EP4294889A1

EP4294889A1 - Crystalline compound

Info

Publication number: EP4294889A1
Application number: EP22706100.9A
Authority: EP
Inventors: Henry James Snaith; Harry SANSOM; Giulia LONGO; Benjamin PUTLAND; Matthew Jonathan ROSSEINSKY; Michael Pitcher; John Claridge; Matthew Dyer
Original assignee: Oxford University Innovation Ltd
Current assignee: Oxford University Innovation Ltd
Priority date: 2021-02-19
Filing date: 2022-02-21
Publication date: 2023-12-27
Also published as: WO2022175687A1; GB202102360D0

Abstract

The invention relates to a crystalline compound comprising one or more first cations, A; one or more second cations, B; one or more third cations, B'; one or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B', and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B', wherein at least some of the one or more third cations, B', are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment.

Description

CRYSTALLINE COMPOUND FIELD OF THE INVENTION The invention relates to a crystalline compound, and to a semiconductor device comprising the crystalline compound. The invention also relates to processes for preparing the crystalline compound. BACKGROUND TO THE INVENTION Hybrid lead perovskites APb²⁺X₃ (A = FA⁺/MA⁺/Cs⁺, X = Br-/I-) continue to be intensely studied as solar absorbers for photovoltaic (PV) applications due to their high absorption coefficients suitable for thin film technology, long charge-carrier diffusion lengths, and high radiative efficiencies. In single-junction devices the current certified record power conversion efficiency (PCE) stands at 25.5%. This is close to matching the very highest efficiencies delivered by silicon PV cells, and high PCEs are a way to minimise the cost of energy from PV. However, this becomes increasingly difficult as heavily optimised systems approach their maximum theoretical efficiency limits. A crucial strategy to overcome this limitation is to combine wider band gap (E_g) materials (E_g from 1.6–2.0 eV) with well-established c-Si (E_g ≈ 1.1 eV) technology to construct tandem cells, which can achieve much higher PCEs than single-junction cells. The tuneability of the band gap in mixed iodide–bromide lead-halide perovskites has opened up the possibility of multi-junction solar cells with c-Si, currently delivering a record PCE of 29.5%, with efficiency improvements to over 32% feasible. There remain a number of compromises which could be improved upon with the discovery of new wide band gap, stable, lead-free, inorganic solar absorber materials. These include the yet unresolved challenge of obtaining band gap stable, low- defect I–Br mixed halide perovskites; the reliance upon organic ammonium cations to deliver a crystallographically phase stable lead-halide perovskite compound, which lead to lower thermal stability than conventional inorganic semiconductors; and finally these materials contain lead, which requires careful management due to the known toxicological issues. In single junction devices the current certified record power conversion efficiency (PCE) stands at 25.2%. Combining the heavily studied mid band gap (1.4–1.6 eV) materials with lower (1.0–1.2 eV) and higher (1.8–2.0 eV) band gaps allows the construction of tandem solar cells with the potential to achieve higher PCEs than for single-junction solar cells. While the bandgap tunability of hybrid perovskites has opened up the possibility of multi-junction solar cells with over 30% PCE, the achievable efficiencies are still limited by the poor electronic quality and poor stability of the mixed iodide-bromide perovskites. For these reasons, development of new materials in the bandgap range 1.8-2.0 eV is particularly important. There are a number of ways to chemically engineer higher band gap materials, including; substituting Br- for I- in APb(I_1-yBr_y)₃ perovskites, using lower dimensional phases such as the 2D A’₂A_n-1Pb²⁺ nI_3n+1 Ruddlesden–Popper phases (where A’= phenylethylammonium (PEA)⁺, butylammonium (BA)⁺ and A = FA⁺/MA⁺/Cs⁺), and lead-free materials with 3D octahedral networks such as double perovskite Cs₂AgBiBr₆. APb²⁺X₃ phases with mixed halides Br- and I- suffer from halide or phase segregation during illumination, and exhibit instabilities towards light in the presence of oxygen, moisture, and heat which all involve a breakdown of the material usually by decomposition of the organic cation, leaving soluble and toxic PbI₂. The Ruddlesden–Popper phases studied have also predominantly contained lead, and 2D octahedral networks have the added complication of charge-carrier confinement which means for high performance devices careful control of crystallization is required to grow the film with a preferred orientation in which the 2D perovskite layers are aligned orthogonally to the substrate. Candidates to replace Pb²⁺ (Sn²⁺, Ge²⁺, Sb³⁺ and Bi³⁺) are currently being investigated, the last of which is isoelectronic with Pb²⁺. Many Sn²⁺ and Ge²⁺ perovskites have been investigated. The ASnI₃ perovskites (A = Cs⁺, MA⁺, FA⁺)² have band gaps of 1.2–1.4 eV and the AGeI₃ perovskites (with A = Cs⁺, MA⁺, FA⁺) have band gaps of 1.6–2.2 eV. However, Sn²⁺ and Ge²⁺ readily oxidize to Sn⁴⁺ and Ge⁴⁺ in air which significantly lowers the device performance. There are also reports that the decomposition products of the tin-containing perovskites should be considered as toxic as those from lead-based materials. Antimony is considered toxic, whereas bismuth is not. The use of BiOCl and BiC₆H₅O₇ (bismuth citrate) in cosmetics and BiOCl in drugs is allowed by the U.S. Food and Drug Administration (FDA). One strategy for replacing Pb²⁺ is with isoelectronic Bi³⁺. Bismuth bromide and chloride networks have been synthesised as double perovskites (MA)₂KBiCl₆ (3.04 eV), (MA)₂AgBiBr₆ (2.02 eV), Cs₂AgBiCl₆ (2.77 eV) and Cs₂AgBiBr₆ (2.19 eV), however they currently remain unsuitable to be useful in tandem cells: in the Cl-containing materials the band gaps are too wide to be combined efficiently with current efficient technologies (hybrid Pb perovskites, Si, CdTe); Cs₂AgBiBr₆ has an unsuitable absorption coefficients containing a minima after the initial absorption onset; and (MA)₂AgBiBr₆ contains the volatile CH₃NH₃ cation. Bismuth iodides A₃Bi₂ ³⁺I₉ (A = K⁺, Rb⁺, Cs⁺, MA⁺, NH₄ ⁺) have been reported with large band gaps as 2D perovskites (A = NH₄ ⁺, K⁺, Rb) or as 0D isolated [Bi₂I₉]³⁺ units (A = Cs⁺, MA⁺) which, based on the current understanding and chemical manipulation of the 2D Pb-containing hybrid perovskites, are not ideal for isotropic charge transport and carrier mobility. This leaves suitable 3D networks of [BiI₆]^3- octahedra very rare. Hypothetical bismuth iodide double perovskites, such as Cs₂AgBiI₆, would possess a lower more ideal band gap but so far have not been stable enough to be synthesised, apart from in nanocrystal form, which identifies a clear opportunity for new materials discovery. Therefore, there is a clear need for materials with a lower band gaps than those of the bismuth-containing perovskites mentioned above, which are also thermally stable and easy to manufacture. Searching for other possible bismuth iodide networks with suitably lower band gaps leads to BiI₃ and the ternary compounds Ag_1-3xBi_1+xI₄ and CuBiI₄. BiI₃ has been reported with an indirect band gap of 1.67(1) eV and devices have reached PCEs of 1.0%. Ag_1-3xBi_1+xI₄ and CuBiI₄ have been reported with suitable band gaps of 1.64–1.93 eV; the variation arising from composition, sample type, and assuming direct or indirect band gaps. Cu-containing CuBiI₄ films have also recently been processed into devices reaching PCEs of 1.1%. However, CuBiI₄ is not a stable phase and decomposes when standing at room temperature. Devices based on a x = −0.33 Ag1-3xBi1+xI4 (Ag3BiI6) solar absorber have reached PCEs of 4.3%, and introducing small amounts of sulfur to the layer has recently been shown to increase the J_sc of devices, increasing the maximum PCE to 5.44(7)%. Various structures based on edge sharing AgI₆ and BiI₆ octahedra fitting the general formula A_aB_bX_x where x=a+3b have been investigated, such as Ag₂BiI₅. Cu-doped Ag₂BiI₅ in which the Cu is said to replace Ag⁺ at the octahedral site of the hexagonal Ag₂BiI₅ structure is disclosed in J. W. Park et al, Sustainable Energy Fuels, 2021, DOI: 10.1039/D0SE01563F. However, these materials have the disadvantage of containing large amounts of expensive silver, and also may be photosensitive due to the large number of Ag-I bonds present. There therefore exists a need to provide new materials with tuneable electronic properties, which address the limitations noted for perovskite materials and which have good stability and which are inexpensive to produce. SUMMARY OF THE INVENTION The present invention provides compounds which have useful electronic and optical properties and which do not have the disadvantages associated with the prior art. In particular, the inventors have surprisingly found that a new class of quaternary compounds address the problems outlined above. Further, these compounds typically do not have the conventional perovskite structure widely investigated in relation to electronic devices (e.g. semiconductor and optoelectronic devices, such as photovoltaics). Instead, these compounds have cations in both octahedral and tetrahedral coordination environments. By expanding this family of materials to the new quaternary systems with different structures to the conventional perovskite materials that are the subject of the prior art, an extra degree of chemical tuneability is gained, which can be further optimised to increase performance and stability of the material. In particular, the invention provides compounds that (i) have a band gap of a suitable size for optoelectronic applications, in particular a band gap small enough for use in tandem cells, (ii) may be made from easily available, environmentally friendly, non-toxic materials (i.e. for instance do not require toxic lead compounds) and (iii) have a good stability profile. These compounds also exhibit good charge-carrier mobility, high absorption coefficients, long average photoluminescence lifetimes and low exciton binding energies, and can be manufactured as thin films suitable for integration into optoelectronic devices. For instance, the new compound Cu₂AgBiI₆ has been synthesised as crystals, powder, and solution-processed thin films. Cu₂AgBiI₆ represents the use of Ag⁺ to stabilise CuBiI₄, and the use of Cu⁺, which is typically too small to include in the larger 8 and 12 coordinate sites in conventional perovskite materials, to reduce the content of expensive Ag⁺ compared to Ag1-3xBi1+xI4 compounds. Simulations involving this material indicate that there is the potential to deliver over 30% power conversion efficiency (PCE) when it is incorporated into a Cu₂AgBiI₆-on-Si tandem cell. The present invention provides a crystalline compound comprising (i) One or more first cations, A; (ii) One or more second cations, B; (iii) One or more third cations, B’; (iv) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment. The invention also provides a film comprising a compound as described herein. The invention also provides a semiconductor device comprising a compound as described herein or a film as described herein. The invention also provides a process for producing a film comprising a crystalline compound comprising (i) One or more first cations, A; (ii) One or more second cations, B; (iii) One or more third cations, B’; (iv) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, said process comprising disposing a film-forming solution on a substrate wherein the film- forming solution comprises the one or more first cations, A, the one or more second cations, B, the one or more third cations, B’, and the one or more anions, X. The invention also provides a process for producing a film comprising a crystalline compound comprising (i) One or more first cations, A; (ii) One or more second cations, B; (iii) One or more third cations, B’; (iv) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, said process comprising: exposing the substrate to a vapour, which vapour comprises one or more reactants for producing said compound and allowing deposition of the vapour onto the substrate to produce a film of said crystalline compound thereon. The invention also provides the use of a compound as described herein as an optoelectronic material, preferably as a photovoltaic material or as an electroluminescent material. BRIEF DESCRIPTION OF THE FIGURES Figure 1(a) shows the Cu₂AgBiI₆ structure solved from 100K SCXRD data with the composition constrained in line with the average composition Cu_2.15(16)Ag_1.04(5)Bi_0.92(7)I_6.00(11) from TEM EDX. Figure 1(b) shows layer 1 containing the sites Oct1 (occupied by 34.6% Ag⁺ and 30.6% Bi³⁺) and Cu1 with coordination environments shown in Figure 1(c) and Figure 1(d), respectively. Figure 1(e) shows that layer 2 contains site Cu2 with coordination environment shown in (f). Sites Cu1 and Cu2 are both occupied by 17.9% Cu⁺. The I–I distances and I–I–I angles of the cubic close- packed (CCP) iodide sub-lattice in the bc and ab planes are shown in Figure 1(g) and Figure 1 (h), respectively. The green, blue, grey, and pink spheres/polyhedra represent I-, Cu⁺, Ag⁺ and Bi³⁺ ions, respectively. Figure 1(i) shows the Pawley fit performed on room temperature laboratory PXRD data. Figure 2(a) shows the absorption coefficient of Cu₂AgBiI₆ thin films (black) measured by a combination of Fourier Transform Infra-Red (FTIR) spectroscopy and Photothermal Deflection Spectroscopy (PDS). This is compared to the reported absorption coefficients of MAPbI₃ (blue dotted) and Cs₂AgBiBr₆ (blue dashed), reproduced from Davies et al., and Longo et al., respectively (Longo et al, Understanding the Performance-Limiting Factors of Cs₂AgBiBr₆ Double- Perovskite Solar Cells. ACS Energy Lett.2020, 2200-2207; Davies et al., Bimolecular Recombination in Methylammonium Lead Triiodide Perovskite is an Inverse Absorption Process. Nat. Commun. 2018, 9 (1), 293; Wright et al., Electron–Phonon Coupling in Hybrid Lead Halide Perovskites. Nat. Commun.2016, 7 (1), 11755). Also shown is the photoluminescence (PL) spectrum of Cu₂AgBiI₆. Figure 2(b) shows Elliott model fitting of the absorption coefficient spectrum, giving a band gap of 2.06(1) eV and exciton binding energy of 25(2) meV. Figure 2(c) shows the partial density of states of Cu₂AgBiI₆ computed with density functional theory for configuration of cations with the lowest computed energy. The cumulative contributions from each species are shown along with the total density of states for energies relative to the computed Fermi energy. Figure 3(a) shows the natural logarithm of the absorption coefficient measured by Photothermal Deflection Spectroscopy(PDS), after it was scaled to match the FTIR absorption coefficient data (raw data and scaled data shown in Figures 14a and b, respectively), showing sub-band gap states down to 1.25 eV. (b) Transient photoluminescence measurement on Cu₂AgBiI₆ (red) fitted by a stretched exponential function (black) with an average lifetime of 33 ns, compared to that of Cs₂AgBiBr₆ (blue) with an average lifetime pf 10 ns, reproduced from Longo et al. Understanding the Performance-Limiting Factors of Cs₂AgBiBr₆ Double-Perovskite Solar Cells. ACS Energy Lett. 2020, 2200-2207. The Cu₂AgBiI₆ and Cs₂AgBiBr₆ films were measured at excitation fluences of 200 nJ cm^-2 and 240 nJ cm^-2, respectively. Figure 4(a) shows the structure of the cell in the optical modelling LiF/ITO/SnO₂/C₆₀/Cu₂AgBiI₆/PolyTPD/ITO/nc-SiO_x:H/(i)a-Si:H/c-Si/(i)a-Si:H/(p)a- Si:H/AZO/Ag. Figure 4(b) shows the photovoltaic (PV) band gap, defined as the energy of the inflection point on the absorption edge, for Cu₂AgBiI₆ films of varying thicknesses. Figure 4(c) shows the external quantum efficiency (EQE) and Figure 4(d) shows the J-V curve of a simulated Cu₂AgBiI₆ on c-Si tandem solar cell, using a transfer matrix optical model, coupled with detailed balance. The diode parameters used were extracted from lead halide perovskite and c-Si J-V curves reported in literature. Figure 5 shows the Pawley fit to room temperature laboratory PXRD data of CuBiI4. The PXRD pattern is fitted using a cubic unit cell with space group Fd3̅m, previously reported by Fourcroy et al. Structure du Tetraiodure de Cuivre(I) et de Bismuth(III), CuBiI₄. Acta Crystallogr. C 1991, 47 (10), 2023-2025. The extracted lattice parameter is a = 12.1580(2) Å. Figure 6 shows SEM EDX of the powder corresponding to nominal composition x = 0 Cu_1-3xBi_1+xI₄, giving an average composition of Cu_1.21(5)Bi_1.11(7)I_4.00(9), and occasionally referred to as CuBiI₄ throughout the text. Figure 7 shows the PXRD patterns of (a) CuBiI₄ as synthesised (b) CuBiI₄ kept in the dark in air at room temperature for three weeks, and (c) CuBiI₄ kept in the dark at -20°C for three weeks. The main peak of the BiI₃ decomposition phase is marked by the asterisk. Figure 8 shows the PXRD patterns of AgBiI₄ and Cu₂AgBiI₆ powders after being exposed to one week in the solar spectrum, sealed in capillaries in air. Also shown are the PXRD of the controls, which were kept in the dark in air. The range shown includes where the largest peaks would appear for possible decomposition phases AgI (black tick marks), CuI (red tick marks) and BiI₃ (blue tick marks), should they have been present. Figure 9(a) shows the Raman spectra of AgBiI₄ and Cu₂AgBiI₆ control samples, showing the characteristic two peaks, which occur at low wavenumbers. The Raman spectra of AgBiI₄ and Cu₂AgBiI₆ powders after being exposed to the AM1.5 solar spectrum in different atmospheres are shown in Figures 9(b) and (c), respectively. There are no significant changes in the spectra of either materials. The change in ratio of the two peaks in the AgBiI₄ He 1 week spectra is likely due to the disorder of the Ag⁺ and Bi³⁺ cations. Figure 10 shows TEM EDX compositions of particles scraped off from the solution processed film with nominal composition of Cu₂AgBiI₆ in solution. Figure 11 shows a Pawley fit of PXRD pattern for Cu₂AgBiI₆ films spincoated on FTO glass, fitted to a rhombohedral phase associated with a quaternary phase (black tick marks), a rhombohedral phase consistent with a Cu_3xBi_1-xI₃ phase (red tick marks), and SnO₂ (blue tick marks). Figure 12 shows SEM images of (a, b) rough, large, dendritic grains obtained during the early stages of solution processing the Cu₂AgBiI₆ films. Figure 13(a, b) show SEM images of the uniform, smooth morphology of the optimised Cu₂AgBiI₆ film at x1001 and x13362 magnification, respectively. Figure 14(a) shows the raw PDS data of Cu₂AgBiI₆ thin films. (b) The PDS data (red) is scaled to match the absorption coefficient measured by FTIR (black). Figure 15 shows the partial density of states plots for the lowest energy configurations arising from each of the four independent ChemDASH calculations shown in Figure 24. The electronic structure is computed using the meta-GGA functional, SCAN,(Sun, J. et al., Strongly Constrained and Appropriately Normed Semilocal Density Functional. Phys. Rev. Lett.2015, 115 (3), 036402) and including spin-orbit coupling. Shown on each plot is the value of the smallest gap between occupied and unoccupied bands. The plot for configuration 1, the lowest energy configuration overall is also shown in Figure 2c, but is included again here for comparison. Although changing the configuration of Ag⁺, Bi³⁺ and Cu⁺ cations within the computed cell does affect the position and shape of peaks in the density of states, in all configurations the Cu 3d-states dominate at the valence band edge, and the conduction band consists of mixed Bi 6p-states and I 5p-states. Figure 16 shows the reported photoluminescence (red) and absorption coefficient (black) of MAPbI₃ thin films, reproduced from Wright et al. and Davies et al. respectively (Wright, A. D. et al., Electron–Phonon Coupling in Hybrid Lead Halide Perovskites. Nat. Commun.2016, 7 (1), 11755; Davies, C. L. et al., Bimolecular Recombination in Methylammonium Lead Triiodide Perovskite is an Inverse Absorption Process. Nat. Commun.2018, 9 (1), 293.). Figure 17(a) The J-V curves of the forward (FB) to short circuit (SC) and the SC-FB scans for the device using Cu₂AgBiI₆ as the solar absorber. (b) The steady state performance, measured at the maximum power point, presenting good stability with both the current density and the PCE increasing over time. Inset is the device architecture used. Figure 18 shows the optical response of a Cs₂AgBiBr₆ on silicon tandem was modelled using the transfer matrix method (Katsidis, C. C.; Siapkas, D. I., General Transfer-Matrix Method for Optical Multilayer Systems with Coherent, Partially Coherent, and Incoherent Interference. Appl. Opt. 2002, 41 (19), 3978-3987). The thickness of the Cs₂AgBiBr₆ (100 nm – 1000 nm) and the anti- reflective coating LiF (10nm – 150 nm) was varied using a differential evolution algorithm till the limiting tandem current was maximised. The optimised device structure was LiF (110 nm)/ITO (80 nm)/SnO2 (5 nm)/C60 (10 nm)/Cs₂AgBiBr₆ (989 nm)/PolyTPD (5 nm)/ITO (25 nm)/nc-SiO_x:H (110 nm)/i-a-Si:H (5 nm)/i-c-Si (250 um)/i-a-Si:H (5 nm)/p-a-Si:H (5 nm)/AZO (70 nm)/Ag (400 nm). The c-Si sub-cell (32.9 mA/cm²) generates six times as much as current as the Cs₂AgBiBr₆ cell (5.5 mA/cm²). This huge current mismatch makes Cs₂AgBiBr₆ impractical as the top cell of a Si based tandem. Figure 19 shows the external quantum efficiencies (EQE) (left) and the J-V curves (right) of simulated Cu₂AgBiI₆ films of varying thickness on c-Si tandem cells, using a transfer matrix optical model, coupled with detailed balance. The diode parameters used were extracted from lead halide perovskite and c-Si J-V curves reported in literature (see SI for more details). The search bounds for Cu₂AgBiI₆ film thicknesses were 1200–1400 nm, 800–1000 nm, and 400–530 nm for the films that optimised at thicknesses of 1389 nm, 999.9 nm and 529.6 nm, respectively. LiF film thickness optimised at 97.3 nm, 96.8 nm, 95.3 nm for the 1389 nm, 999.9 nm and 529.6 nm Cu₂AgBiI₆ film thicknesses, respectively. Figure 20(a) shows the TEM EDX of particles from the Cu₂AgBiI₆ powder synthesis, showing how quenching from 350°C, rather than cooling slowly to room temperature, gives a compositionally homogeneous sample. (b) The samples are also more compositionally homogeneous if the melt is avoided. Figure 21(a) shows the SEM image of the Cu₂AgBiI₆ crystal used for the SCXRD structural solution. The crystal was picked out of the Cu₂AgBiI₆ powder synthesis. Figure 21(b) shows the SEM EDX showing the composition of the crystal to be within error of the powder. Figure 22 shows PXRD patterns for the compositional screening of Cu_3xBi_1-xI₃ carried out between 0.18 ≤ x ≤ 0.32. A pure CuBiI₄ powder was obtained for the nominal composition corresponding to x = 0.25. For values of x < 0.25 BiI₃ impurities were seen (black asterisks). For values of x > 0.25 small amounts of CuI impurities were seen which appear on the shoulders of the CuBiI₄ peaks (red asterisks). The expanded view of the peak at 1.8 Å^-1 in the PXRD patterns of x > 0.25, shows the amount of CuI increases with increasing x above 0.25. Figure 23 shows the “synthetic” extinction co-efficient for Cu₂AgBiI₆ which was used as input for the optical calculations, as described in the optical modelling in example 1. Figure 24 shows the optimised geometry of the lowest energy configurations arising from each of the four independent ChemDASH calculations computed using the van der Waals functional, optB86b-vdW (Klimeš, J. et al., Van der Waals Density Functionals Applied to Solids. Phys. Rev. B 2011, 83 (19), 195131). The computed energy for each configuration is given relative to configuration 1, the lowest energy overall. The green, blue, grey and pink spheres/polyhedra represent I⁻, Cu⁺, Ag⁺ and Bi³⁺ ions respectively. These low energy configurations represent different ordered arrangements of the cations consistent with the average structure determined experimentally. Figure 25 (a) shows the known ternary and new synthesized quaternary phases in the CuI–AgI–BiI₃ phase field and figure 25 (b) the forms in which they have been synthesized. Figure 25 (c) shows the average compositions of the materials synthesized in this investigation, measured with either SEM EDX (AgBiI₄, CuBiI₄) or TEM EDX (CuAgBiI₅, Cu₂AgBiI₆), with 1σ errors (blue areas). The red spots in Figure 25(c) are the stoichiometric compositions chosen to represent the materials and are within error to the measured compositions. Figure 25 (d) shows the areas of the CuI–AgI– BiI₃ phase field explored during the investigation, in example 2. Figure 26 shows the unit cell nomenclature and cubic close-packed (CCP) iodide sub-lattices for the ternary and quaternery compounds in the CuI–AgI–BiI₃ phase field. The unit cell types of the ternary and quaternary compounds referred to throughout the main text are highlighted in bold. Ternary compounds can crystallize in different structure types each associated with a different unit cell. Figure 26 (a) shows the cubic unit cell of Ag_1-3xBi_1+xI₄ (x≥0) and CuBiI₄, and Figure 26 (b) shows the small trigonal unit cell of Ag_1-3xBi_1+xI₄ (x≤0), CuBiI₄ and Cu₂BiI₅. It is helpful to directly compare the structures by transforming both these cells in to a large trigonal unit cell using the transformation matrices shown. Figure 26 (c) shows a large trigonal cell that has a and b directions double that of the small trigonal cell with a volume four times as large, and is √3/2 the volume of the cubic unit cell. The quaternary compounds CuAgBiI₅ and Cu₂AgBiI₆ crystallize in the large and small trigonal unit cells, respectively. Figure 26 (d) shows rhombohedral strain is defined in the case of the small and large trigonal cells. It can be described as extending or contracting the otherwise cubic structure along the body diagonal (111)_cubic, shown by the red arrows. The green spheres represent the iodide ions. Figure 27 shows the nomenclatures used for the different octahedral network of ternary and quaternary compounds in the CuI–AgI–BiI₃ phase field. Figure 27a) shows the defect-spinel octahedral motif consists of a disordered spinel motif (Figure 49). The octahedral motif is a 3D network of edge-sharing octahedra and consists of an overall occupancy of octahedral (O) sites of ½. Using the transformation matrix shown, it can be represented in the large trigonal unit cell where it can be considered as alternating between layers of 3/4 and 1/4 O site occupancy, maintain the overall 1/2 O site occupancy. Figure 27b) shows the 2D CdCl₂ octahedral motif consists of alternating between layers of full O site occupancy, and empty layers, giving overall 1/2 O site occupancy. Figure 27c) shows the i-CdCl₂ octahedral motif consists of every possible O site being occupied, with layered ordering. The layered ordering means that the layers alternate between two different octahedral sites. Purple and yellow octahedra represent octahedral sites Oct1 and Oct2, respectively. Green spheres represent iodide ions. Unit cells are drawn with solid black lines. Figure 28a) shows the tetrahedral site in the spinel structure. One channel in the spinel octahedral motif is highlighted in yellow (top), and the same channel is selected after a 90° anti-clockwise rotation (bottom). The orange and red spheres represent Mg²⁺ and O^2- ions, respectively. The blue octahedra represent Al–O octahedra. Figure 28b) shows the Cu sites in the CuAgBiI₅ defect-spinel octahedral motif, showing that the site has partial layered ordering by only occupying the layers with 1/4 O site occupancy (top). The channel highlighted in red shows that Cu site is different from that in the spinel (bottom). Figure 28c) shows the Cu sites in the small trigonal unit cell (Cu₂BiI₅ and Cu₂AgBiI₆) showing how they have partial layered ordering and occupy every possible T site. The Cu⁺ and I- ions are blue and green, respectively. The purple octahedra represent the Bi–I octahedra. Figure 29(a) shows the new CuBiI₄ structure in the small trigonal cell refined using laboratory powder X–ray diffraction (PXRD) data. Figure 29(b) shows the octahedral coordination environment for site Bi1, occupied by 50% Bi³⁺. Figure 29(c and d) show the tetrahedral coordination environments for sites Cu1 and Cu2, respectively. Cu1 and Cu2 are occupied by 15% and 9% Cu⁺, respectively. Figure 29(e and f) show the I–I distances and I–I–I angles of I- sub- lattice, along the c-direction, and in the ab plane, respectively. The green, blue, and pink spheres represent I-, Cu⁺ and Bi³⁺ ions, respectively. Figure 30(a) shows the Pawley fit of high resolution synchrotron CuAgBiI₅ PXRD data with a cubic unit cell, showing the fit is not good for high Q-spacing due to rhombohedral strain. Figure 30(b) shows the Pawley fit of the same data using a large trigonal cell. Figure 30(c, d, e) show the measured, calculated and difference of the Rietveld refinement of the CuAgBiI₅ structure to a combination of high resolution synchrotron CuAgBiI₅ PXRD (MAC I11, Diamond Light Course, Oxfordshire, UK), HRPD bank 1 NPD , and HRPD bank 2 NPD (ISIS neutron and Muon Source, Oxfordshire, UK), respectively, shown in d-spacing Figure 31 (a-i) shows the new CuAgBiI₅ room temperature structure refined using combined high resolution synchrotron PXRD data and neutron powder diffraction (NPD). Figure 31(a-ii, a-iii) show the octahedral coordination environment for sites Oct1 and Oct2, respectively. Oct1 and Oct2 are both occupied by 42.1(1)% Bi³⁺ and 41.4(2)% Ag⁺. Figure 31 (a-iv) shows the tetrahedral coordination environment for site Cu1, occupied by 17.3(2)% Cu⁺. Figure 31(a-v, a-vi) show the I– I distances and I–I–I angles of I- sub-lattice, along the c-direction, and in the ab plane, respectively. The green, blue, gray, and pink spheres represent I-, Cu⁺, Ag⁺ and Bi³⁺ ions, respectively. Figure 32(a) shows the relationship between the type of octahedral motif (dimensionality) and overall octahedral (O) site occupancy (composition) for CuAgBiI₅, Cu₂AgBiI₆, CuBiI₄ and reported structures AgBi₂I₇, Ag₂Bi₃I₁₁, AgBiI₄, Ag₂BiI₅, Ag₃BiI₆. Increasing overall O site occupancy takes the octahedral motif from 2D to 3D. Figure 32(b) shows the octahedral motif of ternary and quaternary phases in the CuI–AgI–BiI₃ phase field along the solid solution lines between BiI₃ to AgI (green), AgBiI₄ to CuI (blue) and BiI₃ to CuI (bottom axis). Figure 32(c) shows the structures along the AgBiI₄ to CuI solid solution line and the extra control given by the quaternary system by moving in different directions along the solid solution line, which can balance the need for decreasing expensive and photosensitive Ag⁺, increasing dimensionality, and increasing stability. Figure 33 shows PXRD patterns for the compositional screening of Cu_3xBi_1-xI₃ carried out between 0.18 ≤ x ≤ 0.32. A pure CuBiI₄ powder was obtained for the nominal composition corresponding to = 0.25. For values of x < 0.25 BiI₃ impurities were seen (black asterisks). For values of x > 0.25 small amounts of CuI impurities were seen which appear on the shoulders of the CuBiI₄ peaks (red asterisks). The expanded view of the peak at 1.8 Å^-1 in the PXRD patterns of x > 0.25, shows the amount of CuI increases with increasing x above 0.25. Figure 34 shows a Pawley fit to laboratory PXRD data of the synthesis with nominal composition Cu(Ag_1.5Bi_0.5)I₄. The red tick marks correspond to the Cu_0.45Ag_0.55I impurity peaks, and the black tick marks to the rhombohedral Ag-rich x < 0 Ag_1-3xBi_1+xI₄ phases. Figure 35 shows a Pawley fit to the laboratory PXRD pattern of nominal composition Cu_1.5Ag_0.5BiI₅. The tick marks for the CuI impurity, BiI₃ impurity, and the cubic Fd3^തm phase are black, red and blue, respectively. The lattice parameter of the Fd3^തm phase (a = 12.2050(3) Å) is larger than that reported for CuBiI4 (a = 12.1580(2) Å) and smaller than that reported for AgBiI4 (a = 12.21446(4) Å), suggesting a quaternary phase containing both Cu and Ag. Figure 36 shows the laboratory PXRD pattern of the CuxAg1-xBiI4 series. These PXRD patterns are indexed to a mixture of three phases; a zinc blende Cu_1-yAg_yI phase (black tick marks and asterisks), BiI₃ (red tick marks and asterisks), and a cubic Fd3^തm phase (blue tick marks and asterisks). The peaks marked around 2.9 Å^-1 are used to show the presence of all three phases in all of the patterns. Figure 37 shows refined lattice parameters of (a) the Fd3^തm cubic phase and (b) zinc blende Cu_1- _yAg_yI phase over the Cu_xAg_1-xBiI₄ solid solution series. Lattice parameters were refined against an internal LaB₆ standard using a Pawley fit. Figure 38 shows the Cu_4x(AgBi)_1-xI₄ solid solution corresponds to a line of compositions between the known phases CuI and AgBiI₄. Samples were synthesized at x = 0.14, 0.15, 0.16, 0.18, 0.2, 0.21, 0.23, 0.24, 0.26, 0.27, 0.29, 0.31, 0.33, 0.35, 0.37, 0.43, 0.5. The refined lattice parameters of the main cubic Fd3^തm phase and Cu_1-yAg_yI impurity (multiplied by 2) are shown in this figure. For 0.16 < x < 0.18 there could be a pure quaternary phase. For 0.29 ≤ x ≤ 0.31 Cu_4x(AgBi)_1-xI₄ the Cu_1- _yAg_yI impurity lattice parameters are half that of the quaternary phase peaks, showing that these samples could be pure, but alternatively it could be that the Cu_1-yAg_yI impurity peaks cannot be resolved from the main phase peaks. Figure 39(a) shows a selected Q range of the PXRD pattern for the 0.14 ≤ x ≤ 0.16 Cu_4x(AgBi)_1-xI₄ _{samples. The red asterisks highlight peaks due to a rhombohedral R3} ^ത _{m Ag-rich Ag1-3xBi1+xI4} impurity phase, no Cu_1-yAg_yI impurity is seen. Figure 39 (b) shows a selected Q range of the PXRD pattern for the 0.18 ≤ x ≤ 0.20 samples. The blue asterisks highlight peaks due to a Cu_1-yAg_yI impurity phase, no rhombohedral R3^തm Ag-rich Ag_1-3xBi_1+xI₄ impurity phase is seen. Figure 39 (c) shows TEM EDX showing the measured composition of particles from the x = 0.16 and 0.18 samples, showing them to be compositionally inhomogeneous along the solid solution line. Figure 40 shows the Cu_4x(AgBi)_1-xI₄ solid solution corresponds to a line of compositions between the known phases CuI and AgBiI₄. Samples were synthesized at x = 0.14, 0.15, 0.16, 0.18, 0.2, 0.21, 0.23, 0.24, 0.26, 0.27, 0.29, 0.31, 0.33, 0.35, 0.37, 0.43, 0.5. (a) The Pawley fit of the peak which would include the largest CuI impurity peak for 0.27 ≤ x ≤ 0.37. The black line is the measured intensity, the red line is the calculated PXRD pattern, the blue line is the contribution of the CuI impurity phase. For x > 0.27 and x = 0.31 the inclusion of CuI only slightly improves peak shape; which makes the samples be interpreted as pure. (b) The TEM EDX composition of particles from the 0.27 ≤ x ≤ 0.33 samples, showing them to be compositionally inhomogeneous along the Cu_4x(AgBi)_1-xI₄ solid solution line. Figure 41(a) shows the TEM EDX of particles from the x = 0.31 samples, showing how quenching from 350°C, rather than cooling slowly to room temperature, gives a compositionally homogeneous sample. Figure 41(b) shows the samples are also more compositionally homogeneous if the melt is avoided. These results are shown here for Cu₂AgBiI₆ but are also true for CuAgBiI₅. Figure 42(a) is a SEM image of the CuAgBiI₅ crystal used for the SCXRD structure. the crystal was picked out of the CuAgBiI₅ powder sample. Figure 42(b) shows the SEM EDX measurements of the crystal (red points) compared to the TEM EDX measurements of powder samples (black). Figure 43(a) shows the SEM image of the Cu₂AgBiI₆ crystal used for the SCXRD structural solution. The crystal was picked out of the Cu₂AgBiI₆ powder synthesis. Figure 43(b) shows the SEM EDX showing the composition of the crystal to be within error of the powder. Figure 44(a) shows SEM EDX measurements of large melt-grown and CVT-grown crystals. Figure 44(b) is an SEM image of melt-grown crystal. Figure 44(c) is an SEM image of CVT-grown crystal. (d) SEM image of the close-up of CVT-grown crystal surface. The crystals lack a well- defined shape as would be expected from a conventional crystal. Figure 45(a) shows the TEM EDX compositions of particles taken from the CuAgBiI₅ sample and repeats synthesized as explained in example 2. The a and c lattice parameters obtained by fitting a R3^തm large trigonal unit cell in a Pawley fit are shown in Figure 45(b) and (c), respectively. These samples consisted of 0.25g batches which were combined in to a larger batch for NPD. Figure 46 shows the binary compounds CuI, AgI and BiI₃ have all been reported to crystallize with a close-packed iodide sub-lattice.^24-26 CuI crystallizes in a cubic unit cell with a cubic-close packed (CCP) iodide sub-lattice. At room temperature AgI crystallizes in a hexagonal unit cell with a hexagonal-close packed (HCP) iodide sub-lattice. BiI₃ crystallizes in a trigonal unit cell with a HCP iodide sub-lattice. Figure 47(a) shows the zinc blende structure, Figure 47(b) shows the tetrahedral coordination, Figure 47(c) shows the I–I distances and I–I–I angles of the I- sub-lattice of CuI. In CuI the Cu⁺ is tetrahedrally-coordinated and fully ordered. The Cu⁺ tetrahedra corner-share to form the zinc blende structure with the F4^ത3m space group. The CCP sub-lattice has two tetrahedral holes per iodide anion, so that for the composition of CuI only 1/2 of the tetrahedral (T) sites are occupied. Figure 47(d) shows the wurtzite structure, Figure 47(e) shows the tetrahedral coordination, Figure 47(f) shows the I–I distances and I–I–I angles of the I- sub-lattice of AgI. The Ag⁺ is fully ordered on a site that is tetrahedrally coordinated. The tetrahedra corner share to form the wurtzite structure. Half of the total T sites are occupied for composition AgI. Figure 47(g) shows the BiI₃ structure, Figure 47(h) shows the tetrahedral coordination, Figure 47(i) shows the I–I distance and I–I–I angles of the I- sub-lattice of BiI₃. The Bi³⁺ is octahedrally coordinated, which edge-share to create a layered 2D network of octahedra. BiI₃ has an octahedral motif consisting of layers of 2/3 octahedral (O) site occupancy, separated by empty layers of no O site occupancy, giving an overall O site occupancy of 1/3 (Figure 54). Green, blue, gray and pink spheres represent the I-, Cu⁺, Ag⁺ and Bi³⁺ ions, respectively. Figure 48 shows the 2D BiI₃ octahedral motif consisting of layers of 2/3 octahedral (O) site occupancy, separated by empty layers of no O site occupancy, giving an overall O site occupancy of 1/3. Green and purple spheres represent the I- and Bi³⁺ ions, respectively. Figure 49 shows the spinel octahedral motif is a 3D network of edge-sharing octahedra and consists of an overall occupancy of octahedral (O) sites of ½. Red and purple colors represent the O^2- and Al³⁺ ions, respectively. Figure 50 shows the three different tetrahedral sites in the previously reported CuBiI₄.structure. Figure 50a) The Cu1 site is located where the layers of ¾ and ¼ O site occupancy meet, which is eight times in the unit cell, giving a multiplicity of eight. The Cu1 tetrahedra shares faces with four neighbouring octahedra. Figure 50b) The empty channels left by the 50% O site occupancy give rise to 12 linear channels per unit cell, which connect to make a 3D network. Each channel has four Cu⁺ ions which are not shared by any other channel. The 12 channels of four Cu⁺ ions give the multiplicity of 48. Figure 50c) The eight Cu⁺ ions of site Cu3 are at the eight places that four separate channels cross over in the unit cell. Figure 50c) highlights one of the Cu⁺ ions, and the four channels that it is shared between. Figure 50d) Cu2 and Cu3 sites share the channels. Figure 50e) All three sites occupy every possible T site in the CCP sub-lattice. Figure 51 shows the Rietveld refinements of laboratory PXRD data collected at room temperature for the CuBiI₄ material, fitted to (a) the reported cubic Fd3^തm CuBiI₄ structure with a defect-spinel octahedral motif and (b) the new CuBiI₄ structure with the CdCl₂ octahedral motif. Figure 52 shows the CuBiI₄ structure as reported by Fourcroy et al. transformed in to a structure with the CdCl₂ octahedral motif, using the transformation matrix shown. Figure 53(a) shows the CuBiI₄ structural model corresponding to a 50% defect-spinel and 50% CdCl₂ octahedral motif. This is achieved by fixing the octahedral occupancy of octahedral sites common to both octahedral motifs. Then the octahedral sites that are only present in the defect- spinel and CdCl₂ octahedral motifs are fixed to half this value. Octahedral sites shared by both types of octahedral motifs are in black, the defect-spinel in orange, and the CdCl₂ in red. The Cu sites (blue) fill every tetrahedral hole facilitated by the iodide sub-lattice (green). Figure 53(b) shows the change in goodness of fit parameter for Rietveld refinements with respect to the ratio of defect-spinel:CdCl₂ cation motif in the model used in the refinement. Figure 54 shows the PXRD patterns of AgBiI₄, CuAgBiI₅ and Cu₂AgBiI₆ powders after being exposed to one week in the solar spectrum, sealed in capillaries in air. Also shown are the PXRD of the controls, which were kept in the dark in air. The range shown includes where the largest peaks would appear for possible impurity AgI (black tick marks), CuI (red tick marks) and BiI₃ (blue tick marks), should they have been present. Figure 55(a) shows the Raman spectra of AgBiI₄, CuAgBiI₅ and Cu₂AgBiI₆ control samples, showing the characteristic two peaks, which occur at low wavenumbers. The Raman spectra of AgBiI₄, CuAgBiI₅ and Cu₂AgBiI₆ powders after being exposed to the solar spectrum in different atmospheres are shown in Figure 55b, c and d, respectively. An additional peak appears in the CuAgBiI₅ spectrum after one week in the AM1.5 solar spectrum in all atmospheres (*). Figure 56 shows the TEM EDX compositions of particles scraped off from the solution processed film with nominal composition of Cu₂AgBiI₆ in solution. Figure 57 shows the Pawley fit of PXRD pattern for Cu₂AgBiI₆ films spincoated on FTO glass, fitted to a rhombohedral phase associated with a quaternary phase (green tick marks), a rhombohedral phase consistent with a Cu_3xBi_1-xI₃ phase (blue tick marks), and SnO₂ (pink tick marks). Measured using Cu wavelength. Figure 58 shows SEM images of (a, b) rough, large, dendritic grains obtained during the early stages of solution processing the Cu₂AgBiI₆ films. Figure 59(a) shows the raw PDS data of Cu₂AgBiI₆ thin films. Figure 59(b) shows the PDS data (black) is scaled to match the absorption coefficient measured by FTIR (red). Figure 60 shows the transient photoluminescence measurement on Cu₂AgBiI₆ fitted by a stretched exponential function (red) with an average lifetime of 33 ns. Measured at 720 nm using TCSPC, as described in the Experimental Methods section of example 2. Figure 61 shows the density of states at the valence band maximum (VBM) of CuAgBiI₅ and Cu₂AgBiI₆ powders, as measured by XPS. Cu₂AgBiI₆ has some more states just below the VBM than CuAgBiI₅, suggesting the extra states are due to Cu. Figure 62(a) shows the Pawley fit of XRD data collected on a CuAgBiI₅ thin film deposited on a microscope slide, fitted to a large trigonal cell associated with CuAgBiI₅ with refined lattice parameters a = 8.724(1) Å and c = 20.800(5) Å. Figure 62(b) shows the composition of 9 points of the film measured by SEM EDX, with an average composition of Cu_0.82(5)Ag_0.96(9)Bi_1.07(4)I_3.98(13) corresponding to an iodine deficit of 20(3)%. Figure 62 (c) is an image of a CuAgBiI5 film, showing uniformity but a rough surface caused by the morphology seen in the SEM images in Figures 62 (d) and (e). Figure 62(f) is an image of the film when backlit by a white LED, showing the transmission of red light. Figure 63(a) shows the steady-state optical absorption coefficient measured for three thin films of CuAgBiI₅ using a Bruker Vertex 80v Fourier-Transform Infrared (FTIR) spectrometer. The shaded area indicates the standard error in the measurements from variation in thickness across the films. The inset shows the same data, without the error, on a logarithmic scale to make the onset at 900 nm clearer. Figure 63(b) shows the steady-state photoluminescence spectra measured for three fresh thin films of CuAgBiI₅ in vacuum following excitation by a continuous-wave laser at 398 nm with an excitation density of 40.5 Wcm^-2. Measured data are shown in crosses and lines show a ten- point moving average. The sharp peak at 800 nm is due to the second reflection of the laser signal from the diffraction grating in the detection setup. Figure 63(c) shows the time-resolved photoluminescence decays for a fresh thin film of CuAgBiI₅ measured in vacuum using Time- Correlated Single Photon Counting (TCSPC) following excitation by a 398 nm pulsed laser at a repetition rate of 5 MHz. The lowest-fluence decay was fitted with a stretched exponential, shown in yellow, yielding an average lifetime of 0.73 ns. Figure 63(d) shows the time-resolved photoluminescence decays for the same sample as in (c), measured in air after the time-resolved emission spectra shown in Figure 65. The lowest-fluence decay was again fitted with a stretched exponential, giving an average lifetime of 17.9 ns. Figures 63(e,f) show charge-carrier mobilities measured across for two thin films of CuAgBiI₅ using Optical-Pump Terahertz-Probe spectroscopy at fluences of 4.9, 12.5, 25 μJcm^-2. The measured values and their experimental errors are shown in red with error bars, and the mean and standard error are shown as the black line and grey shaded area. Figure 64(a) shows the relationship between octahedral (O) site occupancy and type of octahedral motif formed, giving chemical control over dimensionality of the octahedral network. CuBiI₄ does not fit the trend but it found to be metastable at room temperature. Figure 64(b) shows the same relationship shown in the CuI-AgI-BiI₃ phase space, where the colour map and red contour lines represent O site occupancy. We note that the O site occupancy in (b) is not representative of materials which contain tetrahedral Ag⁺ such as the room temperature structure of AgI. Figure 65(a) shows the time-resolved emission spectra measured in vacuum using TCSPC. The vertical and horizontal lines indicate spectral and transient decay slices, respectively, which are shown in Figures 65(b) and (c). The measured values in Figure 65(b) are shown as crosses, with the lines showing a five-point moving average. Figures 66(a, b) show steady-state photoluminescence spectra measured for two thin films of CuAgBiI₅ after being left in air and darkness for 20, 60, 90 minutes. The PL was measured following excitation by a continuous-wave laser at 398 nm with an excitation density of 39.0 Wcm^-2. The spectra after 20 minutes have the same shape as those measured in vacuum, shown in Figure 63(b). After prolonged exposure to air the PL spectra blue-shift and increase significantly in intensity for both samples, although on slightly different timescales. For clarity, we only plot lines showing the ten-point moving average of the measured data. Figures 66(c, d) show five steady-state photoluminescence spectra for a thin film of CuAgBiI₅ measured under continuous illumination over 15 s by a 398 nm laser at an intensity of 39.0 Wcm^-2 after being left in air for 20 (c) and 90 (d) minutes, respectively. In both cases the spectra do not change shape but do decrease in intensity over 15 s of continuous illumination. Figure 67 shows images and SEM images of solution-processed Cu(Ag_1.5Bi_0.5)I₄ films in Example 3. Figure 68 shows the Pawley fit to a XRD pattern of a Cu(Ag_1.5Bi_0.5)I₄ film in Example 3, fitted to a cubic unit cell, space group Fd3̅m, lattice parameter a = 12.131(2) Å. Figure 69 shows SEM EDX measurements of nominal Cu(Ag_1.5Bi_0.5)I₄ in Example 3 giving an average composition of Cu_0.98(4)Ag_1.46(3)Bi_0.52(2)I_3.24(10) (blue). The compositional spread in Cu, Ag and Bi are smaller for those measured for CuAgBiI₅ (red) and Cu₂AgBiI₆ (green) films. Figure 70 shows the absorption and photoluminescence spectra measured on Cu(Ag_1.5Bi_0.5)I₄ films in Example 3. Figure 71 shows aSEM EDX measurements and average compositions of Cu_4x(AgBi)_1-xI₄ films prepared in Example 4 for nominal compositions x = 0 (AgBiI₄), 0.09 (Cu_0.4AgBiI_4.4), 0.2 (CuAgBiI₅), 0.33 (Cu₂AgBiI₆), 0.6 (Cu₆AgBiI₁₀). Figure 72 shows (from top to bottom) Pawley fits to XRD patterns of Cu_4x(AgBi)_1-xI₄ films for nominal compositions x = 0 (AgBiI₄), 0.09 (Cu_0.4AgBiI_4.4), 0.2 (CuAgBiI₅), 0.33 (Cu₂AgBiI₆), 0.6 (Cu₆AgBiI₁₀). Black tick marks correspond to the main trigonal phases for each fitting. Tick marks of a secondary trigonal phase, AgI, BiI₃, and CuI are shown in orange, red, blue and green respectively. The broad unfitted peaks at 9.7° and 19.2° are due to decomposition product BiOI. There are unidentified peaks in the XRD pattern of Cu₆AgBiI₁₀. Figure 73 shows SEM images of Cu4x(AgBi)1-xI4 films for nominal compositions (a) x = 0 (AgBiI₄), (b) 0.09 (Cu_0.4AgBiI_4.4), (c) 0.2 (CuAgBiI₅), (d) 0.33 (Cu₂AgBiI₆), (e) 0.6 (Cu₆AgBiI₁₀). Figure 74 shows the measured mobility plotted against the closest value of x Cu_4x(AgBi)_1-xI₄ calculated using the average compositions measured by SEM EDX. Figure 75 shows the measured absorption coefficients of Cu_4x(AgBi)_1-xI₄ films. Figure 76 shows (from left to right) images of solution processed Cu_4x(AgBi)_1-xI₄ films for nominal compositions x = 0 (AgBiI₄), 0.09 (Cu_0.4AgBiI_4.4), 0.2 (CuAgBiI₅), 0.33 (Cu₂AgBiI₆), 0.6 (Cu₆AgBiI₁₀). Figure 77 shows annealed (air) and as deposited X-ray diffraction patterns of evaporated Cu₂AgBiI₆ films from Example 5. Data was aligned using the quartz peaks at 16.3° and 33.1° as a reference. Figure 78 shows the top view SEM image of the as deposited film on quartz from Example 5 showing uniform coverage. Figure 79 shows the top view SEM image of uniform film on quartz from Example 5 annealed post-deposition at 110°C in air. Figure 80 shows the top view SEM image of a Cu2AgBiI6 film from Example 5 annealed post- deposition at 130°C in air. The lighter coloured material is silver resulting from beam damage when high electron acceleration voltage and high magnification is used. Upon reducing the magnification quickly, silver is absorbed back into the bulk lattice. Figure 81 shows the large area top view SEM image of a Cu2AgBiI6 film on quartz from Example 5 annealed post-deposition at 150°C in air showing pinholes Figure 82 shows the UV-VIS absorbance of as deposited and annealed Cu₂AgBiI₆ on quartz from Example 5. Figure 83 shows the photoluminescence measurements of as deposited and annealed Cu₂AgBiI₆ film on quartz showing peak photon emission at 720 nm. Figure 84 shows absorption and photoluminescence spectra comparing a Cu₂AgBiI₆ film from Example 5 exposed to 400-700 nm light for one week to a film stored in the dark. Figure 85 shows absorption and photoluminescence spectrum comparing a Cu₂AgBiI₆ film exposed to 400-700 nm light for three weeks to a film stored in the dark. Figure 86 is a photograph of the metallic silver lining the quartz crucibles used to deposit AgI. The yellow powder is trapped silver iodide. Figure 87 shows XRD patterns of fresh AgI powder compared to the residue left in the quartz crucible post-deposition. The XRD pattern belonging to the residue shows the presence of metallic silver. The peak positions of AgI (P 63 mc) and Ag (^^3^ത^) were taken from references [11] and [12] in Example 5 respectively. Figure 88 shows cross sectional SEM images of Cu₂AgBII₆ devices from Example 5 showing voids at the PEDOT:PSS/Cu2AgBII6 interface. The n-i-p device comprises of Glass//ITO//PEDOT:PSS//Cu₂AgBiI₆//PCBM//BCP//Ag. The dashed red box highlights void formation at the PEDOT:PSS//Cu-Ag-Bi-I interface when using quartz crucibles. Figure 89 shows cross section SEM images of Cu₂AgBiI₆ films deposited from quartz crucibles on SnO₂//ITO//Glass from Example 5. The red dashed box highlights the interfacial voids at the SnO₂//Cu₂AgBiI₆ interface. Figure 90 shows cross sectional SEM images of evaporated Cu₂AgBiI₆ films deposited from alumina crucibles on SnO₂//ITO//Glass from Example 5. The Cu₂AgBiI₆ is homogenous across the interface and no voids are observed. Figure 91 shows the fitting of (a) Cu 2p, (b) Ag 3d, (c) Bi 4f, and (d) I 3d core levels of X-ray photoelectron spectroscopy (XPS) data collected for CuAgBiI₅ bulk sample. The fitting of Cu 2p_3/2 is complicated by the presence of the I 3p_1/2, however Cu 2p_1/2 was fit first, and the known spin orbit split energy used. The identification of the charge state of the Cu as Cu⁺ is supported by the lack of any satellite feature of the Cu 2p. Figure 92 shows the valence band maximum (VBM) and conduction band minimum (CBM) positions with respect to vacuum measured for CuAgBiI₅ and Cu2AgBiI₆ bulk samples. The ionisation potential are measured on bulk samples and the optical band gap measured on thin films. Figure 93 shows five steady-state photoluminescence spectra for a thin film of CuAgBiI5 measured under continuous illumination over 15 s by a 398 nm laser at an intensity of 39.0 Wcm^-2 after being left in air for 20 (a) and 90 (b) minutes, respectively. In both cases the spectra do not change shape but do decrease in intensity over 15 s of continuous illumination. Figure 94 shows structures in the CuI-AgI-BiI₃ phase space: binaries CuI, AgI and BiI₃; ternaries Ag₃BiI₆, Ag₂BiI₅, AgBiI₄, AgBi₂I₇, Ag₂Bi₃I₁₁, CuBiI₄, Cu₂BiI₅; and quaternary Cu₂AgBiI₆. Also included is CuAgBiI₅. All these compounds consist of a close-packed iodide sub-lattice with varying arrangements of the cations filling the octahedral and tetrahedral interstitial sites to form the structures shown. Figure 95 shows (a) The three Cu⁺ sites in CuBiI₄ (spinel) colour coded as yellow (site 1), orange (site 2) and red (site 3). A channel in the spinel octahedral (Oct) motif is highlighted in blue, which a cross-section is taken of to show the tetrahedral (Tet) sites inside (sites located behind the channel, which appear to be inside due to the 2D representation of the 3D structure, have been crossed out). The red site (site 3) is the same in spinel. Some, but not all, of the orange sites (site 2) are occupied in CuAgBiI₅. The Tet sites in CuAgBiI₅ can be considered as a reflection of the spinel sites with the mirror plane down the centre of the channel (blue dashed line) (b) The two Cu⁺ sites in the small trigonal unit cell (Cu₂AgBiI₆ and CuBiI₄ (CdCl₂)) showing layered ordering. Also shown are the connectivity’s of the Tet sites, which give a 3D Tet network (c) The layered ordering of Cu⁺ sites in CuAgBiI₅ means they are only in layers with 1/4 Oct interstitial occupancy and do not occupy all the sites associated with tetrahedral site 2 in CuBiI₄ (spinel). (d) The connectivity of tetrahedra in CuAgBiI₅ and spinel, which give 2D and 0D Tet networks, respectively. Figure 96 shows the absorption coefficient and PL measured on CuAgBiI₅ (solid lines) and Cu₂AgBiI₆ (dashed line) thin films. The data for Cu₂AgBiI₆ is taken from Sansom et al. The PL spectra of CuAgBiI₅ and Cu₂AgBiI₆ were measured in vacuum and air, respectively. (b) The density of states of the valence band measured on CuAgBiI₅ (black) and Cu₂AgBiI₆ (red) powders, measured by XPS. (c) The shift and increase in the PL signal of CuAgBiI₅ thin films exposed to air. (d) TRPL of CuAgBiI₅ thin films measured in vacuum (black) and air (blue), compared to Cu₂AgBiI₆, measured in air. Figure 97 shows (a) The room temperature crystal structure of CuAgBiI₅ from Rietveld refinement of combined PXRD and NPD datasets, including coordination environments. (b) A low d-spacing region of the fit, showing that a trigonal cell (ii) fits better than a cubic cell (i). (c) The fits to high resolution synchrotron PXRD (MAC detector, I11, Diamond Light Source, Oxfordshire, UK) and high-resolution NPD (banks 1 and 2, HRPD, ISIS Neutron and Muon Source, Oxfordshire, UK) datasets. Figure 98 shows a cross sectional sketch of the iodine annealing setup, consisting of a metal base plate (1), a rubber gasket ring (2), a glass cover (3), a hole in the baseplate (4) and a valve connected to that hole (5). The samples (6) and the elemental solid iodine (7) are placed under the cover other before closing, evacuating and heating the setup. Figure 99 shows experimental XRD patterns of as-deposited and annealed Cu₂AgBiI₆ films in ambient air. Annealed films were heated to 110°C, 130°C, and 150°C for 15 minutes using a hotplate. Small impurity peaks belonging to BiI₃ and a second quaternary Cu-Ag-Bi-I peak are labelled with (*) and (#) respectively. Peaks highlighted with (+) at 16.3° and 33.2° belong to the quartz substrate. There is no difference in the phases present in films annealed in air (Fig.99) and iodine (Fig.100). Figure 100 shows experimental XRD patterns of as-deposited and annealed Cu₂AgBiI₆ films in iodine vapour. Annealed films were heated to 110°C, 130°C, and 150°C for 15 minutes using a hotplate. There is no difference in the phases present in films annealed in air (Fig.99) and iodine (Fig.100). Figure 101 shows absorption coefficient spectra of 250 nm thick Cu₂AgBiI₆ films on quartz annealed at 110°C, 130°C and 150°C in air. Absorption coefficient data for evaporated 250 nm thick CuI on quartz is shown for comparison. All data was collected using a UV-VIS spectrophotometer. Figure 102 shows absorption coefficient spectra of 250 nm thick Cu₂AgBiI₆ films on quartz annealed at 110°C, 130°C and 150°C in an iodine atmosphere. All data was collected using a UV- VIS spectrophotometer. Figure 103 shows a photoluminescence (PL) spectra of as deposited, air, and iodine-annealed Cu₂AgBiI₆ films on quartz. The PL spectra of CuI is included for reference. Excitation wavelength was 405 nm. Figure 104 shows Time Resolved Photoluminescence (TRPL) spectroscopy of as deposited and annealed films in air. Average lifetimes were extracted by fitting stretched exponentials of the form t_ୡ୦ୟ୰ is the characteristic lifetime, t is time, and β is the stretching factor which accounts for the heterogeneity in decay mechanisms. t_ୟ^ was calculated using where Γ is the gamma function . Excitation wavelength was 405 nm. Figure 105 shows Time Resolved Photoluminescence (TRPL) spectroscopy of as deposited and annealed films in an iodine atmosphere. Average lifetimes were extracted by fitting stretched exponentials of the form where t_ୡ୦ୟ୰ is the characteristic lifetime, t is time, and β is the stretching factor which accounts for the heterogeneity in decay mechanisms. was calculated using where Γ is the gamma function. Excitation wavelength was 405 nm. Figure 106 shows top view SEM images of a) as deposited and 150°C annealed Cu₂AgBiI₆ films in b) air and c) iodine atmospheres. The lighter coloured material is CuI. Composition measured by SEM-EDX is shown inset. Beam voltage was 10KV, spot size 3. DETAILED DESCRIPTION Definitions The term “optoelectronic material”, as used herein, refers to a material which either (i) absorbs light, which may then generate free charge carriers; or (ii) accepts charge, both electrons and holes, which may subsequently recombine and emit light. Such materials may also be referred to as “photoactive materials”. Optoelectronic/photoactive materials may be examples of semiconducting materials. The term “photovoltaic material”, as used herein, refers to a material that absorbs light, then generates free charge carriers. The term “electroluminescent material”, as used herein, refers to a material that accepts charge, both electrons and holes, which subsequently recombine and emit light. The term “photoluminescent material”, as used herein, refers to a material that is able to absorb photons and undergo photoexcitation, then emit photons. A photoemissive material is a material which absorbs light of energies higher than band gap and reemits light at energies at the band gap. The term “electronic material”, as used herein, refers to a material that is able to conduct charge. An electronic material may be a hole conductor material, an electron transporting material, or a material capable of transporting electrons and holes. An electronic material is typically suitable for use in a transistor. The terms “semiconductor” and “semiconducting material”, as used herein, both refer to a material with electrical conductivity intermediate in magnitude between that of a conductor and a dielectric. A semiconductor or semiconducting material may be an negative (n)-type semiconductor, a positive (p)-type semiconductor or an intrinsic (i) semiconductor. A semiconductor or semiconducting material may have a band gap of from 0.5 to 3.5 eV, for instance from 0.5 to 3.0 eV, from 0.5 to 2.5 eV, from 1.0 to 2.5 eV or from 1.0 to 2.0 eV (when measured at 300 K). The terms “semiconductor” and “semiconducting material” have the same meaning herein and may be used interchangeably. The compounds defined herein are typically semiconductors. The terms “semiconductor device” and “semiconducting device”, as used herein, refer to a device comprising a functional component which comprises a semiconducting material. Examples of semiconductor devices include an optoelectronic device such as a photovoltaic device, a solar cell, a photo detector, a photodiode, a photosensor, a chromogenic device, a transistor, a light-sensitive transistor, a phototransistor, a solid state triode, a battery, a battery electrode, a capacitor, a super- capacitor, a light-emitting device and a light-emitting diode. The terms “semiconductor device” and “semiconducting device” have the same meaning herein and may be used interchangeably. The term “optoelectronic device”, as used herein, refers to devices which source, control, detect or emit light. Light is understood to include any electromagnetic radiation. Examples of optoelectronic devices include photovoltaic devices, photodiodes (including solar cells), phototransistors, photomultipliers, photoresistors, light emitting devices, electroluminescent devices, light emitting diodes, charge injection lasers and X-ray scintillators. Often, an “optoelectronic device” that is referred to herein is a photovoltaic device or an electroluminescent device. The term “crystalline” as used herein indicates a crystalline compound, which is a compound having an extended 3D crystal structure. A crystalline compound is typically in the form of crystals or, in the case of a polycrystalline compound, crystallites (i.e. a plurality of crystals having particle sizes of less than or equal to 1 μm). The crystals together often form a layer. The crystals of a crystalline material may be of any size. Where the crystals have one or more dimensions in the range of from 1 nm up to 1000 nm, they may be described as nanocrystals. The compounds defined herein are crystalline compounds. A crystalline compound may comprise multiple crystalline phases with different crystal structures. The term “n-type region”, as used herein, refers to a region of one or more electron-transporting (i.e. n-type) materials. Similarly, the term “n-type layer” refers to a layer of an electron- transporting (i.e. an n-type) material. An electron-transporting (i.e. an n-type) material could, for instance, be a single electron-transporting compound or elemental material. An electron- transporting compound or elemental material may be undoped or doped with one or more dopant elements. The term “p-type region”, as used herein, refers to a region of one or more hole-transporting (i.e. p- type) materials. Similarly, the term “p-type layer” refers to a layer of a hole-transporting (i.e. a p- type) material. A hole-transporting (i.e. a p-type) material could be a single hole-transporting compound or elemental material, or a mixture of two or more hole-transporting compounds or elemental materials. A hole-transporting compound or elemental material may be undoped or doped with one or more dopant elements. The term “perovskite”, as used herein, refers to a material with a three-dimensional crystal structure related to that of CaTiO₃ or a material comprising a layer of material, which layer has a structure related to that of CaTiO₃. The structure of CaTiO₃ can be represented by the formula ABX₃, wherein A and B are cations of different sizes and X is an anion. In the unit cell, the A cations are at (0,0,0), the B cations are at (1/2, 1/2, 1/2) and the X anions are at (1/2, 1/2, 0). Thus, the A cation is typically twelve coordinate and the B cation is typically six coordinate (octahedral coordination). The A cation is usually larger than the B cation. The skilled person will appreciate that when A, B and X are varied, the different ion sizes may cause the structure of the perovskite material to distort away from the structure adopted by CaTiO₃ to a lower-symmetry distorted structure. The symmetry will also be lower if the material comprises a layer that has a structure related to that of CaTiO₃. Materials comprising a layer of perovskite material are well known. For instance, the structure of materials adopting the K₂NiF₄-type structure comprises a layer of perovskite material. The skilled person will appreciate that a perovskite material can be represented by the formula [A][B][X]₃, wherein [A] is at least one cation, [B] is at least one cation and [X] is at least one anion. When the perovskite comprise more than one A cation, the different A cations may distributed over the A sites in an ordered or disordered way. When the perovskite comprises more than one B cation, the different B cations may distributed over the B sites in an ordered or disordered way. When the perovskite comprise more than one X anion, the different X anions may distributed over the X sites in an ordered or disordered way. The symmetry of a perovskite comprising more than one A cation, more than one B cation or more than one X cation, will be lower than that of CaTiO₃. For layered perovskites the stoichiometry can change between the A, B and X ions. As an example, the [A]₂[B][X]₄ structure can be adopted if the A cation has a too large an ionic radii to fit within the 3D perovskite structure. The term “perovskite” also includes A/M/X materials adopting a Ruddlesden-Popper phase. Ruddlesden-Popper phase refers to a perovskite with a mixture of layered and 3D components. Such perovskites can adopt the crystal structure, where A and A’ are different cations and n is an integer from 1 to 8, or from 2 to 6. The term “perovskite” also includes A/M/X materials adopting a Dion-Jacobson phase. Dion-Jacobson phase refers to a perovskite with a mixture of layered and 3D components. Such perovskites can adopt the crystal structure, A where A and A’ are different cations and q is an integer from 1 to 8, or from 2 to 6. The term “mixed 2D and 3D” perovskite is used to refer to a perovskite film within which there exists both regions, or domains, of AMX₃ and perovskite phases. The term “halide” as used herein indicates the singly charged anion of an element in group VII of the periodic table. “Halide” includes fluoride, chloride, bromide and iodide. The term “chalcogenide” as used herein refers to an anion of group 6 element, i.e. of a chalcogen. Typically, chalcogenide refers to an oxide anion, a sulphide anion, a selenide anion or a telluride anion. The term “consisting essentially of” refers to a composition comprising the components of which it consists essentially as well as other components, provided that the other components do not materially affect the essential characteristics of the composition. Typically, a composition consisting essentially of certain components will comprise greater than or equal to 95 wt% of those components or greater than or equal to 99 wt% of those components. Crystalline compound The present invention provides a crystalline compound comprising (i) One or more first cations, A; (ii) One or more second cations, B; (iii) One or more third cations, B’; (iv) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment. The one or more first cations, A, the one or more second cations, B, and the one or more third cations, B’, therefore each correspond to a different cation or group of cations. In other words, the crystalline compound must always contain a minimum of three different types of cation. Thus present invention provides quaternary compounds in which some cations occupy tetrahedral sites and other cations occupy octahedral sites. This occupancy pattern permits a greater variety of compounds with useful optical and electronic properties to be accessed beyond the conventional group of perovskites, which do not contain cations occupying tetrahedral coordination environments. In perovskite materials, the smaller cations usually occupy octahedral sites in the anion lattice whilst the larger cations occupy twelve coordinate cuboctahedral sites. Thus, typically, the crystalline compound is not a perovskite. This occupancy pattern also provides access to structures beyond the previously investigated Ag/Bi/I containing materials mentioned above, in which the cations occupy octahedral sites only. Cu-doped Ag₂BiI₅ in which the Cu is said to replace Ag⁺ at the octahedral site of the hexagonal Ag₂BiI₅ structure is disclosed in J. W. Park et al, Sustainable Energy Fuels, 2021, DOI: 10.1039/D0SE01563F. That document does not therefore disclose compounds in which at least some of the first cations, A, are in a tetrahedral coordination environment. A cations In the crystalline compounds of the invention, at least some of the one or more first cations, A, are in a tetrahedral coordination environment. Thus, it is not the case that the compounds of the invention are simple modifications of known perovskite or Ag/Bi/I containing materials, where additional cations are simply substituted for atoms on the existing octahedral sites. Typically, at least some of the one or more first cations, A, form AX₄ tetrahedra. In some instances, all of the one or more first cations, A, are in a tetrahedral coordination environment. In other words, all of the one or more first cations, A, form AX₄ tetrahedra. In other instances, at least some of the one or more first cations, A, are in a tetrahedral coordination environment and at least some of the one or more first cations, A, are in a non-tetrahedral coordination environment, for instance an octahedral coordination environment. Thus, least some of the one or more first cations, A, form AX₄ tetrahedra and at least some of the one or more first cations, A, form AX₆ octahedra. Typically, the one or more first cations, A, are of a size suitable to occupy a tetrahedral coordination environment in the anion lattice. For instance, the one or more first cations, A, have a Shannon ionic radius of less than 1.5Å, preferably less than 1Å for instance less than 0.8Å or less than 0.7Å. Typically, the one or more first cations, A, have a Shannon ionic radius between 0.3 and 1.5Å, for instance from 0.4 to 1.2Å or from 0.5 to 1Å. The one or more first cations, A, may comprise one or more metal cations and/or one or more non- metal cations. For instance, the one or more first cations, A, may comprise one or more metal cations of size suitable to occupy a tetrahedral coordination environment in the anion lattice. The one or more first cations, A, may comprise one or more non-metal cations of size suitable to occupy a tetrahedral coordination environment in the anion lattice. The one or more first cations, A, may comprise one or more metal cations of size suitable to occupy a tetrahedral coordination environment in the anion lattice and one or more non-metal cations of size suitable to occupy a tetrahedral coordination environment in the anion lattice. Typically, the one or more first cations, A, comprise one or more of Mg²⁺, Ga³⁺, Ge⁴⁺, Si⁴⁺, P⁵⁺, Cu⁺, Zn²⁺, Ti⁴⁺, Al³⁺, In³⁺ and Li⁺. For instance, the one or more first cations, A, may comprise two or more or three or more of Mg²⁺, Ga³⁺, Ge⁴⁺, Si⁴⁺, P⁵⁺, Cu⁺, Zn²⁺, Ti⁴⁺, Al³⁺, In³⁺ and Li⁺. The one or more first cations, A, may consist of two cations selected from the group consisting of Mg²⁺, Ga³⁺, Ge⁴⁺, Si⁴⁺, P⁵⁺, Cu⁺, Zn²⁺, Ti⁴⁺, Al³⁺, In³⁺ and Li⁺. The one or more first cations, A, may be a single cation selected from the group consisting of Mg²⁺, Ga³⁺, Ge⁴⁺, Si⁴⁺, P⁵⁺, Cu⁺, Zn²⁺, Ti⁴⁺, Al³⁺, In³⁺ and Li⁺. The one or more first cations, A, may have the electronic configuration Nd¹⁰, wherein N is an integer from 3 to 5. Thus, the one or more first cations, A may have the electronic configuration 3d¹⁰, 4d¹⁰ or 5d¹⁰, preferably the one or more first cations, A may have the electronic configuration 3d¹⁰ or 4d¹⁰. Cations with the Nd¹⁰ configuration may contribute to band edge states in the crystalline compound, thereby providing useful electronic properties. Preferably, the one or more first cations, A, comprise Cu⁺. Thus, the one or more first cations, A, may comprise Cu⁺ and one or more other cations selected from other metal cations of a size suitable to occupy a tetrahedral coordination environment in the anion lattice. The one or more first cations, A, may comprise Cu⁺ and one or more other cations selected from non-metal cations of a size suitable to occupy a tetrahedral coordination environment in the anion lattice. The one or more first cations, A, may comprise Cu⁺ and one or more other cations selected from other metal cations of a size suitable to occupy a tetrahedral coordination environment in the anion lattice and one or more other cations selected from non-metal cations of a size suitable to occupy a tetrahedral coordination environment in the anion lattice. The one or more first cations, A, may comprise or consist of Cu⁺ and one or more other cations selected from the group consisting of Mg²⁺, Ga³⁺, Ge⁴⁺, Si⁴⁺, P⁵⁺, Zn²⁺, Ti⁴⁺, Al³⁺, In³⁺ and Li⁺. The one or more first cations, A, may be a single cation which is Cu⁺. B cations Typically, at least some of the one or more second cations, B, are in an octahedral coordination environment. Thus, typically, at least some of the one or more second cations, B, form BX6 octahedra. In some instances, all of the one or more second cations, B, are in an octahedral coordination environment. In other words, all of the one or more second cations, B, form BX₆ octahedra. The one or more second cations, B, may comprise one or more metal cations and/or one or more non-metal cations. For instance, the one or more second cations, B, may comprise one or more metal cations of size suitable to occupy a octahedral coordination environment in the anion lattice. The one or more second cations, B, may comprise one or more non-metal cations of size suitable to occupy a octahedral coordination environment in the anion lattice. The one or more second cations, B, may comprise one or more metal cations of size suitable to occupy a octahedral coordination environment in the anion lattice and one or more non-metal cations of size suitable to occupy a octahedral coordination environment in the anion lattice. The one or more second cations, B, may comprise one or more alkali metal cations. The one or more second cations, B, may comprise one or more transition metal cations. The one or more second cations, B, may comprise one or more alkali earth metal cations. The one or more second cations, B, may comprise one or more alkali metal cations and/or one or more alkali metal earth cations and/or one or more transition metal cations. For instance, the one or more second cations, B, may comprise one or more alkali metal cations and one or more transition metal cations. Typically, the one or more second cations, B, comprise one or more of Ag⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cr³⁺, V²⁺, Cd²⁺, Cu⁺, Au⁺, Pt²⁺, Pt⁴⁺, Mg²⁺, Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺, In⁺, Ca²⁺, Sr²⁺, La³⁺ and Y³⁺. For instance, the one or more second cations, B, may comprise two or more, or three or more of Ag⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cr³⁺, V²⁺, Cd²⁺, Cu⁺, Au⁺, Pt²⁺, Pt⁴⁺, Mg²⁺, Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺, In⁺, Ca²⁺, Sr²⁺, La³⁺ and Y³⁺. For instance, the one or more second cations, B, may consist of two cations selected from the group consisting of Ag⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cr³⁺, V²⁺, Cd²⁺, Cu⁺, Au⁺, Pt²⁺, Pt⁴⁺, Mg²⁺, Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺, In⁺, Ca²⁺, Sr²⁺, La³⁺ and Y³⁺. The one or more second cations, B, may be a single cation selected from the group consisting of Ag⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cr³⁺, V²⁺, Cd²⁺, Cu⁺, Au⁺, Pt²⁺, Pt⁴⁺, Mg²⁺, Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺, In⁺, Ca²⁺, Sr²⁺, La³⁺ and Y³⁺. Preferably, the one or more second cations, B, comprise Ag⁺. Thus, the one or more second cations, B, may comprise Ag⁺ and one or more other cations selected from other metal cations of a size suitable to occupy an octahedral coordination environment in the anion lattice. The one or more second cations, B, may comprise Ag⁺ and one or more other cations selected from non-metal cations of a size suitable to occupy an octahedral coordination environment in the anion lattice. The one or more second cations, B, may comprise Ag⁺ and one or more other cations selected from other metal cations of a size suitable to occupy an octahedral coordination environment in the anion lattice and one or more other cations selected from non-metal cations of a size suitable to occupy an octahedral coordination environment in the anion lattice. The one or more second cations, B, may comprise or consist of Ag⁺ and one or more other cations selected from the group consisting of Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cr³⁺, V²⁺, Cd²⁺, Cu⁺, Au⁺, Pt²⁺, Pt⁴⁺, Mg²⁺, Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺, In⁺, Ca²⁺, Sr²⁺, La³⁺ and Y³⁺. For instance, the one or more second cations, B, may comprise or consist of Ag⁺, Na⁺ and In⁺. The one or more second cations, B, may comprise or consist of Ag⁺ and Na⁺. The one or more second cations, B, may be a single cation which is Ag⁺. The one or more second cations may be a single cation which is Na⁺. The molar ratio of the one or more first cations, A, to the one or more second cations, B, in the crystalline compounds as described herein is preferably at least 1:9. In this context, the molar ratio is defined as being the molar ratio between all of the one or more first cations, A, and all of the one or more second cations, B, in the event that multiple types of A and B cation are present. For instance, the molar ratio of the one or more first cations, A, to the one or more second cations, B, in the crystalline compound may be at least 1:8, or at least 1:7, or at least 1:6. Typically, the molar ratio of the one or more first cations, A, to the one or more second cations, B, is at least 1:5, for instance at least 1:1. Therefore, the compound may comprise a larger molar % of the one or more first cations, A, than of the one or more second cations, B. The molar ratio of the one or more first cations, A, to the one or more second cations, B, may be at least 2:1, at least 3:1, at least 4:1 or at least 5:1. The molar ratio of the one or more first cations, A, to the one or more second cations, B, may be no more than 20:1, typically no more than 10:1, for instance no more than 5:1. Thus, typically, the molar ratio of the one or more first cations, A, to the one or more second cations, B, is from 1:9 to 20:1, preferably from 1:5 to 10:1. B’ cations At least some of the one or more third cations, B’, are in an octahedral coordination environment. Thus, typically, at least some of the one or more third cations, B’, form B’X₆ octahedra. In some instances, all of the one or more third cations, B’, are in a octahedral coordination environment. In other words, all of the one or more third cations, B’, form B’X₆ octahedra. The one or more third cations, B’, may have the electronic configuration Ns², wherein N is an integer from 2 to 7. Typically N is an integer from 3 to 6. Thus, the one or more third cations, B’, may have the electronic configuration 3s², 4s², 5s² or 6s². The one or more third cations, B’, may comprise one or more metal cations and/or one or more non-metal cations. For instance, the one or more third cations, B’, may comprise one or more metal cations of size suitable to occupy an octahedral coordination environment in the anion lattice. The one or more third cations, B’, may comprise one or more non-metal cations of size suitable to occupy a octahedral coordination environment in the anion lattice. The one or more third cations, B’, may comprise one or more metal cations of size suitable to occupy a octahedral coordination environment in the anion lattice and one or more non-metal cations of size suitable to occupy a octahedral coordination environment in the anion lattice. Typically, the one or more third cations, B’, comprise one or more of Bi³⁺, Pb²⁺, Sn²⁺, Ge²⁺, In³⁺, Sb³⁺, Te⁴⁺, Tl⁺, In⁺ and Se⁴⁺. For instance, the one or more third cations, B’, may comprise two or more or three or more of Bi³⁺, Pb²⁺, Sn²⁺, Ge²⁺, In⁺, Sb³⁺, Te⁴⁺, Tl⁺, In³⁺ and Se⁴⁺. The one or more third cations, B’, may consist of two cations selected from the group consisting of Bi³⁺, Pb²⁺, Sn²⁺, Ge²⁺, In⁺, Sb³⁺, Te⁴⁺, Tl⁺, In³⁺ and Se⁴⁺. The one or more third cations, B’, may be a single cation selected from the group consisting of Bi³⁺, Pb²⁺, Sn²⁺, Ge²⁺, In⁺, Sb³⁺, Te⁴⁺, Tl⁺, In³⁺ and Se⁴⁺. Preferably, the one or more third cations may be one or more, or two or more, or three or more of Bi³⁺, Sn²⁺, Ge²⁺, In⁺, Sb³⁺, Te⁴⁺, Tl⁺, In³⁺ and Se⁴⁺. Preferably, the one or more third cations, B’, comprise Bi³⁺. Thus, the one or more third cations, B’, may comprise Bi³⁺ and one or more other cations selected from other metal cations of a size suitable to occupy an octahedral coordination environment in the anion lattice. The one or more third cations, B’, may comprise Bi³⁺ and one or more other cations selected from non-metal cations of a size suitable to occupy an octahedral coordination environment in the anion lattice. The one or more third cations, B’, may comprise Bi³⁺ and one or more other cations selected from other metal cations of a size suitable to occupy an octahedral coordination environment in the anion lattice and one or more other cations selected from non-metal cations of a size suitable to occupy an octahedral coordination environment in the anion lattice. The one or more third cations, B’, may comprise or consist of Bi³⁺ and one or more other cations selected from the group consisting of Pb²⁺, Sn²⁺, Ge²⁺, In⁺, Sb³⁺, Te⁴⁺, Tl⁺, In³⁺ and Se⁴⁺. The one or more third cations, B’, may be a single cation with is Bi³⁺. X anions Typically, the one or more anions, X, form a close-packed lattice. Thus, the one or more anions, X, may form a cubic close-packed lattice. The one or more anions, X, may form a hexagonal close- packed lattice. When the one or more anions, X, form a close-packed lattice, at least some of the one or more first cations, A, occupy at least some of the tetrahedral sites, at least some of the one or more second cations, B, may occupy at least some of the octahedral sites, and at least some of the one or more third cations, B’, occupy at least some of the octahedral sites. When the one or more anions, X, form a close-packed lattice, all of the one or more first cations, A, may occupy at least some of the tetrahedral sites, all of the one or more second cations, B, may occupy at least some of the octahedral sites, and all of the one or more third cations, B’, may occupy at least some of the octahedral sites. The one or more anions, X, typically comprise one or more anions selected from halide anions and chalcogenide anions. The one or more anions, X, may consist of halide anions only. The one or more anions, X, may consist of chalcogenide anions only. The one or more anions, X, may consist of both halide and chalcogenide anions. Thus, the one or more anions, X, typically comprise one or more anions selected from F-, Cl-, Br-, I-, O^2-, S^2-, Se^2- and Te^2-. For instance, the one or more anions, X, may comprise two or more or three or more anions selected from F-, Cl-, Br-, I-, O^2-, S^2-, Se^2- and Te^2-. The one or more anions, X, may consist of two anions selected from the group consisting of F-, Cl-, Br-, I-, O^2-, S^2-, Se^2- and Te^2-.The one or more anions, X, may comprise a single anion selected from the group consisting of F-, Cl-, Br-, I-, O^2-, S^2-, Se^2- and Te^2-. For instance, the one or more anions, X, may comprise one or more anions selected from I-, Br-, Cl-, S^2-, Se^2- and O^2- . The one or more anions, X, may comprise two or more anions selected from F-, Cl-, Br-, I-, O^2-, S^2-, Se^2- and Te^2-. For instance, the one or more anions, X, may comprise two or more anions selected from I-, Br-, Cl-, S^2-, Se^2- and O^2-. For instance, the one or more anions, X, may comprise S^2- and one or more anions selected from F-, Cl-, Br- and I-. Preferably, the one or more anions, X, comprise S^2- and I-. Typically, the one or more anions, X, may comprise two or more anions selected from F-, Cl-, Br- and I-. For instance, the one or more anions, X, may comprise F- and one or more anions selected from Cl-, Br- and I-. The one or more anions, X, may comprise Cl- and one or more anions selected from F-, Br- and I-. The one or more anions, X, may comprise Br- and one or more anions selected from F-, Cl- and I-. The one or more anions, X, may comprise I- and one or more anions selected from F-, Cl- and Br-. The one or more anions, X, may comprise I- and Br-. The one or more anions, X, may comprise Br- and Cl-. Preferably, the one or more anions, X, comprise I-. Thus, the one or more anions, X, may comprise I- and one or more other anions selected from halide anions and chalcogenide anions. The one or more anions, X, may comprise I- and one or more other anions selected from F-, Cl-, Br-, I-, O^2-, S^2-, Se^2- and Te^2-. The one or more anions, X, may be a single anion which is I-. Typically, in the crystalline compound, at least some of the one or more first cations, A, are in an tetrahedral coordination environment; at least some of the one or more second cations, B, are in an octahedral coordination environment; and at least some of the one or more third cations, B’, are in an octahedral coordination environment. Thus, in the crystalline compound at least some of the one or more first cations, A, may form AX₄ tetrahedra; at least some of the one or more second cations, B, may form BX₆ octahedra; and at least some of the one or more third cations, B’, may form B’X₆ octahedra. At least some of the one or more first cations, A, may form AX₄ tetrahedra; at least some of the one or more second cations, B, may form BX₆ octahedra; at least some of the one or more third cations, B’, may form B’X₆ octahedra and the one or more anions, X, form a close-packed structure. Typically, at least some of the one or more first cations, A, form AX₄ tetrahedra; all of the one or more second cations, B, form BX₆ octahedra; and all of the one or more third cations, B’, form B’X₆ octahedra. For instance all of the one or more first cations, A, form AX₄ tetrahedra; all of the one or more second cations, B, form BX₆ octahedra; and all of the one or more third cations, B’, form B’X₆ octahedra. Typically, the compound comprises an interconnected network of B’X₆ octahedra. For instance, the compound may comprise a network of edge-sharing B’X₆ octahedra or a network of corner- sharing B’X₆ octahedra. Typically, the compound comprises an interconnected network of B’X₆ octahedra in which the one or more third cations, B’, comprise Bi³⁺ and the one or more anions, X, comprise I-. For instance, the compound may comprises an interconnected network of BiI₆ octahedra, e.g. an network of edge-sharing BiI₆ octahedra or a network of corner-sharing BiI₆ octahedra. The crystalline compound may have a structure based around a cubic-close packed array of the one or more anions, X, in which the one or more second cations, B, and the one or more third cations, B’, occupy edge sharing octahedral sites. The crystalline compound may have a structure based around a cubic-close packed array of the one or more anions, X, in which at least some of the one or more second cations, B, and the one or more third cations, B’, occupy edge sharing octahedral sites which form a two-dimensional network of edge-sharing octahedra. Thus, the crystalline compound may have a structure based on the cadmium chloride structure. The crystalline compound may have a structure based around a cubic-close packed lattice of the one or more anions, X, in which the one or more second cations, B, and the one or more third cations, B’, occupy octahedral sites and the one or more first cations, A, occupy tetrahedral sites. The precise levels of occupancy of the octahedral and tetrahedral sites in the cubic-close packed lattice of the one or more anions, X, will vary depending on the charges on the A, B and B’ cations. Thus, the crystalline compound may have a structure based on the spinel structure. The crystalline compound may be a compound in which the one or more first cations, A, have the electronic configuration Nd¹⁰, wherein N is an integer from 3 to 5 and the one or more third cations, B’, have the electronic configuration Ns², wherein N is an integer from 2 to 7, preferably wherein N is an integer from 3 to 6. Thus, the crystalline compound may be a compound in which the one or more first cations, A, have the electronic configuration 3d¹⁰, 4d¹⁰ or 5d¹⁰ and the one or more third cations, B’, have the electronic configuration 3s², 4s², 5s² or 6s². The crystalline compound may be a compound in which the one or more first cations, A, comprise one or more of Mg²⁺, Ga³⁺, Ge⁴⁺, Si⁴⁺, P⁵⁺, Cu⁺, Zn²⁺, Ti⁴⁺, Al³⁺, In³⁺ and Li⁺;the one or more second cations, B, comprise one or more of Ag⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cr³⁺, V²⁺, Cd²⁺, Cu⁺, Au⁺, Pt²⁺, Pt⁴⁺, Mg²⁺, Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺, In⁺, Ca²⁺, Sr²⁺, La³⁺ and Y³⁺; the one or more third cations, B’, comprise one or more of Bi³⁺, Pb²⁺, Sn²⁺, Ge²⁺, In⁺, Sb³⁺, Te⁴⁺, Tl⁺, In³⁺ and Se⁴⁺; and the one or more anions, X, comprise one or more of I-, Br-, Cl-, S^2-, Se^2- and O^2-. Preferably, the crystalline compound is a compound in which the one or more first cations, A, comprise Cu⁺; the one or more second cations, B, comprise Ag⁺; the one or more third cations, B’, comprise Bi³⁺; and the one or more anions, X, comprise I-. The crystalline compound may be a compound in which the one or more first cations, A, are Cu⁺; the one or more second cations, B, are Ag⁺; the one or more third cations, B’, are Bi³⁺; and the one or more anions, X, are I-. The compound may further comprises a dopant cation. The dopant is an additional cation which may take the site of one of the first (A), second (B) and/or third (B’) cations in the crystal structure. To balance the charge, this may require the presence of charge-balancing defects, e.g. vacancies, in the crystal structure. Typically, the dopant cation is present in low atomic concentration in the crystalline compound. For instance, the dopant cation may be present in an amount of less than 1at.% or an amount of less than 0.5 at.%, typically an amount of less than 0.1at.%. Preferably, the dopant cation is selected from a transition metal cation or a rare earth cation. Typically, the dopant cation is selected from the group consisting of Ce³⁺, Er³⁺, Yb³⁺, Eu²⁺ and Eu³⁺. The crystalline compound may be a compound of formula (I): [A]_4x{[B][B’]}_1-x[X]₄ (I); wherein [A] is the one or more first cations, [B] is the one or more second cations, [B’] is the one or more third cations and [X] is the one or more anions, wherein the one or more first cations, A, are different from the one or more second cations, B, and the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, and wherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (I) may vary by ±30%. As the skilled person would appreciate, when the crystalline compound is a compound of formula (I), the charges on the A, B, B’ and X ions must be such that they can form a compound having the stoichiometry described above. In the compound of formula (I), the one or more first cations, A, may be any first cations as described herein. The one or more second cations, B, may be any second cations as described herein. The one or more third cations, B’, may be any third cations as described herein. The one or more anions, X, may be any anions as described herein. Typically, the one or more first cations, A, comprise Cu⁺, the one or more second cations, B, comprise Ag⁺, the one or more third cations B’, comprise Bi³⁺ and the one or more anions, X, comprise I-. For the compound of formula (I) the stoichiometries deviate by ±30% from the “ideal” values calculated for formula (I) at a certain value of x. Therefore, for a compound where x=0.5, the stoichiometry of the one or more first cations, A, the one or more second cations, B, the one or more third cations, B’, and the one or more anions, X will fall within the following ranges: [A]_(1.4-2.67){[B][B’]}_(0.35-0.67)[X]_(2.8-5.3). For other values of x, the ranges of possible stoichiometries for each ion in formula (I) can be calculated in an analogous way. Typically, the stoichiometry of each of the ions in the compound of formula (II) may vary by ±20%, or by ±10%. Typically, x is from 0.05 to 0.95, preferably from 0.1 to 0.5. x may be from 0.15 to 0.25. x may be from 0.30 to 0.35. For instance x may be 0.2 or x may be 0.33. The compound of formula (I) may be a compound having the formula [A]{[B][B’]}[X]₅ wherein the stoichiometry of each of the ions in the compound may vary by ±30%. The compound of formula (I) may be a compound having the formula [A]₂{[B][B’]}[X]₆ wherein the stoichiometry of each of the ions in the compound may vary by ±30%. The one or more first cations, A, may be any first cations as described herein. The one or more second cations, B, may be any second cations as described herein. The one or more third cations, B’, may be any third cations as described herein. The one or more anions, X, may be any anions as described herein. Typically, the one or more first cations, A, comprise Cu⁺, the one or more second cations, B, comprise Ag⁺, the one or more third cations B’, comprise Bi³⁺ and the one or more anions, X, comprise I-. For instance, the one or more first cations, A, may comprise Cu⁺, the one or more second cations, B, may comprise Ag⁺ and Na+, the one or more third cations B’, may comprise Bi³⁺ and the one or more anions, X, may comprise I-. For instance, the compound of formula (I) may be a compound of the formula Cu₂Ag_yNa_(1-y)BiI₆ where y is from 0 to 1, and wherein the stoichiometry of each of the ions in the compound may vary by ±30%. For example x may be from 0.5 to 0.9, preferably about 0.7. For instance, the compound of formula (I) maybe a compound of formula Cu₆(Ag_1-yNa_y/2In_y/2)BiI₁₀ where y is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound may vary by ±30%. The compound of formula (I) may be a compound of formula (IA): Cu_4x{AgBi}_1-xI₄ (IA); wherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (IA) may vary by ±30%. For the compound of formula (IA) the stoichiometries may deviate by ±30% from the “ideal” values calculated for formula (IA) at a certain value of x. Therefore, for a compound where x=0.5, the stoichiometry of the Cu⁺ cation, the Ag⁺ cation, the Bi³⁺ cation, and I- anion will fall within the following ranges: Cu_(1.4-2.67){AgBi}_(0.35-0.67)I_(2.8-5.3). Of course, for other values of x, the ranges of possible stoichiometries for each ion in formula (IA) can be calculated in an analogous way. Typically, the stoichiometry of each of the ions in the compound of formula (IA) may vary by ±20%, or by ±10%,. Preferably, the compound is Cu_0.4AgBiI_4.4, wherein the stoichiometry of each of the ions in the compound may vary by ±30% (i.e. a compound satisfying the formula Cu_(0.3-0.5)Ag_(0.7-1.3)Bi_(0.7-1.3)I_(3.1- _5.7)). The compound may be Cu_0.4AgBiI_4.4, wherein the stoichiometry of each of the ions in the compound may vary by ±20% or ±10%. Preferably, the compound is CuAgBiI₅, wherein the stoichiometry of each of the ions in the compound may vary by ±30% (i.e. a compound satisfying the formula Cu(0.7-1.3)Ag(0.7-1.3)Bi(0.7-1.3)I(3.5- _6.5). The compound may be CuAgBiI₅, wherein the stoichiometry of each of the ions in the compound may vary by ±20% or ±10%. Preferably, the compound is Cu₂AgBiI₆, wherein the stoichiometry of each of the ions in the compound may vary by ±30% (i.e. a compound satisfying the formula Cu_(1.4-2.6)Ag_(0.7-1.3)Bi_(0.7-1.3)I_(4.2- _7.8)). The compound may be Cu₂AgBiI₆, wherein the stoichiometry of each of the ions in the compound may vary by ±20% or ±10%. Preferably, the compound is Cu₆AgBiI₁₀, wherein the stoichiometry of each of the ions in the compound may vary by ±30% (i.e. a compound satisfying the formula Cu_(4.2-7.8)Ag_(0.7-1.3)Bi_(0.7-1.3)I_(6- ₁₃₎). The compound may be Cu₆AgBiI₁₀, wherein the stoichiometry of each of the ions in the compound may vary by ±20% or ±10%. The crystalline compound may be a compound of formula (II): [A]_1+2x{[B]_3x[B’]_1-x}₂[X]_7+2x (II); wherein [A] is the one or more first cations, [B] is the one or more second cations, [B’] is the one or more third cations and [X] is the one or more anions, wherein the one or more first cations, A, are different from the one or more second cations, B, and the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, and wherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (I) may vary by ±30%. As the skilled person would appreciate, when the crystalline compound is a compound of formula (II), the charges on the A, B, B’ and X ions must be such that they can form a compound having the stoichiometry described above. In the compound of formula (II), the one or more first cations, A, may be any first cations as described herein. The one or more second cations, B, may be any second cations as described herein. The one or more third cations, B’, may be any third cations as described herein. The one or more anions, X, may be any anions as described herein. Typically, the one or more first cations, A, comprise Cu⁺, the one or more second cations, B, comprise Ag⁺, the one or more third cations B’, comprise Bi³⁺ and the one or more anions, X, comprise I-. Thus for the compound of formula (II) the stoichiometries may deviate by ±30% from the “ideal” values calculated for formula (I) at a certain value of x. Therefore, for a compound where x=0.25, the stoichiometry of the one or more first cations, A, the one or more second cations, B, the one or more third cations, B’, and the one or more anions, X will fall within the following ranges: [A]_(1.05- _2.0){[B]_(0.53-0.98)[B’]_(0.53-0.98)}₂[X]_(5.25-9.98). Of course, for other values of x, the ranges of possible stoichiometries for each ion in formula (II) can be calculated in an analogous way. Preferably, in the compound of formula (II) x is from 0.05 to 0.95, for instance from 0.1 to 0.6. x may be from 0.20 to 0.30. x may be from 0.45 to 0.55. Preferably x is 0.25 or x is 0.5. For instance, the compound of formula (II) may be a compound having the formula [A]{[B][B’]}[X]₅ wherein the stoichiometry of each of the ions in the compound may vary by ±30%. The compound of formula (II) may be a compound having the formula [A]{[B]_1.5[B’]_0.5}[X]₄ wherein the stoichiometry of each of the ions in the compound may vary by ±30%. The one or more first cations, A, may be any first cations as described herein. The one or more second cations, B, may be any second cations as described herein. The one or more third cations, B’, may be any third cations as described herein. The one or more anions, X, may be any anions as described herein. Typically, the one or more first cations, A, comprise Cu⁺, the one or more second cations, B, comprise Ag⁺, the one or more third cations B’, comprise Bi³⁺ and the one or more anions, X, comprise I-. The compound of formula (II) may be a compound of formula (IIA): Cu_1+2x{Ag_3xBi_1-x}₂I_7+2x (IIA); wherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (IA) may vary by ±30%. For the compound of formula (IIA) the stoichiometries may deviate by ±30% from the “ideal” values calculated for formula (IIA) at a certain value of x. Therefore, for a compound where x=0.25, the stoichiometry of the Cu⁺ cation, the Ag⁺ cation, the Bi³⁺ cation, and I- anion will fall within the following ranges: Cu_(1.05-2.0){Ag_(0.53-0.98)Bi_(0.53-0.98)}₂I_(5.25-9.98). Of course, for other values of x, the ranges of possible stoichiometries for each ion in formula (IIA) can be calculated in an analogous way. Typically, the stoichiometry of each of the ions in the compound of formula (IIA) may vary by ±20%, or by ±10%. Preferably, the compound is CuAg_1.5Bi_0.5I₄, wherein the stoichiometry of each of the ions in the compound may vary by ±30% (i.e. a compound satisfying the formula Cu_(0.7-1.3)Ag_(1.05-1.95)Bi_(0.35- _0.65)I_(2.8-5.2)). The compound may be CuAg_1.5Bi_0.5I₄, wherein the stoichiometry of each of the ions in the compound may vary by ±20% or ±10%. Typically, the crystalline compound is a semiconductor. Preferably the crystalline compound is a semiconductor having a band gap of from 0.5 to 3.5 eV. The crystalline compound may be a compound which is a semiconductor having a band of from 0.5 to 3.0 eV, from 0.5 to 2.5 eV, from 1.0 to 2.5 eV or from 1.0 to 2.0 eV or from 1.5 to 2.5 eV. Film/device The invention also provides a film comprising a compound as described herein. Typically the film is a film having a thickness of 2 µm or less. For instance, the film may have a thickness of from 1 to 2 µm. Alternatively, the film may have a thickness on the nanoscale, i.e. less than 1 µm. For instance, the thickness of the film may be from 1 to 1000 nm, from 1 to 500 nm or from 1 to 250 nm. The thickness of the film may be from 25 to 750 nm, from 50 to 500 nm or from 75 to 250 nm. The thickness of the film may be varied by changing the concentration of solution (when using solution processing techniques) or by changing evaporation time (when using evaporation processing techniques). The film comprising a compound as described herein may be a passivated film, which comprises: a) a compound as defined herein; and b) a passivating agent. Advantages of employing a passivating agent in the film are evidenced in Example 6 hereinbelow. The passivating agent may be a halogen, sulfur, selenium or an organic amine. Typically the passivating agent is selected from one or more of iodine, sulfur, selenium or methylamine. Preferably the passivating agent is a halogen, and more preferably it is iodine. The passivating agent may be present in the film in, for example, molecular or ionic form. The invention also provides a semiconductor device comprising a compound as or a film as described herein. The semiconductor device may be an optoelectronic device (for instance a photovoltaic device, a solar cell, a photodetector, a photomultiplier, a photoresistor, a charge injection laser, a photodiode, a photosensor, a chromogenic device, a light-sensitive transistor, a phototransistor, a light-emitting device, an electroluminescent device, or a light-emitting diode, an X-ray scintillator), a luminescent device (for instance a phosphor in a display or in lighting), a transistor, a solid state triode, a battery, a battery electrode, a radiation detector, a capacitor or a super-capacitor. Preferably the device is an optoelectronic device. Typically, the optoelectronic device is selected from a photovoltaic device, a light emitting device, a photodetector or an X-ray scintillator. The semiconductor device may be a photodetector, for instance a detector of visible light, infrared light, x-ray radiation and/or gamma ray radiation. The compound may be in any form within the semiconducting device. Typically the compound is in the form of a layer, for instance a photoactive, photoemissive or photoabsorbent, layer. Preferably, the compound is present in the form of a film as described herein. The film may be a passivated film as described herein. Thus, the semiconductor device may comprise a layer of the compound having a thickness of 2 µm or less. The compound often acts as a photoactive component (e.g. a photoabsorbent component or a photoemissive component) within the semiconductor device. The compound may alternatively act as a p-type semiconductor component, an n-type semiconductor component, or an intrinsic semiconductor component in the semiconductor device. For instance, the compound may form a layer of a p-type, n-type or intrinsic semiconductor in a transistor, e.g. a field effect transistor. For instance, the compound may form a layer of a p-type or n-type semiconductor in an optoelectronic device, e.g. a solar cell or an LED. Typically, the semiconductor device comprises: an n-type region comprising at least one n-type layer; a p-type region comprising at least one p-type layer; and, disposed between the n-type region and the p-type region: a layer comprising the compound or the film. For instance, the semiconductor device is often an optoelectronic device, which optoelectronic device comprises: an n-type region comprising at least one n-type layer; a p-type region comprising at least one p-type layer; and, disposed between the n-type region and the p-type region: a layer of semiconducting material which comprises (or consists essentially of) said compound, as defined herein. The optoelectronic device may be a tandem device. The tandem optoelectronic device may comprise an n-type region comprising at least one n-type layer; a p-type region comprising at least one p-type layer; and, disposed between the n-type region and the p-type region: a layer of semiconducting material which comprises (or consists essentially of) said compound, as defined herein, a charge recombination layer, and a layer of a second semiconductor. An n-type layer is typically a layer of an n-type semiconductor. A p-type layer is typically a layer of a p-type semiconductor. The n-type region comprises at least one n-type layer. The n-type region may comprise an n-type layer and an n-type exciton blocking layer. Such an n-type exciton blocking layer is typically disposed between the n-type layer and the layer(s) comprising the semiconducting material. The n- type region may have a thickness of from 50 nm to 1000 nm. For instance, the n-type region may have a thickness of from 50 nm to 500 nm, or from 100 nm to 500 nm. Preferably, the n-type region comprises a compact layer of an n-type semiconductor. The n-type semiconductor may be selected from a metal oxide, a metal sulphide, a metal selenide, a metal telluride, a perovskite, amorphous Si, an n-type group IV semiconductor, an n-type group III- V semiconductor, an n-type group II-VI semiconductor, an n-type group I-VII semiconductor, an n- type group IV-VI semiconductor, an n-type group V-VI semiconductor, and an n-type group II-V semiconductor, any of which may be doped or undoped. Typically, the n-type 41emiconductor is selected from a metal oxide, a metal sulphide, a metal selenide, and a metal telluride. For instance, the n-type region may comprise an inorganic material selected from oxide of titanium, tin, zinc, niobium, tantalum, tungsten, indium, gallium, neodymium, palladium, or cadmium, or an oxide of a mixture of two or more of said metals. For instance, the n-type layer may comprise TiO₂, SnO₂, ZnO, SnO, Nb₂O₅, Ta₂O₅, WO₃, W₂O₅, In₂O₃, Ga₂O₃, Nd₂O₃, PbO, or CdO. The n-type region may comprise an organic electron transporting materials, for instance C₆₀, Phenyl-C61-butyric acid methyl ester (PCBM), Bis-PCBM, or 3,9-bis(2-methylene-(3-(1,1-dicyanomethylene)-indanone))- 5,5,11,11-tetrakis(4-hexylphenyl)-dithieno[2,3-d:2’,3’-d’]-s-indaceno[1,2-b:5,6-b’]dithiophene. The n-type region may comprise an inorganic/organic bilayer such as a TiO₂/ fullerene bilayer, SnO/fullerene bilayer or a ZnO/ fullerene bilayer. Typically, the n-type region comprises SnO₂ or TiO₂, for instance a compact layer of TiO₂ or SnO₂. Often, the n-type region also comprises a layer of a fullerene or a fullerene derivative (for instance C₆₀ or Phenyl-C61-butyric acid methyl ester (PCBM)). Typically, the n-type region comprises TiO₂, SnO₂, ZnO, SnO, C₆₀, PCBM, Bis-PCBM, 3,9-bis(2- methylene-(3-(1,1-dicyanomethylene)-indanone))-5,5,11,11-tetrakis(4-hexylphenyl)-dithieno[2,3- d:2’,3’-d’]-s-indaceno[1,2-b:5,6-b’]dithiophene, or an inorganic/organic bilayer such as a TiO₂/ fullerene bilayer, SnO/fullerene bilayer or a ZnO/fullerene bilayer. Preferably, the p-type region comprises a compact layer of a p-type semiconductor. Suitable p-type semiconductors may be selected from polymeric or molecular hole transporters. The p-type layer employed in the semiconductor device of the invention may for instance comprise spiro-OMeTAD (2,2’,7,7’-tetrakis-(N,N-di-p-methoxyphenylamine)9,9’-spirobifluorene)), P3HT (poly(3-hexylthiophene)), PCPDTBT (Poly[2,1,3-benzothiadiazole-4,7-diyl[4,4-bis(2-ethylhexyl)- 4H-cyclopenta[2,1-b:3,4-b’]dithiophene-2,6-diyl]]), PVK (poly(N-vinylcarbazole)), HTM-TFSI (1- hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide), Li-TFSI (lithium bis(trifluoromethanesulfonyl)imide), tetracene, MeO-TPD (N,N,N′,N′-Tetrakis(4- methoxyphenyl)benzidine), poly-TPD (Poly[N,N’-bis(4-butylphenyl)-N,N’-bisphenylbenzidine]), PTAA (Poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine]) or tBP (tert-butylpyridine). The p-type region may comprise carbon nanotubes. Usually, the p-type material is selected from spiro- OMeTAD, P3HT, PCPDTBT and PVK. Preferably, the p-type layer employed in the optoelectronic device comprises spiro-OMeTAD. In some embodiments, the p-type layer may comprise an inorganic hole transporter. For instance, the p-type layer may comprise an inorganic hole transporter comprising an oxide of nickel, vanadium, copper or molybdenum; Ga₂O₃, CuSCN, NiO, CuI, CuBr, CuSCN, Cu₂O, CuO or CIS; a perovskite; amorphous Si; a p-type group IV semiconductor, a p-type group III-V semiconductor, a p-type group II-VI semiconductor, a p-type group I-VII semiconductor, a p-type group IV-VI semiconductor, a p-type group V-VI semiconductor, and a p-type group II-V semiconductor, which inorganic material may be doped or undoped. The p-type layer may be a compact layer of said inorganic hole transporter. Typically, the p-type layer comprises a p-type material selected from NiO, Ga2O3, CuSCN, CuI, and CuO, spiro-OMeTAD, MeO-TPD, Tetracene, P3HT, Poly-TPD, or PTAA. The semiconductor device typically further comprises one or more first electrodes and one or more second electrodes. The one or more first electrodes are typically in contact with the n-type region, if such a region is present. The one or more second electrodes are typically in contact with the p- type region, if such a region is present. Typically: the one or more first electrodes are in contact with the n-type region and the one or more second electrodes are in contact with the p-type region; or the one or more first electrodes are in contact with the p-type region and the one or more second electrodes are in contact with the n-type region. The first and second electrode may comprise any suitable electrically conductive material. The first electrode typically comprises a transparent conducting oxide. The second electrode typically comprises one or more metals. The second electrode may alternatively comprise graphite. Typically, the first electrode typically comprises a transparent conducting oxide and the second electrode typically comprises one or more metals. The transparent conducting oxide typically comprises fluorine-doped tin oxide (FTO), indium tin oxide (ITO) or aluminium-doped zinc oxide (AZO), and typically ITO. The second electrode typically comprises a metal selected from silver, gold, copper, aluminium, platinum, palladium, or tungsten. Each electrode may form a single layer or may be patterned. The semiconductor device (for instance an optoelectronic device such as a photovoltaic device, or a light emitting device) may comprise the following layers in the following order: I. one or more first electrodes as defined herein; II. an n-type region comprising at least one n-type layer as defined herein; III. a layer of the semiconducting material comprising the crystalline compound as defined herein; IV. a p-type region comprising at least one p-type layer as defined herein; and V. one or more second electrodes as defined herein. The semiconductor device (for instance a photovoltaic device, or a light emitting device) may comprise the following layers in the following order: I. one or more first electrodes which comprise a transparent conducting oxide, preferably FTO; II. an n-type region comprising at least one n-type layer as defined herein; III. a layer of the semiconducting material comprising the crystalline compound as defined herein; IV. a p-type region comprising at least one p-type layer as defined herein; and V. one or more second electrodes which comprise a metal, preferably silver or gold. The one or more first electrodes may have a thickness of from 100 nm to 700 nm, for instance of from 100 nm to 400 nm. The one or more second electrodes may have a thickness of from 10 nm to 500 nm, for instance from 50 nm to 200 nm or from 10 nm to 50 nm. The n-type region may have a thickness of from 50 nm to 500 nm. The p-type region may have a thickness of from 50 nm to 500 nm. The semiconductor device (for instance a photovoltaic device) of the invention may be a single- junction device. Alternatively, it may be a tandem junction or multi-junction device, for instance a tandem junction or multi-junction solar cell. In a tandem junction or multi-junction devices (for instance tandem junction or multi-junction photovoltaic devices) of the invention, the herein disclosed crystalline compounds may be combined with known technologies to deliver optimised performance. Typically, when the photovoltaic device of the invention is a tandem junction photovoltaic device, the device additionally comprises a further photoactive region, i.e. a further region which absorbs light and which may then generate free charge carriers. The further photoactive region is other than the region which comprises the layer comprising the crystalline compound and the adjacent layers comprising charge-transporting materials (electron- and hole- transporting materials, respectively). The further photoactive region is generally outside of the region which comprises the layer comprising the crystalline compound and the adjacent layers comprising charge (electron- and hole-) transporting materials. Thus, the further photoactive region may be disposed between the first electrode and the layer comprising a charge (electron or hole) transporting material, or between the second electrode and the layer comprising a charge (hole or electron) transporting material, in the device of the invention as defined herein. Typically, when the photovoltaic device of the invention is a multi-junction photovoltaic device, the device additionally comprises a plurality of further photoactive regions. Each one of the further photoactive regions may be disposed between the first electrode and the layer comprising a charge (electron or hole) transporting material, or between the second electrode and the layer comprising a charge (hole or electron) transporting material, in the device of the invention as defined herein. Typically, the or each further photoactive region comprises at least one layer of a semiconductor material. The semiconductor material may for instance comprise silicon. It may for instance comprise crystalline silicon. Alternatively, for example, the semiconductor material may comprise copper zinc tin sulphide, copper zinc tin selenide, copper zinc tin selenide sulphide, copper indium gallium selenide, copper indium gallium diselenide or copper indium selenide. Thus, for instance, when the photovoltaic device of the invention is a tandem junction photovoltaic device, the further photoactive region may be a conventional silicon solar cell. The further photoactive region may be a conventional thin film solar cell which may, for instance, comprise crystalline silicon (c-Si) or another thin film technology such as copper zinc tin sulphide, copper zinc tin selenide, copper zinc tin selenide sulphide, copper indium gallium selenide, copper indium gallium diselenide or copper indium selenide. The further photoactive region is preferably a silicon sub-cell. When the photovoltaic device of the invention is a multi-junction photovoltaic device, at least one of the further photoactive regions may be a conventional silicon solar cell. At least one of the further photoactive regions may be a conventional thin film solar cell which may, for instance, comprise crystalline silicon or another thin film technology such as copper zinc tin sulphide, copper zinc tin selenide, copper zinc tin selenide sulphide, copper indium gallium selenide, copper indium gallium diselenide or copper indium selenide. Preferably, the at least one of the further photoactive regions is a silicon sub-cell, typically a silicon sub-cell comprising crystalline silicon. Thus, the photovoltaic device may be a multi-junction photovoltaic device comprising silicon sub- cell comprising crystalline silicon. In one preferred embodiment, however, the optoelectronic device of the present invention is a light- emitting device. It may for instance be a light emitting diode. Process – solution processing The invention also provides a process for producing a film comprising a crystalline compound comprising (i) One or more first cations, A; (ii) One or more second cations, B; (iii) One or more third cations, B’; (iv) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, said process comprising disposing a film-forming solution on a substrate wherein the film-forming solution comprises the one or more first cations, A, the one or more second cations, B, the one or more third cations, B’, and the one or more anions, X. The one or more first cations, A, may be any first cations as described herein. The one or more second cations, B, may be any second cations as described herein. The one or more third cations, B’, may be any third cations as described herein. The one or more anions, X, may be any anions as described herein. Typically, the one or more first cations, A, comprise Cu⁺, the one or more second cations, B, comprise Ag⁺, the one or more third cations B’, comprise Bi³⁺ and the one or more anions, X, comprise I-. Typically, the process comprises dissolving at least one compound comprising the one or more first cations, A, the one or more second cations, B, the one or more third cations, B’, and/or the one or more anions, X, in a solvent to form the film-forming solution. The process may comprise dissolving a compound comprising the one or more first cations, A, dissolving a compound comprising the one or more second cations, B, and dissolving a compound comprising the one or more third cations, B’, in a solvent to form the film-forming solution. Typically, at least one of the compound comprising the one or more first cations, A, the compound comprising the one or more second cations, B, and the compound comprising the one or more third cations, B’, also comprises the one or more anions, X. Therefore, the compound comprising the one or more first cations, A, may be a compound of formula AX_n, where n is an integer based on the charge of the one or more first cations. The compound comprising the one or more second cations, B, may be a compound of formula BX_n, where n is an integer based on the charge of the one or more second cations. The compound comprising the one or more third cations, B’, may be a compound B’X_n, where n is an integer based on the charge of the one or more third cations. The process may also comprise dissolving a compound comprising the one or more anions, X, in a solvent to form the film-forming solution. Typically, the solvent comprises an organic solvent. For instance, the solvent may comprise a polar organic solvent. The solvent may comprise a polar aprotic organic solvent. Preferably, the solvent is selected from dimethylsulfoxide, pyridine, dimethylformamide, N-butylamine, sulfolane (tetrahydrothiophene 1,1-dioxide) and mixtures thereof. For instance the solvent may be a mixture of dimethylsulfoxide and dimethylformamide. In some embodiments, the solvent comprises pyridine. Adding pyridine may improve the morphology of the films of the compound. The process may comprise dissolving a compound comprising the one or more first cations, A, in a first solvent to produce a first solution, dissolving a compound comprising the one or more second cations, B, and a compound comprising the one or more third cations, B’, in a second solvent to produce a second solution, and contacting the first solution with the second solution to form the film-forming solution. The first and/or the second solution may comprise the one or more anions, X. Both the first and second solutions may comprise the one or more anions, X. The process may comprise additionally dissolving a compound comprising the one or more anions, X, in the first solvent and/or second solvent. For instance, the process may comprise dissolving a compound comprising the one or more first cations, A, in a first solvent to produce a first solution, dissolving a compound comprising the one or more second cations, B, and a compound comprising the one or more third cations, B’, and a compound comprising the one or more anions, X, in a second solvent to produce a second solution, and contacting the first solution with the second solution to form the film-forming solution. Typically, the first solvent is different from the second solvent. The first and second solvents are usually organic solvents. For instance, the first and second solvents may be polar organic solvents. The first and second solvents may be polar aprotic organic solvents. Preferably, the first and second solvents are selected from dimethylsulfoxide, pyridine, dimethylformamide, N-butylamine, sulfolane (tetrahydrothiophene 1,1-dioxide) and mixtures thereof. In some embodiments, one of the first or second solvents comprises pyridine. Adding pyridine may improve the morphology of the films of the compound. The process may comprise a step of dissolving a compound comprising the one or more first cations, A, in a first solvent to produce a first solution is performed at a first temperature, wherein the first temperature is between 30 ºC and 250 ºC. Preferably the first temperature is between 100 ºC and 200 ºC. The step of dissolving a compound comprising the one or more second cations, B, and a compound comprising the one or more third cations, B’, in a second solvent to produce a second solution may be performed at a second temperature, wherein the second temperature is between 30 ºC and 250 ºC. Preferably the second temperature is between 75 ºC and 150 ºC. The film-forming solution may be stirred at a third temperature prior to disposing the film-forming solution on the substrate, where the third temperature is between 30 ºC and 200 ºC. Preferably the third temperature is between 75 ºC and 150 ºC. Typically, the substrate is preheated to a fourth temperature, wherein the fourth temperature is between 30 ºC and 200 ºC. Preferably the first temperature is between 100 ºC and 200 ºC. The substrate is typically kept at the fourth temperature whilst the film-forming solution is disposed on the substrate. Typically, the film-forming solution is disposed on the substrate by solution phase deposition. For instance, the film-forming solution may be disposed on the substrate by gravure coating, slot dye coating, screen printing, ink jet printing, doctor blade coating, spray coating, roll-to-roll (R2R) processing, and spin-coating. Preferably the film-forming solution is disposed on the substrate by spin-coating. Process – Vapor deposition The invention also provides a process for producing a film comprising a crystalline compound comprising (v) One or more first cations, A; (vi) One or more second cations, B; (vii) One or more third cations, B’; (viii) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, said process comprising: exposing the substrate to a vapour, which vapour comprises one or more reactants for producing said compound and allowing deposition of the vapour onto the substrate to produce a film of said crystalline compound thereon. Typically, the process comprises evaporating one or more compounds comprising the one or more first cations, A, the one or more second cations, B, the one or more third cations, B’, and the one or more anions, X, to form the vapour. Preferably, the steps of evaporating the one or more compounds and exposing the substrate to the vapour are performed in a vacuum. Preferably, each of the one or more compounds are contained in a crucible that is opaque to visible light. This has the advantage of preventing degradation of the one or more compounds in the case where at least one of the one or more compounds is light sensitive. Preferably wherein the one or more compounds are contained in alumina crucibles. Typically, the process comprises evaporating a first compound comprising the one or more first cations, A, evaporating a second compound comprising the one or more second cations, B, and evaporating a third compound comprising the one or more third cations, B’, to produce the vapour wherein at least one of the first, second and third compounds also comprises the one or more anions, X. In some instances, each of the first, second and third compounds comprises the one or more anions, X. Thus the process may comprise evaporating a first compound of formula AX_n, evaporating a second compound of formula BX_n, evaporating a third compound of formula B’X_n, wherein n is an integer dependent on the valence of the cation A, B or B’. In either process of the invention, the process may further comprises annealing the film. Typically, the process comprises annealing the film at a temperature of from 40 to 250 ºC, for instance from 40 to 200 ºC or from 40 to 150 ºC. Typically, the process comprises annealing the film for 30 seconds to two hours, typically from 5 minutes to one hour. In either process of the invention, the process may further comprise exposing the film comprising a crystalline compound to a passivating agent. More specifically, in either process of the invention, the process may further comprise annealing the film in the presence of a passivating agent to produce a passivated film. Advantages of employing the passivating agent are evidenced in Example 6. Typically, the process comprises annealing the film at a temperature of from 40 ºC to 250 ºC, for instance from 40 ºC to 200 ºC or from 40 ºC to 150 ºC, in the presence of the passivating agent. Preferably, the process comprises annealing the film at a temperature of about 150 ºC, or less than about 150 ºC, or about 130 ºC, or less than about 130 ºC, or about 110 ºC, or less than about 110 ºC, in the presence of the passivating agent. Alternatively, the process comprises annealing the film at room temperature, in the presence of the passivating agent. The process may for instance comprise annealing the film at a temperature of from room temperature to any of the temperatures mentioned above. The process may for instance comprise annealing the film at a temperature of from 15 ºC to 250 ºC, for instance from 15 ºC to 150 ºC or from 15 ºC to 130 ºC, or from 15 ºC to 110 ºC, in the presence of the passivating agent. The process may for instance comprise annealing the film at a temperature of from 80 ºC to 250 ºC, for instance from 100 ºC to 200 ºC or for instance from 120 ºC to 180 ºC, or from 130 ºC to 160 ºC, or from 140 ºC to 190 ºC, in the presence of the passivating agent. Typically, the process comprises annealing the film for 30 seconds to two hours, typically from 5 minutes to one hour. Typically the passivating agent is selected from one or more of a halogen, sulfur, selenium or an organic amine. More typically the passivating agent is selected from one or more of iodine, sulfur, selenium or methylamine. Preferably the passivating agent is iodine. The passivating agent may be present in the resulting passivated film in, for example, molecular or ionic form. Preferably, the passivating agent is applied to the film as a vapour. Accordingly, the process may comprise exposing the film to vapour comprising the passivating agent. More specifically, the process may comprise annealing the film in the presence of a vapour comprising the passivating agent, to produce the passivated film. The process may comprise annealing the film in an atmosphere of evaporated passivating agent, to produce the passivated film. Annealing the film to produce a passivated film may comprise applying the passivating agent to the film at a pressure of less than atmospheric pressure. For example, a pressure of about 5 kPa may be used to apply the passivating agent to the film. Alternatively the passivating agent may be applied to the film under atmospheric pressure, or at a pressure above atmospheric pressure. The passivating agent may be applied to the film in an atmosphere comprising air, or in an atmosphere containing substantially no air. In either process of the invention, the substrate may be a component for a semiconductor device. For instance, the substrate may comprise an first or second electrode, as described herein, and disposed on that electrode, an n-type layer or a p-type layer, as described herein. The substrate may comprise a further photoactive region, as described herein. In that case, the semiconductor device may be a tandem or multi-junction semiconductor device, for instance a tandem or multi- junction photovoltaic device. Thus the invention also provides a process for producing a semiconductor device comprising producing a film comprising a crystalline compound on a substrate using a process as described herein, and disposing one or more further components on the film to produce a semiconductor device. For instance, the process may further comprise disposing one or more n-type or a p-type layers, as described herein, on the film of the crystalline compound, then disposing a first or second electrode, as described herein, on the n-type or p-type layer. The skilled person would be well aware of methods for disposing such layers, for instance by spin-coating. The invention also provides the use of a compound as described herein as an optoelectronic material, preferably as a photovoltaic material or as an electroluminescent material. The invention also provides the use of a compound as described herein as a luminescent material. The invention also provides the use of a compound as described herein as a radiation detecting material, for instance as an X-ray scintillator. For instance, the invention also provides the use of a compound as described herein as a quantum cutter. Quantum cutting (also referred to as down-conversion) is the process in which a photon of higher energy is converted into multiple photons of lower energy. EXAMPLES Example 1 Materials and methods Cu₂AgBiI₆ powder and crystal synthesis The starting materials used were CuI (99.999% Sigma–Aldrich) and AgI (99.999% Alfa Aesar). BiI₃ and AgBiI₄ were synthesised as previously reported.¹ Powder synthesis was carried out in evacuated sealed fused silica tubes. The fused silica tubes had a 6 mm internal diameter, 1 mm thick walls and were sealed at 10^-4 mbar to 15 cm in length. The sample in the tube was cooled in liquid N₂ during the evacuation and sealing process to avoid sublimation of the iodides. CuI (0.0728g), AgI (0.0505g) and BiI₃ (0.1268g) powders (nominal composition Cu_1.85Ag_1.04Bi_1.04I_6.00, just off the Cu_4x(AgBi)_1-xI₄ solid solution line, were sealed in the evacuated fused silica tube. The tube was placed upright in a furnace and heated at 5°C/min to 610°C and kept there overnight before cooling at 5°C/min to 350°C where it was kept for 5 days. It was then cooled to room temperature at 5°C/min. The tube was scored and carefully broken open to retrieve the black powder, which was pelletised in to pellets 5 mm in diameter using a pellet press. The pellets were sealed in an evacuated fused silica tube. The second heating consisted of heating the powder up at 5°C/min to 350°C, keeping it there for 5 days before taking the tube out of the furnace and quenching the bottom half of it (containing the powder) to room temperature in a water bath. It was found to be crucial to quench the material from 350°C and keep the annealing temperatures below the melting point of the mixture to avoid a compositional inhomogeneity which spanned the Cu_4x(AgBi)_1-xI₄ solid solution line (Figures 20a and b). The tube was scored and carefully broken open to retrieve the black powder. A small crystal of approximate dimensions 20 μm × 30 μm × 10 μm, with an average SEM EDX composition of Cu_1.75(6)Ag_1.26(13)Bi_1.08(7)I_6.00(5), was picked out of this sample for structural studies via SCXRD (Figure 21). Cu₂AgBiI₆ thin film deposition 0.2418 g of AgI (1.03 mmol), 0.2381 g of CuI (1.25 mmol) and 0.5172 g of BiI₃ (0.877 mmol), corresponding to nominal composition Cu_1.53Ag_1.26Bi_1.07I_6.00 were dissolved in 2 ml of a mixture of DMSO and DMF with a 3:1 volume ratio. The solution was stirred at 150°C for 30 minutes. The hot solution was filtered with a PTFE syringe filter with 0.22 µm pore size. The filtered solution was then kept at 75°C during the entire spin-coating process. The solution was spin-coated on different substrates in a N₂-filled drybox, with a speed of 4000 rpm for 60 seconds with 4000 rpm/s acceleration. The spin-coated films were then annealed at 50°C in air for 45 minutes, and, after this, at 150°C for 3 minutes. The films had a thickness of 280 nm, measured using a contact profilometer. Cu₂AgBiI₆ device fabrication Fluorine doped tin oxide (FTO) coated glass (15 Ω/sq) were etched using a 2M HCl solution and Zn powder. The patterned FTO substrates were cleaned through sequential sonication in soap, deionized water, acetone and isopropanol. The substrates were dried with a N₂ gun and exposed to O₂ plasma for 10 minutes. The SnO₂ electron transport layer was prepared by spin-coating a solution of SnCl₄·5H₂O in isopropanol (17.5 mg/ml) on top of the FTO coated glasses using 3000 rpm speed for 30 seconds. The so-prepared films were annealed at 100°C for 10 min followed by an annealing at 180°C for 30 min in air. After this the Cu₂AgBiI₆ film was deposited as explained in the thin film deposition. The hole transport material (SpiroOMeTAD, Lumtec) was dissolved in chlorobenzene (85 mg/ml) and doped with 20 µl of LiTFSI (500 mg/ml in butanol) and with 30 µl of tert-butylpyridine. This solution was then deposited on the Cu₂AgBiI₆ layer by dynamic spin- coating in air at 2000 rpm for 45 seconds. The devices were then left overnight in a dry air atmosphere, and then completed by the evaporation of 100 nm silver contacts. The device architecture is shown inset in Figure 4d. Device characterisations were performed in air unless specified otherwise. Powder X-ray diffraction (PXRD) PXRD data were measured on a Panalytical X’Pert Pro diffractometer using Co Kα₁ radiation (λ = 1.7890 Å) in Bragg–Brentano geometry and an X’Celerator detector. Phase identification was carried out using the X’Pert HighScore Plus (Version 2.2a)² with the PDF-2-ICDD database. PXRD data of capillaries used for photostability assessment were measured on a Bruker D8 Advance diffractometer using monochromated Mo Kα₁ radiation (λ = 0.7093 Å). TopasAcademic (Version 5) was used to perform Pawley fittings and Rietveld refinements of the data. VESTA³ was used for graphical representation of the structures. Single crystal X-ray diffraction (SCXRD) Data were collected at 100K on a Rigaku MicroMax-007 HF diffractometer with a molybdenum rotating anode microfocus source and a Saturn 724+ detector using Rigaku Crystal Clear v2.0. Unit- cell indexation, data integration, and reduction were performed using Rigaku CrysAlisPro v171.38.43. The structure was solved and refined using SHELX-2013,⁴ implemented through Olex2.⁵ Compositional Analysis Scanning Electron Microscopy (SEM) and Energy Dispersive X-ray Spectroscopy (EDX) was used to measure the composition as a direct elemental analysis technique. Measurements were carried out using a Hitachi S-4800 SEM with an Oxford Instruments model 7200 EDS X-ray detector. Quantification was carried out using the microanalysis suite of the Inca Suite software (Version 4.15). All powders and crystals were sputtered with 15 nm Au to limit charging effects. Transmission Electron Microscopy (TEM) EDX was carried out using an JEOL JEM 2000FXII TEM microscope operating a W electron source operated at 200 kV, using an EDAX EDX detector, with quantification carried out using EDAX Genesis Spectrum (Version 5.217, 21-Jan- 2008). All samples were prepared by spreading a finely ground powder onto carbon coated Au grids. Beam intensity had to be lowered, by increasing the spot size of the beam, as to not decompose the samples. All compositions calculated from EDX measurements were normalised to the nominal iodide content unless otherwise stated. The one standard deviation (1σ) spread in compositions measured for powder standards CuI and AgI, as well was the synthesised BiI₃, CuBiI₄ and Cu₂AgBiI₆ samples, are shown in Table 1. Table 1. Relative errors of the 1σ standard deviation in the measured compositions obtained by SEM and TEM EDX. CuI and AgI were purchased, the remaining were synthesised. Photostability For the experiments, a Solar Light Model 16S-300-002 Solar Simulator was used which has a spectral output that complies with air mass 1.5 (AM1.5) per the ASTM standard definition. The combination of neutral density filters and lamp-to-sample distance allowed for the tuning of the intensity of the incident light to 1000 W m⁻² as measured by a Solar Light Pyranometer PMA2144 and datalogging radiometer PMA2100. Sample temperatures were monitored using a T-type thermocouple and were found to stay below 35°C. For measurements in sealed atmospheres, powder was loaded into thin-walled (0.01 mm wall thickness) borosilicate capillaries under ambient air, dry synthetic air, and helium, and these were sealed with a gas-oxygen torch. The capillaries were then placed in the solar simulator and subjected to the full solar spectrum at an intensity of 1000 W m⁻². Raman Spectroscopy The measurements were carried out on a Renishaw inVia Reflex with a Leica microscope utilising a 633 nm wavelength red laser with a maximum power of 6.5 mW. The powder samples were measured in borosilicate capillaries and were exposed to 0.5% of the maximum laser power to avoid decomposition of the sample. The spot size was 5 µm. Photoluminescence (PL) Spectroscopy The Cu₂AgBiI₆ film was mounted in a gas-exchange helium cryostat (Oxford Instruments, OptistatCF2) and photoexcited by a 398 nm picosecond pulsed diode laser (PicoHarp, LDH-D-C- 405M). The resultant PL was collected and coupled into a grating spectrometer (Princeton Instruments, SP-2558), which directed the spectrally dispersed PL onto a photon-counting detector (PDM series from MPD), whose timing was controlled with a PicoHarp300 TCSPC event timer. A laser fluence of 200 nJ cm^-2 was used for both the spectral and transient measurements of Cu₂AgBiI₆, which were both taken at a temperature of 295 K. The PL decay trace in Figure 3b was measured at a wavelength of 720 nm. The PL decay trace was fitted by a stretched exponential function I = I0 exp(−(t/τ)^β), where β is the distribution coefficient and τ is the time taken for the PL intensity to drop to I₀/e. Such stretched exponential functions have been used to phenomenologically account for the presence of a local distribution of monoexponential decay rates, whose average lifetime is given by τ_av = (τ/β) Γ(1/β), where Γ is the gamma function.^6-7 The PL spectrum for MAPbI₃ in Figure 16 was measured using the same experimental setup, except it was mounted in a cold-finger cryostat (Oxford Instruments, MicrostatHe) and detected with an iCCD (PI-MAX4, Princeton Instruments), under a laser excitation fluence of 490 nJ cm- ².The spectrum was previously published in Wright et al.⁸ UV-Visible absorption measurements using FTIR spectrometer The UV-Visible absorption spectra in Figure 2 were measured using a Bruker Vertex 80v Fourier transform infrared (FTIR) spectrometer, configured with a tungsten halogen lamp illumination source, a CaF₂ beamsplitter and a silicon detector. The samples were mounted in a gas-exchange helium cryostat (Oxford Instruments, OptistatCF2). The MAPbI₃ absorption spectrum was previously published in Davies et al.⁹ THz photoconductivity An amplified laser system (Spectra Physics, MaiTai–Empower–Spitfire) with a central wavelength of 800 nm, 35 fs pulse duration and 5 kHz repetition rate was used to generate THz radiation via the inverse spin Hall effect, using an emitter made of 2 nm of tungsten / 1.8 nm of Co40Fe40B20/ 2 nm of platinum, supported by a quartz substrate. The transmitted THz radiation was detected using free-space electro-optic sampling with a 1 mm thick ZnTe (110) crystal, a Wollaston prism and a pair of balanced photodiodes. The THz pulse was measured in transmission geometry. The pump beam was frequency-doubled to 400 nm by a β-barium-borate (BBO) crystal. Charge-carrier mobilities were calculated from the initial transmitted signal at time = 0 ps at fluences of 7.8, 15.3 μJcm^ିଶ, as outlined in Wehrenfennig et al.,¹⁰ and measurements were carried out under vacuum (< mbar). J–V characterisation J–V characterisation was measured using a Keithley 2400 sourcemeter and simulated air-mass 1.5 global tilt (AM1.5G) solar irradiation using a Wavelabs Sinus-220 light-emitting diode array, calibrated with a certified Si reference cell. The areas being measured were defined by using a black anodised aluminium mask placed directly in contact with glass side of the substrate and an enclosed sample holder, to shadow the rest of the device. Photothermal deflection spectroscopy (PDS) PDS is an ultrasensitive absorption measurement technique that detects heating of the sample due to the non-radiative relaxation of absorbed light and is insensitive to reflection and scattering. PDS enables the detection of absorbance signals with 5–6 orders of magnitude weaker than the band edge absorption. For the measurements, a monochromatic Pump light beam is shined on the sample (film on Quartz substrate), which on absorption produces a thermal gradient near the sample surface via non-radiative relaxation induced heating. This results in a refractive index gradient in the area surrounding the sample surface. This refractive index gradient is further enhanced by immersing the sample in an inert liquid FC-72 Fluorinert® (3M Company) which has a high refractive index change per unit change in temperature. A fixed wavelength CW laser probe beam is passed through this refractive index gradient producing a deflection proportional to the absorbed light at that particular wavelength, which is detected by a photo-diode and lock-in amplifier combination. Scanning through different wavelengths gives us the complete absorption spectra. Because this technique makes use of the non-radiative relaxation processes in the sample, it is immune to optical effects like interference and scattering. Furthermore, PDS technique is a powerful technique to measure the sub-bandgap tail states in a semiconductor up to an absorption coefficient of 1 cm^-1. Elliot Model Fitting We use Elliott’s formula which expresses the absorption coefficient as a linear combination of the absorption of a bound exciton with the absorption by the joint continuum of states (see Davies et al., Bimolecular Recombination in Methylammonium Lead Triiodide Perovskite is an Inverse Absorption Process. Nat. Commun.2018, 9 (1), 293; Wehrenfennig et al., High Charge Carrier Mobilities and Lifetimes in Organolead Trihalide Perovskites. Adv. Mater.2014, 26 (10), 1584- 1589; Elliott, R. J., Intensity of Optical Absorption by Excitons. Phys. Rev.1957, 108 (6), 1384- 1389; Sell et al., New Analysis of Direct Exciton Transitions: Application to GaP. Phys. Rev. Lett. 1971, 26 (6), 311-314): ^(^) = ^_௫(^) + ^_^^^௧(^) with Where E is the photon energy, E_x is the exciton binding energy, E_g is the band gap, n is the refractive index (at energy E), | _^^௩|^ଶ = |ۦΨ_^|^|Ψ_௩ۧ|^ଶ is the momentum matrix element, ^ is the Dirac delta function, Θ⁽^⁾ is the step function, ^ is the reduced mass, m₀ is the electron rest mass, e is the elementary charge, c is the speed of light in vacuum and ℏ is the reduced Planck’s constant. The experimental absorption spectrum is fitted with the following model: Where A is a fitting constant and the excitonic part of the absorption (^′_௫) and the continuum part (^′_^^^௧) are convoluted with gaussian broadening functions ^_௫ and ^_^^^௧, respectively. Optical Modelling The generalised transfer matrix method was used to model the optical response of the stack.¹¹ The python libraries Numpy and Scipy were used to perform the calculations. Transfer matrix calculations take the complex refractive index spectrum and thickness for each layer as input. The calculation provides us with absorptance of each layer, and the transmittance and reflectance of the stack. We assumed perfect internal quantum efficiency and calculated the short circuit current ^_^^ as the overlap integral of the AM1.5 solar spectrum with the absorptance. The JV curve of each ^ೇ sub-cell was modelled as a single diode: Here, ^ is the ideality factor. The recombination current ^_^ is calculated through the principle of detailed balance: ^_^ = i_{s the blackbody photon flux at 300K, and} ^^^_^^(^) is the absorptance calculated from the transfer matrix calculation. The following diode parameters were assumed for the Cu₂AgBiI₆ sub-cell:¹² ^_ௌ^௨^௧ = 5 ^Ω ⋅ ^^^ଶ, ^_ௌ^^^^^ = 4.2 Ω ⋅ = 0.01. The calculated JV parameters for the Cu₂AgBiI₆ sub-cell were: PCE The following were assumed for the Si sub- cell:¹³ ^ 10 ^Ω ⋅ ^^^ଶ ^ ^ଶ 04 Ω ⋅ ^^ ^ 104 ^^^ 0.0056. The calculated JV parameters for the Silicon sub-cell were: PCE = 12.1%, V_OC = 0.75 V, J_SC = 19.0 mA/cm², FF = 0.84. The thickness of the LiF (20-200nm) the two ITO layers (20-200nm), and Cu₂AgBiI₆ (1200−1800nm) were varied with the indicated bounds using a differential evolution algorithm till a PCE maximum was obtained. The stack used as input for the Transfer Matrix Calculations is given in Table 2. Table 2. The stack used as input for the Transfer Matrix Calculations The source of optical constants for each layer are also cited. For Cu₂AgBiI₆, a “synthetic” absorption co-efficient was created by splicing a 74 meV tail (gradient 1/0.074 eV) between 1.0 eV−1.88 eV, extracted from the gradient of the absorption onset from the combined PDS and FTIR datasets, to the absorption co-efficient obtained from FTIR measurements. The absorption co-efficient was set to zero below 1.0 eV. This was then converted into the extinction coefficient (Figure 23) which was transformed into the refractive index using the Kramer’s Kronig relation:¹⁴ ^(^) = The described method assumes the following: 1. Transfer Matrix Model Limitation: Layer roughness is much smaller than the wavelength of light (~550nm). 2. The “synthetic” absorption coefficient created by appending a tail of energy 74 meV between 1.88-1.45 eV characterises the material well. 3. The same transport materials conventionally used for hybrid perovskites, C60/SnO₂ and PolyTPD are used. This is a relatively minor assumption as most transport layers are thin and poorly absorbing in comparison to the absorber layer, so the precise transport layer does not matter. 4. The Cu₂AgBiI₆ sub-cell can be optimised to bring it to the same radiative efficiency (EQE_EL = 1%) as a well-performing hybrid lead halide perovskite cell. The same shunt and series resistances ( can also be achieved. Density Functional Theory Calculations All periodic density functional theory calculations were performed using version 5.4 of the the VASP code¹⁵ with the projector augmented wave method to describe core electrons.¹⁶ The crystal structure prediction package ChemDASH¹⁷ was used to generate low energy ordered configurations of atoms within (2a+b, a+2b, c) supercells of the disordered experimental structure. Four independent starting configurations were initially generated by randomly occupying the experimental mixed and partially occupied Bi, Ag and Cu sites within the cell. ChemDASH was then used to generate 249 further configurations from each starting configuration by swapping between Bi, Ag and vacancies on the octahedral sites, and Cu and vacancies on the tetrahedral sites. A basin hopping approach using a value for k_BT of 0.01 eV/atom for the Metropolis acceptance criterion was used to ensure that low energy configurations were generated. The geometry of each configuration was optimised using the optB86b-vdW functional¹⁸ to model the important van der Waals interactions between the large iodine anions within the structure. ChemDASH uses a multi-step approach to geometry optimisation, where the final step used here had a plane-wave energy cutoff of 500 eV and forces were minimised to below 0.02 eV/Å. The lowest energy configuration of the 249 generated starting from each of the four starting configurations were then taken forwards to calculate the electronic structure, giving four low energy configurations with significantly different Bi, Ag and Cu orderings. A more accurate geometry was obtained for each of these configurations by performing a further geometry optimisation with the optB86b-vdW functional but with a higher plane-wave energy cutoff of 550 eV and a denser k-point mesh with a k-point spacing of 0.1 Å⁻¹ and to a tighter force threshold of 0.001 eV/Å for convergence (Figure 24). A single-point calculation using the SCAN meta-GGA functional¹⁹ and including the effects of spin-orbit coupling was then used to compute the electronic structure of each of these four configurations, resulting in the partial density of states plots shown in Figure 2 and Figure 15. Reference list for methods section: 1. Sansom, H. C.; Whitehead, G. F. S.; Dyer, M. S.; Zanella, M.; Manning, T. D.; Pitcher, M. J.; Whittles, T. J.; Dhanak, V. R.; Alaria, J.; Claridge, J. B.; Rosseinsky, M. J., AgBiI4 as a Lead- Free Solar Absorber with Potential Application in Photovoltaics. Chem. Mater. 2017, 29 (4), 1538- 1549. 2. Degen, T.; Sadki, M.; Bron, E.; König, U.; Nénert, G., The HighScore Suite. Powder Diffr. 2014, 29 (S2), S13-S18. 3. Momma, K.; Izumi, F., VESTA: a Three-Dimensional Visualization System for Electronic and Structural Analysis. J. Appl. Crystallogr.2008, 41 (3), 653-658. 4. Sheldrick, G. M., A Short History of SHELX. Acta Crystallogr. A 2008, 64, 112-122. 5. Dolomanov, O. V.; Bourhis, L. J.; Gildea, R. J.; Howard, J. A. K.; Puschmann, H., OLEX2: a Complete Structure Solution, Refinement and Analysis Program. J. Appl. Crystallogr.2009, 42 (2), 339-341. 6. Parrott, E. S.; Green, T.; Milot, R. L.; Johnston, M. B.; Snaith, H. J.; Herz, L. M., Interplay of Structural and Optoelectronic Properties in Formamidinium Mixed Tin–Lead Triiodide Perovskites. Adv. Funct. Mater.2018, 28 (33), 1802803. 7. Lindsey, C. P.; Patterson, G. D., Detailed Comparison of the Williams–Watts and Cole– Davidson Functions. J. Chem. Phys.1980, 73 (7), 3348-3357. 8. Wright, A. D.; Verdi, C.; Milot, R. L.; Eperon, G. E.; Pérez-Osorio, M. A.; Snaith, H. J.; Giustino, F.; Johnston, M. B.; Herz, L. M., Electron–Phonon Coupling in Hybrid Lead Halide Perovskites. Nat. Commun. 2016, 7 (1), 11755. 9. Davies, C. L.; Filip, M. R.; Patel, J. B.; Crothers, T. W.; Verdi, C.; Wright, A. D.; Milot, R. L.; Giustino, F.; Johnston, M. B.; Herz, L. M., Bimolecular Recombination in Methylammonium Lead Triiodide Perovskite is an Inverse Absorption Process. Nat. Commun.2018, 9 (1), 293. 10. Wehrenfennig, C.; Eperon, G. E.; Johnston, M. B.; Snaith, H. J.; Herz, L. M., High Charge Carrier Mobilities and Lifetimes in Organolead Trihalide Perovskites. Adv. Mater. 2014, 26 (10), 1584-1589. 11. Katsidis, C. C.; Siapkas, D. I., General Transfer-Matrix Method for Optical Multilayer Systems with Coherent, Partially Coherent, and Incoherent Interference. Appl. Opt. 2002, 41 (19), 3978-3987. 12. Saliba, M.; Matsui, T.; Domanski, K.; Seo, J.-Y.; Ummadisingu, A.; Zakeeruddin, S. M.; Correa-Baena, J.-P.; Tress, W. R.; Abate, A.; Hagfeldt, A.; Grätzel, M., Incorporation of Rubidium Cations into Perovskite Solar Cells Improves Photovoltaic Performance. Science 2016, 354 (6309), 206-209. 13. Masuko, K.; Shigematsu, M.; Hashiguchi, T.; Fujishima, D.; Kai, M.; Yoshimura, N.; Yamaguchi, T.; Ichihashi, Y.; Mishima, T.; Matsubara, N.; Yamanishi, T.; Takahama, T.; Taguchi, M.; Maruyama, E.; Okamoto, S., Achievement of More Than 25% Conversion Efficiency With Crystalline Silicon Heterojunction Solar Cell. IEEE J PHOTOVOLT 2014, 4 (6), 1433-1435. 14. Ohta, K.; Ishida, H., Comparison Among Several Numerical Integration Methods for Kramers-Kronig Transformation. Appl. Spectrosc.1988, 42 (6), 952-957. 15. Kresse, G.; Furthmüller, J., Efficient Iterative Schemes for Ab Initio Total-Energy Calculations using a Plane-Wave Basis Set. Phys. Rev. B 1996, 54 (16), 11169-11186. 16. Kresse, G.; Joubert, D., From Ultrasoft Pseudopotentials to the Projector Augmented-Wave Method. Phys. Rev. B 1999, 59 (3), 1758-1775. 17. Sharp, P. M.; Dyer, M. S.; Darling, G. R.; Claridge, J. B.; Rosseinsky, M. J., Chemically Directed Structure Evolution for Crystal Structure Prediction. Phys. Chem. Chem. Phys. 2020, 22 (32), 18205-18218. 18. Klimeš, J.; Bowler, D. R.; Michaelides, A., Van der Waals Density Functionals Applied to Solids. Phys. Rev. B 2011, 83 (19), 195131. 19. Sun, J.; Ruzsinszky, A.; Perdew, J. P., Strongly Constrained and Appropriately Normed Semilocal Density Functional. Phys. Rev. Lett.2015, 115 (3), 036402. 20. Fourcroy, P. H.; Carre, D.; Thevet, F.; Rivet, J., Structure du Tetraiodure de Cuivre(I) et de Bismuth(III), CuBiI4. Acta Crystallogr. C 1991, 47 (10), 2023-2025. 21. Mazzarella, L.; Lin, Y.-H.; Kirner, S.; Morales-Vilches, A. B.; Korte, L.; Albrecht, S.; Crossland, E.; Stannowski, B.; Case, C.; Snaith, H. J.; Schlatmann, R., Infrared Light Management Using a Nanocrystalline Silicon Oxide Interlayer in Monolithic Perovskite/Silicon Heterojunction Tandem Solar Cells with Efficiency above 25%. Adv. Energy Mater.2019, 9 (14), 1803241. 22. Pan, S. S.; Zhang, Y. X.; Teng, X. M.; Li, G. H.; Li, L., Optical Properties of Nitrogen- Doped SnO2 Films: Effect of the Electronegativity on Refractive Index and Band Gap. J. Appl. Phys. 2008, 103 (9), 093103. 23. Ren, S. L.; Wang, Y.; Rao, A. M.; McRae, E.; Holden, J. M.; Hager, T.; Wang, K.; Lee, W. T.; Ni, H. F.; Selegue, J.; Eklund, P. C., Ellipsometric Determination of the Optical Constants of C60 (Buckminsterfullerene) Films. Appl. Phys. Lett.1991, 59 (21), 2678-2680. 24. Shibata, M.; Sakai, Y.; Yokoyama, D., Advantages and Disadvantages of Vacuum- Deposited and Spin-Coated Amorphous Organic Semiconductor Films for Organic Light-Emitting Diodes. J. Mater. Chem. C 2015, 3 (42), 11178-11191. 25. Holman, Z. C.; Filipič, M.; Descoeudres, A.; De Wolf, S.; Smole, F.; Topič, M.; Ballif, C., Infrared Light Management in High-Efficiency Silicon Heterojunction and Rear-Passivated Solar Cells. J. Appl. Phys. 2013, 113 (1), 013107. 26. Green, M. A., Self-Consistent Optical Parameters of Intrinsic Silicon at 300K including Temperature Coefficients. Sol. Energy Mater. Sol. Cells 2008, 92 (11), 1305-1310. 27. Treharne, R. E.; Seymour-Pierce, A.; Durose, K.; Hutchings, K.; Roncallo, S.; Lane, D., Optical Design and Fabrication of Fully Sputtered CdTe/CdS Solar Cells. J. Phys. Conf. Ser. 2011, 286, 012038. 28. Ciesielski, A.; Skowronski, L.; Trzcinski, M.; Szoplik, T., Controlling the Optical Parameters of Self-Assembled Silver Films with Wetting Layers and Annealing. Appl. Surf. Sci. 2017, 421, 349-356. Example 1 - Results Cu₂AgBiI₆ Crystal Structure We synthesised Cu₂AgBiI₆ powders and crystals by solid state-synthesis in evacuated fused-silica ampoules as described in the methods section above. We solved the structure of Cu₂AgBiI₆ using single crystal X-ray diffraction data (SCXRD) collected at 100 K. The observed reflections could be fitted with a complex twinning of four trigonal unit cells with space groups R3^തm and lattice parameters a = 4.2749(3) Å, c = 20.9395(16) Å, which is metrically cubic within 2σ error ( ^{^} ଶ_^ = ^5.998(1) ≈ √6). Listed in Table 3 are the contribution of each twin and the twinning matrices.

Table 3. The twinning matrix and contribution of each twin indexing the Cu₂AgBiI₆100K SCXRD data with a 4-fold twin of smaller trigonal unit cells. Twin 1 ൩ Twin 2 Refined mass 25.2(7) Isolated reflec 729 Overlapped re 733 −0.3335 −0.6669 −0.1666 ^ 1.0003 1.0005 −0.0001൩ Twin matrix 2.6652 −2.6675 0.3338 Twin 3 ൩ Twin 4 Refined mass fraction (%) 24.9(7) Isolated reflections 745 Overlapped reflections 733 0.9982 0.9973 −0.0001 T^{win matrix} _^ −0.6647 −0.3313 0.1663 ൩ 2.6594 −2.6646 0.3327 The twinning has been reported in more detail for AgBiI₄.25 The structure shown in Figure 1 and Table 6 was refined against the combined reflection listing of all twins. Refined parameters are in Table 4 and further details of the refinement are in the methods section above. Table 4. Cu₂AgBiI₆ structural refinement of single crystal X-ray data collected at 100 K. Identification code Cu2AgBiI6_SCXRD_100K Empirical formula Cu_2.15Ag_1.04Bi_0.92I_6.00 Formula weight 1202.46 Temperature (K) 100(2) Crystal system trigonal S_{pace group R3} ^ത _m a (Å) 4.2749(3) c (Å) 20.9395(16) Volume (Å³) 331.40(5) Z 1.19988 ρ_calc (g/cm³) 6.025 μ (mm^-1) 31.009 F(000) 506.0 Crystal size (mm) 0.02 × 0.03 × 0.01 Radiation MoKα (λ = 0.71073) 2θ range for data collection (°) 5.836 to 55.51 Index ranges -5 ≤ h ≤ 5, -5 ≤ k ≤ 5, -26 ≤ l ≤ 27 Reflections collected 501 Independent reflections 501 [R_sigma = 0.0204] Data/restraints/parameters 501/0/12 Goodness-of-fit on F² 1.130 Final R indexes [I≥2σ] R₁ = 0.0441, wR₂ = 0.1024 Final R indexes [all data] R₁ = 0.0495, wR₂ = 0.1053 Largest diff. peak/hole (eÅ^-3) 2.95/-4.31 The Cu₂AgBiI₆ structure consists of alternating layers of partially filled and vacant octahedral sites in a cubic close packed (CCP) iodide array. This can be considered a disordered CdCl₂ octahedral motif. The Ag+ and Bi3+ cations are shared by the octahedral site in a disordered fashion. The Bi3+ and Ag+ occupancies were constrained to the average TEM EDX pow-der composition of Cu2.15(16)Ag1.04(5)Bi0.92(7)I6.00(11) as 30.6% and 34.7%, respectively. The compositional constraint was required due to the high number of correlated parameters in the refinement deriving from cation disorder and the four twin components. Electron density in the difference Fourier map shows two Cu⁺ sites (Cu1 and Cu2) with equal occupancy. They occupy every possible tetrahedral site facilitated by the CCP iodide sub-lattice. The Cu⁺ occupancies were fixed to occupancies of 17.9%, in line with the measured composition. The bond distance and angles are summarised in Table 5. Table 5. Bond distances and bond angles for the Cu₂AgBiI₆ structure refined against 100K SCXRD data. Cu₂AgBiI₆100K SCXRD i-CdCl₂ Bond distances (Å) Bond angles (°) Oct1 I1 3.0471(9) I1 Oct1 I1 89.09(4) Cu1 I1 2.81(4) I1 Oct1 I1 90.91(4) Cu1 I1 2.583(12) I1 Cu1 I1 107.1(9) Cu2 I1 2.73(4) I1 Cu1 I1 111.7(8) Cu2 I1 2.558(11) I1 Cu2 I1 105.2(9) I1 I1 4.206(3) I1 Cu2 I1 113.3(8) I1 I1 4.2749(3) I1 I1 I1 58.96(4) I1 I1 4.343(3) 1 59.46(2) 60.519(19) 60.000(5) The structure can be considered as having the same structure type as the delithiated LiVO₂ structure with composition Li_0.22VO₂.³⁴ We performed a Pawley fit on room temperature laboratory PXRD data yielding lattice parameters of a = 4.3151(2) Å and c = 21.141(1) (Figure 1i), which is metrically cubic within error ( ^{^} ଶ_^ = ^6.0005(10) ≈ √6). Table 6. Refined structural data for the newly solved Cu₂AgBiI₆ structure solved from 100K SCXRD data with the composition constrained to match the average composition Cu_2.15(16)Ag_1.04(5)Bi_0.92(7)I_6.00(11) from TEM EDX. Bulk Stability Powders of previously reported CuBiI₄ and AgBiI₄ were synthesised by solid state synthesis in evacuated fused-silica ampoules as described in the methods section above. The PXRD pattern of CuBiI₄ was fitted to a cubic unit cell with lattice parameters a = 12.1580(2) Å, larger than the a = 12.134(6) Å reported by Fourcroy et al. (Figure 5).³⁵ SEM EDX confirmed a composition of Cu_1.21(5)Bi_1.11(7)I_4.00(9), within 3σ error of CuBiI₄ (Figure 6). We find that CuBiI₄ is a metastable material which decomposes back in to starting materials BiI₃ and CuI at room temperature, even in the dark (Figure 7). We could slow the rate of decomposition of CuBiI₄ by storing the powder at −20°C. In contrast, we find that Cu₂AgBiI₆ powder is stable when kept in the dark, in air, at room temperature. We exposed synthesised Cu₂AgBiI₆ and AgBiI₄ powders to simulated AM1.5 solar spectrum for one week, sealed in capillaries with synthetic (dry) air, laboratory air, and He atmospheres. AgBiI₄ and Cu₂AgBiI₆ showed no colour change after one week in the solar spectrum and showed no signs of decomposition by PXRD (Figure 8) or Raman spectroscopy (Figure 9). The Cu₂AgBiI₆ composition therefore represents the stabilisation of a Cu-containing bismuth iodide solar absorber and is as stable as AgBiI₄ under the investigated conditions. This is promising since unencapsulated devices using AgBiI₄ absorber layers have been shown to retain 96% of their initial PCE after 1000 hours of storage in air at 26% relative humidity.³⁶ Optical Properties We solution-processed Cu₂AgBiI₆ into thin films for optical property measurements and device fabrication, as we describe in the methods section above. In Figure 10 we present the nominal composition in the solution and the composition of the film measured by TEM EDX as Cu2.52(9)Ag1.02(7)Bi0.82(11)I6.00(20). TEM EDX showed that the films were Cu-rich compared to the nominal composition in the solution and contained Cu_3xBi_1-xI₃ impurities. We performed a Pawley fit of PXRD data collected on the film (Figure 11) which shows it to have a trigonal unit cell (R3^തm) with lattice parameters of a = 4.3476(8) Å, c = 20.868(9) Å and a small trigonal (R3^തm) impurity phase with lattice parameters a = 4.322(1) Å, c = 20.80(1) Å. This is consistent with the two phases identified in the TEM EDX. We found that the film deposition was very sensitive to annealing temperature which we optimised to a two-step anneal to improve film morphology from large rough dendritic crystallites (Figure 12), to a more uniform smooth film (Figure 13). We determined the absorption coefficient spectra of Cu₂AgBiI₆ thin films by a combination of using a Fourier Transform Infra-Red (FTIR) spectrometer to accurately determine the above band gap absorption spectra and Photothermal Deflection Spectroscopy (PDS) to accurately measure the low energy parts of the spectrum. The raw PDS data (Figure 14a) was scaled to match the FTIR data, then the two datasets were combined as shown in Figure 14b. Unusually, for this broad family of compounds, Cu₂AgBiI₆ presents a strong absorption coefficient profile presenting the step increase typical of a direct band gap semiconductor (Figure 2a). The absorption strength that we measure for Cu₂AgBiI₆ at the first peak just above the band edge (1.0 × 105 cm^-1) is considerably stronger than that for MAPbI3 films measured here (0.3 × 105 cm^-1) and reported in literature,³⁷ which is already a very strongly absorbing semiconductor near the band edge. Crucially, Cu₂AgBiI₆ has a much more suitable absorption profile compared to the alternative wide band gap, lead-free double perovskite Cs₂AgBiBr₆ which consists of an initial peak in the absorption spectrum centred at 2.8 eV, followed by a minimum. In the literature, it is unresolved whether this absorption peak in Cs₂AgBiBr₆ is attributed to excitonic contributions or to the nature of the density of states near the band edge.15, ^41-43 Although Tauc analysis is often used to approximate the band gap of lead halide perovskites, this is unphysical since Tauc analysis assumes that the absorption at the band edge is directly into the continuum of states, and neglects the exciton contribution which can dominate features near the band edge. The more accurate approach is a fit according to Elliott theory,^{39, 44} which accounts for contributions from the excitonic and continuum of states. In Figure 2b, we show a fit to the absorption coefficient based on the Elliott model which reveals a band gap of 2.06(1) eV and an exciton binding energy (E_B) of 25(2) meV. This value of the exciton binding energy is higher than that determined for MAPbI₃, but very similar to that determined for CsPbBr₃ and notably comparable to thermal energy at room temperature.^45,46 This indicates that under light absorption at room temperature, free carriers, as opposed to bound excitons, will be generated. The lowering of the band gap and exciton binding energies, in comparison to Cs₂AgBiBr₆ (E_g ≈ 2.2eV, E_B ≈ 220 meV),⁴⁷ is likely to originate from the presence of a stable crystalline framework in which the iodide network is supported. Both the strong absorption properties and low exciton binding energy are very encouraging for potential use as solar absorbers, compared to the previously reported double perovskites. To gain some insight into the nature of the electronic transitions underlying optical absorption, we have performed density functional theory calculations on ordered structural models of Cu₂AgBiI₆. Partial density of states plots (Figures 2c and 15) show that the bottom of the conduction band is dominated by Bi 6p and I 5p states, similar to AgBiI₄ and BiI₃.²⁵ In contrast, Cu 3d-states dominate at the top of the valence band in Cu₂AgBiI₆, mixed with the I 5p states which dominate when Cu is absent. Optical transitions near the band gap energy of Cu₂AgBiI₆ will involve considerable Cu 3d to Bi 6p/I 5p character, in contrast to the I 5p to Bi 6p/I 5p transitions present in Ag_1-3xBi_1+xI₄ and BiI_3. In Figure 2a we also show the photoluminescence (PL) of the Cu₂AgBiI₆ thin film, which we fit to a pseudo-Voigt function (convolution of a Gaussian and Lorentzian function) with a full width half maximum (FWHM) of 289 meV. The PL peak of Cu₂AgBiI₆ is centred at 1.71 eV corresponding to a Stokes shift of 350 meV compared to the estimated direct band gap. For comparison, we show the absorption and emission profile to that of MAPbI₃ in Figure 16. We can fit the PL of MAPbI₃ to a Gaussian function with a FWHM of 96 meV and a Stokes shift of 10 meV. Although the Stokes shift for Cu₂AgBiI₆ is larger than in MAPbI₃, it is still substantially less than the 1 eV separation between the direct gap energy and PL peak in the indirect band gap Cs₂AgBiBr₆.⁴¹ Due to the disordered nature of the Cu₂AgBiI₆ crystal structure, the sub-band gap states of the thin film were investigated using PDS. PDS is a scatter-free, surface-sensitive absorption measurement capable of assessing the presence of sub-band gap states. Interestingly, the PDS measurement reveals absorption at lower energies down to 1.25 eV due to sub-band gap states (Figure 3a). We recorded time resolved PL transients for the Cu₂AgBiI₆ thin films (Figure 3b) and fitted the decays to a stretched exponential function, yielding an average lifetime of 33 ns. This function phenomenologically accounts for a superposition of monoexponential decays⁴⁸ which may be a result of inhomogeneous trap distributions.⁴⁹ Longer charge-carrier lifetimes are more favourable for photovoltaic applications since they allow more time for the charge-carriers to reach the contacts and be extracted, but are sensitive to the trap density in the films, and hence their processing conditions: for MAPbI₃, monomolecular charge-carrier lifetimes ranging from 4 ns to over 1 μs have been reported.⁴⁹ PL lifetime measurements of Cs₂AgBiBr₆ have also been made and are reproduced from Longo et al. in Figure 3b,⁴¹ and a stretched exponential was found to well- describe the long-time decays in this material as well, highlighting the heterogeneity of recombination processes in both Cs₂AgBiBr₆ and Cu₂AgBiI₆. Both materials have an initial fast PL decay but Cu₂AgBiI₆ shows a higher proportion of signal from long-lived (>200 ns) PL compared to Cs₂AgBiBr₆, which is reflected in the lower average PL lifetime (10 ns) of the latter decay. To gain an insight into charge-carrier mobilities in Cu₂AgBiI₆, we performed transient THz photoconductivity measurements using optical-pump, terahertz-probe spectroscopy, which gave a value for the electron-hole sum mobility of 1.7(5) cm²V^-1s^-1. This value is higher than that measured for the double perovskite Cs₂AgBiBr₆ (0.8 cm²V^-1s^-1),⁵⁰ though not quite as high as found in MAPbBr₃ or in current best-in-class hybrid perovskites (8–70 cm²V^-1s^-1).^49,51 Charge-carrier mobilities can be limited by intrinsic factors such as couplings of charge carriers to phonons, but can also be influenced significantly by extrinsic factors such as crystallinity and energetic disorder.⁴⁹ Given the already-promising value measured here, improved understanding of both of these influences could lead to further enhancements of charge-carrier mobilities in Cu₂AgBiI₆, as was found with regards to crystallinity in the mixed-cation, mixed-halide lead halide perovskites.⁵¹ Good charge-carrier diffusion lengths are critical to efficient solar cell operation, and a simplified calculation of , using the values measured here and neglecting higher-order recombination, yields a value of 530 nm, showing a charge-carrier diffusion length suitable for charge extraction, despite the high cation disorder in the structure. Cu₂AgBiI₆ Single Junction Photovoltaic Devices In order to assess if this material does function as an absorber layer in a photovoltaic cell, we fabricated “n-i-p” planar heterojunction devices incorporating a compact SnO₂ n-type charge extraction layer, and spiro-OMeTAD hole-extraction layer. We fully describe the cell preparation and measurements in the methods section above, with the device architecture shown inset in Figure 17. The cell did function, and delivered a PCE of 0.43%, a J_sc of 1.54 mA/cm², a V_oc of 0.47 V and a fill factor of 59.6% (Figure 17a). The device shows hysteresis between the forward bias (FB)-to- short circuit (SC) and the SC-to-FB scan, with the first showing higher performances. However, it is interesting to note that the steady state performances, measured at the maximum power point, present good short-term stability, with both the current density and the PCE increasing over time (Figure 17b). The results show that this device architecture can deliver photocurrent and photovoltage, but it is clear that effort will be required to further optimise the devices. Here we have chosen the archetypical charge extraction materials and device configuration for lead-halide perovskite cells, and it is likely that a new selection of charge extraction layers will be required in order to reach the full potential for this material. In addition, we expect that improved understanding of the optoelectronic properties, and passivating defects, will also be important for device development. Optical Modelling of Cu₂AgBiI₆ on c-Si Tandem We have demonstrated a certain degree of PV operation from this new material, yet the fundamental optical and electronic properties appear to promise significantly higher performance. Prior to expending significant effort upon materials and device optimisation, however, it is important to estimate the ultimate potential efficiency achievable for this material. With knowledge of the above- and below-band gap optical properties of a solar absorber material, it is possible to model its performance in a PV cell, following a detailed balance thermodynamic approach,⁵² as we describe in the materials and methods section. In the thermodynamic assessment of a solar cell, Shockley and Queisser introduced an idealised step-function absorption profile, where the band gap is clearly defined.⁵³ For a real material, the absorption onset is never infinitely steep, and the “PV band gap” is defined as the steepest point of the absorptance curve, which is easily deduced by taking the maximum of the differential of the external quantum efficiency spectrum. This PV band gap is therefore not an intrinsic property of the material, but a property of the PV cell, which is influenced by both absorber layer thickness, its optical absorption properties and the overall optical structure of the solar cell. We constructed an optical and electronic model for a Cu₂AgBiI₆-on-Si tandem cell of the following structure LiF/ITO/SnO₂/C60/Cu₂AgBiI₆/PolyTPD/ITO/nc-SiO_x:H/ (i)a-Si:H/c-Si/(i)a-Si:H/(p)a-Si:H/AZO/Ag (Figure 4a). In Figure 4b, we show the PV band gap of the Cu₂AgBiI₆ junction, as a function of thickness of this layer. Very encouragingly, although the Elliott model band gap is 2.06(1) eV, the PV band gap drops from 2.0 eV, all the way down to 1.7 eV, for film thicknesses from 100 nm to 1200 nm. A band gap of 1.7 eV is close to optimal for combining with Si in a tandem cell, which will allow current matching of the two junctions.¹¹ For comparison, we show in Figure 18 that Cs₂AgBiBr₆ cannot be current matched with Si due to very low absorption in the red-end of the visible spectrum. We simulated the performance of a Cu₂AgBiI₆-on-Si tandem solar cell, using a transfer matrix optical model, coupled with a detailed balance approach for simulating the current voltage curves. We determined the diode parameters for simulating the current voltage curves via fitting lead halide perovskite and Si J-V curves, reported in literature (see materials section above for more details and assumptions made during the modelling). Our main assumption is that electronically, we can optimise the Cu₂AgBiI₆ to work as well as a lead-halide perovskite cell where radiative recombination accounts for 1% of the total recombination events (1% external radiative efficiency), which is well below the world record lead- halide perovskite cell, and so a reasonably conservative assumption, but with the optical properties, including sub gap absorption onset, of our experimentally measured Cu₂AgBiI₆ thin films. Our results suggest that with a thickness of 1710 nm, the Cu₂AgBiI₆ thin-film is capable of being the top-cell in a Si tandem, yielding a matched current density of 19.0 mA/cm² (Figure 4c), a V_OC of 1.92 V, FF of 83% and a corresponding PCE of 30.2% (Figure 4d). We show the influence of varying the Cu₂AgBiI₆ absorber layer thickness upon the tandem cell photovoltaic performance in Figure 19. Thus, this material could genuinely compete on efficiency with lead-halide perovskites integrated into multi-junction PV cells. Conclusion In summary, we have synthesised a stable compound, Cu₂AgBiI₆, as powder, crystals and solution processed thin films. The structure is based on a 2D edge-sharing octahedral network. Octahedral sites are occupied by Ag⁺ and Bi³⁺ in a disordered fashion and Cu⁺ occupies all possible tetrahedral sites located in the cubic close packed iodide sub-lattice. Fitting the absorption profile using the Elliott model shows a band gap of the continuum of states of 2.06(1) eV and an exciton binding energy of only 25(2) meV, and the PV band gap, which we determine from the spectral response of modelled PV cells, can be tuned between 1.7–2.0 eV by varying film thicknesses from 1200–100 nm. The steep rise in absorption from the band edge to a high absorption coefficient of 1.0 × 10⁵ cm^-1 just above the onset, several times higher than for MAPbI₃ (0.3 × 10⁵ cm^-1), indicates great promise for use as a thin-film absorber material in PV devices. With our experimentally determined optical data as input, we simulated the PV properties of a Cu₂AgBiI₆-on-Si tandem cell, and demonstrated the feasibility of this material delivering over 30% power conversion efficiency in such a device. The performance of this cation-decorated cubic close packed iodide array containing Bi³⁺ emphasises the scope for this chemistry to control optoelectronic properties without lead and beyond the perovskite structural family. In contrast, the highly studied lead-free double perovskite Cs₂AgBiBr₆ has a much less suitable absorption profile and exciton binding energy, and a band gap too large to be combined with c-Si in a tandem cell. Cu₂AgBiI₆ may also prove fruitful for other applications such as light emission and radiation detection. References for example 1 – results section 1. Leguy, A. M. A.; Azarhoosh, P.; Alonso, M. I.; Campoy-Quiles, M.; Weber, O. J.; Yao, J.; Bryant, D.; Weller, M. T.; Nelson, J.; Walsh, A.; van Schilfgaarde, M.; Barnes, P. R. F., Experimental and Theoretical Optical Properties of Methylammonium Lead Hal-ide Perovskites. Nanoscale 2016, 8 (12), 6317-6327. 2. Stoumpos, C. C.; Malliakas, C. D.; Kanatzidis, M. G., Semiconducting Tin and Lead Iodide Perovskites with Organic Cations: Phase Transitions, High Mobilities, and Near-Infrared Photoluminescent Properties. Inorg. Chem. 2013, 52 (15), 9019-9038. 3. Sun, S.; Salim, T.; Mathews, N.; Duchamp, M.; Boothroyd, C.; Xing, G.; Sum, T. C.; Lam, Y. M., The Origin of High Efficiency in Low-Temperature Solution-Processable Bilayer Organometal Halide Hybrid Solar Cells. Energy Environ. Sci.2014, 7 (1), 399-407. 4. Johnston, M. B.; Herz, L. M., Hybrid Perovskites for Photovoltaics: Charge-Carrier Recombination, Diffusion, and Radiative Efficiencies. Acc. Chem. Res.2016, 49 (1), 146-154. 5. Dong, Q.; Fang, Y.; Shao, Y.; Mulligan, P.; Qiu, J.; Cao, L.; Huang, J., Electron-Hole Diffusion Lengths > 175 μm in Solution-Grown CH₃NH₃PbI₃ Single Crystals. Science 2015, 347 (6225), 967-970. 6. Ponseca, C. S.; Savenije, T. J.; Abdellah, M.; Zheng, K.; Yartsev, A.; Pascher, T.; Harlang, T.; Chabera, P.; Pullerits, T.; Stepanov, A.; Wolf, J.-P.; Sundström, V., Organometal Halide Perovskite Solar Cell Materials Rationalized: Ultrafast Charge Generation, High and Microsecond-Long Balanced Mobilities, and Slow Recombination. J. Am. Chem. Soc.2014, 136 (14), 5189-5192. 7. Shi, D.; Adinolfi, V.; Comin, R.; Yuan, M.; Alarousu, E.; Buin, A.; Chen, Y.; Hoogland, S.; Rothenberger, A.; Katsiev, K.; Losovyj, Y.; Zhang, X.; Dowben, P. A.; Mohammed, O. F.; Sargent, E. H.; Bakr, O. M., Low Trap-State Density and Long Carrier Diffusion in Organolead Trihalide Perovskite Single Crystals. Science 2015, 347 (6221), 519-522. 8. National Renewable Energy Laboratory, Best Research‐Cell Efficiency Chart. https://www.nrel.gov/pv/cell‐efficiency.html (accessed May 2020). 9. Eperon, G. E.; Hörantner, M. T.; Snaith, H. J., Metal Halide Perovskite Tandem and Multiple- Junction Photovoltaics. Nat. Rev. Chem.2017, 1 (12), 0095. 10. Leijtens, T.; Bush, K. A.; Prasanna, R.; McGehee, M. D., Opportunities and Challenges for Tandem Solar Cells Using Metal Halide Perovskite Semiconductors. Nat. Energy 2018, 3 (10), 828- 838. 11. Hörantner, M. T.; Leijtens, T.; Ziffer, M. E.; Eperon, G. E.; Christoforo, M. G.; McGehee, M. D.; Snaith, H. J., The Potential of Multijunction Perovskite Solar Cells. ACS Energy Lett. 2017, 2 (10), 2506-2513. 12. Mahesh, S.; Ball, J. M.; Oliver, R. D. J.; McMeekin, D. P.; Nayak, P. K.; Johnston, M. B.; Snaith, H. J., Revealing the Origin of Voltage Loss in Mixed-Halide Perovskite Solar Cells. Energy Environ. Sci.2020, 13 (1), 258-267. 13. Wei, F.; Deng, Z.; Sun, S.; Xie, F.; Kieslich, G.; Evans, D. M.; Carpenter, M. A.; Bristowe, P. D.; Cheetham, A. K., The Synthesis, Structure and Electronic Properties of a Lead-Free Hybrid Inorganic-Organic Double Perovskite (MA)₂KBiCl₆ (MA = Methylammonium). Mater. Horiz.2016, 3 (4), 328-332. 14. Wei, F.; Deng, Z.; Sun, S.; Zhang, F.; Evans, D. M.; Kieslich, G.; Tominaka, S.; Carpenter, M. A.; Zhang, J.; Bristowe, P. D.; Cheetham, A. K., Synthesis and Properties of a Lead-Free Hybrid Double Perovskite: (CH₃NH₃)₂AgBiBr₆. Chem. Mater.2017, 29 (3), 1089-1094. 15. McClure, E. T.; Ball, M. R.; Windl, W.; Woodward, P. M., Cs₂AgBiX₆ (X = Br, Cl): New Visible Light Absorbing, Lead-Free Halide Perovskite Semiconductors. Chem. Mater. 2016, 28 (5), 1348-1354. 16. Slavney, A. H.; Hu, T.; Lindenberg, A. M.; Karunadasa, H. I., A Bismuth-Halide Double Perovskite with Long Carrier Recombination Lifetime for Photovoltaic Applications. J. Am. Chem. Soc.2016, 138 (7), 2138-2141. 17. Saparov, B.; Hong, F.; Sun, J.-P.; Duan, H.-S.; Meng, W.; Cameron, S.; Hill, I. G.; Yan, Y.; Mitzi, D. B., Thin-Film Preparation and Characterization of Cs₃Sb₂I₉: A Lead-Free Layered Perovskite Semiconductor. Chem. Mater.2015, 27 (16), 5622-5632. 18. Harikesh, P. C.; Mulmudi, H. K.; Ghosh, B.; Goh, T. W.; Teng, Y. T.; Thirumal, K.; Lockrey, M.; Weber, K.; Koh, T. M.; Li, S.; Mhaisalkar, S.; Mathews, N., Rb as an Alternative Cation for Templating Inorganic Lead-Free Perovskites for Solution Processed Photovoltaics. Chem. Mater. 2016, 28 (20), 7496-7504. 19. Lehner, A. J.; Fabini, D. H.; Evans, H. A.; Hébert, C.-A.; Smock, S. R.; Hu, J.; Wang, H.; Zwanziger, J. W.; Chabinyc, M. L.; Seshadri, R., Crystal and Electronic Structures of Complex Bismuth Iodides A₃Bi₂I₉ (A = K, Rb, Cs) Related to Perovskite: Aiding the Rational Design of Photovoltaics. Chem. Mater.2015, 27 (20), 7137-7148. 20. Hoye, R. L. Z.; Brandt, R. E.; Osherov, A.; Stevanović, V.; Stranks, S. D.; Wilson, M. W. B.; Kim, H.; Akey, A. J.; Perkins, J. D.; Kurchin, R. C.; Poindexter, J. R.; Wang, E. N.; Bawendi, M. G.; Bulović, V.; Buonassisi, T., Methylammonium Bismuth Iodide as a Lead-Free, Stable Hybrid Organic–Inorganic Solar Absorber. Chem. Eur. J 2016, 22 (8), 2605-2610. 21. Sun, S.; Tominaka, S.; Lee, J.-H.; Xie, F.; Bristowe, P. D.; Cheetham, A. K., Synthesis, Crystal Structure, and Properties of a Perovskite-Related Bismuth Phase, (NH₄)₃Bi₂I₉. APL Mater. 2016, 4 (3), 031101. 22. Hebig, J.-C.; Kühn, I.; Flohre, J.; Kirchartz, T., Optoelectronic Properties of (CH3NH3)3Sb2I9 Thin Films for Photovoltaic Applications. ACS Energy Lett.2016, 1 (1), 309-314. 23. Savory, C. N.; Walsh, A.; Scanlon, D. O., Can Pb-Free Halide Double Perovskites Support High-Efficiency Solar Cells? ACS Energy Lett.2016, 1 (5), 949-955. 24. Creutz, S. E.; Crites, E. N.; De Siena, M. C.; Gamelin, D. R., Colloidal Nanocrystals of Lead- Free Double-Perovskite (Elpasolite) Semiconductors: Synthesis and Anion Exchange to Access New Materials. Nano Lett.2018, 18 (2), 1118-1123. 25. Sansom, H. C.; Whitehead, G. F. S.; Dyer, M. S.; Zanella, M.; Manning, T. D.; Pitcher, M. J.; Whittles, T. J.; Dhanak, V. R.; Alaria, J.; Claridge, J. B.; Rosseinsky, M. J., AgBiI₄ as a Lead-Free Solar Absorber with Potential Application in Photovoltaics. Chem. Mater.2017, 29 (4), 1538-1549. 26. Hamdeh, U. H.; Nelson, R. D.; Ryan, B. J.; Bhattacharjee, U.; Petrich, J. W.; Panthani, M. G., Solution-Processed BiI3 Thin Films for Photovoltaic Applications: Improved Carrier Collection via Solvent Annealing. Chem. Mater.2016, 28 (18), 6567-6574. 27. Lehner, A. J.; Wang, H.; Fabini, D. H.; Liman, C. D.; Hébert, C.-A.; Perry, E. E.; Wang, M.; Bazan, G. C.; Chabinyc, M. L.; Seshadri, R., Electronic Structure and Photovoltaic Application of BiI₃. Appl. Phys. Lett.2015, 107 (13), 131109. 28. Brandt, R. E.; Kurchin, R. C.; Hoye, R. L. Z.; Poindexter, J. R.; Wilson, M. W. B.; Sulekar, S.; Lenahan, F.; Yen, P. X. T.; Stevanović, V.; Nino, J. C.; Bawendi, M. G.; Buonassisi, T., Investigation of Bismuth Triiodide (BiI₃) for Photovoltaic Applications. J. Phys. Chem. Lett.2015, 6 (21), 4297-4302. 29. Shao, Z.; Le Mercier, T.; Madec, M. B.; Pauporté, T., AgBi₂I₇ Layers with Controlled Surface Morphology for Solar Cells with Improved Charge Collection. Mater. Lett.2018, 221, 135-138. 30. Zhang, B.; Lei, Y.; Qi, R.; Yu, H.; Yang, X.; Cai, T.; Zheng, Z., An In-Situ Room Temperature Route to CuBiI₄ Based Bulk-Heterojunction Perovskite-Like Solar Cells. Sci. China Mater.2018. 31. Turkevych, I.; Kazaoui, S.; Ito, E.; Urano, T.; Yamada, K.; Tomiyasu, H.; Yamagishi, H.; Kondo, M.; Aramaki, S., Photovoltaic Rudorffites: Lead-Free Silver Bismuth Halides Alternative to Hybrid Lead Halide Perovskites. ChemSusChem 2017, 10 (19), 3754-3759. 32. Pai, N.; Lu, J.; Gengenbach, T. R.; Seeber, A.; Chesman, A. S. R.; Jiang, L.; Senevirathna, D. C.; Andrews, P. C.; Bach, U.; Cheng, Y.-B.; Simonov, A. N., Silver Bismuth Sulfoiodide Solar Cells: Tuning Optoelectronic Properties by Sulfide Modification for Enhanced Photovoltaic Performance. Adv. Energy Mater.2019, 9 (5), 1803396. 33. Hu, Z.; Wang, Z.; Kapil, G.; Ma, T.; Iikubo, S.; Minemoto, T.; Yoshino, K.; Toyoda, T.; Shen, Q.; Hayase, S., Solution-Processed Air-Stable Copper Bismuth Iodide for Photovoltaics. ChemSusChem 2018, 11 (17), 2930-2935. 34. de Picciotto, L. A.; Thackeray, M. M.; David, W. I. F.; Bruce, P. G.; Goodenough, J. B., Structural Characterization of Delithiated LiVO₂. Mater. Res. Bull.1984, 19 (11), 1497-1506. 35. Fourcroy, P. H.; Carre, D.; Thevet, F.; Rivet, J., Structure du Tetraiodure de Cuivre(I) et de Bismuth(III), CuBiI₄. Acta Crystallogr. C 1991, 47 (10), 2023-2025. 36. Lu, C.; Zhang, J.; Sun, H.; Hou, D.; Gan, X.; Shang, M.-h.; Li, Y.; Hu, Z.; Zhu, Y.; Han, L., Inorganic and Lead-Free AgBiI₄ Rudorffite for Stable Solar Cell Applications. ACS Appl. Energy Mater.2018, 1 (9), 4485-4492. 37. De Wolf, S.; Holovsky, J.; Moon, S.-J.; Löper, P.; Niesen, B.; Ledinsky, M.; Haug, F.-J.; Yum, J.-H.; Ballif, C., Organometallic Halide Perovskites: Sharp Optical Absorption Edge and Its Relation to Photovoltaic Performance. J. Phys. Chem. Lett.2014, 5 (6), 1035-1039. 38. Longo, G.; Mahesh, S.; Buizza, L. R. V.; Wright, A. D.; Ramadan, A. J.; Abdi-Jalebi, M.; Nayak, P. K.; Herz, L. M.; Snaith, H. J., Understanding the Performance-Limiting Factors of Cs₂AgBiBr₆ Double-Perovskite Solar Cells. ACS Energy Lett.2020, 2200-2207. 39. Davies, C. L.; Filip, M. R.; Patel, J. B.; Crothers, T. W.; Verdi, C.; Wright, A. D.; Milot, R. L.; Giustino, F.; Johnston, M. B.; Herz, L. M., Bimolecular Recombination in Methylammonium Lead Triiodide Perovskite is an Inverse Absorption Process. Nat. Commun.2018, 9 (1), 293. 40. Wright, A. D.; Verdi, C.; Milot, R. L.; Eperon, G. E.; Pérez-Osorio, M. A.; Snaith, H. J.; Giustino, F.; Johnston, M. B.; Herz, L. M., Electron–Phonon Coupling in Hybrid Lead Halide Perovskites. Nat. Commun.2016, 7 (1), 11755. 41. Schade, L.; Wright, A. D.; Johnson, R. D.; Dollmann, M.; Wenger, B.; Nayak, P. K.; Prabhakaran, D.; Herz, L. M.; Nicholas, R.; Snaith, H. J.; Radaelli, P. G., Structural and Optical Properties of Cs₂AgBiBr₆ Double Perovskite. ACS Energy Lett.2019, 4 (1), 299-305. 42. Bartesaghi, D.; Slavney, A. H.; Gélvez-Rueda, M. C.; Connor, B. A.; Grozema, F. C.; Karunadasa, H. I.; Savenije, T. J., Charge Carrier Dynamics in Cs₂AgBiBr₆ Double Perovskite. J. Phys. Chem. C 2018, 122 (9), 4809-4816. 43. Filip, M. R.; Hillman, S.; Haghighirad, A. A.; Snaith, H. J.; Giustino, F., Band Gaps of the Lead-Free Halide Double Perovskites Cs₂BiAgCl₆ and Cs₂BiAgBr₆ from Theory and Experiment. J. Phys. Chem. Lett.2016, 7 (13), 2579-2585. 44. Elliott, R. J., Intensity of Optical Absorption by Excitons. Phys. Rev. 1957, 108 (6), 1384- 1389. 45. Galkowski, K.; Mitioglu, A.; Miyata, A.; Plochocka, P.; Portugall, O.; Eperon, G. E.; Wang, J. T.-W.; Stergiopoulos, T.; Stranks, S. D.; Snaith, H. J.; Nicholas, R. J., Determination of the Exciton Binding Energy and Effective Masses for Methylammonium and Formamidinium Lead Tri-Halide Perovskite Semiconductors. Energy Environ. Sci.2016, 9 (3), 962-970. 46. Yang, Z.; Surrente, A.; Galkowski, K.; Miyata, A.; Portugall, O.; Sutton, R. J.; Haghighirad, A. A.; Snaith, H. J.; Maude, D. K.; Plochocka, P.; Nicholas, R. J., Impact of the Halide Cage on the Electronic Properties of Fully Inorganic Cesium Lead Halide Perovskites. ACS Energy Lett.2017, 2 (7), 1621-1627. 47. Wu, C.; Zhang, Q.; Liu, Y.; Luo, W.; Guo, X.; Huang, Z.; Ting, H.; Sun, W.; Zhong, X.; Wei, S.; Wang, S.; Chen, Z.; Xiao, L., The Dawn of Lead-Free Perovskite Solar Cell: Highly Stable Double Perovskite Cs2AgBiBr6 Film. Adv. Sci.2018, 5 (3), 1700759. 48. Patel, J. B.; Milot, R. L.; Wright, A. D.; Herz, L. M.; Johnston, M. B., Formation Dynamics of CH₃NH₃PbI₃ Perovskite Following Two-Step Layer Deposition. J. Phys. Chem. Lett.2016, 7 (1), 96-102. 49. Herz, L. M., Charge-Carrier Dynamics in Organic-Inorganic Metal Halide Perovskites. Annu. Rev. Phys. Chem.2016, 67 (1), 65-89. 50. Hutter, E. M.; Gélvez-Rueda, M. C.; Bartesaghi, D.; Grozema, F. C.; Savenije, T. J., Band- Like Charge Transport in Cs₂AgBiBr₆ and Mixed Antimony–Bismuth Cs₂AgBi_1–xSb_xBr₆ Halide Double Perovskites. ACS Omega 2018, 3 (9), 11655-11662. 51. Rehman, W.; McMeekin, D. P.; Patel, J. B.; Milot, R. L.; Johnston, M. B.; Snaith, H. J.; Herz, L. M., Photovoltaic Mixed-Cation Lead Mixed-Halide Perovskites: Links Between Crystallinity, Photo-Stability and Electronic Properties. Energy Environ. Sci.2017, 10 (1), 361-369. 52. Nayak, P. K.; Mahesh, S.; Snaith, H. J.; Cahen, D., Photovoltaic Solar Cell Technologies: Analysing the State of the Art. Nat. Rev. Mater.2019, 4 (4), 269-285. 53. Shockley, W.; Queisser, H. J., Detailed Balance Limit of Efficiency of p‐n Junction Solar Cells. J. Appl. Phys.1961, 32 (3), 510-519. Example 2 Introduction Exploration of the quaternary phases in the Cu–Ag–Bi–I field gives the opportunity to combine the distinct structural chemistries of Ag⁺ and Cu⁺ with iodide to tune the dimensionality of the octahedral network, improve phase-, and photo-stability, and can be used to reduce the amount of photo-sensitive and expensive Ag in the system. Cu⁺ is tetrahedrally coordinated in CuI and CuBiI₄, where Bi³⁺ is octahedral in a defect spinel structure. Ag⁺, while also tetrahedral in the binary AgI, is octahedral in AgBiI₄, where diffraction methods cannot differentiate 2D and 3D arrangements of the occupied sites. Here we synthesise the previously reported Cu-containing ternary CuBiI₄, revisit its reported crystal structure and synthesise the new quaternary CuAgBiI₅ compound. We have previously reported the structure and properties of Cu₂AgBiI₆ but here we, crucially, describe the challenging exploratory synthesis which led to the discovery of CuAgBiI₅ and Cu₂AgBiI₆. We add CuAgBiI₅ to the known compounds in a summary of the CuI–AgI–BiI₃ phase field in Figure 25a and aim to unify the understanding of the structural features of the currently known materials, showing how composition relates to structure. Optoelectronic properties are carried out on solution processed thin films of CuAgBiI₅ that we optimised for spectroscopic investigations and compare to those reported for Cu₂AgBiI₆. Experimental Methods Synthetic Protocols and Phase Identification The starting materials used were CuI (99.999% Sigma–Aldrich) and AgI (99.999% Alfa Aesar). BiI₃ and AgBiI₄ powders were synthesized as previously reported.¹ Powder synthesis and crystal growth via chemical vapour transport (CVT) were carried out in evacuated sealed fused silica tubes. The fused silica tubes had a 6 mm internal diameter, 1 mm thick walls and were sealed at 10- ⁴ mbar to 15 cm in length. The sample in the tube was cooled in liquid N₂ during the evacuation and sealing process to avoid sublimation of the iodides. The full synthetic procedures for each compound are described in the following sections. CuBiI₄ powder A compositional screening of the Cu_3xBi_1-xI₃ composition range was carried out between 0.18 ≤ x ≤ 0.32, at 0.02 intervals. For this, CuI and BiI₃ powders were sealed in fused silica tubes. The sealed tubes were placed upright inside a furnace with enclosed heating elements and heated to 610°C to melt and mix the powders overnight. The furnace was cooled to 350°C at 5°C/min and kept there for 5 days. Afterwards, the tubes were quickly taken out of the furnace and the bottom half of each tube (containing the powder) was quenched to room temperature in a water bath. When the tubes were taken out of the furnace at 350°C and quenched, the red/purple vapour filling the tubes condensed above the water line away from the powder. The reactions formed a black powder at the bottom of the tube and tiny black crystals around the inside of the tubes. The small black crystals were too fragile to retrieve and would smear upon contact. The black powders were mechanically separated by carefully scoring and snapping open the tubes. CuAgBiI₅ powder and crystals The synthesis of CuAgBiI₅ was optimized from the procedure used in the exploratory synthesis. It was found to be crucial to quench the material from 350°C to avoid the compositional inhomogeneity which spanned the Cu_4x(AgBi)_1-xI₄ solid solution line (Figure 42a). Also, by using 350°C melting was avoided, and the composition was further homogenized (Figure 42b). CuI (0.0318g) AgI (0.0653g) and BiI₃ (0.1530g) powders (nominal composition Cu_0.68Ag_1.14Bi_1.06I_5.00, just off the Cu_4x(AgBi)_1-xI₄ solid solution line, were ground together in a pestle and mortar and pressed in to a pellet with a 5 mm diameter. The pellet was sealed in a fused silica tube. The tube was placed upright in a furnace and heated at 5°C/min to 350°C and kept there for 5 days. The tube was taken out after 5 days and the bottom half of it (containing the pellet) was quenched to room temperature in a water bath. The tube was scored and carefully broken open to retrieve the pellet which was then ground in to a black powder. A small crystal approx.20 μm × 15 μm × 10 μm, with an average SEM EDX composition of Cu_0.77(1)Ag_1.27(3)Bi_1.10(5)I_5.00(4), was picked out of this sample for structural studies via SCXRD (Figure 43). Cu₂AgBiI₆ powder and crystals The synthesis of Cu₂AgBiI₆ was optimized from the procedure used in the exploratory synthesis. As with CuAgBiI₅, it was also found to be crucial to quench the material from 350°C and keep the annealing temperatures below the melting point of the mixture to avoid a compositional inhomogeneity which spanned the Cu_4x(AgBi)_1-xI₄ solid solution line (Figures 42a and 42b). CuI (0.0728g), AgI (0.0505g) and BiI₃ (0.1268g) powders (nominal composition Cu_1.85Ag_1.04Bi_1.04I_6.00, just off the Cu_4x(AgBi)_1-xI₄ solid solution line, were placed in a fused silica tube. The tube was placed upright in a furnace and heated at 5°C/min to 610°C and kept there overnight before cooling at 5°C/min to 350°C where it was kept for 5 days. It was then cooled to room temperature at 5°C/min. A second heating consisted of heating the powder up at 5°C/min to 350°C, keeping it there for 5 days before taking the tube out of the furnace and quenching the bottom half of it (containing the powder) to room temperature in a water bath. The tube was scored and carefully broken open to retrieve the black powder. A small crystal of approximate dimensions 20 μm × 30 μm × 10 μm, with an average SEM EDX composition of Cu_1.75(6)Ag_1.26(13)Bi_1.08(7)I_6.00(5), was picked out of this sample for structural studies via SCXRD (Figure 46). CuAgBiI₅ film deposition 156.5 mg AgI and 393.1 mg BiI₃ were dissolved in 0.8 ml DMSO at 100°C over 15 minutes with constant stirring. In a separate vial, 127 mg of CuI powder were dissolved in 0.5 ml pyridine at 100°C over 15 minutes with constant stirring. The AgBiI in DMSO and CuI in pyridine solutions were quickly filtered through a 0.22 μm pore-size, 13 mm diameter PTFE filter and combined into a single vial. The solutions were kept stirring at 100°C during deposition. Microscope slide glass substrates were cut to size (approx. 24mm x 24 mm and sonicated in soap and DI water, acetone and IPA for 15 minutes, dried using an N₂ gun, then further cleaned in a U.V.-Ozone generator. After which, the substrates were quickly transferred to a dry box and heated to 150°C for deposition.100 ul of the hot CuAgBiI solution was deposited dymanically on to the hot substrate and spun at 4000 rpm for 45 seconds on a spincoater. The films were then annealed at 50°C for 30 minutes in air, over which the films turned from a clear orange film to a dark red-brown film. The film was then annealed at 150°C for 15 seconds in air to form the dark black film. Films were transferred to and kept in a N₂ filled glovebox until used for measurements. Powder X-ray diffraction (PXRD) Initial compositional screening and solution processing experiments made use of PXRD data measured on a Panalytical X’Pert Pro diffractometer using Co Kα₁ radiation (λ = 1.7890 Å) in Bragg–Brentano geometry and an X’Celerator detector. Phase identification was carried out using the X’Pert HighScore Plus (Version 2.2a)² with the PDF-2-ICDD database. PXRD data of capillaries used for photostability assessment were measured on a Bruker D8 Advance diffractometer using monochromated Mo Kα₁ radiation (λ = 0.7093 Å). PXRD patterns used for detailed structural analysis were collected at room temperature using the MAC detectors on the I11 beamline at Diamond Light Source (RAL, Oxfordshire, U.K.) with a wavelength of λ = 0.825898 Å. Samples were mixed with 50 vol% amorphous boron to reduce absorption effects and contained within 0.3 mm diameter borosilicate capillaries. TopasAcademic (Version 5) was used to perform Pawley fittings and Rietveld refinements of the data. VESTA³ was used for graphical representation of the structures. X-ray diffraction patterns collected on CuAgBiI₅ films were measured using a Panalytical X’Pert powder diffractometer, using radiation from a Cu-Kα1 source, across 2^ values ranging from 5-40^o. Neutron powder diffraction High-resolution time-of-flight (ToF) data were collected on the HRPD instrument at ISIS Neutron and Muon Source (RAL, Oxfordshire, U.K.) of powder samples at room temperature packed in to vanadium cans with a 6 mm diameter. Data from bank 1 and bank 2 were used for Rietveld refinement in combination with the PXRD I11 dataset. Single crystal X-ray diffraction (SCXRD) Data were collected at 100K on a Rigaku MicroMax-007 HF diffractometer with a molybdenum rotating anode microfocus source and a Saturn 724+ detector using Rigaku Crystal Clear v2.0. Unit-cell indexation, data integration, and reduction were performed using Rigaku CrysAlisPro v171.38.43. The structure was solved and refined using SHELX-2013,⁴ implemented through Olex2.⁵ Compositional Analysis Scanning Electron Microscopy (SEM) and Energy Dispersive X-ray Spectroscopy (EDX) was used to measure the composition as a direct elemental analysis technique. Measurements were carried out using a Hitachi S-4800 SEM with an Oxford Instruments model 7200 EDS X-ray detector. Quantification was carried out using the microanalysis suite of the Inca Suite software (Version 4.15). All powders and crystals were sputtered with 15 nm Au to limit charging effects. Transmission Electron Microscopy (TEM) EDX was carred out using an JEOL JEM 2000FXII TEM microscope operating a W electron source operated at 200 kV, using an EDAX EDX detector, with quantification carried out using EDAX Genesis Spectrum (Version 5.217, 21-Jan- 2008). All samples were prepared by spreading a finely ground powder onto carbon coated Au grids. Beam intensity had to be lowered, by increasing the spot size of the beam, as to not decompose the samples. All compositions calculated from EDX measurements were normalised to the nominal iodide content unless otherwise stated. Photostability Experiment For the experiments, a Solar Light Model 16S-300-002 Solar Simulator was used which has a spectral output that complies with air mass 1.5 (AM1.5) per the ASTM standard definition. The combination of neutral density filters and lamp-to-sample distance allowed for the tuning of the intensity of the incident light to 1000 Wm⁻² as measured by a Solar Light Pyranometer PMA2144 and datalogging radiometer PMA2100. Sample temperatures were monitored using a T-type thermocouple and were found to stay below 35°C. For measurements in sealed atmospheres, powder was loaded into thin-walled (0.01 mm wall thickness) borosilicate capillaries under ambient air, dry synthetic air, and helium, and these were sealed with a gas-oxygen torch. The capillaries were then placed in the solar simulator and subjected to the full solar spectrum at an intensity of1000 Wm⁻². Raman Spectroscopy The measurements were carried out on a Renishaw inVia Reflex with a Leica microscope utilising a 633 nm wavelength red laser with a maximum power of 6.5 mW. The powder samples were measured in borosilicate capillaires and were exposed to 0.5% of the maximum laser power to avoid decomposition of the sample. The spot size was 5 µm. X-ray Photoelectron Spectroscopy (XPS) Measurements were conducted on powders using a SPECS monochromatic Al Kα (1486.6 eV) X-ray source and a PSPMCD5 analyser calibrated to the Ag 3d_5/2 peak. For sample work function measurements, a 20 V bias was applied to the sample to shift the work function from the analyser work function. Samples were measured at a base pressure of 10^-10 mbar. Further details, including spectrometer calibration, can be found elsewhere.⁶ Charge neutralisation at the surface was achieved by means of a low energy electron flood gun and subsequent correction of the binding energy scale to the adventitious C 1s peak (284.8 eV). UV-Visible Absorption Absorption spectra were taken using a Bruker Vertex 80v Fourier-Transform Infrared (FTIR) spectrometer, with a tungsten halogen source and a silicon diode detector. Measurements were carried out under low vacuum (< 5 mbar). To calculate the absorption coefficient, film thicknesses were measured using a Veeco Dektak 150 profilometer. Absorption coefficients were calculated as: where is the sample thickness, blank quartz substrate and silver mirror were used as references for the transmission and reflection, respectively. Steady-State Photoluminescence Measurements Photoluminescence (PL) spectra were measured following excitation by a 398 nm continuous wave laser (PicoHarp, LDH-D-C-405M) at a power density of 40.5 Wcm^-2. The emitted PL was collected and coupled into a grating spectrometer (Princeton Instruments SP-2558), after which light was detected by an iCCD camera (PI-MAX4, Princeton Instruments). Measurements on three fresh films were carried out under vacuum (< 5 × 10^ିଶ mbar). Subsequently, two fresh films were exposed to air over 90 minutes, and measurements were carried out after 20, 60 and 90 minutes at a power density of 39.0 Wcm^-2. The measurements involved five acquisitions being taken, every 3 s, under constant illumination over 15 s. Time-Resolved Photoluminescence Measurements Time-Correlated Single Photon Counting was carried out using the same laser as above to photoexcite the thin films, but in pulsed excitation mode with a repetition rate of 5 MHz at fluences of 1480, 410, 180, 50 nJcm^-2. Photoluminescence was collected using the same monochromator, with a photon-counting detector (PDM series from MPD). Timing is controlled electronically using a PicoHarp300 event timer. Measurements were carried out on under vacuum (< 5 × 10^ିଶ mbar) and air. Time-Resolved Photoluminescence Fitting A least-squares fit was carried out on the highest-fluence PL decay using a stretched exponential of between 0 – 150 ns, and the average decay lifetime is calculated as ^_^௩ = n, D. C., Stretched Exponential Relaxation Arising from a Continuous Sum of Exponential Decays. Phys. Rev. B 2006, 74 (18), 184430) Optical-Pump Terahertz-Probe Spectroscopy An amplified laser system (Spectra Physics, MaiTai - Ascend - Spitfire) with a central wavelength 800 nm, 35 fs pulse duration and 5 kHz repetition rate was used to generate THz radiation via the inverse spin hall effect (see Seifert et al., Efficient Metallic Spintronic Emitters of Ultrabroadband _{Terahertz Radiation. Nature Photon 2016, 10 (7), 483-488) and was detected using free-space} electro-optic sampling with a 1 mm-thick ZnTe (110) crystal, a Wollaston prism and a pair of balanced photodiodes. The THz pulse was measured in transmission geometry. The pump beam was frequency-doubled to 400nm by a β-barium- borate (BBO) crystal. Measurements were carried out under low vacuum (< 10^-2 mbar). Charge-carrier mobilities were calculated from the initial transmitted signal at time t = 0 ps, at fluences of 4.9, 12.5, 25 μJcm^ିଶ. Calculation of Charge-Carrier Mobility At low fluences (< 50 μJcm^ିଶ) there is a linear relationship between the charge-carrier mobility ^ and the change in photoconductivity of the sample ΔS: The change in photoconductivity can in turn be related to the fractional change in terahertz transmission using a standard formula for thin-film samples:⁶ We can calculate the number of photoexcited charge carriers using: By combining equations S3, S2 and S1, we can thus calculate the effective charge-carrier mobility as: Given that the branching ratio 0 ≤ ^ ≤ 1, the calculated effective charge-carrier mobility is always an underestimate; only in the case of full conversion of photons to free charges does our value reflect the true mobility. In addition, our calculated value of ^ arises from changes in photoconductivity due to both electrons and holes, meaning we calculate an overall sum mobility. References for experimental methods section of example 2: 1. Sansom, H. C.; Whitehead, G. F. S.; Dyer, M. S.; Zanella, M.; Manning, T. D.; Pitcher, M. J.; Whittles, T. J.; Dhanak, V. R.; Alaria, J.; Claridge, J. B.; Rosseinsky, M. J., AgBiI₄ as a Lead- Free Solar Absorber with Potential Application in Photovoltaics. Chem. Mater.2017, 29 (4), 1538- 1549. 2. Degen, T.; Sadki, M.; Bron, E.; König, U.; Nénert, G., The HighScore Suite. Powder Diffr. 2014, 29 (S2), S13-S18. 3. Momma, K.; Izumi, F., VESTA: a Three-Dimensional Visualization System for Electronic and Structural Analysis. J. Appl. Crystallogr.2008, 41 (3), 653-658. 4. Sheldrick, G. M., A Short History of SHELX. Acta Crystallogr. A 2008, 64, 112-122. 5. Dolomanov, O. V.; Bourhis, L. J.; Gildea, R. J.; Howard, J. A. K.; Puschmann, H., OLEX2: a Complete Structure Solution, Refinement and Analysis Program. J. Appl. Crystallogr. 2009, 42 (2), 339-341. 6. Whittles, T. J.; Burton, L. A.; Skelton, J. M.; Walsh, A.; Veal, T. D.; Dhanak, V. R., Band Alignments, Valence Bands, and Core Levels in the Tin Sulfides SnS, SnS₂, and Sn₂S₃: Experiment and Theory. Chem. Mater.2016, 28 (11), 3718-3726. Example 2 - Results 2.1. Phase Identification and Composition The optimized synthetic procedures needed to attain homogenous powder samples of CuBiI₄, CuAgBiI₅ and Cu₂AgBiI₆ can be found in the materials section above. For the exploratory synthesis of the CuI–Ag–BiI₃ phase field CuI, AgI and BiI₃ powders were ground together in a pestle and mortar and pressed in to a pellet with a 5 mm diameter. The pellet was placed in a fused silica tube. The sealed tube was placed upright inside a furnace with enclosed heating elements and heated to 610°C to melt and mix the powders overnight. The furnace was cooled to 350°C at 5°C/min and kept there for 5 days. The furnace was then cooled to room temperature at 5°C/min. The tube was scored and carefully broken open to retrieve the pellet, which was then ground, resulting in a black powder. This was based on the same method used to previously synthesize AgBiI₄ powders.⁴⁴ The nominal compositions, regions of the phase field explored, and any modifications to this procedure for the final optimised procedures (such as the importance of avoiding the melt and quenching in the quaternary systems) are discussed in this section. To identify new and targeted compounds phase identification is carried out using Pawley fits to powder X-ray diffraction (PXRD) data to extract lattice parameters of phases (using LaB₆ as an internal standard) and comparing to those reported in the literature, and by SEM and TEM EDX as a direct measurement of elemental composition (using CuI, AgI and BiI3 powders as standards). Close stoichiometric representations are used to refer to the compounds, with their average compositions measured by SEM or TEM EDX shown in Figure 25c and Table 7. Table 7: the average compositions of the materials synthesized in this investigation, measured with either SEM EDX (CuBiI₄) or TEM EDX (CuAgBiI₅, Cu₂AgBiI₆), with 1σ errors. The stoichiometric compositions chosen to represent the materials are within error to the measured compositions. To understand the extracted lattice parameters and unit cells found in the exploratory synthesis, the reported unit cells for the ternary systems need to be collected and introduced. For ternary compound g i ( g i ) the unit cell has been reported to be cubic with lattice parameter a_c (and crystallize in the space group), or a metrically cubic small trigonal unit c_{ell with lattice parameters at =} _{, ct = √3^ୡ (and crystallize in the R3} ^ത _{m space group)} (Figures 26a and b, respectively). The two different structures associated with these unit cells are indistinguishable by PXRD.⁴⁴ The structure with the cubic unit cell is also indistinguishable from a 4-fold twinning of the structure with the metrically cubic small trigonal unit cell by single crystal X-ray diffraction (SCXRD). There is a third type of unit cell, the large trigonal cell, depicted in Figure 26c. This cell is useful to compare structures with the cubic unit cell or the small trigonal unit cell. The structure in the cubic unit cell can be transformed using the matrix in Figure 26 to an equivalent structure in the large trigonal unit cell with lattice parameters a_T = a_c/_√2, c_T = _√3^_^. The large trigonal cell has a volume smaller than the volume of the cubic unit cell, _்^ = _√3 _^^/2, where _^^ is the volume of the cubic unit cell. The small trigonal unit cell can also be transformed into the large trigonal unit cell by doubling it in the a and b directions using the transformation matrix in Figure 26. The large trigonal cell therefore has lattice parameters a_T = 2a_t, ^_T = ^_௧ , and a volume _T^ = 4 _௧^, where _௧^ is the volume of the small trigonal unit cell. The small and large trigonal unit cells can be metrically cubic when ^ _{√ √} respectively. If they are not, then rhombohedral strain is defined as a percentage, where S_{t, rhomb} = ×100 and ×100, for the small and large trigonal unit cells, respectively. Rhombohedral strain can be thought of as extending or contracting the cubic unit cell along the body diagonal (111)_cubic (Figure 26d). The Bi-rich x > 0 Ag_1-3xBi_1+xI₄ compounds (AgBi₂I₇, Ag₂Bi₃I₁₁) and CuBiI₄ have been reported with a cubic unit cell, ^{43, 47-48, 62}whereas the Ag-rich x < 0 Ag_1-3xBi_1+xI₄ compounds (Ag₂BiI₅, Ag₃BiI₆), have been reported in the small trigonal unit cell.^{46, 48, 62} All the reported ternary phases have cubic close-packed (CCP) iodide sub-lattices. 2.1.1. CuBiI₄ A compositional screening of the Cu_3xBi_1-xI₃ composition range was carried out between 0.18 ≤ x ≤ 0.32, at 0.02 intervals. The PXRD patterns of the series showed a pure powder was obtained for the nominal composition corresponding to x = 0.25 Cu_3xBi_1-xI₃ CuBiI₄ (Figure 33). Quenching was found to be crucial in attaining pure CuBiI₄ as slowly cooling to room temperature promoted the growth of CuI and BiI₃ impurities. BiI₃ impurities were seen for x > 0.25 samples, and CuI impurities for x < 0.25 samples. The PXRD pattern of the pure sample indexed to a cubic cell in the Fd3^തm space group as reported by Fourcroy et al. in 1991.⁴³ A Pawley fit gives a lattice parameter of a = 12.1592(3) Å, which is larger than the previously reported a = 12.134(6) Å. SEM EDX measurements gave an average composition of Cu_1.21(5)Bi_1.11(7)I_4.00(9), within 3σ error to CuBiI₄ (Figure 25c): this stoichiometric composition will be used from now on. The 1σ errors expressed in the parentheses gives an indication of the spread in composition and the relative errors compared to the starting materials CuI and BiI₃ are shown in Table 8. There is a comparable spread in the I mol% distribution between the starting materials and the product, however Cu and Bi mol% show more inhomogeneity which can suggest a disordered cation distribution in a well-defined iodide sub-lattice. Table 8. Relative errors of the 1σ standard deviation in the measured compositions obtained by SEM EDX. CuI and AgI were purchased, the remaning were synthesized. 2.1.2. CuAgBiI₅ and Cu₂AgBiI₆ Quaternary compounds were searched for in four locations in the CuI–AgI–BiI₃ phase field; compositions Cu(Ag_1.5Bi_0.5)I₄ and Cu_1.5Ag_0.5BiI₅, as well as the solid solution lines Cu_xAg_1-xBiI₄ (between Cu_1/4Bi_1/4I and Ag_1/4Bi_1/4I) and Cu_4x(AgBi)_1-xI₄ (between CuI and Ag_1/4Bi_1/4I )(Figure 25d). Cu(Ag_1.5Bi_0.5)I₄ is the composition where the material would have the same tetrahedral and octahedral occupancies as spinel MgAl₂O₄ whilst satisfying charge balance. The PXRD pattern showed that no quaternary Cu–Ag–Bi–I phases formed under the synthesis conditions used. Two phases were identified, and their lattice parameters were refined using a Pawley fit (Figure 34). The cubic F ^ത i blende phase had a lattice parameter of a = 6.2868(5) Å. Using the data from Chessin et al.⁶³ this lattice parameter corresponds to a mixed Cu⁺ and Ag⁺ y = 0.45 Cu_1-yAg_yI zinc blende (Cu_0.45Ag_0.55I). The lattice parameters of a small trigonal cell with a = 4.3552(4) Å and c = 20.761(2) Å are within the range reported for rhombohedral Ag-rich x < 0 Ag1-3xBi1+xI4 phases.^{47-48, 62} Cu_1.5Ag_0.5BiI₅ is a composition which gives a 1:1 tetrahedral-to-octahedral cation ratio. A Pawley fit to the PXRD pattern shows the presence of CuI and BiI₃ impurities and a Fd3^തm cubic phase (Figure 35). The lattice parameter of the Fd3^തm phase (a = 12.2050(3) Å) is larger than that reported for CuBiI₄ (a = 12.1580(2) Å) and smaller than that reported for AgBiI₄ (a = 12.21446(4) Å), suggesting a quaternary phase containing both Cu⁺ and Ag⁺. However, the large amounts of impurity phases show that the nominal composition is far from the new quaternary phase. The Cu_xAg_1-xBiI₄ solid solution (x = 0.25, 0.33, 0.50, 0.66, 0.75) corresponds to a line of compositions between the known materials CuBiI₄ and AgBiI₄. The PXRD patterns of samples _{synthesized can all be fitted to a mixture of a F4} ^ത _{3m Cu1-yAgyI zinc blende, BiI3 and a cubic Fd3} ^ത _m phase (Figure 36). The lattice parameters of the Cu_1-yAg_yI zinc blende and the cubic Fd3^തm phase across the series were refined with respect to a LaB₆ internal standard (Figure 37). The lattice parameter of the cubic Fd3^തm phase is refined as a = 12.2034(3) Å at nominal x = 0.25 (Cu_0.25Ag_0.75BiI₄) suggesting a quaternary phase containing both Cu⁺ and Ag⁺. The lattice parameter grows as x increases which indicates an increase in Ag⁺ content until it levels off with a lattice parameter of a = 12.2143(5) Å for nominal compositions x ≥ 0.66 (Cu_0.66Ag_0.33BiI₄), consistent with that reported for AgBiI₄ (a = 12.21446(4) Å). The lattice parameter of the new quaternary phase for nominal composition Cu_1.5Ag_0.5BiI₅ lies within this range of lattice parameters. The lattice parameter of the Cu_1-yAg_yI zinc blende phase also shows a small increase over the series suggesting a small amount of Ag⁺ is being substituted into the structure for Cu⁺ (< 5 mol% at x = 0.75). The extra CuI used in the reaction as x increases is therefore accounted for by an increase in the weight percentage of the zinc blende impurity phase. The range of lattice parameters over the series suggests that a range of quaternary Cu–Ag–Bi–I phases may exist under these synthetic conditions, however, since all samples contained impurity phases the targeted phases do not lie on the Cu_xAg_1-xBiI₄ solid solution line. The Cu_4x(AgBi)_1-xI₄ solid solution corresponds to a line of compositions between the known phases CuI and AgBiI₄. Samples were synthesized at x = 0.14, 0.15, 0.16, 0.18, 0.2, 0.21, 0.23, 0.24, 0.26, 0.27, 0.29, 0.31, 0.33, 0.35, 0.37, 0.43, 0.5. For 0.14 ≤ x ≤ 0.16 samples the PXRD data is fitted to a mixture of a cubic unit cell with space group _Fd3 ^ത _{m and a small trigonal unit cell with space group R3} ^ത _{m. The cubic Fd3} ^ത _{m phase has a constant} lattice parameter for 0.14 ≤ x ≤ 0.16, with an average value of a = 12.206(2) Å, suggesting a quaternary phase containing both Cu⁺ and Ag⁺ (Figure 38). The trigonal R3^തm phase also has constant lattice parameters with average values of a = 4.3530(4) Å, c = 20.764(2) Å, consistent with the reported x < 0 Ag_1-3xBi_1+xI₄ phases (Figure 39a). For 0.18 ≤ x ≤ 0.37 samples there remains a new quaternary cubic Fd3^തm phase with lattice parameter a = 12.209(1) Å. There are no PXRD peaks associated with the x < 0 Ag_1-3xBi_1+xI₄ phases however there is a Cu_1-yAg_yI zinc blende impurity which increases in weight percentage for higher values of x (Figure 39b) suggesting a pure quaternary phase between 0.16 < x < 0.18. Additionally, an interesting point in the series lies between 0.29 ≤ x ≤ 0.31 in which the PXRD appears to only show a new quaternary cubic Fd3^തm phase and the inclusion of the Cu1-yAgyI zinc blende impurity only slightly improves the Pawley fit by improving peak shapes (Figure 40a). The lattice parameter of the Cu_1-yAg_yI zinc blende impurity phase changes with composition and for 0.29 ≤ x ≤ 0.31 the impurity peaks are aligned with the main phase peaks and cannot be deconvoluted from the PXRD pattern during the Pawley fit. For this region, the Cu_1-yAg_yI zinc blende lattice parameter is interpolated to be half the lattice parameter of the main cubic Fd3^തm phase (Figure 38). As the Cu_1-yAg_yI zinc blende phase could not be identified, it is possible that this is a region where a pure quaternary phase also exists. TEM EDX was carried out on particles from samples x = 0.16, 0.18 (Figure 39c) and x = 0.27, 0.29, 0.31, 0.33 (Figure 40b) in an attempt to find the composition of the main cubic Fd3^തm phases. It was found that every individual sample exhibits particles with compositions spread along the Cu_4x(AgBi)_1-xI₄ solid solution line showing a large degree of compositional inhomogeneity. The observation that a single phase can fit the XRD pattern despite the compositional inhomogeneity seen in the EDX indicates that the response of the unit cell size to this inhomogeneity is limited. The synthetic protocols had to be improved to isolate two new compositionally homogeneous phases around these regions identified as pure by PXRD (x = 0.2 CuAgBiI₅ and x = 0.33 Cu₂AgBiI₆). It was found to be crucial to quench the CuAgBiI₅ and Cu₂AgBiI₆ powders from 350°C to avoid the compositional inhomogeneity which spanned the Cu_4x(AgBi)_1-xI₄ solid solution line (Figure 41a). Also, by using 350°C melting was avoided, and the composition was further homogenized (Figure 41b). The average composition of the CuAgBiI₅ powder was measured as Cu_0.88(17)Ag_1.10(6)Bi_0.98(8)I_5.00(11) by TEM EDX. The PXRD pattern indexes to a large trigonal cell in the R3^തm space group. A small crystal approx.20 μm × 15 μm × 10 μm, with an average SEM EDX composition of Cu_0.77(1)Ag_1.27(3)Bi_1.10(5)I_5.00(4), was picked out of this sample for structural studies via SCXRD (Figure 42). The average composition of the Cu₂AgBiI₆ powder was measured as Cu_2.15(16)Ag_1.04(5)Bi_0.92(7)I_6.00(11) by TEM EDX. The PXRD pattern indexes to a small trigonal cell in _{the R3} ^ത _{m space group. A small crystal of approximate dimensions 20 μm × 30 μm × 10 μm, with an} average SEM EDX composition of Cu_1.75(6)Ag_1.26(13)Bi_1.08(7)I_6.00(5), was picked out of this sample for structural studies via SCXRD (Figure S11). Larger millimetre-sized crystals of CuAgBiI₅ and Cu₂AgBiI₆ could be grown via CVT and melt methods (Figure 44), however they were compositionally inhomogeneous, highly twinned, highly absorbing of X-rays, and ultimately not suitable for SCXRD studies. Direct scale up of the CuAgBiI₅ powder synthesis from 0.25g to 2.5g led to a small decrease in compositional homogeneity. Therefore, to produce large samples for neutron powder diffraction (NPD) experiments, 0.25g batches of CuAgBiI5 were screened for their lattice parameter, purity by PXRD, and average TEM EDX composition and then combined (Figure 45). This demonstrates the reproducibility of the synthetic protocol. 2.2. Crystal Structures The iodide sub-lattices, structures, and octahedral motif of the binary compounds CuI, AgI and BiI₃ are described in Figures 46, 47 and 48, respectively. It is convenient to collect and explain the different structures of the previously reported ternary Ag_1-3xBi_1+xI₄ and CuBiI₄ materials by describing the unit cell (and whether it exhibits rhombohedral strain), iodide sub-lattice, octahedral motif formed by the octahedral (O) sites, and the occupancy of the tetrahedral (T) sites. The nomenclatures for these different structural features are described in Figures 25, 26, 27, 49 and 50, are summarized in Table 9, and will be referred to throughout this text. Table 9. A summary of the structural features of binary, ternary and quaternary compounds in the CuI–AgI–BiI₃ phase field. Nomenclature referred to are described in the main text and shown in the corresponding figures. The columns contain the composition, unit cell type, octahedral motif type, tetrahedral site type, iodide sub-lattice packing, structure types, and references of the previously reported, and novel compounds. The arrangement of the octahedra in the ternary Ag_1-3xBi_1+xI₄ and CuBiI₄ materials can be grouped in to three different octahedral motifs. For x = 0 AgBiI₄ the octahedral motif is reported to be the same as that of a spinel MgAl₂O₄ (Figure 49).⁶⁴ For the composition x = 0 AgBiI₄, each octahedral site is occupied by 50% Ag⁺ and 50% Bi³⁺, giving the octahedra full occupancy. However, as the composition moves to Bi-rich, x > 0 Ag_1-3xBi_1+xI₄ (AgBi₂I₇, Ag₂Bi₃I₁₁), vacancies arise on the octahedral site as there is now a less than 2:1 cation-to-anion ratio. Therefore, this octahedral motif will be referred to as defect-spinel, referring to its ability to contain vacancies. The reported structure of CuBiI₄ also has the defect-spinel octahedral motif, with octahedra occupied by 50% Bi³⁺ and 50% vacant, as the Cu⁺ cations occupy tetrahedral sites. The defect-spinel octahedral motif is a 3D network of edge-sharing octahedra and has an overall occupancy of O sites of 1/2. An alternative description is to consider the defect-spinel octahedral motif as a vacancy-ordered rock salt in which the occupied O sites are arranged in the defect-spinel motif. Using the transformation matrix in Figure 27a, the defect-spinel motif can be represented in the large trigonal unit cell. The lowering of the symmetry of the unit cell means that there are now two octahedral sites, which must remain equally occupied to maintain the symmetry of the original cubic structure. The transformation shows that the defect-spinel octahedral motif can be considered as alternating between layers of 3/4 and 1/4 O site occupancy, maintaining the overall 1/2 O site occupancy. The x = 0 AgBiI₄ structure can also be solved by diffraction techniques to have a CdCl₂ octahedral motif, which is represented in the small trigonal unit cell (Figure 27b). It consists of alternating between layers of full O site occupancy, and empty layers. This gives a 2D layered compound and maintains the overall 1/2 O site occupancy. The Ag-rich x < 0 Ag_1-3xBi_1+xI₄ compounds (Ag₂BiI₅, Ag₃BiI₆), have been reported to have the i-CdCl₂ octahedral motif (Figure 27c).^{46, 48, 62} This is represented in the small trigonal unit cell and consists of every possible O site being occupied, with layered ordering and atomic occupancies less than 1. The layered ordering means that the layers alternate between two different edge-sharing octahedral sites. For Ag2BiI5 (refined as Ag_1.92Bi_0.83I₅),⁴⁶ the octahedral site in one of the layers is full with occupancy 67% Ag⁺ and 33% Bi³⁺, and the octahedral site in the neighboring layer has an occupancy of 9.6% Ag⁺. For new structures reported in this work, the site with higher occupancy of the ions will be referred to as site Oct1 (and ions labelled Ag1/Bi1), and the lower occupancy site Oct2 (and ions labelled Ag2/Bi2). This octahedral motif has only been reported for Ag₂BiI₅ and Ag₃BiI₆ compositions for which the octahedral cations-to-anion ratio is greater than two. For the ternary Ag_1-3xBi_1+xI₄ materials, the Ag⁺ and Bi³⁺ cations are octahedrally coordinated, this contrasts with AgI which has tetrahedral Ag⁺. For the reported cubic CuBiI₄, however, Cu⁺ occupies tetrahedral sites, like in CuI. There are three different tetrahedral sites in CuBiI₄. Site Cu1 has a multiplicity of eight, and Hermann–Mauguin (HM) point group 4^ത3m (Figure 50a). To understand the location of site Cu1 in the structure, the octahedral motif must be considered as alternating layers of 3/4 and 1/4 O site occupancy. The tetrahedral Cu1 site is in the layer with 3/4 O site occupancy and shares faces with the octahedra in the layer with 1/4 O site occupancy. Therefore, it is located at the points where the octahedra from the 1/4 O site layers edge share with the octahedra from the ¾ O site layers, which happens eight times in the unit cell. In CuBiI₄ site Cu1 has a Cu occupancy of 18%. Site Cu2 has a multiplicity of 48, and HM point group 2mm (Figure 50b). To understand the location of site Cu2, consider the empty channels left by the 1/2 O site occupancy. There are 12 linear channels per unit cell, which crossover to make a 3D network of channels. Each channel has four Cu⁺ ions which are not shared by any other channel. The 12 channels of four Cu⁺ ions gives the multiplicity of 48. In CuBiI₄ site Cu2 has a Cu occupancy of 12%. Site Cu3 has a multiplicity of eight, and HM group (Figure 50c). These eight Cu⁺ ions are at the eight places that four separate channels cross over in the unit cell. Figure 50c points to one of the Cu⁺ ions, and highlights the four channels that it is in. In CuBiI₄ site Cu3 has a Cu occupancy of 9%. Site Cu3 is the same site that is occupied in the spinel structure (Figure 28a). Cu2 and Cu3 sites share the channels as shown in Figure 50d. Together, all three sites occupy every possible T site in the CCP sub-lattice (Figure 50e). 2.2.1. CuBiI₄ Structure A Rietveld refinement of the previously reported cubic CuBiI₄ structure,⁴³ described in the Section 3.1, on the laboratory PXRD pattern of synthesized CuBiI₄ powder fits the reported structure well, with a refined lattice parameter of a = 12.1592(3) Å (Table 10). Table 11 shows the refined bond lengths and angles for the CuBiI₄ structure.

Table 10. CuBiI₄ Rietveld refinement results for the laboratory PXRD data of the Fd3^തm CuBiI₄ structure with the defect-spinel octahedral motif reported by Fourcroy et al.²⁷ The occupancy and atom positions were fixed as reported, with the lattice parameters and isotropic thermal parameters refined. This model was transformed in to the new R3^തm CuBiI₄ structure with the CdCl₂ octahedral motif, and the lattice parameters and isotropic thermal parameters were refined. Atomic positions are given as fractional coordinates, atom positions and occupancies with no errors indicate that these values were fixed. ത^ത Table 11. Bond distances and bond angles for the reported CuBiI₄ structure with the defect-spinel octahedral motif and the new structure with the CdCl₂ octahedral motif refined against room temperature laboratory PXRD data. We show here that the same structural ambiguity between a structure with a defect-spinel octahedral motif and a CdCl₂ octahedral motif as found in the AgBiI₄ system means that the PXRD of CuBiI₄ can alternatively be indexed with a small trigonal unit cell in the R3^തm space group. This is because reflections that would distinguish the two structures from each other, the [220]_cubic reflections of the Fd3^തm cubic unit cell associated with the defect-spinel octahedral motif structure, have zero intensity (Figure 51). Like for AgBiI₄,⁴⁴ the CuBiI₄ structure with a defect-spinel octahedral motif was therefore transformed into a structure with the 2D CdCl₂ octahedral motif, using the transformation shown in Figure 52. The structures with the defect-spinel octahedral motif and the CdCl2 octahedral motif were found to fit the PXRD equally well, with the refined parameters shown in Table 10. This alternative CuBiI₄ structure (Figure 29) has a small trigonal unit cell and lattice parameters of a = 4.2990(1) Å, c = 21.060(1) Å, which are metrically cubic c/2a = 2.4494(1) = _√6. The octahedral sites of the CdCl₂ octahedral motif are 50% occupied by Bi³⁺ and 50% vacant. There are two Cu⁺ sites (Cu1 and Cu2) with occupancies of 15% and 9%, respectively. These Cu sites occupy every possible T site in the I- sub-lattice (nomenclature T_t, Figure 28c). Partial layered ordering of the sites means that site Cu1 is in the layer with octahedral site Bi1 and Cu2 is in the layer with no O site. Both Cu sites have point symmetry 3m and are distorted along the c direction; the Cu1 Cu–I bond length pointing in the c direction is elongated (2.712(3) Å) compared to the other three bonds (2.660(1) Å) showing distortion of the Cu⁺ towards an apex of the tetrahedron, and the Cu2 Cu–I bond length pointing in the c direction is shortened (2.553(3) Å) compared to the other three bonds (2.6071(8) Å) showing distortion of the Cu⁺ towards the base of the tetrahedron. The bond distances and angles of the refined structure are found in Table 11. In contrast to the reported CuBiI₄ structure with a defect-spinel octahedral motif, this interpretation of the PXRD data presents a new possible 2D layered structure for CuBiI₄. The structure containing the 2D CdCl₂ octahedral motif can be considered as a defect version of a delithiated LiVO₂ structure with composition Li_0.22VO₂.⁶⁵ In order to see if a solution to the PXRD pattern could be attributed to a statistical mix of the two aforementioned structure types, the two structures were transformed in to the same-sized large trigonal cell with a = 8.5980(2) Å, c = 21.060(1) Å (large cell defined in Figure 26a, structure shown in Figure 53a) and refined against the PXRD data. The ratio of the two structures was controlled via fixing the octahedral occupancies. The Cu⁺ occupancies were fixed during the refinements to ensure the composition was maintained. The refinements show a monotonic increase in χ² up to a maximum at the 50:50 ratio (Figure 53b), showing clearly that the structure must adopt either the previously reported cubic CuBiI₄ structure with a 3D defect-spinel octahedral motif, or an alternative structure with a 2D CdCl₂ octahedral motif, but is not a statistical mixture of the two. 2.2.2. CuAgBiI₅ Structure The CuAgBiI₅ structure was solved by Rietveld refinement of complementary combined room temperature high resolution synchrotron PXRD (MAC detector, I11, Diamond Light Source, Oxfordshire, UK) and high resolution NPD (HRPD, ISIS Neutron and Muon Source, Oxfordshire, UK) datasets, with information also gathered from SCXRD data. SCXRD data collected at 100K showed that the solution was a structure with a defect-spinel octahedral motif, however trying to refine the Cu sites would cause an unstable refinement and therefore this defect-spinel octahedral motif was used to create a starting model to refine against the powder data instead. Although the laboratory PXRD data could be fitted to a metrically cubic cell, high resolution synchrotron PXRD data shows rhombohedral strain at low d-spacings (Figures 30a and b); this is very small and refined as 0.0132(2)% (a = 8.63390(5) Å, c = 21.1398(2) Å). In order to allow rhombohedral symmetry the structure with the defect-spinel octahedral motif was refined in the large trigonal cell _{in the R3} ^ത _{m space group, rather than a cubic symmetry in a cubic unit cell (Figure 26a-iii).} Combined PXRD and NPD datasets were then used to refine Bi³⁺, Ag⁺, and vacancy occupancy on the same site of the defect-spinel octahedral motif without needing to use constraints. The room temperature structural solution for the defect-spinel octahedral motif model is shown in Figure 31a. The Rietveld fit to the combined PXRD and NPD data are shown in Figures 30c-e. The refined parameters are shown in Table S5, and the bond distances and angles in Table 12. Table 12. Bond distances and bond angles for the CuAgBiI₅ structure refined against RT PXRD and NPD combined datasets. CuAgBiI₅ RT PXRD and NPD Defect-spinel The two octahedral sites Oct1 and Oct2 are both occupied by 42.1(1)% Bi³⁺ and 41.4(2)% Ag⁺ giving a 3D octahedral network in both Bi³⁺ and Ag⁺. A Fourier difference map showed residual electron density on one of the possible tetrahedral sites but not all. The tetrahedral Cu⁺ does not occupy all tetrahedral sites, like for the CuBiI₄ structures reported thus far. Site Cu1 refined to an occupancy of 17.3(2)%. The location of Cu1 in the structure is in the empty channels caused by the unoccupied octahedra sites, the same as Cu2 site in the reported CuBiI₄ structure (Figure 50b). The difference is that due to partial layered ordering Cu1 is only present in channels in the ab plane, which are half the total amount of channels shown in Figure 50b. This is also different from the site that is occupied in a classic spinel (Figure 28a). The positioning of the tetrahedral Cu⁺ in CuAgBiI₅ makes it a rhombohedrally-distorted defect CuBiI₄ structure. The Cu⁺ is displaced away from the centre of the tetrahedron, towards the apex in the direction that points along the c-axis, which leads to one Cu–I bond (2.4923(2) Å) being shorter than the others (2.6908(3) Å), significantly distorted I–Cu–I angles of 112.14(4)° and 106.676(2)°. The refined composition of Cu_0.65(1)Ag_1.04(2)Bi_1.05(2)I_5.00 is within error of the average powder composition of Cu_0.88(17)Ag_1.10(6)Bi_0.98(8)I_5.00(11) measured by TEM EDX. 2.3. Bulk Photostability The phases reported here have varying stabilities in their powder forms. CuBiI₄ is a metastable material and decomposes back in to starting materials BiI₃ and CuI at room temperature, even in the dark. The rate of decomposition of CuBiI₄ can be slowed by storing the powder in a freezer at −80°C. CuAgBiI₅ and Cu₂AgBiI₆ are both stable when kept in the dark, in air, at room temperature. CuAgBiI₅, Cu₂AgBiI₆ and AgBiI₄ powders were exposed to the AM1.5 solar spectrum for one week, sealed in capillaries with synthetic air, laboratory air, and He atmospheres. CuAgBiI₅ experienced a color change from black to yellow in all atmospheres. This color change does not correspond to a crystalline decomposition product in the PXRD (Figure 54) but is indicated by an extra peak in the Raman spectrum at 150 cm^-1 (Figure 55). AgBiI₄ and Cu₂AgBiI₆ showed no color change after one week in the solar spectrum and showed no signs of decomposition by PXRD or Raman spectroscopy. The Cu₂AgBiI₆ composition therefore represents the stabilization of a Cu- containing bismuth iodide solar absorber and is as stable as AgBiI₄ under the investigated conditions. This is promising as devices using AgBiI₄ absorber layers have been shown to maintain 96% of their PCE after 1000 h.⁵¹ 2.4 Properties of CuAgBiI₅ and Cu₂AgBiI₆ Thin Films The phases reported here have varying stabilities in their powder forms. CuBiI₄ is a metastable material and decomposes back in to starting materials BiI₃ and CuI at room temperature, even in the dark. The rate of decomposition of CuBiI₄ can be slowed by storing the powder in a freezer at −80°C. CuAgBiI5 was stable when kept in the dark, in air, at room temperature. Therefore, we chose CuAgBiI₅ to process in to thin films for property measurements. Films were processed as described in the S.I. It was particularly challenging to dissolve the powders in solutions concentrated enough to form continuous films, to prevent powder crashing out during deposition, and to obtain a smooth, shiny surface. These were overcome by using a mixed DMSO/pyridine solution, depositing from hot solutions onto a preheated substrate, and using a two-step annealing procedure, respectively. We fitted the PXRD pattern of thin films deposited on microscope slides to a large trigonal cell in the R3^തm space group with lattice parameters a = 8.724(1) Å and c = 20.800(5) Å (Figure 62a). The a and c parameters of the thin films are significantly larger and smaller than those of the CuAgBiI₅ powder, respectively. To find out why, we measured the composition of the films by SEM EDX and found them to have an average composition of Cu_0.82(5)Ag_0.96(9)Bi_1.07(4)I_3.98(13) (Figure 62b). While the composition of the metal cations are within 1σ of those measured for the powder, there is a large iodine deficiency of 20(3)%. It is not clear at which stage in the process the iodine is lost, and further optimisation of film deposition will look to rectify this. The films show a certain level of roughness which can be seen in Figure 62c as the film differs from a perfectly shiny black reflective surface, which arises from the coverage or morphology issues we show in the SEM images in Figures 62d and e. We realise that further optimisation of the film deposition would be needed to prepare efficient photovoltaic devices using CuAgBiI₅, however the films proved sufficient for spectroscopic analysis. XPS measured on CuAgBiI₅ and Cu₂AgBiI₆ powders show an increase in states for Cu₂AgBiI₆ at the top of the valence band, which are likely to be Cu⁺ states, considering their compositions (Figure 61). The ionisation potentials (VBM band positions) of the CuAgBiI₅ and Cu₂AgBiI₆ powders were measured as 5.47(10) eV and 5.21(10) eV, respectively. Optical absorption spectra measured for three thin films of CuAgBiI₅ deposited on z-cut quartz are shown in Figure 63a, all showing a clear onset at approximately 680 nm (1.8 eV), similar to the previously reported for Cu₂AgBiI₆ in Example 1,rising to a value of over 1 x 10⁵ cm^-1 for the absorption coefficient, higher than the corresponding value for MAPbI₃ just above the band gap (see De Wolf et al., Organometallic Halide Perovskites: Sharp Optical Absorption Edge and Its Relation to Photovoltaic Performance. J. Phys. Chem. Lett.2014, 5 (6), 1035-1039). A rough estimate of the band gap can be obtained from the inflexion point of the onset of the absorption coefficient, and this gives a value of 2.02 eV across all three samples, again similar to the value of the band gap reported for Cu₂AgBiI₆. We also note the presence of some sub-gap absorption, seen more clearly in the inset of Figure 1a, between approximately 950 – 700 nm, which is possibly due to sub-gap defect states similar to those observed by Photo-Thermal Deflection Spectroscopy in Cu₂AgBiI₆, although we can not rule out scattering of long-wavelength light from the rough film surfaces leading to lower light transmission in this region, and thus higher measured ‘absorption’, but still allowing for the red transmitted light observed when the films are backlit (Figure 62f). Steady-state photoluminescence (PL) spectra were measured on the same thin films under continuous wave excitation in vacuum, with the samples of CuAgBiI₅ showing weak, broad emission peaking at approximately 760 nm (1.63 eV), giving a Stokes shift of several hundred meV, indicative of the presence of energetic disorder or sub-gap defect levels. The weakness of the emission is apparent in the PL spectra shown in Figure 63b, which have a low signal-to-noise even at a very high excitation intensity of 40.5 Wcm^-2, and the broad, weak PL is consistent with films that have a significant amount of energetic disorder and high trap density. This is also consistent with weak PL emission seen for the Ag_1-3xBi_1+xI₄ compounds. The sharp PL peak at 800 nm, observed in all three films, is from the second diffraction of the excitation laser signal from the diffraction grating in the detection setup. To gain an insight into the charge-carrier lifetimes in CuAgBiI₅, time-resolved PL measurements were carried out in vacuum using Time-Correlated Single Photon Counting (TCSPC). The transient decays shown in Figure 63c, measured in vacuum on a fresh sample, show a very fast initial decay, on the order of one ns with no fluence dependence to the decays across three orders of magnitude. The lowest-fluence decay was fitted with a stretched exponential, yielding an average lifetime of ^_^௩ = 0.73 ns and a stretching exponent of ^ = 0.32 (see Johnston, D. C., Stretched Exponential Relaxation Arising from a Continuous Sum of Exponential Decays. Phys. Rev. B 2006, 74 (18), 184430). The low value of ^ is indicative of a highly heterogeneous decay, very similar to that observed in both Cs₂AgBiBr₆ (see Schade et al., Structural and Optical Properties of Cs₂AgBiBr₆ Double Perovskite. ACS Energy Lett.2019, 4 (1), 299-305) and Cu₂AgBiI₆, and is likely due to a distribution of trap states with slightly varying trapping dynamics. The very short lifetime and lack of fluence dependence of the decays are indicative of a high trap density in the CuAgBiI₅ films, leading to fast trap-mediated recombination and scarcity of radiative band-to-band recombination, consistent with the very weak steady-state PL emission. This finding is further supported by time- resolved emission spectra, also measured in vacuum using TCSPC on a fresh sample and shown in Figure 63d. We also see evidence of a high-energy emission around 600 nm over the first 1 ns (Figure 65), where the emission band at 600 nm decreases in intensity relative to the main peak, and the transient decay at 600 nm is faster than that at 720 nm. The emission around 600 nm is likely from band-to-band recombination which is rapidly quenched by very fast trapping, consistent with the high disorder and trapping evident from the steady-state PL measurements. In order to understand the potential impact of atmospheric and light-induced effects on CuAgBiI₅, two fresh samples were left in air and darkness for 90 minutes, during which steady-state PL spectra were measured after 20, 60 and 90 minutes (Figures 66a and b). The thin films were only illuminated for very brief (c.a.15 s) periods during the PL measurements, during which acquisitions were taken every 3 s, after 20, 60 and 90 minutes, respectively. The results show PL spectra after 20 minutes that are similar to those measured on fresh films in vacuum, but which subsequently display a clear blue-shift and large rise in PL intensity under prolonged exposure to air, although the time taken for this to occur varied between the two samples. Similar variation of PL spectra with atmosphere has been widely reported for conventional metal-halide perovskites (Motti et al., Photoinduced Emissive Trap States in Lead Halide Perovskite Semiconductors. ACS Energy Lett.2016, 1 (4), 726-730; Brenes et al., The Impact of Atmosphere on the Local Luminescence Properties of Metal Halide Perovskite Grains. Adv. Mater.2018, 30 (15), 1706208; Galisteo-López et al., Environmental Effects on the Photophysics of Organic–Inorganic Halide Perovskites. J. Phys. Chem. Lett.2015, 6 (12), 2200-2205; et al., Defect Activity in Lead Halide Perovskites. Adv. Mater.2019, 31 (47), 1901183; Knight et al., Electronic Traps and Phase Segregation in Lead Mixed-Halide Perovskite. ACS Energy Lett.2019, 4 (1), 75-84.). Exposure to air has been observed to lead to significant increases in PL intensity across lead-iodide and - bromide perovskites, and this behaviour has been ascribed to the passivation of defects by oxygen (Motti et al., Photoinduced Emissive Trap States in Lead Halide Perovskite Semiconductors. ACS Energy Lett.2016, 1 (4), 726-730; Galisteo-López et al., Environmental Effects on the Photophysics of Organic–Inorganic Halide Perovskites. J. Phys. Chem. Lett.2015, 6 (12), 2200- 2205; Motti et al., Defect Activity in Lead Halide Perovskites. Adv. Mater.2019, 31 (47), 1901183.). The behaviour observed here for CuAgBiI₅ films in air could follow a similar process, where deeper trap states are passivated by oxygen over time, de-activating non-radiative recombination pathways and leading to higher-energy emission and an increase in the PL intensity. This is confirmed by transient decays measured on a film after exposure to air (Figure 8d) for which the decays show a slightly stronger fluence dependence, with longer lifetimes at lower fluences, and a stretched exponential fit to the lowest-fluence decay gives an average lifetime of ^_^௩ = 17.9 ns, much longer than for the fresh sample measured in vacuum. In order to determine whether the observed changes in PL were caused by light-induced effects, PL spectra were recorded at 3 second intervals under continuous illumination by the laser after 20 and 90 minutes for one sample (Figures 66c and d). When measured under constant illumination the spectra of CuAgBiI₅ show a drop in intensity but no change in spectral shape, a process sometimes described as ‘photodarkening’, implying that light-induced effects are not the source of the blue- shift of the spectrum and increase in PL intensity. Photodarkening has been observed in lead-halide perovskites, under both vacuum and nitrogen (Motti et al., Photoinduced Emissive Trap States in Lead Halide Perovskite Semiconductors. ACS Energy Lett.2016, 1 (4), 726-730; Brenes et al., The Impact of Atmosphere on the Local Luminescence Properties of Metal Halide Perovskite Grains. Adv. Mater.2018, 30 (15), 1706208), which in one case has been ascribed to an increased density of hole traps forming under constant illumination (Motti et al., Defect Activity in Lead Halide Perovskites. Adv. Mater.2019, 31 (47), 1901183.). However, in our case the very high laser excitation intensity required to measure PL spectra, of approximately 40 Wcm^-2, leads us to believe that the observed decrease in intensity is likely due to the degradation of the sample region under illumination leading to the creation of point defects in the region under illumination. This is supported by the observation of small burn marks on the thin films after the PL measurements. Finally, optical-pump terahertz-probe spectroscopy was used to measure the effective charge- carrier mobility for two thin films of CuAgBiI₅, yielding values of 1.7 ± 0.2 and 1.3 ± 0.2 cm²V- ¹s^-1, as shown in Figures 63e and f. These values are comparable to charge-carrier mobilities measured for both Cs₂AgBiBr₆ (Hutter et al., Band-Like Charge Transport in Cs2AgBiBr6 and Mixed Antimony–Bismuth Cs2AgBi1–xSbxBr6 Halide Double Perovskites. ACS Omega 2018, 3 (9), 11655-11662) and Cu₂AgBiI₆, and are lower than values reported across conventional metal- halide perovskites (Herz, L. M., Charge-Carrier Mobilities in Metal Halide Perovskites: Fundamental Mechanisms and Limits. ACS Energy Lett.2017, 2 (7), 1539-1548; Herz, L. M., How Lattice Dynamics Moderate the Electronic Properties of Metal-Halide Perovskites. J. Phys. Chem. Lett.2018, 9 (23), 6853-6863). Charge-carrier mobility is influenced by intrinsic effects, such as scattering off of ionised impurities or couplings between charge carriers and the crystal lattice, and extrinsic effects such as poor crystallinity and high energetic disorder or scattering off defects. Given the high trap density that is apparent from the other spectroscopic measurements, it is possible that a reduction in trap density, along with enhanced crystallinity and reduced energetic disorder could lead to an improvement in the charge-carrier mobilities for CuAgBiI₅, although the low values for charge-carrier mobilities reported across a variety of silver-bismuth compositions could be indicative of more fundamental limitations to charge-carrier mobilities in these materials (Hutter et al., Band-Like Charge Transport in Cs2AgBiBr6 and Mixed Antimony–Bismuth Cs2AgBi1–xSbxBr6 Halide Double Perovskites. ACS Omega 2018, 3 (9), 11655-11662; Bartesaghi et al., Charge Carrier Dynamics in Cs2AgBiBr6 Double Perovskite. J. Phys. Chem. C 2018, 122 (9), 4809-4816; Longo et al., Understanding the Performance-Limiting Factors of Cs2AgBiBr6 Double-Perovskite Solar Cells. ACS Energy Lett.2020, 2200-2207.). The ternary and quaternary Cu–Ag–Bi–I materials can be synthesized as powders, crystals and thin films via solid state, vapour deposition and solution processing methods. They are prone, however, to compositional inhomogeneity and therefore TEM EDX was used to probe composition of individual particles of the sample. Compositionally homogeneous quaternary CuAgBiI₅ and Cu₂AgBiI₆ powders were eventually realised by solid state synthesis, however during the optimisation there was evidence for many different compositions along the Cu_4x(AgBi)_1-xI₄ solid solution line which may be attainable via different synthetic routes such as using different temperatures in solid state synthesis, solution processing, or chemical vapour transport. The range of compositions may be possible due to the fact that the ternary and quaternary materials are all based on a well-defined close-packed I- sub-lattice with a heavily disordered cation filling of tetrahedral Cu⁺, and octahedral Ag⁺ and Bi³⁺ which can be arranged in many different ways (they have a high configurational entropy). The Ag⁺ in these materials has been shown to have octahedral coordination, in contradiction to the room temperature structure of AgI, for which Ag⁺ has tetrahedral coordination. The reported I–I distances in the AgI wurtzite structure are 4.58 Å and 4.59262 Å (Figure 47f), whereas the I–I distances in the compounds with octahedrally coordinated Ag⁺ are shorter, close to 4.3 Å. Therefore, the I–I distances of the I- sub-lattices in the ternary and quaternary compositions are too small to host tetrahedral Ag⁺, being more similar to those of the I- sub-lattice of BiI₃ and CuI. The detailed structural investigation carried out in this work allows for us to suggest a relationship between overall O site occupancy (based on composition) and the octahedral motif (structure), over the whole phase field. Figure 64 shows the type of octahedral motif against the O site occupancy. For high O site occupancies above 50%, which can only be obtained in the Ag-rich x < 0 Ag_1- _3xBi_1+xI₄ compositions (Ag₂BiI₅, Ag₃BiI₆), any O site occupancy more than 50% is between already fully occupied octahedral layers. Technically, this is a 3D octahedral motif because every interlayer O site will connect the adjacent layers. These high octahedral occupancies are reached by substituting x Bi³⁺ for 3x Ag⁺ to maintain charge balance. At O site occupancies of 50%, we have previously reported that AgBiI₄ is indistinguishable by diffraction methods between the 3D defect- spinel octahedral motif and a 4-fold twin of the CdCl₂ octahedral motif (see Sansom et al., AgBiI₄ as a Lead-Free Solar Absorber with Potential Application in Photovoltaics. Chem. Mater.2017, 29 (4), 1538-1549). We have also shown here that the same ambiguity exists for CuBiI₄. In Figure 64 we have also put the Bi-rich x > 0 Ag_1-3xBi_1+xI₄ compositions (Ag₂Bi₃I₁₁, AgBi₂I₇) in to this category although no investigation comparing the fits to diffraction data of a structure with the 3D defect-spinel octahedral motif and a 4-fold twin of the structure with the CdCl₂ octahedral motif have been performed for these specific compositions. They have been reported as having cubic unit cells which means they will very likely be in this category. By substituting in x Bi for 3x Ag, O site occupancies reach as low as 45.5% and 42.9% for reported Ag₂Bi₃I₁₁ and AgBi₂I_7, respectively. To obtain O site occupancies lower than this tetrahedral Cu⁺ is added. On the solid solution line between AgBiI₄ and CuI this corresponds to substituting in 4x Cu⁺ for every x(Ag⁺ + Bi³⁺) removed (Cu_4x(AgBi)_1-xI₄), i.e. equal amounts of octahedral Ag⁺ and Bi³⁺ are removed. For CuAgBiI₅, with an O site occupancy of 40%, a 3D defect-spinel octahedral motif with a trigonal unit cell is obtained. For Cu₂AgBiI₆, with an O site occupancy of 33%, the CdCl₂ octahedral motif is obtained showing that at some O site occupancy the octahedral motif becomes 2D. The outlier in this relationship between O site occupancy and octahedral motif is CuBiI₄, which has either the 3D octahedral motif or twinned 2D CdCl2 octahedral motif, with an O site occupancy of only 25%. However, CuBiI₄ was found to decompose in to BiI₃ and CuI upon standing at room temperature showing that it is not a stable phase, likely due to the low O site occupancy. Therefore, CuAgBiI₅ and Cu₂AgBiI₆ represent phase stable Cu- and Bi-containing compounds in the CuI–AgI–BiI₃ phase field, achieved by taking advantage of the chemical tuning made available by the quaternary system to increase O site occupancy of CuBiI₄; for CuAgBiI₅ this corresponds to adding 4x Ag⁺ for x (Cu⁺ + Bi³⁺). For Cu₂AgBiI₆ this corresponds to adding x (Cu⁺ + 2Ag⁺) for every x Bi³⁺ removed. 3. Conclusion Through challenging exploratory synthesis, we have discovered a 3D octahedral network in quaternary compound CuAgBiI₅ and use its structure, with the rest of the known compounds in the CuI–AgI–BiI₃ phase space, to suggest a relationship between the composition (octahedral occupancy) and dimensionality of the octahedral network. We find that substituting tetrahedral Cu⁺ in for octahedral Ag⁺ and Bi³⁺ on the AgBiI₄ to CuI solid solution line leads to a decrease in octahedral site occupancy which changes the network from ambiguously 3D (AgBiI₄) to unambiguously 3D (CuAgBiI₅) to 2D (Cu₂AgBiI₆). We also synthesise CuBiI₄ powder and revisit its structure to alternatively describe it as a 2D CdlCl₂-octahedral motif, similar to the two possible quasi-homometric structural descriptions of AgBiI₄. We show that the outlier to the proposed relationship between octahedral site occupancy and octahedral motif is CuBiI₄, which is metastable at room temperature and decomposes in to CuI and BiI₃. We find no evidence of tetrahedral Ag⁺ in the structures and put this down to the smaller I–I distances of the CCP iodide sub-lattice of the ternary and quaternary compounds, which are more similar to those found in BiI₃ and CuI rather than AgI, in which Ag⁺ is tetrahedrally coordinated. We find evidence via TEM EDX measurements for phases across the AgBiI₄ to CuI line which may be accessible via different synthetic routes. We have also measured the optical properties on solution processed CuAgBiI₅ films and show that the 3D octahedral network does not necessarily increase charge carrier mobilities and PL lifetimes compared to the 2D octahedral network of Cu₂AgBiI₆, however these values could be limited by trap states as we have not implemented passivation strategies here. The PL shows ultra-fast trapping of charge-carriers and evidence of oxygen passivation when measured in air. We find that the high absorption coefficients and useful band gaps of ternary and quaternary compounds on the CuI–AgI–BiI₃ phase space are also found here and find no significant difference between CuAgBiI₅ and Cu₂AgBiI₆. We hope that with increased understanding of the challenging synthesis and solution processing, complex crystal structures, optoelectronic properties, and opportunities, that we encourage further exploration and optimization of the potentially useful semiconductors in the CuI–AgI–BiI₃, and related, phase space. Figure 91 shows the fitting of (a) Cu 2p, (b) Ag 3d, (c) Bi 4f, and (d) I 3d core levels of X-ray photoelectron spectroscopy (XPS) data collected for CuAgBiI₅ bulk sample. The fitting of Cu 2p_3/2 is complicated by the presence of the I 3p_1/2, however Cu 2p_1/2 was fit first, and the known spin orbit split energy used. The identification of the charge state of the Cu as Cu⁺ is supported by the lack of any satellite feature of the Cu 2p. Figure 92 shows the valence band maximum (VBM) and conduction band minimum (CBM) positions with respect to vacuum measured for CuAgBiI₅ and Cu2AgBiI₆ bulk samples. The ionisation potential are measured on bulk samples and the optical band gap measured on thin films. Figure 93 shows five steady-state photoluminescence spectra for a thin film of CuAgBiI₅ measured under continuous illumination over 15 s by a 398 nm laser at an intensity of 39.0 Wcm^-2 after being left in air for 20 (a) and 90 (b) minutes, respectively. In both cases the spectra do not change shape but do decrease in intensity over 15 s of continuous illumination. Figure 94 shows structures in the CuI-AgI-BiI₃ phase space: binaries CuI, AgI and BiI₃; ternaries Ag₃BiI₆, Ag₂BiI₅, AgBiI₄, AgBi₂I₇, Ag₂Bi₃I₁₁, CuBiI₄, Cu₂BiI₅; and quaternary Cu₂AgBiI₆. Also included is CuAgBiI₅. All these compounds consist of a close-packed iodide sub-lattice with varying arrangements of the cations filling the octahedral and tetrahedral interstitial sites to form the structures shown. Figure 95 shows (a) The three Cu⁺ sites in CuBiI₄ (spinel) colour coded as yellow (site 1), orange (site 2) and red (site 3). A channel in the spinel octahedral (Oct) motif is highlighted in blue, which a cross-section is taken of to show the tetrahedral (Tet) sites inside (sites located behind the channel, which appear to be inside due to the 2D representation of the 3D structure, have been crossed out). The red site (site 3) is the same in spinel. Some, but not all, of the orange sites (site 2) are occupied in CuAgBiI₅. The Tet sites in CuAgBiI₅ can be considered as a reflection of the spinel sites with the mirror plane down the centre of the channel (blue dashed line) (b) The two Cu⁺ sites in the small trigonal unit cell (Cu₂AgBiI₆ and CuBiI₄ (CdCl₂)) showing layered ordering. Also shown are the connectivity’s of the Tet sites, which give a 3D Tet network (c) The layered ordering of Cu⁺ sites in CuAgBiI₅ means they are only in layers with 1/4 Oct interstitial occupancy and do not occupy all the sites associated with tetrahedral site 2 in CuBiI₄ (spinel). (d) The connectivity of tetrahedra in CuAgBiI₅ and spinel, which give 2D and 0D Tet networks, respectively. Figure 96 shows the absorption coefficient and PL measured on CuAgBiI₅ (solid lines) and Cu₂AgBiI₆ (dashed line) thin films. The data for Cu₂AgBiI₆ is taken from Sansom et al. The PL spectra of CuAgBiI₅ and Cu₂AgBiI₆ were measured in vacuum and air, respectively. (b) The density of states of the valence band measured on CuAgBiI5 (black) and Cu₂AgBiI₆ (red) powders, measured by XPS. (c) The shift and increase in the PL signal of CuAgBiI₅ thin films exposed to air. (d) TRPL of CuAgBiI₅ thin films measured in vacuum (black) and air (blue), compared to Cu₂AgBiI₆, measured in air. Figure 97 shows (a) The room temperature crystal structure of CuAgBiI₅ from Rietveld refinement of combined PXRD and NPD datasets, including coordination environments. (b) A low d-spacing region of the fit, showing that a trigonal cell (ii) fits better than a cubic cell (i). (c) The fits to high resolution synchrotron PXRD (MAC detector, I11, Diamond Light Source, Oxfordshire, UK) and high-resolution NPD (banks 1 and 2, HRPD, ISIS Neutron and Muon Source, Oxfordshire, UK) datasets. Example 3 - Solution processing of Cu(Ag _.5)I₄ films Cu(Ag_1.5Bi_0.5)I₄ is the composition where the material would have the same tetrahedral and octahedral occupancies as spinel MgAl₂O₄ whilst satisfying charge balance. The aim of this is to make a more ordered version of a quaternary CuAgBiI material, in the hope that the tetrahedral Cu⁺ occupies those in a spinel material. We have solution processed films (Figure 67) which are identified as a pure cubic phase (a = 12.131(2) Å) by PXRD (Figure 68) with nominal composition Cu(Ag_1.5Bi_0.5)I₄. The average SEM EDX composition is measured as Cu_0.98(4)Ag_1.46(3)Bi_0.52(2)I_3.24(10) (Figure 69). These measurements show a cubic unit cell and a much smaller compositional spread in the Cu, Ag, Bi than the films made along the Cu_4xAgBi_1-xI₄ line (CuAgBiI₅, Cu₂AgBiI₆). The band gap is measured as 2.05 eV and the PL peak position at 1.8 eV (Figure 70). These values correspond to a smaller Stokes shift (250 meV) than is observed for Cu₂AgBiI₆ (310 meV) and CuAgBiI₅ (360 meV). Method 235 mg AgI and 197 mg BiI₃ were dissolved in 0.8 ml DMSO at 140°C over 15 minutes with constant stirring. In a separate vial, 127 mg of CuI powder were dissolved in 0.5 ml pyridine at 140°C over 15 minutes with constant stirring. The AgBiI in DMSO and CuI in pyridine solutions were quickly filtered through a 0.22 μm pore-size, 13 mm diameter PTFE filter and combined into a single vial. The solutions were kept stirring at 100°C during deposition. Microscope slides were cut to size (approx.24mm x 24 mm) and sonicated in soap and DI water, acetone and IPA for 15 minutes, dried using an N₂ gun, then further cleaned in a U.V.-Ozone generator. After which, the substrates were quickly transferred to a dry box and heated to 150°C for deposition.100 ul of the hot CuAgBiI solution was deposited dymanically on to the hot substrate and spun at 4000 rpm for 45 seconds on a spincoater. The films were then annealed at 50°C for 30 minutes in air, over which the films turned from a clear orange film to a dark red-brown film. The film was then annealed at 150°C for 15 seconds in air to form the brown film. Films were transferred to and kept in a N₂ filled glovebox until used for measurements. Example 4 - Solution processing of films along Cu_4x(AgBi)_1-xI₄ line To investigate the effect of adding tetrahedral Cu⁺ in for octahedral Ag⁺ and Bi³⁺ on the Cu_4x(AgBi)_1-xI₄ line, films were solution processed corresponding to nominal x = 0 (AgBiI₄), 0.09 (Cu_0.4AgBiI_4.4), 0.2 (CuAgBiI₅), 0.33 (Cu₂AgBiI₆), 0.6 (Cu₆AgBiI₁₀). The composition of films measured by SEM EDX are shown in Figure 71. The average compositions are also reported, indicating some of the films contain iodine deficits. The Pawley fits to XRD patterns are shown in Figure 72. The XRD pattern of x = 0 AgBiI4 is fitted to a small trigonal cell with the R3^തm space group. The XRD pattern of x = 0.09 Cu_0.4AgBiI_4.4 is fitted to a small trigonal cell with the R3^തm space group, and a CuI impurity. The XRD pattern of x = 0.2 CuAgBiI₅ is fitted to a large trigonal cell with the R3^തm space group associated with the CuAgBiI₅ structure. The XRD pattern of x = 0.33 Cu₂AgBiI₆ is fitted to two small trigonal cells. Two cells were needed to fit the peak splitting seen at ~12.5°. The XRD pattern of x = 0.6 Cu₆AgBiI₁₀ is fitted to a small trigonal cell, AgI, BiI₃ and CuI however, there are unindexed peaks corresponding to an unidentified phase and potentially new material (e.g. at ~8.5°). The broad unfitted peaks at ~9.7° and 19.2° corresponds to BiOI impurities which can appear when the samples are not stored in a N₂ atmosphere. Increasing the amount of Cu in pyridine solution leads to a lower coverage of the film on the substrate and change in morphologies as shown in the SEM images in Figure 73. Importantly, Figure 74 shows an order of magnitude increase in the mobility of the film with composition Cu_5.87(12)Ag_0.79(11)Bi_0.90(8)I_8.66(22), compared to the film with composition Ag_1.05(3)Bi_1.07(7)I_4.00(7). Figure 75 shows that there are very small changes in the band gap of the series of films. Method AgI, Bi and I₂ powders (Table 13) were dissolved in 0.8 ml DMSO at 140°C over 15 minutes with constant stirring. In a separate vial CuI powders were dissolved in pyridine at 140°C over 15 minutes with constant stirring. The AgBiI in DMSO and CuI in pyridine solutions were quickly filtered through a 0.22 μm pore-size, 13 mm diameter PTFE filter and combined into a single vial. The solutions were kept stirring at 100°C during deposition. Microscope slides were cut to size (approx. 24mm x 24 mm) and sonicated in soap and DI water, acetone and IPA for 15 minutes, dried using an N₂ gun, then further cleaned in a U.V.-Ozone generator. After which, the substrates were quickly transferred to a dry box and heated to 150°C for deposition.100 ul of the hot CuAgBiI solution was deposited dynamically on to the hot substrate and spun at 4000 rpm for 45 seconds on a spincoater. The films were then annealed at 50°C for 30 minutes in air, over which the films turned from a clear orange film to a dark red-brown film. The film was then annealed at 150°C for 15 seconds in air to form the films shown in Figure 76. Films were transferred to and kept in a N₂ filled glovebox until used for measurements. Table 13. Amounts of starting materials used to prepare Cu_4x(AgBi)_1-xI₄ compounds Example 5 – Triple-Source Vapor Deposition Evaporation of Cu-Ag-Bi-I films Thin films of Cu₂AgBiI₆ were fabricated by co-depositing bismuth(III) iodide (Puratronic, 99.999%), silver(I) iodide (Premion, 99.999%) and copper(I) iodide (Puratronic, 99.998%) precursor powders (Alpha Aesar) inside a BOC Edwards Auto 306 thermal evaporator. The three power sources were custom made by Moorfield Nanotechnology and controlled by an Inficon SQC-310 deposition controller and the evaporation rates measured by three quartz crystal microbalances (QCM). The substrates were protected during the heating and cooling process by a mechanical shutter, and the substrates were rotated during deposition to improve substrate surface coverage. Prior to the deposition, precursor tooling factors were calculated to correct the divergence between the true rate and thickness and the rate and thickness measured by the QCM. This was done by separately depositing 100 nm of each precursor at 0.7 Ås^-1 (as measured by the QCM and SQC-310) onto 30 x 30 mm glass. By measuring the thickness using a Dektak profilometer, a calibrated tooling factor was calculated using equation 1. X-ray diffraction was used to ensure only the precursor was evaporated and no impurities were deposited. All depositions were carried out under vacuum (2 x 10^-6 mbar). Prior to being loaded into the vacuum chamber, the substrates were sonicated in 200 ml of Decon and DI water (2% concentration), 200 ml of DI water, 200 ml acetone, and 200 ml isopropanol for 15 minutes each, then O₂ plasma cleaned for 10 minutes, and UV-Ozone cleaned for 15 minutes. For the co- deposition, three 2.4 CC alumina crucibles were filled with 1.2 CC of either CuI, AgI, or BiI₃ and evaporated onto quartz or glass substrates at varying rates (CuI, AgI, BiI₃ = 0.4, 0.2, 0.2 Ås^-1; 0.26, 0.2, 0.55 Ås^-1; 0.2, 0.2, 0.26 Ås^-1; 0.26, 0.2, 0.37 Ås^-1). Although the deposition rate ratios often do not correspond to the measured stoichiometry, they are correlated. No intentional substrate heating was applied. The final film thicknesses were measured by a Dektak Profilometer and varied between 200 nm – 400 nm. Films were annealed in ambient air post-deposition between 90°C and 150°C for 10 minutes or were left as deposited. PXRD data were measured on a PANalytical X’Pert Pro diffractometer using Cu K-α radiation (λ=1.5406 Å). Diffractograms for 200 nm films deposited at rates CuI, AgI, BiI₃ = 0.2, 0.2, 0.55 Ås^-1 are shown in Figure 77. A Pawley Fit analysis revealed the coexistence of two rhombohedral ternary or quaternary phases (^3^ത^) and small amounts of BiI₃ impurities. Due to the complexity of the Cu-Ag-Bi-I phase space, their closely related diffraction patterns and lattice parameters with the ternary and quaternary compositions (AgBi₂I₇ [1], Ag₂BiI₅ [2], Ag₃BiI₆ [3], AgBiI₄ [4], Cu₂BiI_5, CuBiI₄ [5], CuAgBiI₅), identification of the exact composition of the rhombohedral phases by XRD and SEM-EDX alone is difficult. Further analysis using TEM-EDX is necessary. Morphology and Composition Morphology and elemental composition were investigation using Scanning Electron Microscopy (SEM) and Energy Dispersive X-ray Spectroscopy (EDX). Measurements were taken on an FEI- Quanta 600 FEG microscope and elemental composition quantitively analysed using Inca Suite software. Accelerating voltage was 10KV and spot size was 3. Top-view SEM images can be seen in Figures 78-81. Grain size was measured to be approximately 50-200 nm and increased slightly with annealing temperature. Similarly, the density of pinholes was found to increase with increasing annealing temperature, but no differences were observed for films annealed in ambient air and N₂. Films were found to degrade at annealing temperatures greater than 150°C. The composition was measured to be Cu_2.25Ag_0.81Bi_1.11I_6.00. Absorption and Photoluminescence UV-Visible absorption was measured in ambient air using a UV Spectrophotometer (Cary 300). Measurements of 200 nm thick evaporated Cu₂AgBiI₆ films on quartz substrates are shown in Figure 82. Absorbance data shows the evaporated Cu₂AgBiI₆ films strongly absorb in the UV- Visible region. Films that were not annealed showed the greatest absorbance, whilst absorbance decreased for films annealed at higher temperatures, possibly due to increasing pinhole density. Photoluminescent Spectroscopy (PL) was measured at in ambient air using a PicoQuant FluoTime 300 and analysed in EasyTau 2 software. The measurements were taken on 200 nm thick Cu₂AgBiI₆ films on quartz substrates with a laser excitation wavelength of 405 nm. The PL spectrum in Figure 83 shows the Cu₂AgBiI₆ films are emissive with peak wavelength emission at 720 nm, in line with that found for solution processed films. Films annealed at higher temperatures are more emissive, whilst films annealed at 110°C or less emit weakly. A small shoulder at 690 nm for films annealed at 130°C and 150°C may correspond to the second rhombohedral phase found in the PXRD data. Stability and Light Soaking The photostability of a quaternary Cu₂AgBiI₆ film was investigated by comparing the absorption and PL intensity of films left in the dark to those exposed to an 80W, 400-700 nm LED array. The experiment was conducted in an N₂ glovebox in the absence of O₂ and H₂O (< 5 ppm). The film was evaporated onto SnO₂/ITO substrates and annealed at 110°C in air. One half of the sample was placed under the LED array, and the other half placed in a dark container. The absorbance and PL spectrum were measured after one week, and after three weeks. Figures 84-85 report the changes in absorption and PL after one and three weeks respectively. After one week, the absorption of the illuminated film is higher than the film stored in the dark, possibly due to increased scattering or increased sub-band gap absorption by trap states. The dark sample’s PL peak emission is at 718 nm, whilst the illuminated sample’s peak emission is blue-shifted to 707 nm and has lower intensity. The FWHM decreases from 360 meV to 340 meV. After three weeks, a further small increase in the illuminated film’s absorption is observed, but the relative difference in intensity between the illuminated and dark film decreases. The illuminated sample’s PL peak emission blueshifts further to 698 nm, whilst the dark samples PL peak emission blueshifts further to 701 nm. The illuminated film’s FWHM further narrows to 335 meV, and the dark sample’s FWHM narrows to 330 meV, suggesting any photodegradation has stabilised. There were no visual differences between the films throughout the study. Overall, the evaporated Cu₂AgBiI₆ films appear to be stable in the presence of light, as is observed for the material in its powder form. It is noted that that the PL intensities of samples measured after one week cannot reliably be compared to those measured after three weeks due to changing PL set-up parameters (e.g. unavoidable laser output fluctuations and changing alignment of the PL signal onto the spectrometer). Alumina vs Quartz Crucibles - Photodissociation Initially, the precursors CuI, AgI, and BiI₃ were evaporated using quartz crucibles but inspection of the crucible containing silver iodide post-deposition presented two concerns. Firstly, the base of the crucible cracked after each deposition. Due to the high deposition temperature of AgI (500°C), care must be taken to choose a suitably long ramping time to increase and decrease the temperature to avoid thermal shock and cracking. Secondly, visual observation of the silver iodide crucible post- deposition showed the presence of metallic silver (Figure 86). This was confirmed by comparing the XRD patterns of fresh silver iodide precursor powder to the post-deposition crucible residue as shown in Figure 87. It is proposed that silver originates from the photodissociation of silver iodide. Silver-halides have been intensely studied for their application to photography [6]–[10] . Exposure to above bandgap (2.83 eV or 438 nm) radiation dissociates silver iodide to form a latent image of Ag⁰ particles and iodine I⁰. The reaction proceeds as follows: Upon illumination, an iodide valance band electron is promoted to the conduction band generating an e-h pair that are free to move through the AgI lattice. The pair can either recombine on an iodine site or become trapped at defect sites. Electrons trapped at defect sites can combine with mobile silver ions causing reduction of Ag⁺ to Ag⁰. The presence of Ag⁰ catalyses further reductions leading to clusters of Ag⁰ which are used to develop a latent image. In this set-up, radiation is generated from a heated tungsten wire (T>2800K) used to heat the crucible and precursor, and the direct absorption of above bandgap photons by AgI is facilitated by the transparent quartz crucible. Pure silver evaporates at much greater temperature than the AgI deposition temperature and remains in the crucible, in contrast to iodine that rapidly sublimes onto the substrates. However, in replacing the quartz crucible with an opaque alumina (Al₂O₃) crucible, silver is not produced. The alumina crucibles also offer greater resistance to cracking. Void Formation Cross-section SEM micrographs of Cu₂AgBiI₆ films deposited from quartz crucibles on PEDOT:PSS (Figure 88) and SnO₂ (Figure 89) respectively revealed the presence of interfacial voids. Voids are commonly observed in solution processed films and usually form from the imperfect evaporation of solvents, but it’s less commonly observed in evaporated films. The voids were unrelated to annealing atmosphere and temperature and were present in all as-deposited films implying formation during deposition/cooling. When replacing the quartz crucibles with alumina crucibles, the voids were eliminated (Figure 90). Although the exact reaction pathway is unknown and requires further investigation, it may be related to the iodine produced from the photodissociation of silver iodide, and the radiation generated by the tungsten wire that was incident on the substrate. Several studies have reported I₂ triggered degradation in the hybrid perovskites MAPbI₃, FAPbI₃ and FA_0.8Cs_0.2PbI₃ [13], [14] but to the best of our knowledge, it hasn’t been reported for inorganic materials. Wang et al. exposed 50 nm of MAPbI₃ to I₂ vapor in a N₂-filled glovebox with and without light exposure and found that in both films, the organic cation decomposes leading to a yellow PbI₂ phase [13]. Those concurrently exposed to light were found to degrade faster. Although this degradation pathway cannot be replicated in Cu₂AgBiI₆ films, it suggests a similar mechanism involving light and iodine could be detrimental to inorganic compounds such as Cu₂AgBiI₆. Bibliography for Example 5 [1] Y. Kim et al., “Pure Cubic-Phase Hybrid Iodobismuthates AgBi2I7for Thin-Film Photovoltaics,” Angew. Chemie - Int. Ed., vol.55, no.33, pp.9586–9590, 2016. [2] I. Turkevych et al., “Photovoltaic Rudorffites: Lead-Free Silver Bismuth Halides Alternative to Hybrid Lead Halide Perovskites,” ChemSusChem, vol.10, no.19, pp.3754–3759, 2017. [3] T. Oldag, T. Aussieker, H. L. Keller, C. Preitschaft, and A. Pfitzner, “Solvothermale synthese und bestimmung der kristallstrukturen von AgBiI 4 und Ag3BiI6,” Zeitschrift fur Anorg. und Allg. Chemie, 2005. [4] H. C. Sansom et al., “AgBiI4 as a Lead-Free Solar Absorber with Potential Application in Photovoltaics,” Chem. Mater., vol.29, no.4, pp.1538–1549, 2017. [5] H. Yu et al., “Gradient formation and charge carrier dynamics of CuBiI4based perovskite- like solar cells,” Sustain. Energy Fuels, vol.4, no.6, pp.2800–2807, 2020. [6] S. Grzesiak, J. Belloni, and J. L. Marignier, “Development kinetics of silver clusters on silver halides,” Radiat. Phys. Chem., vol.77, no.6, pp.713–727, 2008. [7] J. J. Ramsden, “The photolysis of small halide particles,” Proc R Soc L. A, vol.392, no. 1803, pp. 427–444, 1984. [8] G. Burley, “Photolytic behavior of silver iodide,” J. Res. Natl. Bur. Stand. Sect. A Phys. Chem., vol.67A, no.4, p.301, 1963. [9] R. C. Baetzold, “Properties of silver clusters adsorbed to silver bromide,” J. Phys. Chem. B, vol.105, no.17, pp.3577–3586, 2001. [10] T. Tani, “Review of mechanisms of photographic sensitivity,” Imaging Sci. J., vol.55, no. 2, pp.65–79, 2007. [11] G. Burley, “Structure of hexagonal silver iodide,” J. Chem. Phys., 1963. [12] L. Vegard, “ IX. The structure of silver crystals ,” London, Edinburgh, Dublin Philos. Mag. J. Sci., vol.31, no.181, pp.83–87, 1916. [13] S. Wang, Y. Jiang, E. J. Juarez-Perez, L. K. Ono, and Y. Qi, “Accelerated degradation of methylammonium lead iodide perovskites induced by exposure to iodine vapour,” Nat. Energy, 2017. [14] F. Fu et al., “I2 vapor-induced degradation of formamidinium lead iodide based perovskite solar cells under heat-light soaking conditions,” Energy Environ. Sci., 2019. Example 6 – Passivation Iodine annealing Scanning Electron Microscopy Energy Dispersive X-ray Spectroscopy (SEM-EDX) reveals that evaporated Cu₂AgBiI₆ films are up to 5% iodine deficient. To investigate the potential passivation effects of iodine (I₂), Cu₂AgBiI₆ films were annealed at 110°C, 130°C, and 150°C in an I₂ atmosphere for 15 minutes. These films were compared to films annealed at the same temperature and for the same time in air. The annealing set up is shown in Figure 98. Prior to loading the samples, the base plate was pre-heated to the target temperature using a hotplate. Once the target temperature was reached, 250 nm thick Cu₂AgBiI₆ samples and 1g of solid iodine crystals were positioned on the base plate before sealing the atmosphere with a rubber gasket, glass cover, and pumping down to a pressure of 5 KPa using a rotary vacuum pump. To prevent iodine prematurely subliming during the pump down procedure, iodine crystals were placed on stacked microscope slides. XRD Measurements Experimental XRD patterns of as-deposited and annealed Cu₂AgBiI₆ films in a) ambient air and b) iodine vapour are shown in Figures 99 and 100. Absorption coefficient spectra Absorption coefficient spectra of as-deposited and annealed Cu₂AgBiI₆ films in a) ambient air and b) iodine vapour are shown in Figures 101 and 102. An increasing red shift in absorption onset is observed with increasing annealing temperature. Using Tauc plots and assuming a direct transition, the bandgap was estimated to be 2.02 eV, 1.96 eV, 1.93 eV, 1.92 eV for as-deposited, and films annealed in air (110°C, 130°C and 150°C respectively). These values are in good agreement with films annealed in iodine (1.98 eV, 1.91 eV, and 1.90 eV) showing no significant difference in absorption properties at the band edge. However, films annealed at temperatures greater than 110°C in both air and iodine develop an absorption peak at approximately 3eV, which can be attributed to the generation of CuI. Absorption data for 250 nm evaporated CuI on quartz is also shown in Figure 101 for comparison. Photoluminescence spectra The temperature sensitivity of Cu₂AgBiI₆ is also observed in the photoluminescence spectra shown in Figure 102, where PL emission increases in intensity and blue shifts with increasing annealing temperature. Peak positions were estimated by fitting a two-term Gaussian to experimental data. As deposited films, and films annealed at 110°C, 130°C , 150°C emit at 780 nm, 752 nm, 747 nm, and 743 nm respectively, whilst films annealed in an iodine atmosphere emit at 771 nm, 743 nm, and 736 nm respectively. Photoluminescence is significantly higher in films annealed in iodine, suggesting potential passivation effects and hence higher radiative efficiency. However, it is likely that the development of this peak is due to the increasing production of CuI with temperature: despite CuI possessing a bandgap of 3eV, an iodine-vacancy assisted radiative transition at approximately 720 nm is commonly observed [1]. Time Resolved Photoluminescence (TRPL) spectra Average charge lifetime was found to increase with increasing annealing temperature, with the largest lifetimes seen for films annealed in an iodine atmosphere (Figures 103 and 104). The lifetimes for as deposited, and 110°C, 130°C, and 150°C films in air were estimated to be 15.6 ns, 13.4 ns, 24.7 ns, and 34.3 ns respectively, whilst films annealed at the same temperature in an iodine atmosphere were estimated at 10.9 ns, 40.4 ns, and 51.8 ns respectively [2]. Due to the strong overlap between the CuI and Cu₂AgBiI₆ emission wavelength, it is not clear whether this increase in lifetime is due to photo excited charges generated in Cu₂AgBiI₆, or CuI. Bibliography for Example 6 [1] G Lin et al., “Luminescence Properties and Mechanisms of CuI Thin Films Fabricated by Vapor Iodization of Copper Films,” Materials, 9(12), p.990, 2016. [2] D Johnston, “Stretched exponential relaxation arising from a continuous sum of exponential decays,” Physical Review B, 74(18), 2006.

Further aspects of the invention Further aspects of the invention are set out in the following numbered clauses: 1. A crystalline compound comprising (i) One or more first cations, A; (ii) One or more second cations, B; (iii) One or more third cations, B’; (iv) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment. 2. A crystalline compound according to clause 1, wherein the compound is not a perovskite, and/or wherein the molar ratio of the one or more first cations, A, to the one or more second cations, B, in the crystalline compounds is at least 1:9, preferably wherein the molar ratio of the one or more first cations, A, to the one or more second cations, B, in the crystalline compounds is at least 1:5. 3. A crystalline compound according clause 1 or clause 2 wherein at least some of the one or more first cations, A, form AX₄ tetrahedra. 4. A crystalline compound according to any preceding clause wherein the one or more first cations, A, have the electronic configuration Nd¹⁰, wherein N is an integer from 3 to 5. 5. A crystalline compound according to any preceding clause wherein the one or more first cations, A, have an ionic radius of less than 1Å. 6. A crystalline compound according to any preceding clause wherein the one or more first cations, A, comprise one or more of Mg²⁺, Ga³⁺, Ge⁴⁺, Si⁴⁺, P⁵⁺, Cu⁺, Zn²⁺, Ti⁴⁺, Al³⁺, In³⁺ and Li⁺. 7. A crystalline compound according to any preceding clause wherein the one or more first cations, A, comprise Cu⁺. 8. A crystalline compound according to any preceding clause wherein at least some of the one or more second cations, B, are in an octahedral coordination environment. 9. A crystalline compound according to any preceding clause wherein at least some of the one or more second cations, B, form BX₆ octahedra. 10. A crystalline compound according to any preceding clause wherein the one or more second cations, B, comprise one or more of Ag⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cr³⁺, V²⁺, Cd²⁺, Cu⁺, Au⁺, Pt²⁺, Pt⁴⁺, Mg²⁺, Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺, In⁺, Ca²⁺, Sr²⁺, La³⁺ and Y³⁺. 11. A crystalline compound according to any preceding clause wherein the one or more second cations, B, comprise Ag⁺. 12. A crystalline compound according to any preceding clause wherein at least some of the one or more third cations, B’, form B’X₆ octahedra. 13. A crystalline compound according to any preceding clause wherein the one or more third cations, B’, have the electronic configuration Ns², wherein N is an integer from 2 to 7, preferably wherein N is an integer from 3 to 6. 14. A crystalline compound according to any preceding clause wherein the one or more third cations, B’, comprise one or more of Bi³⁺, Pb²⁺, Sn²⁺, Ge²⁺, In⁺, Sb³⁺, Te⁴⁺, Tl⁺, In³⁺ and Se⁴⁺. 15. A crystalline compound according to any preceding clause wherein the one or more third cations, B’, comprise Bi³⁺. 16. A crystalline compound according to any preceding clause wherein the one or more anions, X, comprise one or more anions selected from halide anions and chalcogenide anions. 17. A crystalline compound according to any preceding clause wherein the one or more anions, X, comprise one or more of I-, Br-, Cl-, S^2-, Se^2- and O^2-. 18. A crystalline compound according to any preceding clause wherein the one or more anions, X, comprise I-. 19. A crystalline compound according to any preceding clause wherein: - at least some of the one or more first cations, A, form AX₄ tetrahedra; - at least some of the one or more second cations, B, form BX₆ octahedra; and - at least some of the one or more third cations, B’, form B’X₆ octahedra. 20. A crystalline compound according to any preceding clause wherein the crystalline compound has a structure based on the cadmium chloride structure. 21. A crystalline compound according to clauses 1 to 19 wherein the crystalline compound has a structure based on the spinel structure. 22. A crystalline compound according to any preceding clause wherein the one or more first cations, A, have the electronic configuration Nd¹⁰, wherein N is an integer from 3 to 5 and wherein the one or more third cations, B’, have the electronic configuration Ns², wherein N is an integer from 2 to 7, preferably wherein N is an integer from 3 to 6. 23. A crystalline compound according to any preceding clause wherein the one or more first cations, A, comprise one or more of Mg²⁺, Ga³⁺, Ge⁴⁺, Si⁴⁺, P⁵⁺, Cu⁺, Zn²⁺, Ti⁴⁺, Al³⁺, In³⁺ and Li⁺; the one or more second cations, B, comprise one or more of Ag⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cr³⁺, V²⁺, Cd²⁺, Cu⁺, Au⁺, Pt²⁺, Pt⁴⁺, Mg²⁺, Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺, In⁺, Ca²⁺, Sr²⁺, La³⁺ and Y³⁺; the one or more third cations, B’, comprise one or more of Bi³⁺, Pb²⁺, Sn²⁺, Ge²⁺, In⁺, Sb³⁺, Te⁴⁺, Tl⁺, In³⁺ and Se⁴⁺; and the one or more anions, X, comprise one or more of I-, Br-, Cl-, S^2-, Se^2- and O^2-. 24. A crystalline compound according to any preceding clause wherein the compound comprises an interconnected network of B’X₆ octahedra, preferably wherein the compound comprises an interconnected network of BiI₆ octahedra. 25. A crystalline compound according to any preceding clause wherein: the one or more first cations, A, comprise Cu⁺; the one or more second cations, B, comprise Ag⁺; the one or more third cations, B’, comprise Bi³⁺; and the one or more anions, X, comprise I-. 26. A crystalline compound according to any preceding clause wherein the compound further comprises a dopant cation, preferably wherein the dopant cation is selected from a transition metal cation or a rare earth cation. 27. A crystalline compound according to any preceding clause wherein the compound is a compound of formula (I): [A]_4x{[B][B’]}_1-x[X]₄ (I); wherein [A] is the one or more first cations, [B] is the one or more second cations, [B’] is the one or more third cations and [X] is the one or more anions, wherein the one or more first cations, A, are different from the one or more second cations, B, and the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, andwherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (I) may vary by ±30%. 28. A crystalline compound according to clause 27 wherein x is from 0.05 to 0.95, preferably wherein x is from 0.1 to 0.5. 29. A crystalline compound according to clause 27 or clause 28 wherein x is from 0.15 to 0.25 or x is from 0.30 to 0.35, preferably wherein x is 0.2 or x is 0.33. 30. A crystalline compound according to any one of clauses 27 to 29 wherein the compound of formula (I) is a compound of formula (IA): Cu_4x{AgBi}_1-xI₄ (IA); wherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (IA) may vary by ±30%. 31. A crystalline compound according to any preceding clause wherein the compound is CuAgBiI₅, wherein the stoichiometry of each of the ions in the compound may vary by ±30%. 32. A crystalline compound according to any preceding clause wherein the compound is Cu₂AgBiI₆, wherein the stoichiometry of each of the ions in the compound may vary by ±30%. 33. A crystalline compound according to any one of clauses 1 to 26 wherein the compound is a compound of formula (II): [A]_1+2x{[B]_3x[B’]_1-x}₂[X]_7+2x (II); wherein [A] is the one or more first cations, [B] is the one or more second cations, [B’] is the one or more third cations and [X] is the one or more anions, wherein the one or more first cations, A, are different from the one or more second cations, B, and the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, and wherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (I) may vary by ±30%. 34. A crystalline compound according to clause 33 wherein x is from 0.05 to 0.95, preferably wherein x is from 0.1 to 0.6. 35. A crystalline compound according to clause 33 or clause 34 wherein x is from 0.20 to 0.30 or where x is from 0.45 to 0.55, preferably wherein x is 0.25 or x is 0.5. 36. A crystalline compound according to any one of clauses 33 to 35 wherein the compound of formula (II) is a compound of formula (IIA): Cu_1+2x{Ag_3xBi_1-x}₂I_7+2x (IIA); wherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (IA) may vary by ±30%. 37. A crystalline compound according to any one of clauses 1 to 26 and 33 to 36 wherein the compound is CuAg_1.5Bi_0.5I₄ , wherein the stoichiometry of each of the ions in the compound may vary by ±30%. 38. A crystalline compound according to any preceding clause wherein the crystalline compound is a semiconductor, preferably wherein the crystalline compound is a semiconductor having a band gap of from 0.5 to 3.5 eV. 39. A film comprising a crystalline compound as defined in any preceding clause. 40. A passivated film comprising: a) a crystalline compound as defined in any one of clauses 1-38; and b) a passivating agent. 41. A passivated film according to clause 40, wherein the passivating agent is selected from one or more of iodine, sulfur, selenium or methylamine. 42. A semiconductor device comprising a compound as defined in any one of clauses 1 to 38 or a film as defined in any one of clauses 39-41. 43. A semiconductor device according to clause 42 wherein the device is an optoelectronic device, optionally wherein the semiconductor device is an optoelectronic device selected from a photovoltaic device, a light emitting device, a photodetector or an X-ray scintillator optionally wherein the semiconductor device is a photovoltaic device selected from a single- junction photovoltaic device, a tandem junction photovoltaic device or a multi-junction photovoltaic device. 44. A semiconductor device according to clause 42 or clause 43, which semiconductor device comprises: an n-type region comprising at least one n-type layer; a p-type region comprising at least one p-type layer; and, disposed between the n-type region and the p-type region: a layer comprising the compound or the film. 45. A process for producing a film comprising a crystalline compound comprising (i) One or more first cations, A; (ii) One or more second cations, B; (iii) One or more third cations, B’; (iv) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, said process comprising disposing a film-forming solution on a substrate wherein the film-forming solution comprises the one or more first cations, A, the one or more second cations, B, the one or more third cations, B’, and the one or more anions, X. 46. A process according to clause 45 wherein the film-forming solution comprises pyridine. 47. A process according to clause 45 or clause 46 wherein the process comprises dissolving a compound comprising the one or more first cations, A, in a first solvent to produce a first solution, dissolving a compound comprising the one or more second cations, B, and a compound comprising the one or more third cations, B’, in a second solvent to produce a second solution, contacting the first solution with the second solution to form the film-forming solution. 48. A process according to clause 47 wherein the first and/or the second solution comprises the one or more anions, X, optionally wherein both the first and second solutions comprise the one or more anions, X. 49. A process according to clause 47 or clause 48 wherein the first solvent is different from the second solvent. 50. A process according to any one of clauses 47 to 49 wherein the first and second solvents are organic solvents, preferably wherein the first and second solvents are polar organic solvents, preferably wherein the first and second solvents are polar aprotic organic solvents. 51. A process according to any one of clauses 47 to 50 wherein the first and second solvents are selected from dimethylsulfoxide, pyridine, dimethylformamide, N-butylamine and mixtures thereof, preferably wherein one of the first or second solvents comprises pyridine. 52. A process according to any one of clauses 47 to 51 wherein the step of dissolving a compound comprising the one or more first cations, A, in a first solvent to produce a first solution is performed at a first temperature, wherein the first temperature is between 30 ºC and 200 ºC, preferably wherein the first temperature is between 100 ºC and 200 ºC. 53. A process according to any one of clauses 47 to 52 wherein the step of dissolving a compound comprising the one or more second cations, B, and a compound comprising the one or more third cations, B’, in a second solvent to produce a second solution is performed at a second temperature, wherein the second temperature is between 30 ºC and 200 ºC, preferably wherein the second temperature is between 75 ºC and 150 ºC. 54. A process according to any one of clauses 45 to 53 wherein the film-forming solution is stirred at a third temperature prior to disposing the film-forming solution on the substrate, wherein the third temperature is between 30 ºC and 200 ºC, preferably wherein the third temperature is between 75 ºC and 150 ºC. 55. A process according to any one of clauses 45 to 54 wherein the substrate is preheated to a fourth temperature, wherein the fourth temperature is between 30 ºC and 200 ºC, preferably wherein the first temperature is between 100 ºC and 200 ºC. 56. A process according to any one of clauses 45 to 55 wherein the film-forming solution is disposed on the substrate by solution phase deposition, preferably wherein the film- forming solution is disposed on the substrate by spin-coating. 57. A process for producing a film comprising a crystalline compound comprising (i) One or more first cations, A; (ii) One or more second cations, B; (iii) One or more third cations, B’; (iv) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, said process comprising: exposing the substrate to a vapour, which vapour comprises one or more reactants for producing said compound and allowing deposition of the vapour onto the substrate to produce a film of said crystalline compound thereon. 58. A process according to clause 57 which process comprises evaporating one or more compounds comprising the one or more first cations, A, the one or more second cations, B, the one or more third cations, B’, and the one or more anions, X, to form the vapour. 59. A process according to clause 57 or 58 wherein the steps of evaporating the one or more compounds and exposing the substrate to the vapour are performed in a vacuum. 60. A process according to clauses 58 or 59 wherein each of the one or more compounds are contained in a crucible that is opaque to visible light, preferably wherein the one or more compounds are contained in alumina crucibles. 61. A process according to any one of clauses 57 to 60 which process comprises evaporating a first compound comprising the one or more first cations, A, evaporating a second compound comprising the one or more second cations, B, and evaporating a third compound comprising the one or more third cations, B’, to produce the vapour wherein at least one of the first, second and third compounds also comprises the one or more anions, X. 62. A process according to clause 61 wherein each of the first, second and third compounds comprises the one or more anions, X. 63. A process according to any one of clauses 45 to 62 which process further comprises annealing the film, preferably which process comprises annealing the film at a temperature of from 40 to 250ºC. 64. A process according to any one of clauses 45 to 63 wherein - the one or more first cations, A, are as defined in any one of clauses 3 to 7; - the one or more second cations, B, are as defined in any one of clauses 8 to 11; - the one or more third cations, B’, are as defined in any one of clauses 12 to 15; - the one or more anions, X, are as defined in any one of clauses 16 to 18; and/or - the compound is as defined in any one of clauses 2 and 19 to 38. 65. A process according to any one of clauses 45 to 64 wherein the substrate is a component for a semiconductor device. 66. A process according to any one of clauses 45 to 65, wherein the process further comprises exposing the film comprising a crystalline compound to a passivating agent. 67. A process according to clause 66, wherein the passivating agent comprises vapour of one or more of iodine, sulfur, selenium or methylamine 68. A process for producing a semiconductor device comprising producing a film comprising a crystalline compound on a substrate using the process as defined in any one of clauses 45 to 67, and disposing one or more further components on the film to produce a semiconductor device. 69. Use of a compound as defined in any one of clauses 1 to 38 as a. an optoelectronic material, preferably as a photovoltaic material or as an electroluminescent material; b. a luminescent material, preferably as a phosphor; c. a scintillator in a radiation detector; or d. a photodetector.

Claims

CLAIMS 1. A crystalline compound comprising (i) One or more first cations, A; (ii) One or more second cations, B; (iii) One or more third cations, B’; (iv) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment.

2. A crystalline compound according to claim 1, wherein the compound is not a perovskite, and/or wherein the molar ratio of the one or more first cations, A, to the one or more second cations, B, in the crystalline compounds is at least 1:9, preferably wherein the molar ratio of the one or more first cations, A, to the one or more second cations, B, in the crystalline compounds is at least 1:5.

3. A crystalline compound according to claim 1 or claim 2 wherein the one or more first cations, A, comprise one or more of Mg²⁺, Ga³⁺, Ge⁴⁺, Si⁴⁺, P⁵⁺, Cu⁺, Zn²⁺, Ti⁴⁺, Al³⁺, In³⁺ and Li⁺, preferably wherein the one or more first cations, A, comprise Cu⁺.

4. A crystalline compound according to any preceding claim wherein at least some of the one or more second cations, B, are in an octahedral coordination environment.

5. A crystalline compound according to any preceding claim wherein the one or more second cations, B, comprise one or more of Ag⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cr³⁺, V²⁺, Cd²⁺, Cu⁺, Au⁺, Pt²⁺, Pt⁴⁺, Mg²⁺, Li⁺, Na⁺, K⁺, Rb⁺, Cs⁺, In⁺, Ca²⁺, Sr²⁺, La³⁺ and Y³⁺, preferably wherein the one or more second cations, B, comprise Ag⁺.

6. A crystalline compound according to any preceding claim wherein the one or more third cations, B’, comprise one or more of Bi³⁺, Pb²⁺, Sn²⁺, Ge²⁺, In⁺, Sb³⁺, Te⁴⁺, Tl⁺, In³⁺ and Se⁴⁺, preferably wherein the one or more third cations, B’, comprise Bi³⁺.

7. A crystalline compound according to any preceding claim wherein the one or more anions, X, comprise one or more anions selected from halide anions and chalcogenide anions, preferably wherein the one or more anions, X, comprise I-.

8. A crystalline compound according to any preceding claim wherein the compound comprises an interconnected network of B’X₆ octahedra, preferably wherein the compound comprises an interconnected network of BiI₆ octahedra.

9. A crystalline compound according to any preceding claim wherein: the one or more first cations, A, comprise Cu⁺; the one or more second cations, B, comprise Ag⁺; the one or more third cations, B’, comprise Bi³⁺; and the one or more anions, X, comprise I-.

10. A crystalline compound according to any preceding claim wherein the compound is a compound of formula (I): [A]_4x{[B][B’]}_1-x[X]₄ (I); wherein [A] is the one or more first cations, [B] is the one or more second cations, [B’] is the one or more third cations and [X] is the one or more anions, wherein the one or more first cations, A, are different from the one or more second cations, B, and the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, and wherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (I) may vary by ±30%.

11. A crystalline compound according to claim 10 wherein the compound of formula (I) is a compound of formula (IA): Cu_4x{AgBi}_1-xI₄ (IA); wherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (IA) may vary by ±30%.

12. A crystalline compound according to any preceding claim wherein the compound is CuAgBiI₅, wherein the stoichiometry of each of the ions in the compound may vary by ±30%.

13. A crystalline compound according to one of claims 1 to 11 wherein the compound is Cu₂AgBiI₆, wherein the stoichiometry of each of the ions in the compound may vary by ±30%.

14. A crystalline compound according to any one of claims 1 to 9 wherein the compound is a compound of formula (II): [A]_1+2x{[B]_3x[B’]_1-x}₂[X]_7+2x (II); wherein [A] is the one or more first cations, [B] is the one or more second cations, [B’] is the one or more third cations and [X] is the one or more anions, wherein the one or more first cations, A, are different from the one or more second cations, B, and the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, and wherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (I) may vary by ±30%.

15. A crystalline compound according to claim 14 wherein the compound of formula (II) is a compound of formula (IIA): Cu_1+2x{Ag_3xBi_1-x}₂I_7+2x (IIA); wherein x is greater than 0 and less than 1, and wherein the stoichiometry of each of the ions in the compound of formula (IA) may vary by ±30%.

16. A crystalline compound according to any one of claims 1 to 9, 14 or 15 wherein the compound is CuAg_1.5Bi_0.5I₄ , wherein the stoichiometry of each of the ions in the compound may vary by ±30%.

17. A film comprising a crystalline compound as defined in any preceding claim.

18. A passivated film comprising: a) a crystalline compound as defined in any one of claims 1-16; and b) a passivating agent.

19. A passivated film according to claim 18, wherein the passivating agent is selected from one or more of iodine, sulfur, selenium or methylamine.

20. A semiconductor device comprising a compound as defined in any one of claims 1 to 16 or a film as defined in any one of claims claim 17 to 19.

21. A semiconductor device according to claim 20 wherein the device is an optoelectronic device, optionally wherein the semiconductor device is an optoelectronic device selected from a photovoltaic device, a light emitting device, a photodetector or an X-ray scintillator optionally wherein the semiconductor device is a photovoltaic device selected from a single- junction photovoltaic device, a tandem junction photovoltaic device or a multi-junction photovoltaic device.

22. A process for producing a film comprising a crystalline compound comprising (i) One or more first cations, A; (ii) One or more second cations, B; (iii) One or more third cations, B’; (iv) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, said process comprising disposing a film-forming solution on a substrate wherein the film- forming solution comprises the one or more first cations, A, the one or more second cations, B, the one or more third cations, B’, and the one or more anions, X.

23. A process for producing a film comprising a crystalline compound comprising (i) One or more first cations, A; (ii) One or more second cations, B; (iii) One or more third cations, B’; (iv) One or more anions, X; wherein the one or more first cations, A, are different from the one or more second cations, B, and to the one or more third cations, B’, and wherein the one or more second cations, B, are different from the one or more first cations, A, and to the one or more third cations, B’, wherein at least some of the one or more third cations, B’, are in an octahedral coordination environment, and wherein at least some of the one or more first cations, A, are in a tetrahedral coordination environment, said process comprising: exposing the substrate to a vapour, which vapour comprises one or more reactants for producing said compound and allowing deposition of the vapour onto the substrate to produce a film of said crystalline compound thereon.

24. A process according to claim 22 or claim 23 wherein - the one or more first cations, A, are as defined in claim 3; - the one or more second cations, B, are as defined in claim 4 or claim 5; - the one or more third cations, B’, are as defined in claim 6; - the one or more anions, X, are as defined in claim 7; and/or - the compound is as defined in any one of claims 2 and 8 to 16.

25. A process according to any one of claims 22 to 24 wherein the substrate is a component for a semiconductor device.

26. A process according to any one of claims 22 to 25, wherein the process further comprises exposing the film comprising a crystalline compound to a passivating agent.

27. A process according to claim 26, wherein the passivating agent comprises vapour of one or more of iodine, sulfur, selenium or methylamine.

28. A process for producing a semiconductor device comprising producing a film comprising a crystalline compound on a substrate using a process as defined in any one of claims 22 to 27, and disposing one or more further components on the film to produce a semiconductor device.

29. Use of a compound as defined in any one of claims 1 to 16 as a. an optoelectronic material, preferably as a photovoltaic material or as an electroluminescent material; b. a luminescent material, preferably as a phosphor; c. a scintillator in a radiation detector; or d. a photodetector.