Technical Field of the Invention

The invention relates to a method of forming a perovskite and photovoltaic elements comprising

perovskites .

Background of the Invention

Lead-based perovskite solar cells have gained research interest due to the rapid rise in power

conversion efficiency (PCE) from 3.8 % in 2009 to 22.1 % in 2016. Many of the improvements have been achieved using methylammonium lead iodide (MAPbls), formamidinium lead iodide (FAPbls) or a mixture of both with small amounts of bromine (Br) and caesium (Cs) or rubidium (Rb) cations.

Despite the promise of lead-based perovskite solar cells, challenges associated with long term thermal stability remain for these perovskites. The toxicity of lead in the perovskite solar cells can also present environment and manufacturing and disposal costs that can prevent the widespread use of lead-based perovskite solar cells. Further, passivation of lead-based perovskites is often required to improve the PCE, but passivation requires additional steps during manufacture, adding to the costs of using lead-based perovskites in solar cells.

Summary of the Invention

A first aspect of the invention provides a method of forming a perovskite having a formula of ABX3, comprising: providing a film of precursor solution for forming the perovskite, the solution comprising a first salt for forming component A, and a mixture of a second salt and a dopant salt for forming component B, wherein component X is formed from anions of the salts; and

heating the film of precursor solution to form the perovskite and to convert the dopant salt to a metal of the dopant salt,

wherein a molar ratio of [the second salt] : [the dopant salt] in the precursor solution is such that the perovskite has a surface enriched with the metal of the dopant salt compared to a bulk of the perovskite.

The term "metal of the dopant salt", "metal of the first salt" and "metal of the second salt" refers to the element that is formed from the dopant/first/second salt is no longer in free in solution in its ionic form and instead is incorporated into the perovskite in a crystal lattice and/or in its metallic (i.e. M°) form.

The molar ratio of [the second salt] : [the dopant salt] in the precursor solution may range from about 99:1 to about 85:15.

A molar ratio of [a metal of the second salt] : [the metal of the dopant salt] at the surface may range from about 95:5 to about 30:70. The second salt may be a divalent species . In some embodiments the second salt is a Pb ²⁺ . The second salt may be Sn ²⁺ .

The dopant salt may include a salt of a group 2 element, or any salt having the properties of a group 2 salt, or a bivalent (divalent) salt. The dopant salt can include Mg ²⁺ , Ba ²⁺ , Sr ²⁺ , Ca ²⁺ , Zn ²⁺ , Mn ²⁺ , Co ²⁺ , Ge ²⁺ _, Pd ²⁺ and/or Sn ²⁺ . For example, a salt having the properties of a group 2 salt may include a salt having a similar atomic radii of a group 2 metal salt, such as Sn ²⁺ . A combination of dopant salts may be used. In some embodiments the dopant salt is a trivalent species .

In an embodiment, a molar ratio of [Pb ²⁺ ] _: [Sr ²⁺ ] in the precursor solution ranges from about 99:1 to about 95:5, or a molar ratio of [Pb ²⁺ ] _: [Ca ²⁺ ] in the precursor solution ranges from about 98:2 to 85:15. In an embodiment, a molar ratio of [Pb ²⁺ ] _: [Sr ²⁺ ] in the precursor solution is 98:2. In an embodiment, a molar ratio of [Pb ²⁺ ] _: [Ca ²⁺ ] in the precursor solution is 95:5.

The first salt may include a salt from a group 1 element or any salt having the properties of a group 1 salt or a monovalent ion. The first salt may include Ag ⁺ , Rb ⁺ , K ⁺ and/or Cs ⁺ . In some embodiments the first salt includes an organic compound. The monovalent ion may be the organic compound. The organic compound may be a methylammonium ion. The anions of the salts may include Cl- , F-, I- and/or Br- .

The precursor solution may comprise a single solvent. The solvent may be a polar organic solvent, such as N- Methyl-2-pyrrolidone (NMP) , dimethylformamide (DMF) or dimethyl sulfoxide (DMSO) . The solvent may comprise a solvent mixture of polar organic solvents . The precursor solution may comprises a mixture of DMF: DMSO. The mixture may range from about 99:1 to about 1:99 v/v DMF: DMSO. In some embodiments, the precursor solution may comprise a mixture of about 2:1 v/v to 4:1 v/v DMF: DMSO. In some embodiments, a DMF: NMP solvent mixture is used as the precursor solution.

The method may further comprise forming the film of precursor solution on a substrate. The precursor solution may be formed via spin coating. The film of the precursor solution may be heated to a temperature ranging from about 100°C to about 300°C. This can help to release excess solvent and form the perovskite. In an embodiment, when the dopant is Sr ²⁺ , the film of the precursor solution may be heated to 100°C. In an embodiment, when the dopant is Ca ²⁺ , the film of the precursor solution may be heated to 300°C .

A second aspect provides a perovskite formed using the method of the first aspect.

A third aspect provides a perovskite having a formula of ABX3 for use in a photovoltaic element, comprising: a first component as component A, a mixture of a second metal and a dopant metal as component B, and one or more halides as component X;

wherein a surface of the perovskite is enriched with the dopant metal compared to a bulk of the perovskite .

The dopant metal may be in its oxide form at the surface. A molar ratio of [the second metal] : [the dopant metal] at the surface may range from about 95:5 to about 30:70. A molar ratio of [the second metal] : [the dopant metal] in the bulk of the perovskite may range from about 96:4 to about 85:15.

The first component metal may be a metal including a group 1 element, or an element capable of having the properties of a group 1 element, or an element capable for forming a monovalent ion, or an organic compound, such as Ag, Rb, K, Cs and/or a methylammonium species.

The dopant metal may include a group 2 element or any element having group 2 properties or a metal capable of forming a bivalent (divalent) ion, including Sn, Mg, Ba,

Sr, Ca and/or Eu.

The one or more halides may include F, Cl, I and/or

Br .

The second metal may be Pb.

A formula of the perovskite may be CsPbi-xSrxl2Br or CsPbi-xCaxl3. A formulate of the perovskite may be MAPbi- xYxl3, where MA is a methylammonium species and Y is Ca or

Sr. A forth aspect provides a photovoltaic device

comprising the perovskite of the third aspect.

A fifth aspect provides a photovoltaic element comprising :

a substrate; and

a first layer comprising a perovskite having a formulate ABX3, wherein component A comprises a first component, component B comprises a mixture of a second metal and a dopant metal, and component X comprises one or more halides;

wherein a surface of the perovskite is enriched with the dopant metal compared to a bulk of the perovskite .

The perovskite may be otherwise as defined according to the third aspect. The surface may be enriched with the dopant metal . Some of the dopant metal present at the surface may be in its oxide form. The surface enriched by the dopant metal may be configured to provide a

passivating effect for the first layer. The photovoltaic element may further comprise an anti-reflective layer. The anti-reflective layer may be MgF2. The photovoltaic element may have a PCE ranging from about 10% to about 13.5%.

A sixth aspect provides a method of forming a photovoltaic device, comprising:

providing a substrate; and

depositing a first solar cell structure on the substrate, the first solar cell structure comprising a perovskite having a formulate ABX3, wherein

component A comprises a first component, component B comprises a mixture of a second metal and a dopant metal, and component X comprises one or more halides; wherein a surface of the perovskite is enriched with the dopant metal compared to a bulk of the perovskite.

Depositing the first layer may comprises the method of the first aspect.

Brief Description of the Drawings

The embodiments of the invention will now be

described, by way of example, with reference to the accompanying drawings in which:

Figure 1 shows X-ray diffraction patterns of

perovskite films made from different CsPbi- _xC a _xl 3 precursors where x is the molar ratio of Ca ²⁺ in the precursor solution. * in Figure 1 denotes Pbl2 impurity peak and # denotes FTO peak.

Figure 2 shows scanning electron microscopy images of CsPbi- _xC a _xl 3 perovskite films where x is the molar ratio of Ca ²⁺ in the precursor solution (a) CsPbl3, (b)

CsPb0.98Ca0.02I3, (c) CsPb0.95Ca0.05I3, (d) CsPb0.93Ca0.07I3, (e)

CsPbo.9Cao.1I3, (f) CsPbo.85Cao.15I3. Scale bar is 5 pm.

Figure 3 shows 3-dimensional atomic force microscopy (AFM) images of the implied CsPbi- _x Ca _x l ₃ perovskite films where x is the molar ratio of Ca ²⁺ in the precursor solution (a) CsPbl3, (b) CsPbo. ₉ sCao.02I ₃ , (c) CsPbo. ₉₅ Cao.05I ₃ , and (d) CsPbo. ₉₃ Cao.07I ₃ .

Figure 4 shows size distributions of complexes in CsPbi- _xC a _xl 3 perovskite precursor solution obtained from dynamic light scattering (DLS) measurements where x is the molar ratio of Ca ²⁺ in the precursor solution.

Figure 5 shows XPS spectra of (a) Ca 2p, (b) Pb 4f,

(c) Cs 3d, (d) O Is, and (e) C Is at the surface of perovskite films fabricated using CsPbi- _xC a _xl 3 precursor solutions . Figure 6 shows (a) absorbance spectra and (b) time- resolved PL (tr-PL) decays of perovskite films fabricated using CsPbi- _x Ca _x l3 precursor solutions.

Figure 7 shows photovoltaic parameters (with standard deviations) of perovskite devices as a function of Ca ²⁺ substitution in the perovskite precursor solution.

Figure 8 shows (a) light J-V characteristics under reverse scan and (b) EQE spectra of champion CsPbi- _x Ca _x l3 devices where x is the molar ratio in the precursor, (c) Light J-V characteristics under reverse scan of the champion device using 5% Ca ²⁺ in precursor before and after MgF2 anti-reflection layer, (d) Stabilised output measured at the maximum power point (0.8V) for the best performing device with MgF2.

Figure 9. Cross-sectional SEM image of a typical device structure FT0/c-Ti0 ₂ /mp-Ti0 ₂ /CsPbl ₃ /P3HT/Au; Scale Bar is 500 nm.

Figure 10. 3-dimensional AFM images of the implied CsPbi- _x Ca _x l3 perovskite film where x is

the molar ratio of Ca ²⁺ in the precursor solution (a)

CsPbI ₃ , (b) CsPb0.98Ca0.02I3, (c) CsPb0.95Ca0.05I3, and (d)

CsPbo.9 ₃ Cao.071 ₃ -

Figure 11. Top-view AFM images of implied CsPbi- _x Ca _x l3 perovskite films where x is the

molar ratio of Ca in the precursor solution (a) CsPbl3, (b) CsPbo.98Ca ₀ .02l ₃ , (c) CsPbo . gsCao .05I3 , and (d) CsPbo.9 ₃ Ca ₀ .07I3 ·

Figure 12. XPS spectra for (a) Ca 2p, (b) Cs 3d, (c) C

Is, and (d) O Is 50nm into the

perovskite films fabricated using CsPbi- _x Ca _x l3 precursor solution .

Figure 13. Ab-initio simulation of the (a, c)

electronic band-structure and (b, d) density of states for a 2x2x2 CsPbo.875Cao.i25l3 (x=0.125) and CsPbl ₃ supercell, respectively, produced using the projector augmented wave (PAW) method.

Figure 14. Distributions of reverse scan PCE of 26 perovskite devices fabricated using 5% Ca ²⁺ substitution in the precursor solution.

Figure 15. Light J-V characteristics of champion Cs perovskite devices using CsPb _0.9 sCa _0.02 I ₃ precursor under forward and reverse scans.

Figure 16. (a) Normalized PCE for encapsulated 5% Ca and 0% Ca CsPbi- _x Ca _x l3 devices (x is the molar ratio in the precursor) as a function of storage time in the dark at room temperature with relative humidity between 50-70%. Normalised Voc, Jsc and FF of the (b) 5% Ca and

(c) 0% Ca device.

Figure 17. Relative humidity of ambient in which perovskite devices were stored.

Figure 18 shows (a) XRD patterns of CsPblABr films and (b) light J-V characteristics of CsPbI2Br devices using different annealing temperature.

Figure 19 shows (a) XRD patterns of low-temperature- processed CsPbi- _x Sr _x l2Br films and SEM images of (b)

CsPbI ₂ Br , (c) CsPbo . ggSro . oil2Br , (d) CsPb ₀ .98Sr ₀ .02l2Br, (e)

CsPbo.97Sro .03l2Br, and (f) CsPbo. gsSro . osl2Br . The inset SEM images are taken in the darker region.

Figure 20 shows XPS spectra for (a) Pb 4f and Sr 3d and (b) I 3d for the CsPbi- _x Sr _x l2Br films.

Figure 21 shows (a) absorbance spectra and (b) time- resolved PL (tr-PL) decay profiles of CsPbi- _x Sr _x l2Br films.

Figure 22 shows (a) light J-V characteristics under reverse scan and (b) EQE spectra and (c) photovoltaic parameters (with standard deviations) of FT0/c-Ti0 ₂ /mp- Ti0 ₂ /CsPbi- _x Sr _x l ₂ Br/P3HT/Au devices as a function of Sr ²⁺ concentration in the perovskite. Figure 23 shows (a) light J-V characteristics and electrical parameters under different scan conditions (red: 0.1 V/s; black: 1 V/s) for the CsPbo.9sSro.o2l2Br champion device, (b) photocurrent density (black) and PCE (red) as a function of time for the champion

CsPbo.98Sro .02l2Br solar cell held at a bias of 0.81 V, (c) normalized Jsc, Voc, FF, and PCE for an encapsulated

CsPbo.98Sro .02l2Br (initial efficiency: 9.5%) perovskite solar cell as a function of storage time, and (d) light J-V characteristics of CsPbo.9sSro.o2l2Br (black) and CsPblhBr (red) cells before and after heat treatment at 100 °C for lh. Solid line: before heat treatment; dashed line: after heat treatment .

Figure 24 shows photovoltaic parameters under reverse scan of FT0/c-Ti0 ₂ /mp-Ti0 ₂ /CsPbl ₂ Br/P3HT/Au devices at two annealing temperatures.

Figure 25 shows SEM top view images of CsPblhBr films annealed at (a) 100 °C and (b) 310 °C. Scale bar is 2pm.

Figure 26 shows XRD patterns of CsPbi- _xS r _xl 2Br films annealed at 310 °C for 10 minutes.

Figure 27 shows top view SEM image of CsPbo.9sSro.o2l2Br film and the corresponding energy dispersive spectroscopy (EDS) elemental mapping of Cs (red) , Pb (blue) , Sr

(green) , I (aqua) and Br (purple) . Scale bar is 10pm.

Figure 28 shows SEM images of (a) CsPbo. ggSro.oil2Br, (c) CsPbo.97Sro .03l2Br films and corresponding SEM images under backscattered mode of the same (b) CsPbo. _ggS ro.oil2Br (d)

CsPbo.97Sro .03l2Br films. Red scale bar is 5 pm.

Figure 29 shows XPS spectra of CsPbi- _xS r _xl 2Br films for Sr3p at the surface (black curve) and at 10 nm depth (red curve) . The peak shift from 269.5 eV (lOnm depth) to 268.5 eV at the surface indicates Sr oxide might have formed on the surface. Figure 30 shows steady-state PL spectra of CsPbi- _x Sr _x l2Br films.

Figure 31 shows distributions of reverse scan PCE of 25 CsPbo.98Sro.o2l2Br devices fabricated in this work.

Figure 32 shows normalized Jsc, Voc, FF, and PCE for CsPbå2Br perovskite solar cells with encapsulation as a function of storage time (initial efficiency is 6.2 %).

Detailed Description of Embodiments

A first aspect of the invention provides a method of forming a perovskite having a formula of ABX3, comprising:

providing a precursor solution comprising a first salt for component A, and a mixture of a second salt and a dopant salt for component B, wherein anions of the salts provides component X; and

heating the precursor solution to form the perovskite ,

wherein a molar ratio of [the second salt] : [the dopant salt] in the precursor solution is such that the perovskite has a surface enriched with a metal of the dopant salt compared to a bulk of the perovskite. By having a surface enriched with the metal of the dopant salt, a passivation layer at the surface of the perovskite may be formed. When incorporated into a photovoltaic element, the passivation layer may help to prevent carrier recombination and improve the efficiency of the photovoltaic element. In some embodiments, the metal of the dopant salt is in its oxide form at the surface. For example, if Ca ²⁺ is used as the dopant salt in the precursor solution, CaO may be present as the oxide at the surface of the perovskite. The oxide layer may be formed by reaction of the dopant salt in the precursor solution during synthesis of the perovskite. The oxide layer may also be formed by reaction of the metal of the dopant salt reacting with oxygen at the surface of the perovskite .

If the molar ratio of [the second salt] : [the dopant salt] in the precursor solution is too low then there may be an insufficient concentration of dopant salt to form the enriched surface. Conversely, if the concentration of the dopant salt is too high, the enriched surface may start to act as an insulator. In the latter case, the insulating layer would decrease the efficiency of a photovoltaic element. The concentration of the dopant may range from about 99:1 to about 85:15. The ratio of [the second salt] : [the dopant salt] can vary depending on the dopant salt. In some embodiments more than one dopant salt can be used, where the total concentration of the dopant salts is used to determine the ratio of [the second salt] : [the dopant salt] .

Because the surface is enriched with the metal of the dopant salt, the molar ratio of [a metal of the second salt] : [the metal of the dopant salt] at the surface will differ from the ratio of [the second salt] : [the dopant salt] used in the precursor solution. In some embodiments, the molar ratio of [a metal of the second salt] : [the metal of the dopant salt] at the surface ranges from about 95:5 to about 30:70. Again, the specific ratio will be

determined by the second salt and the types of dopant salts, their relative ratios, and the perovskite to be formed .

The dopant salt may include a salt of a group 2 element or any salt having the properties of a metal salt from a group 2 element or a bivalent salt. The dopant salt may include Mg ²⁺ , Ba ²⁺ Ca ²⁺ Zn ²⁺ , Mn ²⁺ , Co ²⁺ , Ge ²⁺ _, Pd ²⁺ _, Sn ²⁺ _, and/or Eu ²⁺ . A combination of one or more dopant salts may be used. A valency of the dopant salt may be selected to be the same as a valency of the second salt.

The second salt may be Sn ²⁺ or Pb ²⁺ . By incorporating a dopant salt into the perovskite, the amount of e.g. Pb required to form the perovskite can be reduce. This can be advantageous since reducing the amount of Pb helps to reduce the environmental impacts associate with forming, using, and disposing the perovskite.

In some embodiments Sr ²⁺ or Ca ²⁺ is used as the dopant salt. When Pb ²⁺ is used as the second salt, a molar ratio of [Pb ²⁺ ] _: [Sr ²⁺ ] in the precursor solution may range from about 99:1 to about 95:5. A molar ratio of [Pb ²⁺ ] _: [Ca ²⁺ ] in the precursor solution may range from about 98:2 to 85:15. Molar ratios outside of the 99:1 to about 95:5 for

[Pb ²⁺ ] _: [Sr ²⁺ ] and 98:2 to 85:15 for [Pb ²⁺ ] _: [Ca ²⁺ ] tend to form perovskites with little to no enriched surface or a surface that becomes too insulating. Without being bound by theory, it is thought that surface enrichment with the metal of the dopant salt occurs due to kinetic and thermodynamic effects during perovskite formation, with the perovskite trying to achieve its lowest energy state. Because surface enrichment occurs in situ during formation of the perovskite, may be no need for further surface modification to provide favourably perovskite surface properties. Put another way, the perovskite may self- assemble to form the surface passivation layer. This may be advantageous as it can help to remove the need to perform post-modification steps to afford a passivation layer. In an embodiment, a molar ratio of [Pb ²⁺ ] _: [Sr ²⁺ ] in the precursor solution is 98:2. In an embodiment, a molar ratio of [Pb ²⁺ ] _: [Ca ²⁺ ] in the precursor solution is 95:5.

The first salt may include a salt from a group 1 element or any salt having the properties of a metal salt from a group 1 element or a monovalent ion. The first salt may include Ag ⁺ , Rb ⁺ , K ⁺ , Cs ⁺ and/or a methylammonium ion. A mixture of salts may be used as the first salt. Changing the first salt may help to improve the stability of the perovskite when heating, for example when annealing.

Changing the first salt may help to reduce the annealing temperature ( s ) required to form the perovskite. For example, Cs ⁺ may be advantageous in some embodiments as it can reduce the annealing temperature to around 100° when Sr ²⁺ is used as the dopant salt. The annealing temperature may be about 300°C when Cs ⁺ is used as the first salt and Ca ²⁺ is used as the dopant salt. When the first component is an inorganic component, such as a group 1 element, the perovskite can be considered an inorganic perovskite.

However, in some embodiments, a methylammonium salt or similar may be used as the first component. When a methylammonium salt or similar is used as the first salt, the perovskite can be considered an organic perovskite.

The disclosure is to be interpreted broadly to include both inorganic and organic perovskites unless context clearly indicates otherwise. The anions of the salts may include Cl-, F-, I- and/or Br- .

The precursor solution may be a solvent or a solvent mixture having two or more solvents. In an embodiment, the precursor solution comprises a mixture of DMF:DMSO. Some embodiment use a mixture of DMF:NMP as the solvent for the precursor solution. The mixture may range from about 2:1 v/v to 4:1 v/v DMF:DMSO. In some embodiments a 4:1 DMF:NMP v/v mixture is used. The specific ratio of DMF:DMSO and/or DMF:NMP may be determined by the first, second and dopant salts used to form the perovskite. For example, the solubility of one of the salts in the precursor solution may determine the type(s) of solvent used. The method may further comprise applying the

precursor solution to a substrate. The substrate may be a substrate used to form a photovoltaic cell. For example, the substrate may be mp-Ti0 ₂ . The precursor solution may be applied to the substrate via spin coating. Once applied to the substrate, the precursor solution would then be heated (e.g. annealed) to form the perovskite. Heating may help to accelerate evaporation of the solvent (s) used to form the perovskite.

The presence of the dopant salt may help to reduce colloidal cluster size during the heating step, and this may help to improve film smoothness. This may be

advantageous as smoother films typically exhibit better photovoltaic element properties .

A second aspect provides a perovskite formed using an embodiment of the method as set forth above.

A third aspect provides a perovskite having a formula of ABX ₃ for use in a photovoltaic element, comprising:

a first component for component A, a mixture of a second metal and a dopant metal for component B, and one or more halides for component X;

wherein a surface of the perovskite is enriched with the dopant metal compared to a bulk of the perovskite .

The dopant metal may be in its oxide form at the surface. For example, if Ca or Sr is the dopant metal, the surface of the perovskite may comprise CaO or SrO, respectively. It should be appreciated that not all of the dopant metal at the surface will be in its oxide form, so the surface can comprise a mixture of the dopant metal in its oxide form and other forms.

Because the surface of the perovskite is enriched with the dopant metal, the composition and ratio of the metals that form the perovskite will vary between the surface and a bulk of the perovskite. For example, in an embodiment, a molar ratio of [the second metal] : [the dopant metal] at the surface ranges from about 95:5 to about 30:70, but a molar ratio of [the second metal] : [the dopant metal] in the bulk of the perovskite ranges from about 96:4 to about 85:15. The specific ratios and relative difference in composition between the surface and the bulk will be determined by the first, second and dopant metals and their relative molar ratios to one another .

The first component may include a metal from a group 1 element or any element having group 1 properties and/or any element that is capable of being incorporated into a perovskite structure in the position of a group 1 element. The first component may include Ag, Rb, K and/or Cs . For example, an element having a similar atomic radii of a group 1 element may be used as the first metal so long as it can be incorporated into the perovskite structure in the same position as a group 1 element such as Cs . A mixture of metals may be used as the first metal. In some embodiments the first component is an organic component such as a methylammonium species. A mixture of organic component (s) and metal (s) can be used in some embodiments.

The dopant metal may include a group 2 element, any element having group 2 properties, and/or any element that is capable of being incorporated into a perovskite structure in the position of a group 2 element. The dopant metal may be Mg, Ca, Sr, Ba, Zn, Mn, Co, Ge _, Pd _{, S} n and/or Eu. The one or more halides may include F, Cl, I and/or Br . The second metal may be Pb. Incorporating the dopant metal helps to reduce the amount of Pb in the perovskite and this can be advantageous as it helps to reduce the environmental impacts of the perovskite, its use and subsequent disposal.

In an embodiment, a formula of the perovskite is CsPbi-xSrxå2Br or CsPbi-xCaxl ₃ . A formulate of the perovskite may also include MAPbi-xYxå3, where Y is the dopant metal and MA is a methylammonium species .

Another embodiment provides a photovoltaic element comprising an embodiment of a perovskite as defined above.

Another embodiment provides a photovoltaic element comprising :

a substrate; and

a first layer comprising a perovskite having a formulate ABX3, wherein component A comprises a first component, component B comprises a mixture of a second metal and a dopant metal, and component X comprises one or more halides;

wherein a surface of the perovskite is enriched with the dopant metal compared to a bulk of the perovskite .

The perovskite may be otherwise as defined as set forth above. The surface may be enriched with the dopant metal. The dopant metal may be in its oxide form at the surface . The surface enriched by the dopant metal may be configured to provide a passivating effect for the first layer. Put another way, the surface enriched by the dopant metal may act as a passivation layer. A passivation layer can help to reduce the occurrence of recombination which can help to improve the efficiency of the photovoltaic element. The photovoltaic element may further comprise an anti-reflective layer. The anti-reflective layer may be MgF2. The anti-reflective layer may help to improve the efficiency of the photovoltaic element. The photovoltaic element may have a PCE ranging from about 10% to about 13.5%. The photovoltaic element may be used in a

multilayer photovoltaic cell, such as a tandem cell.

Embodiments will now be described with reference to the following examples.

1. Ca doped CsPble

1.1 Experimental

1.1.1 Device fabrication.

Patterned FTO-coated glass (Nippon Sheet Glass,

TEC10, 10 QD-l) was cleaned by sonication in deionized water with 2 % Hellmanex, acetone and isopropanol for 20 min, respectively. After drying, the substrate was treated by UV ozone cleaner for 20 min. To form the compact Ti0 ₂ blocking layer (c-Ti0 ₂₎ , a solution of titanium

diisopropoxide bis ( acetylacetonate ) in ethanol was

deposited on the clean substrate by spray pyrolysis at 450 °C and the substrate was subsequently annealed on a hot plate at 400 °C. After cooling, 150 nm mp-Ti0 ₂ layer was deposited by spin coating Dyesol 30 NR-T paste with a 1:6 (by weight) dilution in ethanol for 12 s at 4000 rpm.

After spin coating, the substrate was dried at 100°C for 10 min and then annealed at 450 °C for 30 min. Prior to deposition of perovskite film, the substrate was cleaned by a UVO cleaner for another 20 min and was then

transferred to a N2 filled glovebox. The Csl concentration in the CsPbi- _xC a _xl 3 perovskite precursor solution is 1 M, which was prepared by dissolving Csl (Alfa Aesar) , Pbl2 (Alfa Aesar) and Cal2 (Sigma-Aldrich) , stoichiometrically in a mixed solvent of dimethylformamide (DMF) (Sigma- Aldrich) and DMSO (Sigma-Aldrich) with a volume ratio of 4:1. The perovskite film was deposited by gas-quenching method (gas assisted spin coating) . After the perovskite precursor solution was spread on the mesoporous Ti02 layer, the substrate was spun firstly at 1000 rpm for 10 s and then 4000 rpm for 30 s. N2 stream (5.5 bar) was blown over the spinning substrate for 15 s after spinning at 4000 rpm for 5 s. The perovskite film was then annealed at 300 °C for 10 min on a hot plate. The hole transporting solution was prepared by dissolving 10 mg/ml Poly ( 3-hexylthiophene ) (P3HT) (Sigma-Aldrich) in chlorobenzene (Sigma Aldrich) and was deposited on perovskite layer by spin coating at 3000 rpm for 30 s. 100 nm of gold was then thermally evaporated on P3HT to form the top electrode. Finally, the device was encapsulated using polyisobutene (PIB) . For anti-reflection coating, 75 nm of magnesium fluoride (MgF2) was thermally evaporated on the glass side.

1.1.2. Sample preparation for AFM measurement.

The replicate of perovskite surface were fabricated using a one-step soft lithographic replication process, similar to that reported in J. Mater. Chem. A 2017, 5,

969. The silicone elastomer base and its curing agent (Sylgard 184) were purchased from Dow Corning. The

reagents were mixed at a ratio 10:1 (w/w) and degassed for about an hour to prepare the cross-linking silicone elastomer which were then poured onto the perovskite film. After curing for 1 day, the implied perovskite films were obtained by peeling the silicone gel off the perovskite surface .

1.1.3. Characterization.

Dynamic light scattering (DLS) measurements were carried out by Zetasizer Nano ZS. The measurement was done with a quartz cuvette at room temperature. Top view and cross-sectional SEM images were obtained using a field emission SEM (NanoSEM 230) in vacuum. X-ray diffraction (XRD) patterns were measured using a PANalytical Xpert Materials Research diffractometer system with a Cu Ka radiation source (l = 0.1541 nm) at 45 kV and 40 mA in air. The sample was sealed in a dome air-sealed holder during the measurement. AFM measurements were performed on implied perovskite films to evaluate surface roughness using Bruker Dimension ICON SPM. X-ray photoelectron spectroscopy (XPS) study was carried using the Thermo ESCALAB250Xi X-ray photoelectron spectrometer. The optical reflection and transmission spectra were measured using Cary spectrophotometer and the test samples are

encapsulated with PIB. The time-resolve PL (tr-PL) decay traces were measured by the Microtime-200 (PicoQuant) with 470 nm excitation and detection through a 620/40 nm band pass filter. The current density-voltage (J-V)

measurements were performed using a solar cell I-V testing system from PV Measurements Inc. under an illumination power of 100 mW cm ^-2 with a 0.159 cm ² aperture and a scan rate at 0.02 Vs ^-1 . All devices have been light soaked for 30 minutes before the J-V measurement. The external quantum efficiency (EQE) was measured using the PV

Measurement QXE7 spectral response system with

monochromatic light from a xenon arc lamp. Both J-V and EQE measurements were undertaken on encapsulated devices at room temperature in ambient condition.

1.2. Discussion

To investigate the effect of partial substitution of Pb ²⁺ with Ca ²⁺ on the CsPbl3 precursor solution and on the crystallinity and morphology of films subsequently formed, X-ray diffraction (XRD) measurements were carried out on FT0/c-Ti0 ₂ /mp-Ti0 ₂ /CsPbi- _xC a _xl3 test structures where x is the molar ratio of Ca in the precursor. Figure 1 shows the XRD patterns of the CsPbi- _x Ca _x l3 films . They all show typical cubic perovskite phase with the main XRD peaks at 14.5°, 20.8°, and 29.1° corresponding to (100), (110) and

(200) plane, respectively, in good agreement with those previously reported [ACS Energy Lett. 2017, 2, 1901].

However, the reference CsPbl3 film (0% Ca) has a small peak at 12.6° corresponding to a Pbl2 which is typical when there is an incomplete conversion of the precursor into perovskite phase. As the molar ratio of Ca ²⁺ increases, the Pbl2 peak disappears and the crystallinity of perovskite phase improves. According to a previous density functional theory study, XRD peaks were expected to lower 2 theta angle when Pb is partially substituted with alkaline-earth metal such as Mg, Sr, Ba and Ca . However, in the

measurements, there is negligible angle shift, or the shift is too small that is below the measurement

resolution .

Scanning electron microscopy (SEM) surface images of CsPbi- _x Ca _x l3 perovskite films (where x is the molar ratio of Ca in the precursor) are shown in Figure 2. The reference CsPbl3 (0% Ca) film is composed of densely packed small grains, with average domain size of 1 pm. With 2% Ca ²⁺ in the precursor, the perovskite film shows less coverage. However, at 5% Ca ²⁺ , the small grains combine into larger grains with size up to several pm resulting in better coverage. Cross-sectional SEM images of the solar devices made from CsPbl ₃ , CsPbo.9sCao.02I3, and CsPbo.95Cao.05I3 were also taken. Results are shown in Figure 9 revealing dramatic reduction in roughness as the molar ratio of Ca ²⁺ increases to 5%. To measure the surface roughness of the films, atomic force microscopy (AFM) measurement was performed on implied perovskite films which were silicone gel replicates of the CsPbi- _x Ca _x l3 perovskite films as the replicates do not degrade that the original perovskite films otherwise would during the AFM measurements . Details of the replication process can be found in the

Experimental section. Top view AFM images are shown in Figure 10 while the three dimensional AFM images are shown in Figure 3. Table 1 summarises the average roughness (Ra) of the films. Results show that the reference CsPbl3 (0% Ca) film has the rougher surface with an Ra of 67. lnm and the benefits brought about by Ca is most evident when Ca ²⁺ molar ratio is 5% as seen in the CsPb0.95Ca0.05I3 film with an Ra =25.0 nm.

Table 1. Average roughness of the implied perovskite film made from precursors with different Ca ²⁺ molar ratios

0% Ca 2% Ca 5% Ca 7% Ca

Ra (nm) 67.1 90.7 2530 32.5

To understand the reason for these morphology changes, we investigate the effect of partial substitution of Pb ²⁺ with Ca ²⁺ on the CsPbi- _xC a _xl 3 precursor solution using dynamic light scattering (DLS) measurements.

Treating the perovskite precursor solutions as colloidal dispersions, the size of the colloids can be determined from the DLS measurements and results for solutions with different molar ratios of Ca ²⁺ are shown in Figure 4. It can be seen that the radius of the colloids in the reference sample (0% Ca) is predominantly 1.7 pm. As the molar ratio of Ca increases, the colloidal particle size is reduced to 1 pm. This could be due to the Ca ²⁺ being a hard Lewis acid similar to hydrogen, H ⁺ , compared to Pb ²⁺ , which is a soft acid, according to the Lewis acidity. The Ca ²⁺ might work similarly to hydrohalic acids, triggering the dissolution of the colloids in the precursor solution reducing the size of the colloids. The smaller colloid results in a more uniform distribution of the reactants inducing better nucleation and crystallization of the

5 perovskite film which is smoother with larger grain size.

In order to understand the effect of Ca ²⁺ on the surface composition of the perovskite films, an X-ray photoelectron spectroscopy (XPS) study was performed and results are shown in Figure 5. As the molar ratio of Ca 10 increases in the precursor, the presence of Ca ²⁺ at the

surface of the film subsequently formed is evident from the presence of Ca 2p peaks at 347.5 and 351 eV (Figure 5a and its zoomed-in image) . The peak intensity increases as Ca content increases. In addition, the intensities of Pb 15 4f peaks (Figure 5b) and Cs 3d peaks (Figure 5c) reduce

significantly with Ca ²⁺ , especially when Ca ²⁺ > 5%. From the XPS results in Figure 5 and Figure 11, we estimated the Ca/ (Ca+Pb) atomic ratio at the surface and 50nm into the film for the CsPbi- _xC a _xl 3 film where x is the molar ratio of 20 Ca used in the precursor solution. Results are listed in

Table 2.

Table 2. Ca/ (Ca+Pb) atomic ratios of CsPbi- _xC a _xl 3 perovskite films where x is the molar ratio of Ca ²⁺ in the precursor 25 solutions

Ca CsPbla CsPbo.seCa0.02I3 CsPbo.95Cao.05I3 CsPbo.93Cao.07I3 CsPbo.9Cao.1I3

/ (Ca+Pb)

atomic

ratio

Surface N.A N.A 5.20% 32.40% 70.10%0 nm into N.A N.A 4.60% 4.80% 6.60% the film For the CsPbo.93Cao.07I3 film, the Ca atomic ratio is 32% at the surface rather than 7% as anticipated. For the CsPbo.9Cao.1I3 film, the Ca ratio at the surface is 70% rather than the expected 10%. This suggests that the film surface is enriched with Ca ²⁺ after the crystallisation and film formation of perovskite. In addition, the intensities of C Is (Figure 5d) and O Is (Figure 5e) increase with Ca ²⁺ , indicating the presence of an oxide layer with traces of carbon which could be a mixture of CaO and CaC03. CaO could be a result of reaction between Ca ²⁺ and H2O in dimethylformamide (DMF) , while the CaC03 is the result of reaction between CaO and CO2 likely to be from the ambient. As the XRD peaks of CaC03 are located at similar positions to the peaks of CsPbl3, it is hard to distinguish it in the XRD patterns . With appropriate amount of the wider bandgap materials such as CaO (~6.3eV) or CaC03 (7eV) at the surface, recombination is suppressed similar to the passivation effect provide by excess Pbl2 in perovskite. However, excessive amount of Ca ²⁺ (>5%) is detrimental to device performance due to the insulating nature of this oxide layer, which will be discussed later. To

characterise the film further into the bulk, 50 nm was etched away from the surface using Ar ion etching. Results of the XPS performed are shown in Figure 12. The peaks of Ca 2p are present in films when Ca ²⁺ ³5% in the precursor while the Cs 3d peaks remain constant regardless of the Ca ²⁺ content. When Ca ²⁺ < 5% in the precursor, Ca 2p peaks can hardly be detected, indicating that there is

insufficient Ca to be incorporated into the perovskite lattice. For the CsPbo.93Cao.07I3 film, the Ca atomic ratio is around 5%, slightly lower than 7% as anticipated (Table 2) . For the CsPbo.9Cao.1I3 film, the Ca ratio is 7%, again slight lower than the expected 10%. This suggests that the film 50nm into the bulk has slightly less Ca ²⁺ than that in the precursor solution, due to the accumulation of Ca ²⁺ at the surface as shown in previous results.

To investigate the effect of Ca ²⁺ in the precursors on the opto-electronic properties of the perovskite film, reflection and transmission measurements and time-resolved photoluminescence (tr-PL) measurements were carried out on test perovskite film structures. The corresponding absorption curves shown in Figure 6a show blue shifts in the perovskite films from precursors with Ca ²⁺ molar ratio ³ 5%. The bandgap is increased from 1.72 eV to 1.93 eV when Ca ²⁺ molar ratio is increased to 15% in the precursor. We have stimulated the electronic band-structure and density of states for the CsPbo .87sCao .125I3 and CsPbl ₃ supercells using the Projector augmented wave method

(Figure 13) . Bandgap widening is observed for

CsPbo.875Cao.i25l3 compared to CsPbl ₃ which is attributed to the mixing of wider bandgap states introduced by Ca-I bonds into the conduction and valence bands, generally defined by Pb-I bonds, on the substitution of Ca on Pb lattice sites. Wang et al. [Int. 2017, 43, 13101] has also conducted density function theory simulation on CsPbI3 when Pb is substituted by alkaline-metal and also reported an increase in bandgap.

Figure 6b shows the tr-PL decay curves for the perovskite films and Table 3 shows the corresponding carrier lifetime components extracted by fitting the curves with a bi-exponential function. Table 3. Lifetimes extracted from tr-PL decay curves for perovskite films deposited using CsPbl-xCaxI3 precursor solutions .

Lifetime CsPbl ₃ CsPbomsCao CsPbomsCao CsPbo. _93C ao

(ns) .02I3 .05I3 .07I3

Ί GT5 0.93 2T3 GT9

t2 5.1 2.3 8.1 6.0

eff 3.4 1.8 6.6 5.2

Perovskite films made from precursors with 5% and 7% Ca ²⁺ substitution have better effective lifetimes compared to that of the reference CsPbl3 due to the improved film morphology including coverage and smoothness as shown in Figures 2 and 3 and effective passivation provided by the Ca rich oxide layer at the surface. However, if there is insufficient Ca ²⁺ in the precursor, e.g. molar ratio at 2%, there is no apparent benefit as can be seen by the lower effective lifetime (Table 3), poorer coverage of the film (Figure 2b) and rougher surface (Figure 3b and Table 1) . This is because the amount of Ca is not sufficient to be incorporated into the perovskite lattice as shown in the XPS results (Figure 5a and Figure 11) resulting in no bandgap change (Figure 6a) but only a small red-shift as a result of defects levels introduced by the small amount of Ca, similar to what has been observed in

stoichiometrically unbalanced CsPbl2Br.

To confirm the benefits of Ca substitution on device performance, FTO/c-Ti0 ₂ /mp-Ti0 ₂ /CsPbi- _xC a _xl3 /P3HT/Au

photovoltaic devices were fabricated. Figure 7 and Table 4 show the distributions of the PCE and the other parameters of the devices . The parameters of the champion devices are shown in brackets. Using precursors with 5% Ca ²⁺

substitution, the resultant devices give the best

performance with an average PCE of 11% (Figure 14), Figure 16) and a champion PCE of 12.6%. Without Ca substitution, the reference devices have an average PCE of 9.4% and the best PCE of 10.6%. The significant improvement in Ca- incorporated devices comes from improved Voc and FF due to better morphology which in turn results in better

perovskite/HTM interface, higher bandgap and surface passivation. Figure 15 shows light J-V characteristics of champion Cs perovskite devices using CsPbo.98Cao.02I3

precursor under forward and reverse scans . Table 4. Photovoltaic parameters with standard deviations of perovskite devices as a function of Ca ²⁺ concentration in the perovskite precursor solution. Numbers in brackets are champion device parameters .

PCE (%) Jsc (mA/cm ² ) Voc (mV) FF (%)

Reference 9.4 + 0.83 17.4 + 0.47 790 + 64.2 68.7 + 4.36

(10.6) (17.1) (853) (72.6)

2% Ca 9.0 + 0.77 18.1 + 0.83 752 + 35.5 65.4 + 4.07

(9.72) (17.9) (828) (65.7)

5% Ca 11.0 + 0.74 17.1 + 0.38 881 + 27.3 73.0 + 3.52

(12.6) (17.3) (940) (77.7)

7% Ca 9.8 + 0.34 16.1 + 0.15 833 + 49.2 73.4 + 2.29

(10.4) (16.0) (882) (73.9)

10% Ca 8.6 + 1.29 15.1 + 0.57 832 + 84.4 68.2 + 6.54

(10.6) (15.2) (914) (76.0)

15% Ca 6.74 + 0.39 13.6 + 0.57 769 + 32.6 64.5 + 3.35

(7.16) (12.8) (790) (69.8) Cell results confirm that 2% Ca in the precursor is insufficient to bring about any benefits but result in poorer performance (lower FF and lower Voc) due to i) poorer surface coverage and therefore poorer

perovskite/HTM interface, and ii) the absence of surface passivation due to the lack of Ca to enable the formation of Ca rich oxide. The small improvement in JSC could be due to the scattering caused by the rougher perovskite layer. When Ca ²⁺ ³ 5% in the precursor, the Jsc of the corresponding devices decrease due to the enlarged bandgap (Figure 8a) and hence reduced absorption as seen in the measured external quantum efficiencies (EQE) shown in Figure 8b. When Ca ²⁺ substitution is more than 10% in the precursor, the excessive Ca of the subsequent perovskite film and the excessive insulating oxide layer at the perovskite surface inhibit carrier transport between the perovskite and the HTM, leading to reductions in all electrical parameters in the corresponding devices.

However, respectable performance can still be obtained (>6.7% of PCE) when 15% of the Pb is replaced by Ca in the precursor solution.

After applying MgF2 anti-reflection coating, the PCE of the champion cell was 13.5% with a JSC of 17.9 mAcrrr ² ,

Voc of 945 mV, and FF of 80% under reverse scan (Figure 8c) . In addition, a stabilized PCE of 13.3% was achieved (Figure 8d, Figure 16) . These are the best efficiencies and highest fill factor reported for a CsPbl3 perovskite solar cell to-date. The encapsulated device also showed an outstanding stability maintaining 85% of the initial PCE for over two months in ambient conditions (Figure 17) .

This study demonstrates that Ca ²⁺ substitution of lead in CsPbl3 precursor solution produces perovskite film with stabilised CsPbl3 phase and enhanced film quality in multiple aspects. Firstly, Ca ²⁺ reduces the size of the colloids in the precursor solution providing a more uniform distribution of the reactants for nucleation and film crystallisation, leading to larger grains and smoother films. As a result, a better interface between perovskite and HTM is obtained increasing Voc and FF for the corresponding solar devices . Secondly, a Ca rich oxide layer forms at the surface providing a passivating effect, leading to an increase in carrier lifetime. Thirdly, the bandgap is increased with Ca ²⁺ . The amount of the Ca ²⁺ substitution in the precursor solution is optimised in this work. It is found that when the molar ratio of Ca ²⁺ £2%, the amount is insufficient to cause an increase in the bandgap or to form an oxide passivation layer at the surface but results in rougher film with poorer surface coverage and higher defects confirmed by optical, XPS,

SEM, AFM and tr-PL measurements . On the other hand, excessive Ca ²⁺ in the precursor (³10%) causes excessive insulating calcium oxide to be formed inhibiting carrier transport between perovskite and HTM layer. At the optimum concentration (5% Ca ²⁺ in the precursor solution) , the corresponding glass/FTO/compact Ti0 ₂ /mesoporous Ti0 ₂ /CsPbi- _x Ca _x l ₃ /P3HT/Au solar devices achieved an average efficiency of 11.0%. With an anti-reflection coating, the champion cell delivered 13.5% PCE under a reverse scan and a stabilized PCE of 13.3%. Encapsulated device also showed a remarkable stability, maintaining 85% of the initial PCE for over two months. Also, the perovskite film made from 5% Ca ²⁺ substitution in precursor has a lower Pb content and a bandgap of 1.72 eV. The latter makes it as a good candidate as a top cell in a double junction perovs kite/Si tandem device.

A summary of the performance of MgF2/glass/FTO/compact Ti0 ₂ /mesoporous Ti0 ₂ /CsPbi- _x Ca _x l ₃ /P3HT/Au solar cells compared to existing solar cells is given in Table 5. Table 5. Summary of CsPbl3-based devices

0.65 12.2 44 ₃ .5 V0C ^“ >JSC

FT0/c-Ti02/CsPbI3/Spiro/Ag 0.62 11.9 31.3 ₂ . ₃ Jsc ->VoC

ITO /Ca /C60/CsPbl3 /TAPC/TAPC :Mo03/Ag 9.4 ?

1.06 13.8 72 10.5 Voc->Jsc Jsc

FT0/c-Ti02/CsPbl3/P3HT/Au 0.99 13.9 68 9.3 -> Voc

FT0/c-Ti02/CsPbI 3 /Spiro/Au 0.55 0.5 44.2 0.1 V0C ^" >JSC

FT0/c-Ti02/CsPbl3/P3HT/Au 0.71 12.1 67 5.7 ?

1 13.0 68 8.8 Voc ^->J SC

FT0/Sn02/CsPbI 3/Carbon nanotubes /Spiro/Au

⁷ · 8 Stabilized

2. Sr doped CsPbl2Br

5 2.1 Experimental

2.1.1 Device fabrication.

Patterned FTO-coated glass (Nippon Sheet Glass,

TEC10, 10 W III-l) was cleaned by sonication in deionized water with 2% Hellmanex, acetone, and isopropanol for 200 min. After drying, the substrate was treated by UV ozone cleaner for 20 min. To form the compact Ti0 ₂ blocking layer ( _C -TXO2) , a solution of titanium diisopropoxide bis- (acetylacetonate ) in ethanol was deposited on the clean substrates by spray pyrolysis at 450 °C, and the substrate5 was subsequently annealed on a hot plate at 400 °C. After cooling, a 150 nm mp-Ti0 ₂ layer was deposited by spin coating for 12 s for 4000 rpm, using a Dyesol 30 NR-T paste with a 1:6 (by weight) dilution in ethanol. After spin coating, the substrate was dried at 100 °C for 10 min and then annealed at 450 °C for 30 min. Prior to

deposition of the perovskite film, the substrates were cleaned by a UVO cleaner for another 20 min and were then transferred to a N2-filled glovebox. The CsBr

concentration in the perovskite precursor solution is 0.7 M, which were prepared by dissolving CsBr (Alfa Aesar) ,

Pbl2 (Alfa Aesar) , and Srl2 (Sigma-Aldrich)

stoichiometrically in a mixed solvent of DMF (Sigma- Aldrich) and DMSO (Sigma-Aldrich) with a volume ratio of 2:1. The perovskite film was deposited by a gas-quenching method (gas assisted spin coating) . After the perovskite precursor solution was spread on the mesoporous Ti02 layer, the substrate was spun first at 1000 rpm for 10 s and then at 4000 rpm for 30 s. A N2 stream (5.5 bar) was blown over the spinning substrate for 15 s after spinning at 4000 rpm for 5 s. The perovskite film was then annealed at 100 °C for 5 min on a hot plate. The hole transporting precursor solution was prepared by dissolving 10 mg/mL poly (3- hexylthiophene ) (P3HT) (Sigma-Aldrich) in chlorobenzene (Sigma-Aldrich) and was deposited on the perovskite by spin coating at 3000 rpm for 30 s. Gold (100 nm) was then thermally evaporated on the HTM to form the top electrode. Finally, the device was encapsulated using polyisobutene (PIB) .

2.1.2 Characterization

Top-view and cross-sectional SEM images were obtained using a field emission SEM (NanoSEM 230) in vacuum. EDS measurements were carried out by the NanoSEM 230 using a Bruker SDD-EDS detector in vacuum. XRD patterns were measured using a PANalytical Xpert Materials Research diffractometer system with a Cu Ka radiation source (l = 0.1541 nm) at 45 kV and 40 mA in air. The sample used for XRD measurement was ( FTO/CsPbi- _xS r _xl 2Br/poly (methyl methacrylate)). Poly (methyl methacrylate (PMMA) was coated on the film to seal the sample from ambient moisture. The optical reflection and transmission spectra were measured using a Cary spectrophotometer, and the test samples were encapsulated. The current density-voltage (J-V)

measurements were performed using a solar cell I-V testing system from PV Measurements Inc. under an illumination power of 100 mW cm ^-2 with a 0.159 cm ² aperture and a scan rate at either 0.1 or 1 V s ^_1 . All devices were light soaked for 15 min before the J-V measurement, which had an effect of stabilizing the cell performance. The external quantum efficiency (EQE) was measured using the PV

Measurement QXE7 spectral response system with

monochromatic light from a xenon arc lamp. All

measurements were undertaken at room temperature in ambient conditions, and the devices were encapsulated.

2.2 Discussion

CsPbl2Br thin films are obtained by dissolving CsBr and Pbl2 stoichiometrically in a mixed solvent of N,N- dimethylformamide (DMF) and dimethyl sulfoxide (DMSO) . Films are spin-coated and then annealed at 100 or 310 °C on a hot plate to form a 250 nm thick perovskite film. X- ray diffraction (XRD) patterns of low-temperature- and high-temperature-annealed CsPbå2Br thin films are shown in Figure 18a. The Bragg peaks at 2Q = 14.6 and 29.5° show that both films are well-oriented in the cubic (100) direction. CsPbå2Br annealed at low temperature has a broader fwhm than the film annealed at high temperature. This indicates that the crystal size of the low temperature- processed CsPbå2Br is smaller than that from the high-temperature-processed sample.

The top-view scanning electron microscopy (SEM) in Figure 24 shows that the crystal size of the low- temperature-processed CsPbl2Br film is around 200-500 nm, while the crystal size of the high-temperature processed CsPbå2Br film is around 1-2 pm. Solar cells annealed at these low and high temperatures are then fabricated with the structure of FT0/compact-Ti0 ₂ ( c-Ti0 ₂ ) /mesoporous- Ti0 ₂ (mp-Ti0 ₂ ) /CsPblhBr/poly ( 3-hexylthiophene-2 , 5-diyl ) (P3HT)/Au. P3HT is chosen because of better stability compared to the commonly used spiro-OMeTAD . The light current density versus voltage (J-V) characteristics under 1 sun illumination of these devices are shown in Figure 18b. The best device that uses a low-temperature-processed CsPbå2Br has a higher PCE (>7%) than the best device that uses a high-temperature-processed CsPblABr (the

photovoltaic parameters of the devices are shown in Figure 24). To investigate the effect of partially substituting Pb in the low-temperature-processed CsPbl2Br, Sr is

incorporated at different Sr ²⁺ concentrations, and the crystallinity and morphology of the films are studied. XRD patterns of the CsPbi- _x Sr _x l2Br films are shown in Figure 19. The XRD peak intensity of the 14.6° and 29.5° peaks is reduced with increasing Sr ²⁺ due to degradation of the film in the ambient environment during measurement. As the content of Sr ²⁺ increases, no additional phase can be observed, but the XRD peak at 20.4° is stronger,

indicating that the preferred orientation of the

perovskite is altered. However, it is hard to discern change in the perovskite peak location due to the reduced peak intensity and the similar ionic radius of Sr (118 pm) and Pb (119 pm) . The XRD patterns of high-temperature- processed CsPbi- _xS r _xl 2Br films are also measured (see Figure 25) . Identical peaks can be observed. In terms of

crystallinity, the low-temperature CsPbi- _xS r _xl 2Br is comparable to that of the high-temperature-processed counterpart. Figure 26 shows XRD patterns of CsPbi- _xS r _xl 2Br films annealed at 310 °C for 10 minutes.

Figure 26 shows that the XRD patterns of the CsPbi- _xS r _xl 2Br films annealed at high temperature have basically the same pattern as films annealed at low temperature. The difference between the red and black lines is quite similar to that observed in films annealed at low

temperature. The XRD peaks for increasing Sr ²⁺ content are of lower intensity due to the film degradation. A stronger peak is observed at 20.4°, indicating that the preferred orientation of the perovskite has changed.

Figure 19b-f shows the top-view SEM of the CsPbl2Br film and films with increasing Sr content. The reference CsPbl2Br film shows a rough surface and is composed of densely packed crystalline grains, with an average grain size of ~200-500 nm. The addition of Sr ²⁺ drastically changes the morphology of the perovskite and results in the appearance of "snowflakes", which appear brighter under the SEM; see Figure 19. The amount of snowflakes increases with Sr ²⁺ , as shown in Figure 19b-f. Energy dispersive X-ray spectroscopy (EDS) mapping was carried out on the CsPbo.9sSro.o2l2Br film, with results shown in Figure 27. As shown by EDS mapping, all of the elements including Sr were distributed homogeneously throughout the film. This is also verified by the backscattered electron (BSE) images of the CsPbo.99Sro.oil2Br and CsPbo.97Sro.o3l2Br films (see Figure 28), showing negligible difference (same average atomic number) in the film composition in the snowflake regions throughout the bulk of the film In order to study the effect of Sr ²⁺ incorporation on the elemental composition at the surface of CsPbi- _x Sr _x l2Br films, X-ray photoelectron spectroscopy (XPS) was

performed; see Figure 20. As shown in Figure 20a, the Pb 4f spectrum for the CsPbå2Br film shows 4fs _/2 and 4f7 _/2 peaks at 137.9 and 142.8 eV, respectively, corresponding to the Pb ²⁺ cations. The I 3d _3/2 and I 3ds _/2 peaks are also evident in the I 3d spectrum for the CsPbå2Br film; see Figure 20b.

After incorporation of Sr, the Pb 4fs _/2 , Pb 4f7 _/2 , I 3d _3/2 , and I 3ds _/2 peaks shift to higher binding energy, indicating that the chemical structures of the surface have been modified. In addition, Sr ²⁺ cations at the surface of the film are evident with the presence of Sr 3d _3/2 and 3ds _/2 peaks at 134.4 and 136.3 eV, respectively.

The Sr/Pb atomic ratios at the surface of the CsPbi- _x Sr _x l2Br films are summarized in Table 1.

Table 1. Sr/Pb Atomic Ratios of CsPbi- _x Sr _x l2Br Films

Sr/Pb Pbi.ooSro.oo Pbo.99Sro.01 Pbo.98Sro.02 Pbo.97Sro.03 Pbo.95Sro.05 atomic

ratio

(%) *

surface 11.4 20.1 28.4 38.2

3 nm 4.9 6.8 9.8 12.6 etched

*only the ratio of Pb : Sr in CsPbi- _x Sr _x l2Br is shown for clarity

The Sr/Pb atomic ratios at the surface for the

CsPbo.98Sro .02å2Br and CsPbo.95Sro.o5l2Br films are 0.20 and 0.38. At the depth of 3 nm, the Sr/Pb atomic ratio drops dramatically to 0.07 and 0.13 for CsPbo.9sSro.o2l2Br and CsPbo.95Sro .05l2Br films. However, these ratios are higher than the molar ratios used in the synthesis of the films, which are 0.02 and 0.05, respectively, for the

CsPbo.98Sro .02l2Br and CsPbo.95Sro.o5l2Br films. This shows that the film surface is strongly enriched with Sr ²⁺ compared to the bulk of the film. In addition, XPS measurements of the film at the surface and at 10 nm from the surface shown in Figure 29 show that the binding energy of the Sr 3p peak shifted from 268.5 eV (at the surface) to 269.5 eV (depth = 10 nm) , suggesting the possibility Sr oxide formation on the surface because for SrO the Sr 3p peak has been reported to be between 268 eV and 269 eV.

Optical and photoluminescence measurements are also carried out on the CsPbi- _xS r _xl 2Br films; see Figure 21.

Figure 21a shows that the absorption onset of CsPbi- _xS r _xå 2Br is around 656-664 nm (1.87-1.89 eV) and there is no shift in the absorption onset with Sr content. The absorption of the film improves when Sr is incorporated as long as it is limited to less than 5%. The steady-state PL spectra of these films also show peaks at around 660-665 nm (see Figure 30), which are consistent with those in the absorbance spectra. Time-resolved PL (tr-PL) decays for CsPbi- _xS r _xl 2Br/mp-Al203 glass are also measured and shown in Figure 21b. An insulating mp-Al203 layer was coated as the scaffold layer of the perovskite film to eliminate the effect of electron extraction on the results. Using a biexponential decay function, the PL decay traces were fitted to determine the decay times of the fast and slow components, which are summarized in Table 2. Table 2. Typical Lifetimes Extracted from TCSPC for

CsPbi- _x Sr _x l2Br

lifetime Pbi.ooSro.oo Pb0.99Sr0.0 ₁ Pb0.98Sr0.02 Pb0.9 ₇ Sr0.0 ₃ Pb0.95Sr0.05

(ns ) *

n 2.2 2 2.1 1.6 2.3 t ₂ 11.1 13.3 17.1 16.7 9.3

*only the ratio of Pb : Sr in CsPbi- _x Sr _x l2Br is shown for clarity

The presence of the fast component (ii) in the PL decay is commonly assumed to indicate the presence of defect trapping, and the slow component ( 2) corresponds to the effective recombination lifetime. The defect trapping lifetime, ii, of all of the films is relatively the same, with a value of 2 ns. However, as the Sr content increases from 1% to 2%, 2 increases from 11.1 ns to 17.1 ns, suggesting a better effective recombination lifetime. This suggests better surface passivation provided by the Sr- enriched surface, as is evident in the XPS results.

However, as the Sr content increases further, for example, at a concentration of 5%, the excess Sr ²⁺ doping in the perovskite film enhances electron-hole recombination, which will have a detrimental effect on photovoltaic performance of the CsPbo.gsSro.os^Br device.

To confirm the effectiveness of Sr incorporation on device performance, FT0/c-Ti0 ₂ /mp-Ti0 ₂ /CsPbi- _x Sr _x l ₂ Br/P3HT/Au solar cells were fabricated. Results are shown in Figure 22 and Table 3. Table 3. Photovoltaic Parameters of FT0/c-Ti0 ₂ /mp-Ti0 ₂ / CsPbi- _x Sr _x l ₂ Br/P3HT/Au Champion Devices under

Reverse Scan

PCE (%) Jsc (mA/cm ² ) Voc (mV) FF (%)

Reference 7.7 13.4 962 59.8 1% Sr 8.3 14.3 938 62.2 2% Sr 11.2 15.3 1043 69.9 3% Sr 9.8 14.2 999 69.2 5% Sr 6.4 11.2 927 61.3 As the Sr content in the perovskite increases, the short-circuit current density J _sc increases as long as the amount of Sr ²⁺ is small, that is, <5%. The improvement in J _sc is due to better absorption in the perovskite film, as measured and discussed earlier (Figure 21a). The trends between J _sc in Figure 22c and the absorption in Figure 21a are consistent. The incorporation of Sr also improves the open-circuit voltage V _oc and fill factor (FF) of the cells (see Figure 22c and Table 3), resulting in an increase in PCE when x < 0.05. This is due to the better effective lifetime as measured by tr-PL and discussed previously and is likely to be due to the passivation provided by enrichment of Sr at the surface of the CsPbi- _x Sr _x I2Br . The benefits of Sr cease when its content is too high, for example, when x = 0.05. The drop in J _sc and the poorer EQE (Figure 21b) for the CsPbo.gsSro.os^Br devices are due to the lack of improvement in absorption and the degraded lifetime, which also lead to drops in the V _oc and FF, resulting in reduced PCE to values well below those of the control CsPblhBr devices; see Figure 22c.

An average PCE of 25 CsPbo. sSro.o l Br devices is 10.1%

(see Figure 30). The champion CsPbo. sSro.o l Br device achieves the highest PCE at 11.3% and a stabilized PCE at 10.8%. These results are remarkable among reported inorganic perovskite solar cells. In addition, this is the first demonstration of low-temperature-processed Sr- incorporated CsPbl Br solar cells that has a comparable efficiency to that of an inorganic perovskite device annealed at high temperature. The light J-V curves of the champion CsPbo. gsSro. ₀₂ l ₂ Br device scanned at different rates and in opposite directions are shown in Figure 23a. The hysteresis observed in our J-V curves is typical for

CsPbl Br devices. The inclusion of mp-Ti0 ₂ reduces

hysteresis when our device is compared to planar

structure. Therefore, a high stabilized efficiency at 10.8% can be achieved in the best device; see Figure 23b.

In addition to device performance, the stability of the CsPbl Br and CsPbo. gsSro. ₀₂ l ₂ Br devices was also

evaluated. Figure 23c shows the normalized V _oc , J _sc , FF, and PCE as a function of storage (25 °C, relative humidity <

50% in the dark) time of an encapsulated CsPbo . _g sSro . lBr device. During the first week of storage, V _oc and J _sc remained the same, while the FF increased, which led to an increase in PCE due to oxidation of P3HT. It is worthwhile to note that the rate of oxidation in this case was slower than that previously reported due to the use of

encapsulation in the present case [CS Appl. Mater.

Interfaces 2016, 8, 12101-12108]. After the first week, the normalized device parameter remained stable for 3 weeks. The encapsulated CsPblgBr device (see Figure 32) shows a similar trend as the encapsulated CsPbo. gsSro. ₀₂ l ₂ Br device showing good stability. Figure 31 shows

distributions of reverse scan PCE of 25 CsPbo. gsSro. ₀₂ l ₂ Br devices fabricated in this work.

In addition, a preliminary thermal stability test was performed on CsPblgBr and CsPbo. gsSro. ₀₂ l ₂ Br devices. J-V measurements were taken before and after heat treatment at 100 °C in ambient conditions for 1 h. As shown in Figure 23d, there is nearly no difference in the J-V curves of the CsPbo.98Sro.02l2Br after heat treatment, while the

CsPbl2Br reference cell experienced V _oc and FF drops after the heat treatment. This indicates that CsPbo.9sSro.o2l2Br has better thermal stability than CsPbl2Br.

Low-temperature-processed caesium lead halide perovskite solar cells on a mesoporous architecture have been successfully demonstrated using the gas-quenching method. By partially substituting Pb with Sr, CsPbi- _x Sr _x l2Br films were investigated. At optimum concentration (x = 0.02), the average PCE of CsPbi- _x Sr _x l2Br devices increased from 6.6% (for CsPbl2Br) to 10.1%. (for CsPbo.9sSro.02l2Br) . The champion CsPbo.9sSro.o2l2Br cell delivered the highest PCE at 11.3%, with a V _oc of 1.07 V, a J _sc of 14.9 mA cm ^-2 , a FF of 0.71, and a stabilized efficiency at 10.8%. Sr-doped CsPbl2Br showed better thermal stability.

In the claims which follow and in the preceding description of the invention, except where the context requires otherwise due to express language or necessary implication, the word "comprise" or variations such as "comprises" or "comprising" is used in an inclusive sense, i.e. to specify the presence of the stated features but not to preclude the presence or addition of further features in various embodiments of the invention.