Optimal compositions of inorganic perovskites for the fabrication of electronic devices, particularly photovoltaic devices

Abstract

An electronic device is proposed comprising at least one layer made of a material containing an inorganic perovskite: * of stoichiometric composition ABX3 where: - A is an element from Cs and Rb, - B is an element or mixture of elements from Ge, Sn and Pb, and - X is an element or mixture of elements from Cl, Br and I, * and having a Goldschmidt factor t between 0.87 and 0.93, characterizing satisfactory stability. Figure 1

Description

Title of the invention: Optimal compositions of inorganic perovskites for the manufacture of electronic devices, particularly photovoltaic devices

[0001] This disclosure relates to the field of manufacturing electronic devices, in particular photovoltaic devices, for example solar panels for the exploitation of renewable energies.

[0002] It relates more particularly to the fabrication of potentially stable (especially temperature-stable) inorganic perovskite chemical compositions and their use, particularly in the aforementioned field. Indeed, depending on their forbidden energy bands, these materials can be used as absorbers, passivation layers, or other applications, notably for various photovoltaic cell technologies (tandem, multi-junction, or others).

[0003] Halogenated perovskites are easily synthesized (e.g., via the liquid process) and are straightforward to implement (particularly in thin films). Their band gap allows for a wider absorption of the solar spectrum compared to conventional materials (silicon technology, CuInGa(SSe)2 technology or CIGS technology, etc.). They currently achieve efficiencies of 26% and form the basis of tandem cells with efficiencies 33% higher.

[0004] Nevertheless, these perovskites remain unstable and do not currently allow for the consideration of a commercial deployment of this technology. Various aging factors include, in particular, received ultraviolet radiation ("UV" hereafter), the presence of water (H2O) or oxygen (O2), temperature and, of course, the type of techniques used for their manufacture.

[0005] One of the reasons for their instability is related to the organic nature of perovskites. Many solutions have been explored (effects of substitutions, perovskite encapsulation techniques, etc.), but a readily usable technique remains to be found to obtain long-term stability beyond a few thousand hours today and towards the material lifetime of a photovoltaic device, usually a few decades.

[0006] A technique seems to be emerging recently. It is based on the use of inorganic perovskites, from the "ABX3" family (where A = K, Rb and Cs; B = Ge, Sn and Pb; and X = Cl, Br and I), for example. They have yields close to 21%, possess good stability under normal operating conditions, and the impact of molecules such as water on their electronic properties seems less important than for organic perovskites. They are therefore also good candidates for passivation layers.

[0007] Determining suitable chemical compositions for inorganic halogenated perovskites should provide potential performance in terms of temperature stability and tolerance to various contaminants found in ambient air, such as water. Optoelectronic characteristics satisfying the requirements of industrially manufactured photovoltaic cells for conventional outdoor use should then be obtainable.

[0008] One factor of instability of these materials however is their rich sequences of phase transitions in temperature and / or pressure.

[0009] The present invention improves the situation.

[0010] It starts from the observation that these instabilities in this type of material are often linked to the existence of vibration modes or phonons, called "soft modes" or "unstable modes".

[0011] It also stems from the observation that these instabilities are minimal for ranges of inorganic perovskite compositions that have band gaps compatible with photovoltaic applications, for single-junction cells as well as tandem or multi-junction cells with silicon or CIGS. With silicon in particular, the composition of the perovskite used can be chosen to ensure lattice matching.

[0012] Furthermore, it has been observed that some of these inorganic perovskites thus obtained had very high static dielectric constants, thus forming prime candidates for the development of materials with supercapacitive effects, in particular for electronic devices of the supercapacitor type.

[0013] Thus, more generally, it has been observed that the aforementioned unstable vibration modes were much less pronounced in certain particular families of compositions and that their fabrication in thin film in all types of electronic devices therefore did not present any particular difficulties compared to that of other perovskites.

[0014] Thus, what is referred to here is an electronic device comprising at least one layer made of a material comprising an inorganic perovskite: * of stoichiometric composition ABX3 where: - A is one element among Cs and Rb, - B is an element or mixture of elements from Ge, Sn and Pb, and - X is an element or mixture of elements from Cl, Br and I, * and having a Goldschmidt factor t between 0.87 and 0.93.

[0015] An approach was used to calculate phonon modes and to study the amplitude stability of the corresponding vibration modes. A so-called vibration amplitude "unstable" can be explained by a phonon frequency that has been calculated to be negative.

[0016] It was then observed in fact: - on the one hand, a verified correlation between the so-called "Goldschmidt t" tolerance factor (ideally close to or equal to 0.9) and the stability of the inorganic perovskite, and - on the other hand, that it was possible, by this approach of calculating phonon modes, to determine precisely at least the type of cations A which could offer promising perovskite stability (therefore among Cs (cesium) and Rb (rubidium), according to this determination).

[0017] In one embodiment, the material considered is at least partially crystalline, according to a crystallographic structure of the type Pm3m-

[0018] The aforementioned phonon mode calculations could indeed be carried out by considering, for example, a supercell of dimensions that are multiples of an elementary unit cell having this Pmini- symmetry

[0019] The calculation of the aforementioned phonon modes made it possible not only to determine the types of cations A offering the best stability, but also the type of mixture of cations B and / or anions X offering optimal stability.

[0020] Thus, in one embodiment, the inorganic perovskite then has more particularly the stoichiometric composition A(BzB'i.z)(XyX'i.y)3 where: - z and y are between 0 and 1 (greater than or equal to 0 and less than or equal to 1), - A is one element among Cs and Rb, - B and B' are distinct elements among Ge, Sn and Pb, and - X and X' are distinct elements among Cl, Br and I.

[0021] When the cation A is cesium (Cs), it has been observed that the material is devoid of unstable vibration mode when chosen from the compositions: CsGeo25Pb o.75(Bro.5Clo.5)3, CsPb0.nsSno.gvsIs, CsPbo.25Sno.75I3, CsPbo.375Sno.625(Bro.25Clo.75)3, CsPbo.sSn O.5(Bro.25Clo.75)3, CsPbo.5Sno.5Br3, CsPbo.5Sno.5Cl3, CsPbo.625Sno.375C13, and CsPbo.75Sno.25Q3.

[0022] Thus, in one embodiment of the electronic device, the material is devoid of unstable vibration mode and is chosen from one of the nine compositions above.

[0023] Since a plurality of compositions offer good perovskite stability, it is possible to select in particular at least one composition for a intended application.

[0024] For example, in an embodiment where the electronic device is specifically of the photovoltaic type and where the aforementioned layer has photovoltaic properties, the material can be chosen both for: - be devoid of unstable vibration modes, and - have a gap width (or "gap" below) between 1.4eV and 2.3eV, compatible with a photovoltaic application.

[0025] In this case, the material can be selected from the three compositions CsPbo.375 Sn0.625(Bro.25Clo.75)3, CsPbo.5Sno.5(Bro.25Clo.75)3, and CsPbo.5Sno.5Br3.

[0026] Indeed, it was observed that these three compositions had gaps between 1.4eV and 2.3eV and were devoid of unstable vibrational modes.

[0027] In an example embodiment where the device is of the photovoltaic type and has a tandem or multi-junction cell comprising a silicon layer, and where the aforementioned layer has photovoltaic properties with a band gap between 1.4eV and 2.3eV, the inorganic perovskite can then have a composition of the type Rb(PbxSni x)Br3, with x between 0 and 0.7, with a mesh size close to a silicon mesh size to deposit the silicon and perovskite layers one on top of the other.

[0028] Indeed, this other family of rubidium-based compositions, Rb(PbxSni x)Br3, with x between 0 and 0.7, also has (in addition to those based on cesium) the advantage of having: - A limited number of unstable modes (less than 10, whereas other rubidium-based compositions can have more than 30 unstable modes), and - Gaps between 1.4 and 2.3 eV.

[0029] Furthermore, for this rubidium-based family, the lattice size is sufficiently close to that of crystalline silicon so that a relatively low or at least controlled dislocation rate can be obtained in the perovskite / silicon stacking.

[0030] Still in an embodiment where the device is of the photovoltaic type, but where the layer is a passivation layer (therefore with a band gap greater than 2.3eV), the material is devoid of unstable vibration mode if it is chosen from the compositions CsGeO25 Pbo.75(Bro.5Clo.5)3, CsPbo.5Sno.5Cl3, CsPbo.625Sno.375C13 and CsPbo.75Sno.25Q3.

[0031] In an embodiment where the device is of the supercapacitor type, the material can be chosen for supercapacitive effects and is devoid of unstable vibration mode if it is chosen from the compositions: CsGeo.25Pbo.75(Bro.5Clo.5)3, CsPbo.125Sno.875 I3, CsPbo.25Sno.75I3, and CsPbo.5Sno.5(Bro.25Clo.75)3.

[0032] Indeed, the static dielectric constant of stable cesium-based perovskites has also been calculated and very promising results (with static dielectric constants greater than 100) have been obtained for the four compositions above.

[0033] According to another aspect, a computer-based method is referred to for estimating the stability of an inorganic perovskite of stoichiometric composition ABX3 where: - A is at least one element among Li, Na, K, Rb and Cs, - B is at least one element or a mixture of elements from among Ge, Sn and Pb, and - X is at least one element or a mixture of elements from Cl, Br and I, The process involves: - Consider a supercell having dimensions that are multiples (for example 2x2x2) of the size of a unit cell of a crystal of said inorganic perovskite, - For each element A and for different mixtures of cations B and different mixtures of anions X, test phonon modes characterizing vibrations in the supercell to identify vibration modes with unstable amplitude (typically having negative phonon frequencies), and determine for each composition a number of modes with unstable amplitude, - Select at least one composition whose number of unstable amplitude vibration modes is less than a threshold, with a view to using the selected composition for the manufacture of an electronic device comprising at least one layer made in an inorganic perovskite having the selected composition.

[0034] Typically, the aforementioned threshold can be 10, for example, particularly for the rubidium-based family mentioned above.

[0035] In an embodiment where a photovoltaic application is targeted, for example, the aforementioned process may more specifically include: - Select a plurality of compositions whose number of unstable amplitude vibration modes is less than a threshold, and - Estimate at least one bandwidth restriction for each composition to select at least one composition compatible with a targeted photovoltaic application. With a view to using the selected compatible composition to manufacture a photovoltaic type electronic device comprising at least one layer made of an inorganic perovskite having said compatible composition.

[0036] Depending on the intended application for the layer (passivation or photovoltaic properties), the value of its gap can be the criterion for selecting the compatible composition, mentioned above.

[0037] According to another aspect, it also relates to a computer program comprising instructions for implementing the above-mentioned process, when these instructions are implemented by a processor. According to another aspect, a non-transient, computer-readable recording medium is proposed on which such a computer program is recorded.

[0038] Other advantages and features will become apparent upon reading the detailed description of embodiments below, and upon examination of the accompanying drawings on which: - Fig. 1 illustrates a correlation between the stability of the amplitude of the vibration modes of inorganic perovskites and their Goldschmidt factor t, - Figure 2 illustrates the positions of atomic sites in a supercell having 2x2x2 times the size of the primitive cubic unit cell (space group Pmï>m) of an ABX3 type perovskite, - Figure 3 illustrates an algorithmic approach to selecting the best candidate perovskites in terms of stability, - Fig. 4 illustrates the evolution of the band gap energy with the chemical composition of a perovskite of formula Cs(BxB'i_x)(XyX'i_y)3 (with B = Ge, Sn and Pb, and X = Cl, Br and I), - Fig. 5 illustrates the number of unstable modes as a function of the chemical composition of a perovskite with the formula Cs(BxB'i.x)(XyX'i.y)3 (with B = Ge, Sn and Pb, and X = Cl, Br and I), - Fig. 6a illustrates the evolution of the band gap energy with the chemical composition of a perovskite of formula Rb(BxB'i x)(XyX'i y)3 (with B = Ge, Sn and Pb, and X = Cl, Br and I), - Fig. 6b illustrates the number of unstable modes as a function of the chemical composition of a perovskite with the formula Rb(BxB'i x)(XyX'i y)3 (with B = Ge, Sn and Pb, and X = Cl, Br and I), - Fig. 7 schematically illustrates the classic NiP architecture of a single-junction photovoltaic cell comprising the usual thin layers for connection and electron or hole collection, but nevertheless and particularly including an active layer with photovoltaic properties made of an ABX3 type material where A=Cs or Rb, - Figures [Fig. 8a] and [Fig. 8b] illustrate the variation of the average lattice parameter for cubic perovskites, respectively of type Cs(BxB'i_x)(Xy X'j_y)3 and Rb(BxB'i x)(XyX'i y)3, with B = Ge, Sn and Pb, and X = Cl, Br and I.

[0039] In the following description, it is proposed to define firstly the chemical composition of inorganic halogenated perovskites by optimizing their vibration modes so that stability is an intrinsic property of these materials.

[0040] To this end, it is proposed to first exploit atomic modeling methods, based on so-called ab-initio approaches (Density Functional Theory (DFT), hybrid functionals) respecting the first principles of quantum chemistry, in order to study the stability of the performance of different inorganic perovskites under different constraints, in particular the stability of the vibration modes of these materials which guide their phase transitions in temperature.

[0041] In addition to the stability constraint, there will be the constraint of practical use, for example in photovoltaic cells. Typically, in the context of the development of so-called "tandem" cell technologies, or triple junctions and wide-bandgap compounds, the materials must meet various criteria: - To have a forbidden energy band (or "gap") between 1.4 and 2.3 eV, - Avoid phase transitions within the operating temperature range. - Possibly have optoelectronic properties comparable to reference materials (in order not to disrupt cell yields).

[0042] A first approach to this is the search for the stabilization in particular of the cubic phase of perovskites which is one of the phases common to all these materials (the procedure described here can nevertheless be adapted to all phases).

[0043] One of the factors contributing to the instability of these materials is their rich sequences of phase transitions: these are often linked to the existence of vibrational modes or phonons (so-called "soft" or "unstable" modes). It is proposed below to define the chemical composition of complex inorganic perovskites in order to reduce or even eliminate these specific vibrational modes while also providing a qualitative and quantitative description of the structural, electronic, and optoelectronic properties of the materials obtained.

[0044] The 45 basic inorganic perovskites considered to define the optimal chemical compositions are those most commonly studied in the field of photovoltaics, i.e. the general formula ABX3, with: A = Li, Na, K, Rb and Cs, B = Ge, Sn and Pb and X = Cl, Br and I.

[0045] A Hamiltonian has been defined by calculation in order to reproduce the structural, electronic, and dielectric properties of the previously defined materials with good accuracy. agreement with experience: It combines 17.23% of the so-called "Hartree-Fock" exchange with the so-called "PBE" exchange-correlation functional according to the reference: J. Perdew et al., Physical Review Letters, 77, 3865 (1996).

[0046] It turns out that the average errors obtained on the lattice parameters and energy bands of CsBX3 perovskites (therefore with A=Cs for cesium) are 2% and 10% with respect to the available experimental data. Thus, predictive calculations from a Hamiltonian can indeed be performed on these materials.

[0047] With reference to Figure 3 discussed later, a so-called "supercell" approach was then used in order to take into account that the phase transitions are linked to the existence of unstable modes at points G, X, R, and M of the Brillouin zone, considering in particular a size of this "supercell" corresponding to 2 x 2 x 2 times that of the primitive cubic unit cell (space group Pm3m) of the ABX3.

[0048] The nature of the phase transitions and the stability of perovskites is generally correlated with a parameter called "Goldschmidt factor" and denoted "t" below.

[0049] It is usually considered that if 0.8 < t < 1, the perovskite is stable in its cubic phase. However, Table 1 shows that many perovskites satisfying this criterion nevertheless possess unstable modes. This criterion is therefore not sufficient for selecting materials that meet the desired stability criterion.

[0050] Table 1 below presents the Goldschmidt tolerance factors for each pure compound, with criterion Ci corresponding to compliance with the experimental criterion 0.8 < t < 1 for obtaining stable perovskites in the cubic phase. It also presents the high-symmetry k-points of the first Brillouin zone exhibiting imaginary phonon modes. Criterion C2, on the other hand, indicates compounds whose cubic phase is dynamically metastable. Table 1#:

[0051] t Ci A-points with unstable mo des C2 LiGeCl3 0.760 xr, X, M, R x LiGeBr3 0.757 xr, M, RX LiSnI3 0.684 X r, LiPbI3 0.651 X r, X, M, R X NaGeCh 0.891 / r, X, M, R X NaGeBr3 0.881 / r, X, M, R X NaGeI3 0.866 / r, X, M, R X NaSnCl3 0.798 X r, X, M, R X NaSnBr3 0.793 X r, X, M, R X NaSnI3 0.787 X r, X, M, R X NaPbCl3 0.754 X r, X, M, R X NaPbBr3 0.752 X r, X, M, R X NaPbI3 0.749 X r, X, M, R X KGeCl3 0.960 / r, X, M, R X KGeBr3 0.946 / r, X, M, R X KGeI3 0.927 / r, X, M, R X KSnCl3 0.860 / r, X, M, R X KSnBr3 0.853 / r, X, M, R X KSnI3 0.842 / r, X, M, R X KPbCl3 0.813 / r, X, M, R X KPbBr3 0.808 / r, X, M, R X KPbI3 0.801 / r, X, M, R X RbGeCl3 0.983 / r,x, m X RbGeBr3 0.967 / r,x, m X RbGeI3 0.946 / r, X, M, R X RbSnCl3 0.880 / r, m, r X RbSnBr3 0.872 / r, X, M, R X RbSnI3 0.859 / M, R X RbPbCl3 0.832 / M, R X RbPbBr3 0.826 / r, m, r X RbPbI3 0.818 / M, R X CsGeCl3 1.027 X r,x, m X CsGeBr3 1.009 X r,x, m X CsGeI3 0.985 / r,x, m X CsSnCl3 0.920 / r,x, m X CsSnBr3 0.910 / / CsSnI3 0.894 / / CsPbCl3 0.870 / M, R X CsPbBr3 0.862 / M, R X CsPbI3 0.851 / M, R X

[0052]

[0053]

[0054]

[0055]

[0056] To select the relevant materials for defining the chemical compositions of stable complex inorganic perovskites, the t factor of the different perovskites was correlated with the magnitude of the most common structural deformations during the phase transitions of these materials, as illustrated in [Fig.1]. Figure 1 illustrates the magnitude of the different distortions in ABX3 perovskites as a function of the Goldschmidt factor t, in particular for: - the rotation of the octahedra forming the structure of the ABX3 material, in (a), - the displacement of cations A, in (b), - the displacement of cations B, in (c), and - the distortion of the bond length of the octahedra, in (d). It turns out that the optimal family with respect to these amplitudes of different distortions is the one where the cation A is Cs (cesium) or Rb (rubidium). This [Fig. 1] shows that only the perovskites that minimize the differences (the lowest in the curves) are the CsBX3 and RbBX3 families (with B = Ge, Sn and Pb, and X = Cl, Br and I). The t criterion satisfied by these materials is: 0.875 < t < 0.925. These two families are therefore selected to optimize the chemical composition of complex cubic perovskites of the type ABxB'i_x(XyX'i y)3, in order to minimize or eliminate the existence of unstable modes. To refine relevant determinations of x and y in the above composition, the following steps can be carried out: a. The relaxation of each possible compound configuration is studied to select the most stable structure (minimizing the internal energy of the material), b. The gap value is estimated to determine if the material is suitable for a photovoltaic application. c. It is further determined whether the material in question exhibits unstable phonon modes (vibration modes with unstable amplitude), and if not, d. It is determined whether phase segregation is possible, which would be unfavorable (this determination can be based on a comparison of the enthalpies calculated for the phases with and without segregation), and if this is not the case, e. A more detailed determination is carried out, in particular of the gap, optical and electronic properties, as well as the lattice parameter of the material to determine in particular whether a lattice match can be achieved with materials such as silicon for multi-junction cells or tandem with silicon for example, in order to model a photovoltaic cell which can offer the required characteristics.

[0057] Figure 2 shows the substitution sites of ions A, B, and X, for applying step a) above to optimize and determine the dynamically stable phases for each chemical composition. To study the properties of a complex perovskite of the type ABxB'i_x(XyX'i y)3, 288 different chemical compositions were explored, which corresponds to 780 different atomic configurations.

[0058] More specifically, it is represented: - the substitution sites A, B and X in a 2x2x2 type supercell, in (a), - the sites of the cations A for the symmetry Pm3m (independent atomic positions by symmetry operation in this space group Pm3ni), in (b), - the sites of the B cations for this Pm3m symmetry, in (c), and - the sites of the X anions for this Pm3m symmetry, in (d).

[0059] Figure 3 presents possible algorithmic steps for selecting the best candidate compositions based on their stability. Figure 3 can thus summarize the steps of a process of the type presented above, according to an example embodiment. It can also correspond to the flowchart of a possible algorithm for a computer program of the type presented in the introduction above.

[0060] Here, to explore the effects of the chemical composition of solid perovskite solutions on their properties, an approach considering a supercell with dimensions larger than a unit cell was applied. Such an approach is detailed in particular in the document: F. Lafond et al., in the Journal of Physical Chemistry C, 124, 10353 (2020).

[0061] For each "commensurable" chemical composition, all substitution schemes that are not equivalent by symmetry have been taken into account. It is indeed possible to generate several arrangements of atoms that would give the same final chemical composition, but without necessarily having the same properties. For each of them, the lattice parameters and internal coordinates have therefore were optimized, while maintaining the symmetry imposed by the substitution scheme in question.

[0062] For each composition, only the configuration that minimizes the internal energy of the system is retained (the most stable). The calculation of phonons at the high symmetry point Ty is performed in order to estimate the dynamic stability.

[0063] In the case of unstable modes (those with imaginary frequencies), the symmetry has been reduced accordingly, and the structure optimization has been redone to allow shifts corresponding to the phonon eigenvector of such an unstable mode. Since several modes of imaginary phonons can exist, and each can lead to a different reduction in the symmetry order (and therefore to different distortions), each symmetry-independent mode is thus tracked, again generating several configurations for a composition.

[0064] For each composition, the most stable configuration was selected for subsequent structural optimization, followed again by the calculation of vibrational frequencies (as in the previous step) for the final stability analysis. The contribution of phonons (zero-point energy of the first Brillouin zone) was then taken into account for the final evaluation of the relative stability of the configurations.

[0065] It will be understood from the comments in Figure 3 above that the distribution of cations B and / or anions X in the sites of the Pm^n crystal lattice (determining a stoichiometry to be respected) makes it possible to determine configurations offering the desired stability in terms of vibrational modes (eliminating unstable modes or limiting the number of unstable modes to a value below a threshold). It is further explained below that the chemical composition (with a precise stoichiometry) of the perovskites then makes it possible to reduce the number of unstable modes that these materials can possess.

[0066] It has already been observed that only cesium-based perovskites CsBxB'i x(XyX' iy)3 can possess chemical compositions that completely eliminate the existence of unstable modes. It has also been observed that rubidium (Rb)-based perovskites exhibit composition ranges with an overall low number of unstable modes.

[0067] It will then be understood that the precision of a particular perovskite composition (determination of optimal values ​​for x and y in ABxB'i.x(XyX'i.y)3, with A = Cs or Rb) makes it possible firstly to reduce the number of unstable modes that these materials can possess, but also to achieve a desired property (for example a gap value, a lattice match, or others) depending on a given application context (for photovoltaic applications typically).

[0068] Figure 4 shows the evolution of the band gap energy with the chemical composition for the cubic unit cell. Although this energy may change with phase transitions, this figure provides a relevant basis for selecting chemical compositions with a given band gap energy according to the desired technology for photovoltaic applications.

[0069] In particular, [Fig. 4] shows the band gap energy for cubic perovskites of the type Cs(BxB'i_x)(XyX'i_y)3 (with B = Ge, Sn and Pb, and X = Cl, Br and I). The band gap energy limits of interest for photovoltaics are indicated by the 1.4 eV and 2.3 eV lines shown in white.

[0070] Figure 5 also shows the evolution of the number of unstable modes. More In particular, [Fig. 5] shows the influence of the chemical composition of Cs(BxB'i.x)(XyX'i.y)3 type perovskites (with B = Ge, Sn, and Pb, and X = Cl, Br, and I) on the number of unstable modes in the material. Thus, typically, candidate compositions relative to a desired band gap, using the table in [Fig. 4], should be compared with the number of unstable modes according to the table in [Fig. 5] to verify whether a candidate composition, based on its band gap, is stable or not.

[0071] Table 2 below presents, by way of example, nine chemical compositions of different complex inorganic perovskites of the type Cs(BxB'i.x)(XyX'i.y)3 which do not possess unstable modes, with: - their Goldschmidt factor (t), - their forbidden energy band (Eg in eV), - their electronic dielectric constant (ex) and in parentheses the corresponding refractive index (n), - as well as their static dielectric constant (e).

[0072] These compounds all have a t factor equal to (or close to) 0.9, thus confirming better results relative to the compositions in Table 1 above.

[0073] [Tables2] Composition t Eg eœ (n) e CsGeo.25Pbo.75(Bro.5Clo.5)3 0.90 2.32 3.66(1.91) 123. CsPbo.i25Sno.875l3 0.89 1.13 5.89 (2.43) 387. cspbo.25sno.75I3 0.88 1.26 5.59 (2.36) 487. CsPbo.375Sno.625(Bro.25Clo.75)3 0.90 1.88 3.66(1.91) 16. CsPbo.5Sno.5(Bro.25Clo.75)3 0.89 1.94 3.59(1.89) 208. cspbo.5sno.5Br3 0.88 1.93 4.06 (2.01) 28. Cspbo.5sno.5cl3 0.89 2.59 3.23 (1.80) 32. CsPb0.625Sn0.375Cl3 0.89 2.71 3.15(1.77) 31. CsPbo.75Sno.25Ch 0.88 2.86 3.08 (1.75) 7.

[0074] Although some of the materials in Table 1 could exhibit unstable modes, a clear correlation between the stability of the materials obtained and their Goldschmidt factor t (here equal to or close to 0.9) was observed in the development of this Table 2.

[0075] It typically appears that the composition family Cs(PbxSni_x)(BryCliy) has gap values ​​between 1.4 and 2.3eV which are advantageously compatible with photovoltaic applications, provided that x and y are judiciously chosen between 0 and 1 and without risk of obtaining unstable modes according to [Fig.5] discussed later.

[0076] Thus, it is possible to modulate the x,y composition in a given formulation to refine, for example, desired gap or mesh matching properties.

[0077] Other properties may also be sought: for example, the static dielectric constant of the composition CsPbxSni_x(Bro.25Clo.75)3 differs significantly depending on whether x=0.375 or x=0.5 (for which the material may be suitable if supercapacitive effects are sought).

[0078] Although rubidium-based perovskites of the type RbBxB' ix(XyX' iy)3 do not possess a chemical composition without unstable modes (these modes being, however, very limited in number for this family of materials), the band gap energies are of interest for photovoltaic applications as shown: - Figure 6a illustrates the estimated bandgap energy as a function of the parameters x and y in the composition Rb(BxB'i x)(XyX'i y)3 with B = Ge, Sn and Pb, and X = Cl, Br and I, the bandgap energy limits of interest for photovoltaic applications being shown in white; and - The [Fig.6b], illustrating the number of unstable phonon modes for cubic perovskites Rb(BxB'i_x)(XyX'i y)3 (with B = Ge, Sn and Pb, and, X = Cl, Br and I).

[0079] Like cesium-based perovskites, although this band gap energy can change with phase transitions, [Fig.6a] provides a relevant basis for choosing chemical compositions with a given band gap energy according to the desired technology for photovoltaic applications.

[0080] Furthermore, compared to [Fig. 5], it can be observed that the layer in [Fig. 6b] for the Rb-based perovskite is less contrasted than that for the Cs-based perovskite. [Fig. 6b] indeed exhibits more gradual variations in the number of unstable modes as a function of the x,y compositions. Thus, it is possible to search for a candidate material around a given composition relative to its gap for to allow fine-tuning of the desired gap value (or the desired mesh parameter, for example) with reasonable predictability on the stability of the material.

[0081] Typically with reference to [Fig.6b], three light horizontal bands appear for the compositions: - Rb(PbxSni X)C13 with x between 0.4 and 1, - Rb(PbxSni x)Br3 with x between 0 and 0.7, - Rb(PbxSni X)I3 with x between 0 and 0.8.

[0082] The composition Rb(PbxSni X)C13 is unsuitable for a photovoltaic layer due to its band gap incompatibility (too high) but could be used for manufacturing a passivation layer. The two other compositions based on Br and I allow for the selection of different possible band gap values, but only the composition Rb(PbxSn ix)Br3 offers an acceptable lattice match with silicon for use in a tandem cell with this material (as shown in [Fig. 8b] discussed later). This composition Rb(PbxSni x)Br3, with x between 0 and 0.7, offers a wide range of possible band gaps, and this is compatible with a silicon layer for manufacturing a tandem cell.

[0083] The potential applications in fields other than photovoltaics remain the same as those for cesium-based perovskites, particularly if supercapacitive effects are sought.

[0084] In the field of application of a photovoltaic device in the broad sense, depending on their forbidden energy bands, these perovskite-based materials of thus optimized compositions can be used both as "absorbers" (active layer with photovoltaic properties in a photovoltaic cell), or as passivation layers, or others, and this for different possible technologies of photovoltaic cells (tandem, multi-junction, or others).

[0085] Typically: - Materials with band gaps between 1V and 1.3 eV can be considered as alternatives to silicon and used in "single junction" type cells, - Materials with band gaps between 1.3 eV and 1.8 eV can be considered for "tandem" type photovoltaic cells such as Si / perovskite or CIGS / perovskite cells, - Materials with band gaps > 1.8 eV can be considered as passivation layer materials for various photovoltaic cells transparent to solar radiation.

[0086] As presented above, some of these materials can possess high dielectric constants (» 10) which extends their applications to fields such as optoelectronics and microelectronics for applications that can be associated with piezoelectricity (piezoelectric sensors or emitters) or high-density dielectric energy storage (such as hypercapacitors or supercapacitors). Another application is photocatalysis (for CO2 reduction, for example).

[0087] Two main applications in the photovoltaic field are nevertheless presented and detailed below.

[0088] A first application is the manufacture of a single junction cell based on inorganic perovskites of type A(BxB'i_x)(XyX'i y)3 with A=Cs or Rb, B = Ge, Sn and Pb, and X = Cl, Br and I, with an appropriate x,y composition such as one of the nine compositions in Table 2 for example.

[0089] To this end, a frequently used architecture is of the "classic" NiP type, as shown in [Fig. 7] (the electron-extracting layer being placed as close as possible to the substrate, which is generally glass). The glass and transparent conductive oxide (or "TCO," which serves as the "bottom" contact) substrate can be purchased pre-assembled and cleaned. The most frequently used TCO in this architecture is FTO (for "fluorine-doped tin oxide"). The substrate is cleaned in an ultrasonic bath at room temperature, using various solvents (acetone, demineralized water, or others), and subsequently with ultraviolet radiation.

[0090] An electron transport layer (ETL) is then deposited on the TCO layer. Most often, this layer is composed of inorganic materials that allow:

[0091] - good light transmission,

[0092] - a favorable band alignment with the conduction band of the perovskite (and blocking the passage of the holes), and

[0093] - good electron mobility.

[0094] In this respect, many candidates exist, the most commonly used being SnO2, TiO2 and ZnO.

[0095] These are deposited in the form of a solution forming a homogeneous layer, which can subsequently undergo annealing at about one hundred degrees for about thirty minutes.

[0096] The absorber layer containing the selected perovskite is then deposited onto this layer. This layer can be deposited by first preparing a mixture of the precursors AX and BX2 (in salt form), in proportions that allow the desired solid solution to be reproduced. The salts are dissolved in organic solvents (such as NMP, DMF, DMSO) at room temperature, and a concentration of approximately 1 mol / L is reached in the solution. Additives, such as FAAc, can be added to the solution to improve crystallinity.

[0097] The resulting perovskite layer generally has a thickness between 100 and 500 nm. During deposition, regardless of the method chosen, a stream of nitrogen is blown onto the substrate to dry the layer and evaporate the solvent. The deposition is followed by annealing at approximately 100°C to evaporate any remaining traces of solvent. An additional annealing step at a higher temperature (340°C) may be required to remove any remaining traces of additive (if one has been added).

[0098] Perovskite can be formed in several ways: in the form of nanocrystals, single crystals, polycrystals.

[0099] Finally, the hole-carrying layer (HTL) is deposited on top of this layer (over the perovskite). This layer can be composed of spiro-OMeTAD, PTAA, NiOx, CuSCN, Cul, or other materials. Its band alignment must allow the passage of holes and the blocking of electrons, as well as good hole mobility. Following its deposition, the layer is annealed at approximately one hundred degrees Celsius.

[0100] The method for depositing the different layers between the electrodes may be identical or different. For industrialization purposes, it may be preferable to use the same method. Several methods may be considered, namely: - Deposition by "spin coating", or by - “Blade coating”, or by - “Slot-die coating”, or - By inkjet or "Inkjet printing"

[0101] Each of these methods, after calibration, makes it possible to obtain layers of specific depths.

[0102] These deposition methods are generally carried out in a controlled environment (in a "glove box"), but can also be carried out in a dry room (with a humidity level around 20%) or, more rarely, in the open air.

[0103] Finally, a last contact layer (“top”) is added by thermal evaporation. This is generally a good conductor of metal (such as gold, silver or copper). The cell can then be encapsulated.

[0104] Figure 7 schematically illustrates the classic NIP architecture of a photovoltaic cell with layers of the aforementioned type. The glass (base substrate) is located at the bottom (below the TCO). During operation, the solar cell receives light from the bottom of the architecture shown here.

[0105] An "inverted" pin type architecture can also be achieved, starting with a glass + ITO (indium-doped tin oxide) substrate, then depositing the hole transport, absorber and electron transport layers according to the same methods mentioned previously.

[0106] A second possible application is the manufacture of a tandem cell combining this perovskite material and a silicon layer.

[0107] The architecture used is that of a tandem cell known as two-terminal, where there is electrical coupling between the two absorbers.

[0108] The "low" cell can be purchased pre-assembled. This can be made of silicon, but other materials can also be considered (such as CIGS materials). The silicon absorber layer is generally a few hundred micrometers thick. The deposition surface is cleaned with ethanol by spin coating as well as by ultraviolet radiation. A transparent recombination layer (generally IZO, indium-doped zinc oxide) is deposited on top. The HTL, perovskite, and ETL layers are added subsequently, following the procedure detailed previously for a single-junction cell. A contact is added, and finally an anti-reflective layer.

[0109] The synthesis of a four-terminal cell can also be carried out following the same procedure as before: each cell is fabricated independently, and the two tandems are then assembled. The contacts between the two absorber layers added (compared to the 2-terminal model) must be transparent to allow the operation of the two independent cells (silicon and perovskite, respectively).

[0110] The alignment of the lattice parameter between the perovskite and the layers at its interface is also considered a selection criterion. The average lattice parameters in the cubic perovskites A(BB')(XX')3 are shown in Figures 8a and 8b. These values ​​in Figures 8a and 8b are to be compared with the lattice parameter value of silicon, which is 0.543 nm. Furthermore, a lattice deviation tolerance, for example, of 2% (without a buffer layer) remains acceptable for obtaining a perovskite / silicon layer stack with a sufficiently limited number of dislocations so as not to impair the photovoltaic properties of the cell.

[0111] More generally, this description is not limited to the embodiments described above by way of example, but extends to other variants, particularly with regard to manufacturing techniques for electronic devices (especially photovoltaic devices) known as "thin-film" devices. The materials of at least one of these thin films are chosen from the family of inorganic perovskites with general stoichiometric composition ABX3 where: - A is one element among Cs and Rb, - B is an element or mixture of elements from Ge, Sn and Pb, and - X is an element or mixture of elements from Cl, Br and I, as long as their Goldschmidt tolerance factor t is well between 0.87 and 0.93.

Claims

Demands

1. Electronic device comprising at least one layer made of a material comprising an inorganic perovskite: * of stoichiometric composition ABX3 where: - A is an element from Cs and Rb, - B is an element or mixture of elements from Ge, Sn and Pb, and - X is an element or mixture of elements from Cl, Br and I, * and having a Goldschmidt factor t between 0.87 and 0.

93.

2. Device according to claim 1, wherein the material is at least partially crystalline according to a crystallographic structure of the type Pm?>m.

3. Device according to any one of the preceding claims, wherein the inorganic perovskite has a stoichiometric composition A(BzB'i zXXyX'iy)3 where: - z and y are between 0 and 1, - A is an element from Cs and Rb, - B and B' are distinct elements from Ge, Sn and Pb, and - X and X' are distinct elements from Cl, Br and I.

4. A device according to any one of the preceding claims, wherein the material is devoid of an unstable vibration mode and is selected from the compositions: CsGeo.25Pbo.75(Bro.5Clo.5)3, CsPbo.125Sno.s75k, CsPbo.25Sno.75I3, CsPbo.375Sno.625(Bro.25Clo.75)3, CsPbo.5Sno.5(Bro.25Clo.75)3, CsPbo.5Sno.5Brs, CsPbo.5Sno.5Ch, CsPb0.625Sn0.375Cl3, and CsPbo.75Sno.25Cl

5. 3- Device according to any one of the preceding claims, the device being of the photovoltaic type, in which said layer has photovoltaic properties and the material: - is devoid of unstable vibration mode, - has a band gap between 1.4eV and 2.3eV, and - is selected from the compositions CsPbo.375Sno.625(Bro.25Clo.75)3, CsPbo.5Sno.5(Bro.25Clo.75)3, and CsPbo.5Sno.5Brs.

6. Device according to any one of claims 1 to 3, the device being of the photovoltaic type and having a tandem or multi-junction cell comprising a silicon layer, in which said layer has photovoltaic properties with a band gap between 1.4eV and 2.3eV, and inorganic perovskite has a composition of the type Rb(Pbx Sni_x)Br3, with x between 0 and 0.7, with a mesh size close to that of silicon for depositing the silicon and perovskite layers on top of each other.

7. Device according to any one of claims 1 to 4, the device being of the photovoltaic type, in which said layer is a passivation layer and the material: - is devoid of unstable vibration mode, - has a band gap greater than 2.3eV, and - is selected from the compositions CsGeo.25Pbo.75(Bro.5Clo.5)3, CsPb05 Sno.sCh, CsPbo.625Sno.375Cl3 and CsPbo.75Sno.25Ch.

8. A device according to any one of claims 1 to 4, the device being of the supercapacitor type, wherein the material is devoid of unstable vibrational modes and is selected, for supercapacitive effects, from the compositions: CsPbo.25Pbo.75(Bro.5Clo.5)3, CsPbo.125Sno.875l3, CsPbo.25Sno.75k, and CsPbo.5SnoXBro^sClojsh-

9. A computer-based method for estimating the stability of an inorganic perovskite of stoichiometric composition ABX3 where: - A is at least one element from Li, Na, K, Rb, and Cs, - B is at least one element or mixture of elements from Ge, Sn, and Pb, and - X is at least one element or mixture of elements from Cl, Br, and I. The method comprises: - Considering a supercell of dimensions that are multiples of an elementary unit cell size of a crystal of said inorganic perovskite, - For each element A and for different mixtures of cations B and different mixtures of anions X, testing phonon modes characterizing vibrations in the supercell to identify vibrational modes with unstable amplitude, and determining for each composition a number of modes with unstable amplitude, - Selecting at least one composition whose number of vibrational modes with unstable amplitude is less than a threshold.With a view to using the selected composition for the manufacture of an electronic device comprising at least one layer made of an inorganic perovskite having the selected composition.

10. A method according to claim 9, comprising: - Selecting a plurality of compositions whose number of unstable amplitude vibration modes is less than a threshold, and - Estimating at least one band gap of each composition to select at least one composition compatible with a targeted photovoltaic application, with a view to using the selected compatible composition to manufacture a photovoltaic type electronic device and comprising at least one layer made in an inorganic perovskite having said compatible composition.

11. Computer program comprising instructions to implement the method according to any one of claims 9 and 10, when said instructions are implemented by a processor.