FIELD OF THE INVENTION

The present invention relates to a perovskite on silicon multi-junction photovoltaic device that comprises a perovskite material that has both a band gap that makes it suitable for use in multi-junction photovoltaic devices and improved stability, and methods of fabricating such a photovoltaic device.

BACKGROUND OF THE INVENTION

Over the past forty years or so there has been an increasing realisation of the need to replace fossil fuels with more sustainable energy sources. The new energy sources should also have low environmental impact, be highly efficient and be easy to use and cost effective to produce. To this end, solar energy is one of the most promising technologies. However, the high cost of manufacturing devices that capture solar energy, including high material costs, has historically hindered its widespread use.

Every solid has its own characteristic energy-band structure which determines a wide range of electrical characteristics. Electrons can transition from one energy band to another, but each transition requires a specific minimum energy and the amount of energy required will be different for different materials. The electrons acquire the energy needed for the transition by absorbing either a phonon (heat) or a photon (light). The term "band gap" for crystalline materials refers to the energy difference range in a solid where no electron states can exist, and generally means the energy difference (in electron volts) between the top of the valence band and the bottom of the conduction band. The efficiency of a material used in a photovoltaic device, such as a solar cell, under normal daylight conditions is a function of the band gap for that material. If the band gap is too high, most daylight photons cannot be absorbed and are lost to transmission or reflection; if it is too low, then most photons have much more energy than necessary to excite electrons across the band gap, and the rest will be lost to thermalisation. The Shockley-Queisser limit refers to the theoretical maximum efficiency that can be obtained with a single junction solar cell. The highest possible efficiency is about 33% and can be obtained with a 1.34eV band gap. The focus of much of the recent work on photovoltaic devices has been the quest for materials which have a band gap as close as possible to this maximum.

One class of photovoltaic materials that has attracted significant interest has been the hybrid organic-inorganic halide perovskites. Materials of this type form an ABX3 crystal structure which has been found to show a favourable band gap, a high absorption coefficient and long diffusion lengths, making such compounds ideal as an absorber in photovoltaic devices. Early examples of hybrid organic-inorganic metal halide perovskite materials are reported by Kojima, A et al. (2009) Organometal Halide Perovskites as Visible-Light Sensitizers for Photovoltaic Cells. J. Am. Chem. Soc, 131 (17), pp.6050-6051 in which such perovskites were used as the sensitizer in liquid electrolyte based photoelectrochemical cells. Kojima et al. report that a highest obtained solar energy conversion efficiency (or power energy conversion efficiency, PCE) of 3.8%, although in this system the perovskite absorbers decayed rapidly and the cells dropped in performance after only 10 minutes.

Subsequently, Lee, M et al. "Efficient Hybrid Solar Cells Based on Meso-Superstructured Organometal Halide Perovskites", Science, 338(6107), pp.643-647 (2012) reported a "meso- superstructured solar cell" in which the liquid electrolyte was replaced with a solid-state hole- transporting material, (HTM), spiro-MeOTAD. Lee et al. reported a significant increase in the conversion efficiency achieved, whilst also achieving greatly improved cell stability through avoiding the use of a liquid solvent. In the examples described therein, CHsNHsPbb perovskite nanoparticles assumed the role of the sensitizer within the photovoltaic cell, injecting electrons into a mesoscopic T1O2 scaffold and holes into the solid-state HTM. Both the T1O2 and the HTM act as selective contacts through which the charge carriers produced by photoexcitation of the perovskite nanoparticles are extracted.

Further work described in WO2013/171517 disclosed how the use of mixed-anion perovskites as a sensitizer/absorber in photovoltaic devices, instead of single-anion perovskites, results in more stable and highly efficient photovoltaic devices. In particular, this document disclosed that the superior stability of the mixed-anion perovskites is highlighted by the finding that the devices exhibit negligible colour bleaching during the device fabrication process, whilst also exhibiting full sun power conversion efficiency of over 10%. In comparison, equivalent single- anion perovskites are relatively unstable, with bleaching occurring quickly when fabricating films from the single halide perovskites in ambient conditions.

More recently, WO2014/045021 described planar heterojunction (PHJ) photovoltaic devices comprising a thin film of a photoactive perovskite absorber disposed between n-type (electron transporting) and p-type (hole transporting) layers. Unexpectedly it was found that good device efficiencies could be obtained by using a compact (i.e. without effective/open porosity) thin film of the photoactive perovskite, as opposed to the requirement of a mesoporous composite, demonstrating that perovskite absorbers can function at high efficiencies in simplified device architectures. Recently the application of perovskites in photovoltaic devices has focussed on the potential of these materials to boost the performance of conventional silicon-based solar cells by combining them with a perovskite-based cell in a tandem/multi-junction arrangement. In this regard, a multi-junction photovoltaic device comprises multiple separate sub-cells (i.e. each with their own photoactive region) that are "stacked" on top of one another such that the photoactive regions overlie one another, and convert more of the solar spectrum into electricity - increasing the overall efficiency of the device. To do this, the photoactive region of each sub- cell is selected so that the band gap of the sub-cell ensures that it will efficiently absorbs photons from a specific segment of the solar spectrum. Multi-junction solar cells are a way to circumvent the above-mentioned Shockley-Queisser limit if an appropriate band gap combination is used. The combination of multiple photoactive regions/sub-cells with different band gaps provides that on the one hand, a wide range of incident photons can be absorbed in the lower bandgap sub cells, while on the other hand, each photoactive region/sub-cell will be more effective at extracting energy from the photons within the relevant part of the spectrum in the higher bandgap sub cells. In theory, the lowest band gap of a multi-junction photovoltaic device will be lower than that of a typical single junction device, such that a multi-junction device will be able to absorb photons that possess less energy than those that can be absorbed by a single junction device. Furthermore, for those photons that would be absorbed by both a multi-junction device and a single junction device, the multi-junction device will absorb those photons more efficiently, as having band gaps closer to the photon energy reduces thermalization losses.

In a typical multi-junction device the top photoactive region/sub-cell in the stack has the highest band gap, with the band gap of the lower photoactive regions/sub-cells reducing towards the bottom of the device. This arrangement maximizes photon energy extraction as the top photoactive region/sub-cell absorbs the highest energy photons first whilst allowing the transmission of photons with less energy. Each subsequent photoactive region/sub-cell then extracts energy from photons closest to its band gap thereby minimizing thermalization losses. The bottom photoactive region/sub-cell then absorbs all remaining photons with energy above its band gap. When designing multi-junction cells it is therefore important to choose photoactive regions/sub-cells with the right bandgaps in order to optimise harvesting of the solar spectrum. In this regard, for a tandem photovoltaic device that comprises two photoactive regions/sub-cells, a top photoactive region/sub-cell and a bottom photoactive region/sub-cell, it has been shown that the bottom photoactive region/sub-cell should have a band gap of around .1 eV whilst the top photoactive region/sub-cell should have a band gap of around 1.7eV (Coutts, T et al, (2002) "Modelled performance of polycrystalline thin-film tandem solar cells", Progress in Photovoltaics: Research and Applications, 10(3), pp.195- 203).

Today's photovoltaic (PV) market is dominated by single-junction solar cells made of silicon (bandgap = 1.12 eV). A low cost material with a complementing bandgap of about 1.7 eV has been sought for a long time within the scientific community. Consequently, there has been interest in developing hybrid organic-inorganic perovskite solar cells for use in tandem photovoltaic devices given that the band gap of these perovskite materials can be tuned from around 1.5eV to over 2eV by varying the halide composition of organometal halide perovskites (Noh, J. et al, (2013) "Chemical Management for Colorful, Efficient, and Stable Inorganic- Organic Hybrid Nanostructured Solar Cells", Nano Letters, p.1303211 12645008). In particular, by varying the halide composition it is possible to tune the band gap of an organometal halide perovskite to around 1.7eV, such that it is then ideal for use as the top sub-cell in a tandem structure when combined with a crystalline silicon bottom sub-cell.

In this regard, Schneider, B.W. et al (Schneider, B.W. et al., [2014] "Pyramidal surface textures for light trapping and a nti reflection in perovskite-on-silicon tandem solar cells", Optics Express, 22(S6), p.A1422) reported on the modelling of a perovskite-on-silicon tandem cell in which the modelled cell has a 4-terminal, mechanically stacked structure. Loper, P. et al (Loper, P. et al., 2015. Organic-inorganic halide perovskite/crystalline silicon four-terminal tandem solar cells. Physical chemistry chemical physics: PCCP, 17, p.1619) reported on the implementation of a four-terminal tandem solar cell consisting of a methyl ammonium lead triiodide (CH ₃ NH ₃ Pbl ₃ ) (i.e. organometal halide perovskite) top sub-cell mechanically stacked on a crystalline silicon heteroj unction bottom sub-cell. Similarly, Bailie, C. et al. (Bailie, C. et al., 2015. Semi-transparent perovskite solar cells for tandems with silicon and CIGS. Energy Environ. Sci., pp.1-28) reported on a mechanically-stacked tandem solar cell consisting of a methyl ammonium lead triiodide (CH3NH ₃ Pbl3) top sub-cell on a copper indium gallium diselenide (CIGS) or low-quality multi-crystalline silicon bottom sub-cell. Filipic, M. et al. (Filipic, M. et al., 2015. CH ₃ NH ₃ Pbl3 perovskite / silicon tandem solar cells: characterization based optical simulations. Optics Express, 23(7), pp.480-484) reported on the simulation of both mechanically stacked (four terminal) and monolithically integrated (two terminal) tandem devices consisting of a methyl ammonium lead triiodide (CHsNHsPbb) top sub-cell and a crystalline silicon bottom sub-cell. Mailoa, J. P. et al. (Mailoa, J. P. et al., [2015], "A 2-terminal perovskite/silicon multi-junction solar cell enabled by a silicon tunnel junction", Applied Physics Letters, 106(12), p.121 105) then reported on the fabrication of a monolithic tandem solar cell consisting of a methyl ammonium lead triiodide (CHsNHsPbU) top sub-cell and a crystalline silicon bottom sub-cell. In a mechanically stacked multi-junction photovoltaic device the individual sub-cells are stacked on top of each other and are each provided with their own separate electrical contacts, such that the individual sub-cells are connected in parallel and do not require current matching. This contrasts with a monolithically integrated multi-junction photovoltaic device in which the individual sub-cells are electrically connected in series between a single pair of terminals, which results in the need for a recombination layer or a tunnel junction and current matching between adjacent sub-cells. Whilst a mechanically stacked multi-junction photovoltaic device does not require current matching between the sub-cells, the additional size and cost of the additional contacts and substrates, and the high losses due to lateral transport, make mechanically stacked structures less favourable than monolithically integrated structures.

SUMMARY OF THE PRESENT INVENTION

According to a first aspect of the present invention there is provided a multijunction photovoltaic device as specified in claims 1 to 10.

In a first embodiment, the multijunction photovoltaic device comprises a perovskite sub-cell in electrical contact with a crystalline silicon sub-cell, the crystalline silicon sub-cell comprising an n-type layer diffused into a p-type substrate and an electrically insulating passivation layer between the n-type layer and the perovskite sub-cell, the passivation layer having a plurality of openings or vias enabling electrical contact between the crystalline sub-cell and the perovskite sub-cell.

Advantageously, the photovoltaic device has a monolithically integrated structure, for example a tandem structure consisting of the perovskite sub-cell and the crystalline silicon sub-cell.

All previously described monolithic perovskite-silicon tandem cells have been formed using silicon cells made using n-type wafers. The vast majority use silicon heterojunction (SHJ) cells as the lower cell. This technology makes up a relatively small fraction of the installed silicon cell capacity. It is highly desirable to realize a tandem cell on a diffused emitter type cell with a p-type wafer, such as comprises > 80% of the current PV market. One such example was described by Mailoa et al. (Applied Physics Letters vol. 106 (No. 12), 121105 (2015) and US2016/0163904, who fabricated a tunnel junction type structure followed by atomic layer deposition of a layer of titania in order to realise a tandem cell using a silicon cell comprising a homojunction cell on an n type silicon wafer. Werner et al. (Applied Physics Letters vol. 109 (No. 23), 233902 (2016) and US2016/0190377 also disclosed the use of zinc tin oxide as a recombination layer directly upon the boron-diffused emitter. In both cases n-type substrates were used. We are not aware of any work demonstrating the fabrication of a tandem cell on a silicon cell based on a p-type wafer. In addition, in silicon based tandem cells, the electrical interface between the two sub-cells has only previously been exemplified as a full area contact with no electrically insulating passivation layer having openings or vias.

The photovoltaic device of the present invention has a multi-junction structure comprising a first sub-cell disposed over a second sub-cell, the first sub-cell comprising a photoactive region comprising the perovskite material. The photovoltaic device may have a monolithically integrated structure. In a monolithically integrated multi-junction photovoltaic device the two or more photovoltaic sub-cells are deposited directly onto one another and are therefore electrically connected in series. The photovoltaic device may optionally further comprise an intermediate region connecting the first sub-cell to the second sub-cell, wherein each intermediate region comprises one or more electrically conductive interconnect layers.

The photoactive region of the perovskite sub-cell may comprise a layer of the perovskite material without open porosity. Alternatively, the layer of the perovskite material may be in contact with a porous scaffold material that is disposed between an n-type region and a p-type region. The porous scaffold material may comprise or consist essentially of any of a dielectric material and a semiconducting/charge transporting material. The layer of the perovskite material may then be disposed within pores of/be conformal with a surface of the porous scaffold material. Alternatively, the layer of the perovskite material may fill the pores of the porous scaffold material and form a capping layer on the porous scaffold material, which capping layer consists of a layer of the photoactive material without open porosity.

The photovoltaic device having a multi-junction structure may further comprise a first electrode, a second electrode, with the first sub-cell and the second sub-cell disposed between the first and second electrodes.

The first electrode may then be in contact with the n-type region of the first (perovskite) sub- cell, and wherein the first electrode comprises a transparent, semi-transparent or light transmissive electrically conductive material. The first electrode may then be an electron collecting electrode, whilst the second electrode is a hole collecting electrode. In a tandem device, the second electrode will then be in contact with the second (crystalline silicon) sub- cell.

The perovskite sub cell preferably comprises a p-i-n structure, in which the p-type layer is deposited first (to be in contact with the n-type layer of the underlying silicon sub-cell), followed by the intrinsic layer and then a n-type layer on top.

The silicon sub cell may comprise a single crystal (mono-crystalline), poly- or micro-crystalline material.

According to a second aspect of the present invention there is provided a method of fabricating a multi-junction photovoltaic device comprising a perovskite sub-cell in electrical contact with a silicon sub-cell, as specified in claims 11 to 18.

In one embodiment, the method comprises:

a. Forming an electrically insulating passivation layer over a silicon sub-cell, the silicon sub-cell comprising an n-type layer diffused into a p-type single crystal or polycrystalline substrate, said passivation layer having a plurality of openings, b. Forming a perovskite sub-cell over the passivation layer, wherein there is electrical contact between the silicon sub-cell and the perovskite sub-cell through the plurality of openings.

Step a. may comprise selective deposition of a patterned passivation layer, for example by evaporation or sputtering through a shadow mask, or by blanket deposition followed by selective removal of material in specific areas to form openings. The selective removal can be achieved by, for example, laser ablation, or photolithography using a photo-patternable resist followed by wet or dry etching. As an alternative, the passivation layer itself could be photo- patternable so that a photoresist is not required. As yet another alternative, the opening can be formed by printing regions comprising a reactive metal (such as silver particles mixed with a glass frit) on top of the passivation layer and subsequently sintering or firing the printed metal mixture to form a metal contact which replaces the passivation layer in the regions printed with the reactive metal mixture.

The passivation layer can be, for example, silicon nitride, SiN _x , silicon oxynitride, silicon dioxide, SiO _x , alumina, titania etc.. The perovskite layers formed in step b. may be deposited using vacuum thin-film deposition techniques or by wet chemical methods from a solution or suspension.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will now be more particularly described by way of example only with reference to the accompanying drawings, in which:

Figure 1 illustrates schematically a prior art monolithically integrated multi-junction photovoltaic device that comprises a top, perovskite based sub-cell and a bottom, silicon heterojunction (SHJ) sub-cell;

Figure 2 illustrates schematically an example of a prior art silicon heterojunction (SHJ) sub cell;

Figure 3 shows a schematic cross-section of a multijunction photovoltaic device according to a first embodiment of the present invention;

Figure 4 shows a prior art PERC cell in cross section;

Figure 5 illustrates a schematic cross-section of a second embodiment of a multi-junction photovoltaic device according to the present invention, which employs a passivated emitter rear contact (PERC) silicon sub cell;

Figure 6 shows an IV plot for a tandem device according to an embodiment of the present invention using a PERC silicon sub-cell and laser ablated contact holes; and Figure 7 illustrates measured quantum efficiency of the same device.

DETAILED DESCRIPTION

Definitions

The term "photoactive", as used herein, refers to a region, layer or material that is capable of responding to light photoelectrically. A photoactive region, layer or material is therefore capable of absorbing the energy carried by photons in light that then results in the generation of electricity (e.g. by generating either electron-hole pairs or excitons).

The term "perovskite", as used herein, refers to a material with a three-dimensional crystal structure related to that of CaTi0 ₃ or a material comprising a layer of material, which layer has a structure related to that of CaTi03. The structure of CaTi03 can be represented by the formula ABX ₃ , wherein A and B are cations of different sizes and X is an anion. In the unit cell, the A cations are at (0,0,0), the B cations are at (1/2, 1/2, 1/2) and the X anions are at (1/2, 1/2, 0). The A cation is usually larger than the B cation. The skilled person will appreciate that when A, B and X are varied, the different ion sizes may cause the structure of the perovskite material to distort away from the structure adopted by CaTiC>3 to a lower-symmetry distorted structure. The symmetry will also be lower if the material is in the form of a layer that has a structure related to that of bulk CaTi03. Materials comprising a layer of perovskite material are well known. For instance, the structure of materials adopting the K2N1F4 type structure comprises a layer of perovskite material. The skilled person will appreciate that a perovskite material can be represented by the formula [A][B][X] ₃ , wherein [A] is at least one cation, [B] is at least one cation and [X] is at least one anion. When the perovskite comprises more than one A cation, the different A cations may distributed over the A sites in an ordered or disordered way. When the perovskite comprises more than one B cation, the different B cations may distributed over the B sites in an ordered or disordered way. When the perovskite comprise more than one X anion, the different X anions may distributed over the X sites in an ordered or disordered way. The symmetry of a perovskite comprising more than one A cation, more than one B cation or more than one X cation, will often be lower than that of CaTiC>3.

As mentioned in the preceding paragraph, the term "perovskite", as used herein, refers to (a) a material with a three-dimensional crystal structure related to that of CaTiC or (b) a material comprising a layer of material, wherein the layer has a structure related to that of CaTiC . Although both of these categories of perovskite may be used in the devices according to the invention, it is preferable in some circumstances to use a perovskite of the first category, (a), i.e. a perovskite having a three-dimensional (3D) crystal structure. Such perovskites typically comprise a 3D network of perovskite unit cells without any separation between layers. Perovskites of the second category, (b), on the other hand, include perovskites having a two- dimensional (2D) layered structure. Perovskites having a 2D layered structure may comprise layers of perovskite unit cells that are separated by (intercalated) molecules; an example of such a 2D layered perovskite is [2-(1-cyclohexenyl)ethylammonium] ₂ PbBr4. 2D layered perovskites tend to have high exciton binding energies, which favours the generation of bound electron-hole pairs (excitons), rather than free charge carriers, under photoexcitation. The bound electron-hole pairs may not be sufficiently mobile to reach the p-type or n-type contact where they can then transfer (ionise) and generate free charge. Consequently, in order to generate free charge, the exciton binding energy has to be overcome, which represents an energetic cost to the charge generation process and results in a lower voltage in a photovoltaic cell and a lower efficiency. In contrast, perovskites having a 3D crystal structure tend to have much lower exciton binding energies (on the order of thermal energy) and can therefore generate free carriers directly following photoexcitation. Accordingly, the perovskite semiconductor employed in the devices and processes of the invention is preferably a perovskite of the first category, (a), i.e. a perovskite which has a three-dimensional crystal structure. This is particularly preferable when the optoelectronic device is a photovoltaic device.

The perovskite material employed in the present invention is one which is capable of absorbing light and thereby generating free charge carriers. Thus, the perovskite employed is a light- absorbing perovskite material. However, the skilled person will appreciate that the perovskite material could also be a perovskite material that is capable of emitting light, by accepting charge, both electrons and holes, which subsequently recombine and emit light. Thus, the perovskite employed may be a light-emitting perovskite.

As the skilled person will appreciate, the perovskite material employed in the present invention may be a perovskite which acts as an n-type, electron-transporting semiconductor when photo-doped. Alternatively, it may be a perovskite which acts as a p-type hole-transporting semiconductor when photo-doped. Thus, the perovskite may be n-type or p-type, or it may be an intrinsic semiconductor. In preferred embodiments, the perovskite employed is one which acts as an n-type, electron-transporting semiconductor when photo-doped. The perovskite material may exhibit ambipolar charge transport, and therefore act as both n-type and p-type semiconductor. In particular, the perovskite may act as both n-type and p-type semiconductor depending upon the type of junction formed between the perovskite and an adjacent material.

Typically, the perovskite semiconductor used in the present invention is a photosensitizing material, i.e. a material which is capable of performing both photogeneration and charge transportation.

The term "mixed-anion", as used herein, refers to a compound comprising at least two different anions. The term "halide" refers to an anion of an element selected from Group 17 of the Periodic Table of the Elements, i.e., of a halogen. Typically, halide anion refers to a fluoride anion, a chloride anion, a bromide anion, an iodide anion or an astatide anion.

The term "metal halide perovskite", as used herein, refers to a perovskite, the formula of which contains at least one metal cation and at least one halide anion. The term "organometal halide perovskite", as used herein, refers to a metal halide perovskite, the formula of which contains at least one organic cation.

The term "organic material" takes its normal meaning in the art. Typically, an organic material refers to a material comprising one or more compounds that comprise a carbon atom. As the skilled person would understand it, an organic compound may comprise a carbon atom covalently bonded to another carbon atom, or to a hydrogen atom, or to a halogen atom, or to a chalcogen atom (for instance an oxygen atom, a sulphur atom, a selenium atom, or a tellurium atom). The skilled person will understand that the term "organic compound" does not typically include compounds that are predominantly ionic such as carbides, for instance.

The term "organic cation" refers to a cation comprising carbon. The cation may comprise further elements, for example, the cation may comprise hydrogen, nitrogen or oxygen. The term "inorganic cation" refers to a cation that is not an organic cation. By default, the term "inorganic cation" refers to a cation that does not contain carbon.

The term "semiconductor", as used herein, refers to a material with electrical conductivity intermediate in magnitude between that of a conductor and a dielectric. A semiconductor may be an n-type semiconductor, a p-type semiconductor or an intrinsic semiconductor.

The term "dielectric", as used herein, refers to material which is an electrical insulator or a very poor conductor of electric current. The term dielectric, as used herein, typically refers to materials having a band gap of equal to or greater than 3.0 eV, preferably greater than 4 eV.

The term "n-type", as used herein, refers to a region, layer or material that comprises an extrinsic semiconductor with a larger concentration of electrons than holes. In n-type semiconductors, electrons are therefore majority carriers and holes are the minority carriers, and they are therefore electron transporting materials. The term "n-type region", as used herein, therefore refers to a region of one or more electron transporting (i.e. n-type) materials. Similarly, the term "n-type layer" refers to a layer of an electron-transporting (i.e. an n-type) material. An electron-transporting (i.e. an n-type) material could be a single electron- transporting compound or elemental material, or a mixture of two or more electron-transporting compounds or elemental materials. An electron-transporting compound or elemental material may be undoped or doped with one or more dopant elements.

The term "p-type", as used herein, refers to a region, layer or material that comprises an extrinsic semiconductor with a larger concentration of holes than electrons. In p-type semiconductors, holes are the majority carriers and electrons are the minority carriers, and they are therefore hole transporting materials. The term "p-type region", as used herein, therefore refers to a region of one or more hole transporting (i.e. p-type) materials. Similarly, the term "p-type layer" refers to a layer of a hole-transporting (i.e. a p-type) material. A hole- transporting (i.e. a p-type) material could be a single hole-transporting compound or elemental material, or a mixture of two or more hole-transporting compounds or elemental materials. A hole-transporting compound or elemental material may be undoped or doped with one or more dopant elements.

The term "band gap", (sometimes spelled bandgap) as used herein, refers to the energy difference between the top of the valence band and the bottom of the conduction band in a material. The skilled person may readily measure the band gap of a material without undue experimentation.

The term "layer", as used herein, refers to any structure which is substantially laminar in form (for instance extending substantially in two perpendicular directions, but limited in its extension in the third perpendicular direction). A layer may have a thickness which varies over the extent of the layer. Typically, a layer has approximately constant thickness. The "thickness" of a layer, as used herein, refers to the average thickness of a layer. The thickness of layers may easily be measured, for instance by using microscopy, such as electron microscopy of a cross section of a film, or by surface profilometry for instance using a stylus profilometer.

The term "porous", as used herein, refers to a material within which pores are arranged. Thus, for instance, in a porous material the pores are volumes within the body of the material where there is no material. The individual pores may be the same size or different sizes. The size of the pores is defined as the "pore size". The limiting size of a pore, for most phenomena in which porous solids are involved, is that of its smallest dimension which, in the absence of any further precision, is referred to as the width of the pore (i.e. the width of a slit-shaped pore, the diameter of a cylindrical or spherical pore, etc.). To avoid a misleading change in scale when comparing cylindrical and slit-shaped pores, one should use the diameter of a cylindrical pore (rather than its length) as its "pore-width " (Rouquerol, J. et al, (1994) Recommendations for the characterization of porous solids (Technical Report). Pure and Applied Chemistry, 66(8)). The following distinctions and definitions were adopted in previous lUPAC documents (J. Haber. (1991) Manual on catalyst characterization (Recommendations 1991). Pure and Applied Chemistry.): micropores have widths (i.e. pore sizes) smaller than 2 nm; Mesopores have widths (i.e. pore sizes) of from 2 nm to 50 nm; and Macropores have widths (i.e. pore sizes) of greater than 50 nm. In addition, nanopores may be considered to have widths (i.e. pore sizes) of less than 1 nm.

Pores in a material may include "closed" pores as well as open pores. A closed pore is a pore in a material which is a non-connected cavity, i.e. a pore which is isolated within the material and not connected to any other pore and which cannot therefore be accessed by a fluid to which the material is exposed. An "open pore" on the other hand, would be accessible by such a fluid. The concepts of open and closed porosity are discussed in detail in J. Rouquerol et al..

Open porosity, therefore, refers to the fraction of the total volume of the porous material in which fluid flow could effectively take place. It therefore excludes closed pores. The term "open porosity" is interchangeable with the terms "connected porosity" and "effective porosity", and in the art is commonly reduced simply to "porosity". The term "without open porosity", as used herein, therefore refers to a material with no effective porosity. Thus, a material without open porosity typically has no macropores and no mesopores. A material without open porosity may comprise micropores and nanopores, however. Such micropores and nanopores are typically too small to have a negative effect on a material for which low porosity is desired.

In addition, polycrystalline materials are solids that are composed of a number of separate crystallites or grains, with grain boundaries at the interface between any two crystallites or grains in the material. A polycrystalline material can therefore have both interparticle/interstitial porosity and intraparticle/internal porosity. The terms "interparticle porosity" and "interstitial porosity", as used herein, refer to pores between the crystallites or grains of the polycrystalline material (i.e. the grain boundaries), whilst the terms "intraparticle porosity" and "internal porosity", as used herein, refer to pores within the individual crystallites or grains of the polycrystalline material. In contrast, a single crystal or monocrystalline material is a solid in which the crystal lattice is continuous and unbroken throughout the volume of the material, such that there are no grain boundaries and no interparticle/interstitial porosity.

The term "compact layer", as used herein, refers to a layer without mesoporosity or macroporosity. A compact layer may sometimes have microporosity or nanoporosity.

The term "scaffold material", as used herein, therefore refers to a material that is capable of acting as a support for a further material. The term "porous scaffold material", as used herein, therefore refers to a material which is itself porous, and which is capable of acting as a support for a further material.

The term "transparent", as used herein, refers to material or object allows visible light to pass through almost undisturbed so that objects behind can be distinctly seen. The term "semi- transparent", as used herein, therefore refers to material or object which has a transmission (alternatively and equivalently referred to as a transmittance) to visible light intermediate between a transparent material or object and an opaque material or object. Typically, a transparent material will have an average transmission for visible light (generally light with a wavelength of from 370 to 740 nm) of around 100%, or from 90 to 100%. Typically, an opaque material will have an average transmission for visible light of around 0%, or from 0 to 5%. A semi-transparent material or object will typically have an average transmission for visible light of from 10 to 90%, typically 40 to 60%. Unlike many translucent objects, semi-transparent objects do not typically distort or blur images. Transmission for light may be measured using routine methods, for instance by comparing the intensity of the incident light with the intensity of the transmitted light.

The term "electrode", as used herein, refers to a conductive material or object through which electric current enters or leaves an object, substance, or region. The term "negative electrode", as used herein, refers to an electrode through which electrons leave a material or object (i.e. an electron collecting electrode). A negative electrode is typically referred to as an "anode". The term "positive electrode", as used herein, refers to an electrode through which holes leave a material or object (i.e. a hole collecting electrode). A positive electrode is typically referred to as a "cathode". Within a photovoltaic device, electrons flow from the positive electrode/cathode to the negative electrode/anode, whilst holes flow from the negative electrode/anode to the positive electrode/cathode.

The term "front electrode", as used herein, refers to the electrode provided on that side or surface of a photovoltaic device that it is intended will be exposed to sun light. The front electrode is therefore typically required to be transparent, semi-transparent, or at least light transmissive so as to allow light to pass through the electrode to the photoactive layers provided beneath the front electrode. The term "back electrode", as used herein, therefore refers to the electrode provided on that side or surface of a photovoltaic device that is opposite to the side or surface that it is intended will be exposed to sun light.

The term "charge transporter" refers to a region, layer or material through which a charge carrier (i.e. a particle carrying an electric charge), is free to move. In semiconductors, electrons act as mobile negative charge carriers and holes act as mobile positive charges. The term "electron transporter" therefore refers to a region, layer or material through which electrons can easily flow and that will typically reflect holes (a hole being the absence of an electron that is regarded as a mobile carrier of positive charge in a semiconductor). Conversely, the term "hole transporter" refers to a region, layer or material through which holes can easily flow and that will typically reflect electrons. The term "consisting essentially of" refers to a composition comprising the components of which it consists essentially as well as other components, provided that the other components do not materially affect the essential characteristics of the composition. Typically, a composition consisting essentially of certain components will comprise greater than or equal to 95 wt% of those components or greater than or equal to 99 wt% of those components.

The term "bifacial", as used herein, refers to a photovoltaic device/solar cell/sub-cell that can collect light and generate electricity through both of its faces; the front, sun-exposed face and the rear face. Bifacial devices/cells achieve a power gain by making use of diffuse and reflected light as well as direct sunlight. In contrast, the term "monofacial" refers to a photovoltaic device/solar cell/sub-cell that can only collect light and generate electricity through its front, sun-exposed face.

The term "conform", as used herein, refers to an object that is substantially the same in form or shape as an another object. A "conformal layer", as used herein, therefore refers to a layer of material that conforms to the contours of the surface on which the layer is formed. In other words, the morphology of the layer is such that the thickness of the layer is approximately constant across the majority of the interface between the layer and the surface on which the layer is formed.

Device Structure - General

Figure 1 illustrates schematically a prior art monolithically integrated multi-junction photovoltaic device 100 that comprises a first/top sub-cell 1 10 comprising a photoactive region that comprises a perovskite material, whilst the second/bottom sub-cell 120 comprises a silicon heterojunction (SHJ). The multi-junction photovoltaic device 100 has a monolithically integrated structure and therefore comprises just two electrodes, a front/first electrode 101 and a back/second electrode 102, with the first/top sub-cell 1 10 and the second/bottom sub- cell 120 disposed between these two electrodes. In particular, the first sub-cell 110 is in contact with the first/front electrode 101 and the second sub-cell 120 is in contact with the second/back electrode 102. The monolithically integrated multi-junction photovoltaic device 100 typically also comprises a metal grid on the top surface of the front/first electrode 101 as a top contact (not shown). By way of example, the top contact could be provided a metal grid or fingers produced by screen printing of a silver and/or copper paste. In addition, as the monolithically integrated structure comprises just two electrodes, the first and second sub-cells 1 10, 120 are then connected to one another by an intermediate region 130 comprising one or more interconnect layers. For example, the interconnect layer(s) can comprise any of a recombination layer and a tunnel junction. In a monolithically integrated multi-junction photovoltaic device the individual sub-cells are electrically connected in series, which results in the need for a recombination layer or a tunnel junction and current matching between the sub-cells.

The perovskite material in the photoactive region of the first sub-cell 110 is configured to function as a light absorber (i.e. photosensitizer) within the photoactive region. As the top sub- cell in a multi-junction device, the perovskite material therefore preferably has a band gap from 1.50eV to 1.75eV, and more preferably from 1.57eV to 1.67eV. The second sub-cell comprising the silicon heterojunction (SHJ) then preferably has a band gap of around 1.1 eV.

In addition, the perovskite material in the photoactive region of the first sub-cell 1 10 may also be configured to provide charge transport. In this regard, perovskite materials are able to act not only as light absorbers (i.e. photosensitizers) but also as an n-type, p-type or intrinsic (i- type) semiconductor material (charge transporter). A perovskite material can therefore act both as a photosensitizer and as the n-type semiconductor material. The perovskite material may therefore assume the roles both of light absorption and long range charge transport. This structure is described in more detail in WO2016/198898. The perovskite sub cell has an n-i-p arrangement with the n-type layer adjacent the silicon sub-cell.

Figure 2 illustrates schematically an example of the second/bottom sub-cell 120 that in the prior art (WO2016/ 198898) comprises a silicon heterojunction (SHJ). In this regard, the term silicon heterojunction (SHJ) refers to an amorphous silicon/crystalline silicon heterojunction that makes use of a crystalline silicon (c-Si) wafer 121 as a photoactive absorber and hydrogenated amorphous silicon (a-Si:H) thin-films 122, 123, 124, 125 for junction formation and surface passivation. A silicon heterojunction (SHJ) is sometimes also referred to as a heterojunction with intrinsic thin layer (HIT) when any thin layers of intrinsic hydrogenated amorphous silicon (a-Si:H) are present as passivation/buffer layers. A silicon heterojunction (SHJ) therefore typically comprises a p-type a-Si:H emitter 122, an intrinsic a-Si:H passivation/buffer layer 123, an n-type c-Si photoactive absorber 121 , another intrinsic a-Si:H passivation/buffer layer 124, and a back-surface field (BSF) layer made of n-type a-Si 125. Optionally, a silicon heterojunction (SHJ) can further comprise a layer of a transparent conducting oxide (TCO) (e.g. ITO) 26 between the back-surface field (BSF) layer 25 and the back electrode 102. When present, this rear layer of TCO assists in maximising the infrared response by increasing internal reflectance at the rear surface.

The use of a silicon heterojunction (SHJ) as the second/bottom sub-cell 120 has a number of advantages. Firstly, single-junction solar cells based on silicon heterojunction (SHJ) technology have been shown to achieve records energy conversion efficiencies of over 25%, which maximises the potential for a multi-junction device comprising a silicon heterojunction (SHJ) cell to achieve high efficiencies. Secondly, as the silicon heterojunction (SHJ) makes use of an n-type c-Si photoactive absorber 121 with a p-type a-Si emitter 122, the formation of the perovskite-based first sub-cell 110 on the second sub-cell 120 as a substrate is initiated by the deposition of the n-type layers followed by the sequential deposition of the perovskite material and the p-type layers, which have been found to have advantages when processing a monolithically integrated perovskite-on-silicon multi-junction photovoltaic device.

In contrast to the tandem cells described above and in the prior art, the present invention does not use a c-Si photoactive absorber fabricated on an n-type wafer as the second/bottom sub- cell. Instead it comprises a diffused emitter type cell with a p-type wafer as the c-Si photoactive absorber having a diffused n-type layer on top. Suitable silicon sub cells for the present invention include the type of cell is known as an Aluminium Back Surface Field (AI-BSF) cell and the type of cell known as the passivated emitter rear contact (PERC) cell.

Figure 3 shows a schematic diagram showing a cross-section of a first example of a multi- junction photovoltaic device according to the present invention. The silicon sub-cell (220) comprises a p-type silicon substrate (1) having on its upper surface a thin n-type layer (3) on top of this layer is provided an electrically insulating layer (5). This layer is provided with contact holes (23) filled by an electrically conductive material forming vias (25). The electrically conductive material is in contact on its bottom side with the n-type silicon layer, and on the other side with a thin layer (7) of a light transmissive conductive material such as a transparent conductive oxide (TCO) which covers the insulating layer 5 and conductive vias (25). A rear contact 17 is provided by a layer of aluminium. Region 14 is an induced p+ layer adjacent this rear electrode.

The upper layers comprise the perovskite sub-cell (210) and include the thin conductive layer 7, a hole transporting layer (9) in contact with the conductive layer 7, a (nominally intrinsic) perovskite photoactive absorber layer (11) and an electron transporting layer (13). Finally, a further thin light transmissive conductive layer ( 5) is provided on the top surface. A further electrode (not shown) may be deposited on top of the conductive layer 15 - typically a patterned layer comprising a metal having a high conductivity such as Al, Cu or Ag.

An important aspect of the present invention is that the perovskite sub cell 210 has what is known as an "inverted" structure as described in "Organic-Inorganic Halide Perovskite Photovoltaics: From Fundamentals to Device Architectures" edited by Nam-Gyu Park et al., Springer (2016) ISBN-13: 978-3319351124 on pages 307 to 324 - i.e. the p type layer is adjacent the silicon sub cell, and is deposited in the sequence p-i-n.

In the above described multi-junction photovoltaic devices, the hole transporting layer (9) of the first (perovskite) sub-cell typically comprises one or more p-type layers. Often, the p-type region is a p-type layer, i.e. a single p-type layer. In other examples, however, the p-type region may comprise a p-type layer and a p-type exciton blocking layer or electron blocking layer. If the valence band (or highest occupied molecular orbital energy levels) of the exciton blocking layer is closely aligned with the valence band of the photoactive material, then holes can pass from the photoactive material into and through the exciton blocking layer, or through the exciton blocking layer and into the photoactive material, and we term this a p-type exciton blocking layer. An example of such is tris[4-(5-phenylthiophen-2-yl)phenyl]amine, as described in Masaya Hirade, and Chihaya Adachi, "Small molecular organic photovoltaic cells with exciton blocking layer at anode interface for improved device performance" Appl. Phys. Lett. 99, 153302 (2011).

A p-type layer is a layer of a hole-transporting (i.e. a p-type) material. The p-type material may be a single p-type compound or elemental material, or a mixture of two or more p-type compounds or elemental materials, which may be undoped or doped with one or more dopant elements.

A p-type layer may comprise an inorganic or an organic p-type material. Typically, the p-type region comprises a layer of an organic p-type material.

Suitable p-type materials may be selected from polymeric or molecular hole transporters. The p-type layer employed in the photovoltaic device of the invention may for instance comprise spiro-OMeTAD (2,2',7,7'-tetrakis-(N,N-di-p-methoxyphenylamine)9,9'-spirob ifluorene)), P3HT (poly(3-hexylthiophene)), PCPDTBT (Poly[2, 1 ,3-benzothiadiazole-4,7-diyl[4,4-bis(2- ethylhexyl)-4H-cyclopenta[2, 1-b:3,4-b']dithiophene-2,6-diyl]]), PVK (poly(N-vinylcarbazole)), HTM-TFSI (1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide), Li-TFSI (lithium bis(trifluoromethanesulfonyl)imide) or tBP (tert-butylpyridine). The p-type region may comprise carbon nanotubes. Usually, the p-type material is selected from spiro-OMeTAD, P3HT, PCPDTBT and PVK. Preferably, the p-type region consists of a p-type layer that comprises spiro-MeOTAD.

A p-type layer may for example comprise spiro-OMeTAD (2,2',7,7'-tetrakis-(N,N-di-p- methoxyphenylamine)9,9'-spirobifluorene)), P3HT (poly(3-hexylthiophene)), PCPDTBT (Poly[2, 1 ,3-benzothiadiazole-4,7-diyl[4,4-bis(2-ethylhexyl)-4H-cyclop enta[2, 1-b:3,4- b']dithiophene-2,6-diyl]]), or PVK (poly(N-vinylcarbazole)).

Suitable p-type materials also include molecular hole transporters, polymeric hole transporters and copolymer hole transporters. The p-type material may for instance be a molecular hole transporting material, a polymer or copolymer comprising one or more of the following moieties: thiophenyl, phenelenyl, dithiazolyl, benzothiazolyl, diketopyrrolopyrrolyl, ethoxydithiophenyl, amino, triphenyl amino, carbozolyl, ethylene dioxythiophenyl, dioxythiophenyl, or fluorenyl. Thus, a p-type layer employed in the photovoltaic device of the invention may for instance comprise any of the aforementioned molecular hole transporting materials, polymers or copolymers.

Suitable p-type materials also include m-MTDATA (4,4',4"- tris(methylphenylphenylamino)triphenylamine), MeOTPD (N,N,N',N'-tetrakis(4- methoxyphenyl)-benzidine), BP2T (5,5'-di(biphenyl-4-yl)-2,2'-bithiophene), Di-NPB (N,N'-Di- [(1-naphthyl)-N,N'-diphenyl]-1 , 1 '-biphenyl)-4,4'-diamine), a-NPB (N,N'-di(naphthalen-1-yl)- Ν,Ν'-diphenyl-benzidine), TNATA (4,4',4"-tris-(N-(naphthylen-2-yl)-N- phenylamine)triphenylamine), BPAPF (9,9-bis[4-(N,N-bis-biphenyl-4-yl-amino)phenyl]-9H- fluorene), spiro-NPB (N2,N7-Di-1-naphthalenyl-N2,N7-diphenyl-9,9'-spirobi[9H-fluo rene]-2,7- diamine), 4P-TPD (4,4-bis-(N,N-diphenylamino)-tetraphenyl), PEDOT:PSS and spiro- OMeTAD.

A p-type layer may be doped, for instance with tertbutyl pyridine and LiTFSI. A p-type layer may be doped to increase the hole-density. A p-type layer may for instance be doped with NOBF4 (Nitrosonium tetrafluoroborate), to increase the hole-density.

In other examples, a p-type layer may comprise an inorganic hole transporter. For instance, a p-type layer may comprise an inorganic hole transporter comprising an oxide of nickel, vanadium, copper or molybdenum; Cul, CuBr, CuSCN, Cu20, CuO or CIS; a perovskite; amorphous Si; a p-type group IV semiconductor, a p-type group lll-V semiconductor, a p-type group ll-VI semiconductor, a p-type group l-VII semiconductor, a p-type group IV-VI semiconductor, a p-type group V-VI semiconductor, and a p-type group ll-V semiconductor, which inorganic material may be doped or undoped. A p-type layer may be a compact layer of said inorganic hole transporter.

A p-type layer may for instance comprise an inorganic hole transporter comprising an oxide of nickel, vanadium, copper or molybdenum; Cul, CuBr, CuSCN, Cu20, CuO or CIS; amorphous Si; a p-type group IV semiconductor, a p-type group lll-V semiconductor, a p-type group ll-VI semiconductor, a p-type group l-VII semiconductor, a p-type group IV-VI semiconductor, a p- type group V-VI semiconductor, and a p-type group ll-V semiconductor, which inorganic material may be doped or undoped. A p-type layer may for instance comprise an inorganic hole transporter selected from Cul, CuBr, CuSCN, Cu20, CuO and CIS. A p-type layer may be a compact layer of said inorganic hole transporter.

The p-type region may have a thickness of from 5 nm to 1000 nm. For instance, the p-type region may have a thickness of from 50 nm to 500 nm, or from 100 nm to 500 nm. In the above described multi-junction photovoltaic devices, the p-type region 112 of the first sub-cell preferably has a thickness from 10 nm to 50 nm, and more preferably of approximately 20 nm. The p-type region could also comprise a bi-layer or multilayer structure consisting of 2 or more layers having different materials.

Electron transporting layers suitable for use in perovskite photovoltaic sub cells in the present embodiments have recently been described in the review paper "Current status of electron transport layers in perovskite solar cells: materials and properties", Mahmood, Sarwar and Mehran, RSC Adv. 2017.7.17044.

The electron transporting layers typically comprise n-type regions. In the above described multi-junction photovoltaic device, the n-type region of the first sub-cell comprises one or more n-type layers. Often, the n-type region is an n-type layer, i.e. a single n-type layer. In other examples, however, the n-type region may comprise an n-type layer and a separate n-type exciton blocking layer or hole blocking layer.

An exciton blocking layer is a material which is of wider band gap than the photoactive material, but has either its conduction band or valance band closely matched with those of the photoactive material. If the conduction band (or lowest unoccupied molecular orbital energy levels) of the exciton blocking layer are closely aligned with the conduction band of the photoactive material, then electrons can pass from the photoactive material into and through the exciton blocking layer, or through the exciton blocking layer and into the photoactive material, and we term this an n-type exciton blocking layer. An example of such is bathocuproine (BCP), as described in P. Peumans, A. Yakimov, and S. R. Forrest, "Small molecular weight organic thin-film photodetectors and solar cells" J. Appl. Phys. 93, 3693 (2001) and Masaya Hirade, and Chihaya Adachi, "Small molecular organic photovoltaic cells with exciton blocking layer at anode interface for improved device performance" Appl. Phys. Lett. 99, 153302 (2011)}.

The n-type layer (13) is a layer of an electron-transporting (i.e. an n-type) material. The n- type material may be a single n-type compound or elemental material, or a mixture of two or more n-type compounds or elemental materials, which may be undoped or doped with one or more dopant elements.

The electron transporting material employed may comprise an inorganic or an organic n-type material.

A suitable inorganic n-type material may be selected from a metal oxide, a metal sulphide, a metal selenide, a metal telluride, a perovskite, amorphous or nanocrystalline Si, an n-type group IV semiconductor, an n-type group lll-V semiconductor, an n-type group ll-VI semiconductor, an n-type group l-VII semiconductor, an n-type group IV-VI semiconductor, an n-type group V-VI semiconductor, and an n-type group ll-V semiconductor, any of which may be doped or undoped.

More typically, the n-type material is selected from a metal oxide, a metal sulphide, a metal selenide, and a metal telluride.

Thus, an n-type layer may comprise an inorganic material selected from oxide of titanium, tin, zinc, niobium, tantalum, tungsten, indium, gallium, neodymium, palladium, or cadmium, or an oxide of a mixture of two or more of said metals. For instance, an n-type layer may comprise Ti0 ₂ , Sn0 ₂ , ZnO, Nb ₂ 0 ₅ , Ta ₂ 0 ₅ , W0 ₃ , W ₂ 0 ₅ , ln ₂ 0 ₃ , Ga ₂ 0 ₃ , Nd ₂ 0 ₃ , PbO, or CdO.

Other suitable n-type materials that may be employed include sulphides of cadmium, tin, copper, or zinc, including sulphides of a mixture of two or more of said metals. For instance, the sulphide may be FeS ₂ , CdS, ZnS, SnS, BiS, SbS, or Cu ₂ ZnSnS ₄ .

An n-type layer may for instance comprise a selenide of cadmium, zinc, indium, or gallium or a selenide of a mixture of two or more of said metals; or a telluride of cadmium, zinc, cadmium or tin, or a telluride of a mixture of two or more of said metals. For instance, the selenide may be Cu(ln,Ga)Se2. Typically, the telluride is a telluride of cadmium, zinc, cadmium or tin. For instance, the telluride may be CdTe.

An n-type layer may for instance comprise an inorganic material selected from oxide of titanium, tin, zinc, niobium, tantalum, tungsten, indium, gallium, neodymium, palladium, cadmium, or an oxide of a mixture of two or more of said metals; a sulphide of cadmium, tin, copper, zinc or a sulphide of a mixture of two or more of said metals; a selenide of cadmium, zinc, indium, gallium or a selenide of a mixture of two or more of said metals; or a telluride of cadmium, zinc, cadmium or tin, or a telluride of a mixture of two or more of said metals.

Examples of other semiconductors that may be suitable n-type materials, for instance if they are n-doped, include group IV elemental or compound semiconductors; amorphous Si; group lll-V semiconductors (e.g. gallium arsenide); group ll-VI semiconductors (e.g. cadmium selenide); group l-VII semiconductors (e.g. cuprous chloride); group IV-VI semiconductors (e.g. lead selenide); group V-VI semiconductors (e.g. bismuth telluride); and group ll-V semiconductors (e.g. cadmium arsenide).

When an n-type layer is an inorganic material, for instance ΤΊΟ2 or any of the other materials listed above, it may advantageously be a compact layer of said inorganic material. Preferably the n-type layer is a compact layer of ΤΊΟ2.

Other n-type materials may also be employed, including organic and polymeric electron- transporting materials, and electrolytes. Suitable examples include, but are not limited to a fullerene or a fullerene derivative, an organic electron transporting material comprising perylene or a derivative thereof, or poly{[N,N0-bis(2-octyldodecyl)-naphthalene- ,4,5,8- bis(dicarboximide)-2,6-diyl]-alt-5,50-(2,20-bithiophene)} (P(NDI20D-T2)). For example, the n-type region may comprise a n-type layer comprising one or more of C60, C70, C84, C60- PCBM, C70-PCBM, C84-PCBM and carbon nanotubes. It may comprise C60-IPB, C60-IPH, C70-IPB, C70IPH or mixtures thereof. Such materials are commercially available from Solenne BV, Zernikepark 6, 9747AN Groningen, The Netherlands.

The n-type region may have a thickness of from 3 nm to 1000 nm. Where the n-type region comprises a compact layer of an n-type semiconductor, the compact layer has a thickness of from 3 nm to 200 nm. In the above described multi-junction photovoltaic device, the n-type region 1 11 of the first sub-cell 210 preferably has a thickness from 3 nm to 1000 nm, more preferably 5 nm to 40 nm, and yet more preferably of approximately 5 nm. Thin conductive layer 7 may comprise a transparent conductive oxide (TCO) such as indium tin oxide (ITO) or aluminium doped zinc oxide (AZO), carbons (e.g. graphene), metal nanowires etc. Typically, the intermediate region comprises an interconnect layer that consists of indium tin oxide (ITO) that acts as a recombination layer. Preferably the interconnect layer of ITO has a thickness of from 10 nm to 60 nm, and more preferably a thickness of approximately 50 nm.

Device Structure - Transparent Electrode

In the above described multi-junction photovoltaic device, the thin light transmissive conductive layer (15) is the electrode provided on that side or surface of the photovoltaic device that it is intended will be directly exposed to sun light. The first electrode (15) is therefore required to be light transmissive, preferably transparent, so as to maximise the transmission of the light through the electrode to the photoactive layers of the first and second sub-cells provided beneath, whilst also having sufficient electrical conductivity. In particular, for multi-junction devices, the first electrode should transmit a large proportion of light over the complete optical window (i.e. from 300 nm to 1200 nm in wavelength) as transmission of the longer wavelengths is highly important for achieving useful power conversion efficiencies.

The first electrode 15 therefore preferably consists of material that has a sheet resistance (Rs) from 10 ohms per square (Ω/sq) to 100 Ω/sq and an average transmission for visible and infrared light of at least 85% (i.e. transmits at least 85% of light from 300 nm to 1200 nm in wavelength). More preferably, the first electrode 15 consists of material that has a sheet resistance (Rs) of equal to or less than 50 Ω/sq and an average transmission for visible and infrared light of greater than 90%, and preferably has an average transmission for visible and infrared light of at least 95%.

Particularly suitable materials for use as the transparent front electrode include transparent conductive oxides (TCO). Transparent conductive oxides (TCO) are doped metal oxides that are electrically conductive and have a comparably low absorption of light. TCOs can have greater than 80% transmittance of incident light as well as conductivities higher than 10 ⁴ S/cm (i.e. resistivity of ~10 ^~4 Ω-cm) for efficient carrier transport. Examples of suitable TCO materials include indium tin oxide (ITO), hydrogen doped indium oxide (IOH), aluminium doped zinc oxide (AZO), fluorine doped tin oxide (FTO), indium-doped zinc oxide (IZO), niobium-doped titanium dioxide (Nb:Ti02), etc.. The first electrode 15 therefore preferably comprises of a layer of transparent conductive oxides (TCO). By way of example, the first electrode 15 can comprise of a layer of any of indium tin oxide (ITO), aluminium doped zinc oxide (AZO), fluorine doped tin oxide (FTO), indium-doped zinc oxide (IZO), and niobium-doped titanium dioxide (Nb:TiC>2). More preferably, the first electrode 15 therefore preferably consists of a layer of indium tin oxide (ITO). When the first electrode 15 consists of a layer of indium tin oxide (ITO) it is preferable that the layer has a thickness of from 100 nm to 200 nm, and more preferably of 150 nm.

Conventional techniques for fabrication of layers of TCO materials typically involve a magnetron sputtering process. To minimise ion impact damage to underlying layers, an additional protective inorganic buffer layer may be employed. Alternatively atomic layer deposition, or a remotely generated plasma may be used to deposit a layer of TCO, as described in WO2016/198898. Such methods do not require a high temperature annealing step.

The TCO preferably has a sheet resistance (R _s ) equal to or less than 50 ohms per square (Ω/sq) and an average transmission for visible and infrared light of greater than 90% (i.e. transmits at least 90% of light above 300nm in wavelength), and preferably has an average transmission for visible and infrared light of at least 95 %.

Device Structure - Perovskite Material

In the above described multi-junction photovoltaic devices, the first sub-cell 210 comprises a photoactive region that comprises a perovskite material. The perovskite material in the photoactive region of the first sub-cell 210 is configured to function as a light absorber/a photosensitizer within the photoactive region. The perovskite material then preferably has a band gap from 1.50eV to 1.75eV, and more preferably from 1.65eV to 1.70eV. The second sub-cell comprising the silicon heterojunction (SHJ) then preferably has a band gap of around 1.1 eV.

The perovskite material may have general formula (I):

[A][B][X] ₃ (I)

wherein [A] is one or more monovalent cations, [B] is one or more divalent inorganic cations, and [X] is one or more halide anions.

[X] preferably comprises one or more halide anions selected from fluoride, chloride, bromide, and iodide, and preferably selected from chloride, bromide and iodide. More preferably [X] comprises one or more halide anions selected from bromide and iodide. In some examples, [X] preferably comprises two different halide anions selected from fluoride, chloride, bromide, and iodide, and preferably selected from chloride, bromide and iodide, and more preferably comprises bromide and iodide.

[A] preferably comprises one or more organic cations selected from methylammonium (CH ₃ NH ₃ ⁺ ), formamidinium (HC(NH) ₂ )2 ⁺ ), and ethyl ammonium (CHsCh NhV), and preferably comprises one organic cation selected from methylammonium (CH3 H3 ⁺ ) and formamidinium (HC(NH ₂ )2 ⁺ )- [A] may comprise one or more inorganic cations selected from Cs ⁺ , Rb ⁺ , Cu ⁺ , Pd ⁺ , Pt ⁺ , Ag ⁺ , Au ⁺ , Rh ⁺ , and Ru ⁺ .

[B] preferably comprises at least one divalent inorganic cation selected from Pb ²⁺ and Sn ²⁺ , and preferably comprises Pb ²⁺ .

In preferred examples, the perovskite material has the general formula:

AxA'i-xB(X _y XV _y ) ₃ (IA)

wherein A is formamidinium (FA), A' is a caesium cation (Cs ⁺ ), B is Pb ²⁺ , X is iodide and X is bromide, and wherein 0 < x < 1 and 0 < y≤ 1. In these preferred embodiments, the perovskite material can therefore comprise a mixture of two monovalent cations. In addition, in the preferred embodiments, the perovskite material can therefore comprise either a single iodide anion or a mixture of iodide and bromide anions. The present inventors have found such perovskite materials can have band gaps in from 1.50eV to 1.75eV and that layers of such perovskite materials can be readily formed with suitable crystalline morphologies and phases. More preferably, the perovskite material is FAi. _x CsxPbl3-yBr _y .

In order to provide highly efficient photovoltaic devices, the absorption of the absorber should ideally be maximised so as to generate an optimal amount of current. Consequently, when using a perovskite as the absorber in a photovoltaic device or sub-cell, the thickness of the perovskite layer should ideally be in the order of from 300 to 600nm, in order to absorb most of the sun light across the visible spectrum. Typically, therefore, the thickness of the layer of the perovskite material is greater than 100nm. The thickness of the layer of the perovskite material in the photovoltaic device may for instance be from 100 nm to 1000 nm. The thickness of the layer of the perovskite material in the photovoltaic device may for instance be from 200 nm to 700 nm, and is preferably from 300nm to 600nm. In the above described multi-junction photovoltaic devices, the planar layer of perovskite material 11 in the photoactive region of the first/top sub-cell 210 preferably has a thickness from 350 nm to 450 nm, and more preferably of approximately 400nm.

The perovskite layer may be prepared as described in WO2013/171517, WO2014/04502 , WO2016/198889, WO2016/005758, WO2017/089819, and in the reference books "Photovoltaic Solar Energy: From Fundamentals to Applications" edited by Angele Reinders and Pierre Verlinden, Wiley-Blackwell (2017) ISBN-13: 978-1 118927465 and "Organic- Inorganic Halide Perovskite Photovoltaics: From Fundamentals to Device Architectures" edited by Nam-Gyu Park et al., Springer (2016) ISBN-13: 978-3319351 124.

Device Structure - Second Sub-cell Surface Profile

When developing a monolithically integrated perovskite-on-silicon multi-junction photovoltaic device one of the most important considerations is the interface between the perovskite sub- cell and the adjacent crystalline silicon bottom sub-cell. In this regard, as described in Schneider, B.W. et al and Filipic, M. et al. referred to above, conventional commercial crystalline silicon solar cells feature textured surfaces that are designed to reduce reflection and increase the optical path length, with these surface textures usually consisting of randomly distributed pyramids, prepared by etching along the faces of the crystal planes, or regular inverted pyramids. Deposition of perovskite subs cells on such textured surfaces has been described previously in WO2016/198898. Alternatively, silicon sub cells having a flat front surface can be used.

Alternative Device Structure - PERC sub-cell

As an alternative to the Aluminium Back Surface Field (AI-BSF) silicon sub-cell described in the first embodiment above, the silicon sub cell can consist of a passivated emitter and rear contact cell (known as a PERC cell). Such a structure is described, for example, by D. Zielke et al. "21.7 % Efficient PERC Solar Cells with AIOx Tunneling Layer", published in 26th European Photovoltaic Solar Energy Conference and Exhibition [201 1] pp 1 1 15-1 119, ISBN 3-936338-27-2.

A typical conventional PERC cell is shown schematically in cross section in Figure 4. Unlike in the case of the AI-BSF cell described above, the back surface of the p-type silicon wafer (300) is not directly in contact with an aluminium electrode (302) everywhere, but is provided with an electrically insulating layer comprising a bilayer having a first layer of deposited AI2O3, S1O2 or SiOx (304) followed by a subsequent deposited layer of S13N4 (306), This insulating bilayer is then patterned to form contact holes or vias, and then a layer of aluminium is blanket deposited and sintered. For a conventional PERC cell, the opposite surface of the wafer is provided with a thin diffused n-type layer (308) having a thin passivation layer (310), typically of S13N4 blanket deposited on top. On top of this passivation layer, on the side on which the light falls in use, silver electrodes (312) are printed. Typically, the top surface having the silver electrodes will have an uneven surface topography to increase efficiency. In Figure 4 the top surface is flat to more clearly indicate the separate layers.

In the present invention, such a PERC cell is modified such that the top passivation layer is patterned to provide contact holes or vias prior to fabrication of a top sub-cell comprising a p- i-n perovskite sub-cell where the p type layer of the perovskite cell is in electrical contact with the n-type diffused layer beneath the passivation layer (308). A multi-junction (in the present case a tandem cell) according to the present invention is shown schematically in Figure 5. Once again the top layer is shown to be flat to more clearly indicate the separate layers.

In both the PERC cell embodiment and the AL-BSF cell embodiment, the electrically insulating passivation layer on the silicon sub-cell must be provided with a plurality of openings.

In the present invention, the tunnel recombination junction (TRJ) to the top cell is made by selectively opening the passivating top layers of the AI-BSF or PERC cell. This can be achieved a number of techniques.

The first technique is known as laser ablation. Short pulse laser ablation is described, for example in "Extended study on short pulse laser ablation of dielectric layers" by Theobald, Mayerhofer, Grosser, Harney and Schneider, 28 ^th European Photovoltaic Solar Energy Conference and Exhibition, Paris 2014 pp 1231 - 35, or "Influence of ultra-short pulse laser ablation of silicon nitride passivation layers" by Trusheim, Schultz-Ruhtenberg, Smeets, Das and Wieduwilt, 26 ^th European Photovoltaic Solar Energy Conference and Exhibition, pp 1623 - 27.

A second technique is by selective area deposition of a reactive material, such as a printing a silver ink (such as for example Solamet (RTM) PV19a photovoltaic metallization - commercially available from Du Pont (UK) Ltd, Coldharbour Lane, Bristol BS16 1 QD, UK .

The opening of the passivating contact should be less than 5% of the total front area, preferably less than 1 %. A large number, meaning a high density, of small openings (<100 urn diameter, preferably <30 urn) is preferable over a smaller number of larger openings. The density of openings should be >10 cm ^-2 , preferably >200 cm ^-2 . After opening the contact, the perovskite sub-cell can be deposited, which may include the initial deposition of a highly transparent conductive layer (7). In another embodiment, the side of the p-type wafer without the diffused emitter is passivated over its full area by an oxide such as alumina or a nitride or a combination of both deposited by, for example, ALD or PECVD. The recombination contact is then made by selectively opening the passivating layer by either laser ablation or by selective area deposition of a reactive material, such as an aluminium ink. Laser opening of the rear dielectric (AIOx/SiNx) stack is a standard process in PERC cells. Again, the opening of the passivating contact should be less than 5% of the total front area, preferably less than 1 % (same details apply as above).

A device according to the present invention using a PERC bottom cell and a perovskite top cell was fabricated using laser ablation to form the contacts through a 50nm SiN _x passivation layer. The contacts were 30 microns in diameter and were formed in a regular hexagonal array having a 600 micron pitch. Experimental results for this tandem cell are shown in Figures 6 and 7.

The quantum efficiency measurements of Figure 7 were performed using a Bentham PVE300 system. The detector responsivity was calibrated using both Si and Ge reference cells between 300-1200nm. Measurements were made without external voltage bias. The response of the perovskite sub-cell was measured under red light ( >800nm) illumination and the silicon sub-cell was measured under blue light (<450nm) illumination.

The IV measurements of Figure 6 were made with a 2651A Keithley source meter in conjunction with a TriSol TSS300 AAA solar simulator from OAI. Illumination density was calibrated with a filtered reference cell. The voltage was scanned at 0.04375 V intervals at a slew rate of 170mV/s.

The underlying PERC silicon sub-cell had an open circuit voltage of 657 millivolts, a short circuit current density of 34.6 milliamps per square centimetre, a fill-factor of 79.2% and an efficiency of 18.0%. The fabricated tandem cell had an open circuit voltage of 1.63 volts and a similar efficiency. Efficiency will be able to be increased with optimization of the device stack to provide a higher efficiency than the single cell. The advantage of the present invention is to provide an easy, simple and inexpensive way of integrating a perovskite top layer into a fabrication scheme for a known PERC device.

In the above described examples, the multi-junction photovoltaic device could be considered to be monofacial, such that it is configured to only collect light and generate electricity through its front, sun-exposed face. However, the majority of the features described above are equally applicable to a bifacial multi-junction photovoltaic device that can collect light and generate electricity through both of its faces; the front, sun-exposed face and the rear face. In particular, the present inventors have recognised that the multi-junction photovoltaic device could be configured into a bifacial architecture.

Additionally, if desired a further perovskite-based sub-cell may be provided provided beneath the second sub-cell in order to boost the energy conversion efficiency of the second sub-cell with respect to the light absorbed from the rear side of the device.

It will be appreciated that individual items described above may be used on their own or in combination with other items shown in the drawings or described in the description and that items mentioned in the same passage as each other or the same drawing as each other need not be used in combination with each other.

Furthermore, although the invention has been described in terms of preferred embodiments as set forth above, these embodiments are illustrative only. Those skilled in the art will be able to make modifications and alternatives in view of the disclosure which are contemplated as falling within the scope of the appended claims.