 eV. The best power conversion efficiency of 0.96% was measured for PSC with perovskite film of 300 nm thickness, exhibiting the Voc of 0.51 V, Jsc of 5.41 mA cm −2 and FF of 35. The inefficient electron/hole extraction towards the electrodes was explained by the mismatched energybarrier between TiO2/Perovskite/HTL layers resulting in poor efficiency of ~1% as shown in Figure 24a-c.  A Ti-based perovskite solar cell was fabricated first by Chen et al. [87]. The insertion of C 60 between TiO 2 -ETL layer and Cs 2 TiBr 6 resulted in promising efficiency of 3.3% with V oc of 1.02 V, J sc of 5.69 mA cm −2 and FF of 0.564 in a reverse scan.

Conclusions and Perspective
Despite the significant power conversion efficiency of Pb-based perovskites for PV applications, two main drawbacks including the instability and toxicity of Pb hinder their large-scale fabrication and commercialization. To develop non-toxic and air-stable photo-absorbers for PSCs, double halide perovskites with a formula of A 2 M + M 3+ X 6 were suggested and investigated as a new design strategy.
In this review, theoretical and experimental results of Cs-based Pb-free double halide perovskites, influence of metal doping/alloying on selected perovskites along with current potential applications are highlighted. A thorough review of Cs 2 M + M 3+ X 6 structures with special emphasis on (Bi 3+ , In 3+ , Sb 3+ ) as M 3+ elements is carried out. In addition, Cs-based vacancy ordered double halide perovskites with formula of Cs 2 M 4+ X 6 are also reviewed.
Based on the obtained results from different research methodologies and procedures, it is clear that Cs-based Pb-free double halide perovskites have enhanced thermal and ambient stability, but the bandgap values of these materials are not optimal for PV applications. However, many studies have demonstrated that bandgap of these materials could be tuned through metal doping/alloying.
Among the various studied Cs 2 M + M 3+ X 6 structures, Bi 3+ -based double perovskites have shown promising features for PV applications with an estimated indirect bandgap varying from ≈1.7 eV [56] to 2.89 eV [59] which could be tuned from indirect to direct by metal doping [91]. Also, to the best of our knowledge, the latest study based on Cs 2 AgBiBr 6 showed the best photovoltaic efficiency of 3.11% [66] with unique features such as high environmental stability, low toxicity and long carrier recombination lifetime. Moreover, significant optical and electronic properties with remarkable performances, turn the Cs 2 AgBiBr 6 into an outstanding candidate for other optoelectronic applications such as LEDs, UV detectors, X-ray detectors etc. Processes related to employing appropriate coating methods and managing bulk engineering of Cs 2 AgBiBr 6 structure, along with managing interface engineering by for example introducing interfacial layers between ETL/Cs 2 AgBiBr 6 /HTL have to be concentrated for further enhancement of efficiency of PSCs based on this promising material.
Studies have revealed that In 3+ -based double perovskites have large direct bandgaps varying from ≈3.23 eV to 3.57 eV [85,124]. It was indicated that the parity-induced forbidden transition in Cs/In 3+ -based double halide perovskites causes the optical transition obstruction and makes these materials unsuitable for solar devices, but these materials showed inherent environmental stability against air/moisture. Also, it was shown through alloying of metal ions, not only the PL properties of Cs/In-based double halide perovskites were significantly enhanced, an important feature for LEDs, but also the large direct band gap of this compound was tuned from~3.5 eV to 2.54 eV [125].
The Cs/Sb 3+ double halide perovskites were shown to have large indirect bandgap varying from ≈1.64 eV (single crystals) to 3.0 eV [62,68], that hinders them of showing desirable PV performance. Cs/Sb 3+ double halide perovskites are thermodynamically unstable. Amongst vacancy-ordered lead-free double halide perovskites reviewed, Cs 2 TiBr 6 showed the relatively higher efficiency of 3.3% [87] along with high stability under ambient environment.
Our review shows that by employing different theoretical and experimental approaches in the studies has had a significant influence on the results which describe optoelectronic characteristics of Cs-based double halide perovskites. For example, it was shown that synthetic methodologies and conditions (temperature, the concentration of precursors, time, solvents) and employing different types of characterization instruments (UV-Vis spectroscopy vs. DRS) led to obtaining varying photophysical and structural results, so in order to reduce/eliminate these discrepancies, as this review clearly indicates, a combined theoretical and experimental approach will be the best way to get in-depth understanding of lead-free double halide perovskite materials. It also helps to address the issues like bandgap alignment, carrier and interfacial dynamics between perovskite and transporting layers.
Furthermore, plenty of studies in order to employ suitable electrodes, HTL and ETL layers along with efficient interfacial layers could be carried out since these layers have a significant effect on PSCs efficiency. Exploring more on Cs-based double perovskite materials especially on Cs/Bi 3+ -based lead-free double halide perovskites may help the researchers to overcome the toxicity and instability challenges faced with lead-based perovskite materials for different applications in PV, photocatalysis, photodetectors, lightemitting devices etc.