Field of invention

The invention relates to thin film photovoltaic modules, especially modules involving a discontinuous layer of fibers deposited on a substrate, a substantially continuous transparent conductive oxide film provided on the substrate and on the fiber layer and a silicon layer deposited on the substantially continuous transparent conductive oxide film. Such modules show an increased light harvesting into the semiconductor layer or layers of the thin film photovoltaic module.

Background of the invention

Thin film photovoltaics (TFPV) is an important application converting solar energy into electricity. In the TFPV modules (or TFPV cells; terms "module" and "cell" are used interchangeably in the present application to denote means for photovoltaic generation) a glass substrate typically works as a deposition substrate onto which the necessary layers are deposited. The layers typically include an index matching layer (typically silicon oxynitride or silicon oxycarbide), the transparent conductive oxide (TCO) layer, at least one semiconductor layer or silicon layer (terms "silicon layer" and "semiconductor layer" are used

interchangeably in this text to denote the active parts of a photovoltaic cell exhibiting the photovoltaic effect) for the actual photoelectric conversion, and the back contactors. An antireflective coating may be applied on the glass to increase solar radiation harvesting.

The semiconductor of the TFPV module may be made from amorphous silicon (a-Si), micro or nanocrystalline (μ/η) silicon (μ/η-Si) or cadmium telluride (CdTe). The cells of fourth TFPV type, Cu-ln-Ga-Se (CIGS) thin film solar cells, are built on glass, metal foil or plastic substrates and do not require a TCO layer between the substrate and the semiconductor.

The advantage of the silicon based TFPV modules is that they use about one- hundredth of the amount of silicon used by crystalline silicon PV. The a-Si solar cells unfortunately suffer from low efficiency and thus the CdTe modules, which provide the lowest cost-per-megawatt and the CIGS modules, which grant the highest efficiency are serious competitors to a-Si modules.

Stacking thin layers of Si-based semiconductor layers is a valid method for improving the competitive edge of the a-Si-modules. Such multi-junction cells apply either an underlying a-SiGe or μ/η-Si layer below the a-Si layer. This structure allows better light trapping and thus the efficiency of the cell increases.

The TCO layer obviously needs to carry out the electrons generated in the semiconductor layer of the PV cell. Thus the electrical resistivity of the TCO layer needs to be low. There is, however, an intrinsic level of the conductivity which arises from the fact that the light absorption into the TCO layer manufactured from fluorine doped tin oxide (FTO) increases together with the electrical conductivity: if the TCO layer is made very thick, the absorption increases simply due to the increased layer thickness. If the carrier concentration of the TCO layer is increased, there will be a substantial absorption at higher wavelengths due to the plasma oscillations, although the visual transmission is good. Thus the FTO layer thickness generally needs to be lower than 1 micrometer, the carrier concentration less than 3 x 10 ²⁰ cm ^"3 and the carrier mobility around 30 cm ² Vs. This relates to a typical square resistance of 1 0 - 14 D/D.

Carrying out the electrons is exactly what the TCO layer is for In the CdTe modules. In the Si-based thin film PV modules, the TCO layer has, however, also another function: light trapping. The light absorption of the Si-films is rather poor but can be improved by guiding the light so that it travels parallel to the semiconductor surface rather than opposite to it. This can be realized by scattering the light by the structure of the TCO-layer/semiconductor interface. A TCO made from fluorine doped tin oxide (FTO) layer may be deposited such that the grown crystal itself provides the required structure for visible light scattering. However, scattering longer IR-wavelengths requires larger discontinuities at the TCO/semiconductor interface. In the prior art such structure has been realized by depositing μ/η-particles or by growing larger inclusions with modified tin oxide layer deposition chemistry.

FTO can be manufactured from various precursors including e.g. stannic and stannous chlorides and organotin compounds such as monobityltin trichloride (MBTC), fluorine acid, trifluoroacetic acid, etc. The characteristics of the FTO film, e.g. the surface morphology are influenced by the precursor used. The reaction chemistry is also greatly influenced by additives, e.g. the role of water vapor on the reaction kinetics is widely recognized.

Description of the state of the art The Japanese patent application JP 2001-322552, Asahi Glass Company, Ltd.,

October 19, 2001 , describes a substrate with a transparent conductive oxide film, comprising a substrate and transparent conductive oxide film provided on the substrate and constituted by a plurality of ridges and plurality of flat portions, wherein the surfaces of the ridges and the flat portions have many continuous micron-size protrusions. The application also describes a process for producing, on a transparent substrate, discontinuous small ridges by an atmospheric pressure chemical vapor deposition (CVD) method. The structure is aimed to increase light scattering at a long wavelength, as it is effective to increase the surface roughness of the TCO film.

The term "ridge" is generally understood to imply a substantial length associated with a raised feature. However, the application JP 2001-322552 has no teaching of substantial length to the ridges and the micrographs presented in the application appear to show hill-like structures, i.e. hemispherical structures, not "ridges". Also the manufacturing method described would produce such hemispherical structures, not ridges.

Hemispherical structures having a size of roughly the wavelength of light scatter light at longer wavelength region of solar radiation. However, as such structures miss the freedom of controlling or tuning the width and length of the structure individually, they are not optimal for controlling the scattering properties. Also, as the micron-size structures are quite thick, they show a certain level of light absorption, even if they are manufactured from electrically non-conducting material. Thus the prior art shows considerable technical problems. The combination of fibers and conducting metal oxide films as such is known from prior art. United States Patent Application publication US 2010/0002357 A1 , II Doo Kim, et al., 7.1.2010, describes a porous conducting metal oxide electrode prepared by depositing a porous conducting metal oxide film comprising a conducting metal oxide film layer having a network structure of nanofibers, comprising nanograins or nanoparticles, on at least one surface of a current collector, and a conducting metal oxide coating layer on the network layer of the porous conducting metal oxide. Said publication is related to supercapacitor applications, and does not teach anything related to photovoltaic cells. In particular, the publication does not teach anything on the advantages which can be achieved by using a fiber layer in in TFPV modules. Summary of the invention

The aim of the current invention is to introduce a product which solves the problems of the prior art. The aim of the current invention is also to introduce a process for producing such a product.

The invented product comprises a module with a transparent conductive oxide film. At least one single discontinuous layer of fibers is deposited on a substrate and the layer may be at least partially embedded into the substrate. A substantially continuous transparent conductive oxide film is provided on the substrate and on the fiber layer. Additionally, a refractive index matching layer may be provided on the substrate and on the fiber layer and/or a diffusion barrier layer may be provided on the substrate. The aim of the

discontinuous layer of fibers is to effectively scatter light to the semiconductor layer of the thin film photovoltaic cell. The aim of the refractive index layer is to improve light harvesting into the structure. The aim of the diffusion barrier layer is to reduce alkaline ion diffusion into the transparent conductive oxide layer. At least one silicon layer is deposited on the substantially continuous transparent conductive oxide film to achieve the photovoltaic effect. It is specifically beneficial to scatter the long-wavelength solar radiation by the discontinuous layer of fibers. For this reason, in one embodiment of the invention, the diameter of the fiber is beneficially close to the light wavelength, typically less than 2500 nm, which is effectively the long wavelength limit of the usable solar radiation. The structure comprising a discontinuous layer of fibers deposited on the substrate and a substantially continuous transparent conductive oxide film provided on the substrate and on the fiber layer has a higher haze than a structure without a discontinuous layer of fibers deposited on the substrate and with a substantially continuous transparent conductive oxide film provided on the substrate, especially at a wavelength region from 800 to 1000 nm. Thus the invented structure effectively scatters light at long solar wavelengths. According to this embodiment, the discontinuous layer of fibers consists essentially on single layer of individual fibers. In another embodiment of the invention the diameter of the fiber is beneficially much less than the infrared wavelength, i.e. the fiber diameter is beneficially 50 - 500 nm and more beneficially between 100 and 300 nm. Such fibers are deposited as a mesh of fibers and the discontinuous layer of fibers is formed by multiple layers of these thin fibers, e.g.

approximately ten layers of fibers with a diameter of about 200 nm. According to this embodiment, the discontinuous layer of fibers consists on multiple layers of individual fibers.

The aspect ratio of the fiber, i.e. the ratio of the fiber length to its diameter, is a parameter which may be used to optimize the light scattering from the fiber layer. As the light scattering depends on other parameters as well, some of which will be described below, the aspect ratio (fiber length: fiber diameter) may show wide variation, i.e. from 1 :1 to °°:1 , i.e. to practically infinite fibers, and preferably from 2:1 to 100: 1.

The scattering properties and especially the light absorption properties of the discontinuous fibers depend greatly on the fiber material and on the filling of the fibers. Thus the fibers may show at least a partially hollow or at least a partially multilayered or porous structure.

The material of the fibers in the substantially discontinuous fiber layer is preferably metal oxide, most preferably a metal oxide or a combination of metal oxides with low light absorption. Such metal oxides may include e.g. oxides of Mg, Ca, Sr, Ti, Zr, Hf, Ce, Zn, Al, Ga, In, Si, Ge, Sn or Pb. Most preferable metal oxides are oxides of Zn, In and Sn. It is beneficial to deposit the fibers of the substantially discontinuous layer of fibers in such a way that the fibers have an essentially random orientation. This produces angular scattering patterns which are much smoother than scattering patterns from oriented fibers or essentially spherical ridges. The coverage of the substrate by the fibers should be such that light is effectively scattered but the light absorption to the fiber layer is minimized. Thus the coverage of the substrate by the discontinuous layer of fibers is preferably between 10 and 70%. On the contrary, the coverage of the continuous transparent conductive oxide layer is essentially 100%.

The substantially continuous transparent conductive oxide film provided on the substrate and on the fiber layer is beneficially provided with a surface structure where the outcroppings have a height between 30 nm and 200 nm and the distance between the adjacent outcropping tops is preferably between 30 nm and 200 nm.

The TFPV module with the discontinuous layer of fibers and the substantially continuous transparent conductive oxide film deposited on the substrate also comprises a silicon layer deposited over the substantially continuous transparent conductive oxide film.

The silicon layer is preferably deposited with an atmospheric pressure chemical vapor deposition (APCVD) process. However, it may obviously be deposited by various other processes, such as by Plasma Enhanced Chemical Vapor Deposition (PECVD) or by sputtering.

Brief description of the drawings

In the following, the invention will be described in more detail with reference to the appended schematic drawings, in which

Fig. 1 shows the principle drawing (side view) of the invented module with a transparent conductive oxide film, comprising a discontinuous layer of fibers deposited on the substrate, a substantially continuous transparent conductive oxide film provided on the substrate and on the fiber layer and a silicon layer deposited on the substantially continuous transparent conductive oxide film;

Fig. 2 shows a top view of the invented module without the silicon layer; Fig. 3 shows an embodiment of the process for manufacturing the discontinuous layer of fibers; and

Fig. 4 shows an embodiment of the process for manufacturing the transparent conductive oxide film.

For the sake of clarity, the figures only show the details necessary for understanding the invention. The structures and details which are not necessary for understanding the invention and which are obvious for a person skilled in the art have been omitted from the figures in order to emphasize the characteristics of the invention.

Detailed description of preferred embodiments

Figure 1 shows a thin film photovoltaic module 1 with a transparent conductive oxide film 3. Module 1 comprises a discontinuous layer of fibers 2 deposited on the substrate 1 and a substantially continuous transparent conductive oxide film 3 provided on the substrate and on the fiber layer 2. The average length of the fibers is L and the average diameter of the fibers is D. The fiber length L is longer than the diameter D, but the fiber length can be varied in a large range depending on the structure of the overall system, possibly including a diffusion barrier 6, index-matching layer 5 and a silicon layer 7. The index-matching layer 5 and the diffusion barrier 6 should be understood as functional layers, and both functions can be integrated into a single physical layer.

The average diameter D and the average length L of the fiber 2 is selected such that the fiber layer 2 effectively scatters light, especially at the upper end of solar radiation, i.e. at wavelengths higher than 800 nm. Such effective scattering may be proven by measuring the haze of the substrate comprising the discontinuous layer of fibers 2 and comparing that with the haze measured from a substrate which does not comprise the discontinuous layer of fibers 2. The haze measurement from the substrate with fibers shows a higher haze at wavelengths 800 nm - 1000 nm than the substrate without fibers. The haze is preferably measured by applying the ASTM standard D 1003-00.

The preferred average fiber diameter D is less than 2500 nm and more preferably less than 1000 nm. The fiber length L is longer than the fiber diameter D. The fiber length L may be practically infinite (in relation to light scattering so that the effect of the fiber ends to the scattering are minimal), or - if the fiber end effects are utilized for scattering - the ratio of the fiber length L to the fiber radius D (aspect ratio) may be as short as 2:1. Preferably the aspect ratio is between 2: 1 and 100: 1 , but it is outlined that the optimal average fiber length depends on the overall design of the TFPV module.

The material of the fibers 2 is selected so that the light absorption to the fiber materials is minimized. Suitable materials are metal oxides, preferably oxides of Mg, Ca, Sr, Ti, Zr, Hf, Ce, Zn, Al, Ga, In, Si, Ge, Sn, P, and most preferably oxides of Zn, In and Sn. Preferably the fibers 2 are hollow which still reduces light absorption to the fiber layer. The fibers 2 may also show a multilayer structure, which may be used to reduce light reflection in hollow fibers or the fiber may have a porous structure. Processes for realizing such structures are described below. Typically the fiber layer is held on the substrate by the TCO layer deposited on the fiber layer. However, the adhesion of the fiber 2 layer to the substrate 1 may still be improved by partially embedding the fibers 2 into the substrate 1. This may effectively be realized by heating at least the surface of the substrate 1 after the deposition of the fiber layer 2. Heating may be carried out in a furnace or by radiation or convection directed essentially towards the surface of the substrate 1. Most preferably heating is carried out with electromagnetic radiation having a wavelength of approximately 8-12 μιτι, such as the radiation from a carbon dioxide (C0 ₂ ) laser.

The transparent conductive oxide film 3 provided on the substrate and on the fiber layer is preferably provided with a surface structure where the outcroppings 4 have a height h between 30 nm and 200 nm and the average distance S between the adjacent outcropping tops is preferably between 30 nm and 200 nm. Such a structure effectively scatters the visible light of solar radiation. The average thickness of the TCO layer, H, is preferably between 400 nm and 1000 nm.

Figure 2 shows the top view of the invented substrate without showing the silicon layer. It is beneficial to deposit the fibers 2 in the substantially discontinuous layer of fibers so that the fibers have an essentially random orientation. This produces angular scattering patterns which are much smoother than scattering patterns from oriented fibers or essentially circular ridges. The coverage of the substrate by the fibers 2 should be such that light is effectively scattered but the light absorption to the fiber layer is minimized. Thus the coverage of the substrate by the discontinuous layer of fibers is preferably between 10% and 70% viewed from the top as in Figure 2.

Figure 3 shows a principle drawing of a preferred process for producing substrate 1 with a discontinuous layer of fibers 2 deposited on the substrate and with a substantially continuous transparent conductive oxide film 3 provided on the substrate and on the fiber layer. The discontinuous layer of fibers 2 is produced by electrospinning fibers comprising organic and inorganic material on the substrate 1 in the phase A and removing the organic material from the fibers 2 by substrate heating in the phase B. This produces a substrate 1 with inorganic fibers 2, as shown in Figure 3C.

Electrospinning 100 is a process by which organic fibers 2*, typically polymer fibers, with diameters D* ranging from a few nanometers to several micrometers can be produced using a electrostatically driven jet of a solution. The high voltage electrostatic field 103 charges the surface of the solution 101 and induces the ejection of a liquid jet 102 through a spinneret. In a typical process, an electrical potential HV is applied between the liquid jet 102 and the grounded target substrate 1. When the applied electric field 103 overcomes the surface tension of the jet 102, a charged fibers 2* are ejected. The jet 102 exhibits bending instabilities caused by repulsive forces between the charges carried with the jet 102. The jet 102 extends through spiraling loops and as the loops increase in diameter the fibers 2* grow longer and thinner until they solidify and are collected on the substrate 1. The spiraling mechanism as well as the spinneret dimensions and electrostatic field can be used to control the diameter D* and length L ^* of the organic fibers 2*. When producing the fibers 2, essential for the current invention, a further component, such as an acetate compound of a metal is added to the solution 101. After the

electrospinning process the fibers 2 ^* are provided to a heat treatment (calcinations), which turns fibers 2* into inorganic oxide fibers 2. Such further component/ organic component solutions include, but are not limited to, aluminum di-sec-butoxide / ethylacetoacetate (for Al ₂ 0 ₃ ), titanium tetraisopropoxide / polyvinyl pyrrolidone (for Ti0 ₂ ), zinc acetate / polyvinyl acetate (for ZnO). Also ZnCI can be used as a precursor for fabricating ZnO fibers as well as SnCI ₂ 2H ₂ 0 for Sn0 ₂ fibers. The method is not limited to the fabrication of single metal oxides, but e.g. Sn0 ₂ -ZnO multicomponent fibers can be fabricated, too. The heat treatment is typically carried out in air at 400-800°C, preferably at around 500°C. As described earlier, fibers with hollow cores are preferred for low light absorption into the fiber layer. Such hollow fibers can be produced e.g. by using a spinneret with two coaxial capillaries, through which heavy mineral oil and the precursor solution are co-spun to form fibers with core/sheath structure. The oil can be extracted with octane or all organic material may be removed during heat treatment, which leaves hollow metal oxide fibers 2. Alternatively, inorganic microtubes with precisely defined nanoscale walls are produced by employing atomic layer deposition (ALD) on electrospun polymer fibers. In one embodiment, hollow Al ₂ 0 ₃ fibers 2 where produced by first electrospinning polyvinylalcohol (PVA) fibers 2*, Coating the fibers 2* with an Al ₂ 0 ₃ layer using AI(CH ₃ ) ₃ (TMA) and H ₂ 0 as precursors in a conventional ALD process and removing the organic part by heat treating the fibers 2* in air at 400°C for several hours. It is obvious for a person skilled in the art that such a procedure, as well as the spinneret with various coaxial capillaries, may also be used to produce multilayer fibers 2.

The organic part of the precursor solution for the electrospinning process has a great effect on the properties of the fiber 2. When miscible solvents (with immiscible polymers) are used, highly porous fibers 2 can be obtained. Such porous fibers provide an alternative degree of freedom for designing the scattering pattern of the fiber layer 2.

Figure 4 shows an embodiment, where aerosol assisted chemical vapor deposition (AACVD) process is used for producing a transparent conductive oxide layer 3 on the substrate 1 and on the non-continuous layer of fibers 2. Liquid raw material is fed from a conduit 214 into an atomizer 206 producing ultra small liquid droplets 217, The raw material 203 is a combination of monobutyl tin chloride (MBTC) - trifluoroacetate (TFA) - methanol in a weight ratio of 3: 1 :5. The feed rate of the raw material is 20 ml/min. As the atomizing gas, nitrogen N ₂ is conducted into a gas conduit 208. The volume flow of nitrogen gas is about 20 l/min. Fine droplets 217 are discharged from the nozzle. The droplets are further guided onto the surface of glass 1 , whose temperature is approximately 550°C. The droplets evaporated in the vicinity of and pyrolyzed on the surface of the glass 1 , whereby a fluorine- doped tin oxide coating 3 is provided onto the surface of the substrate 1 and on the fiber layer 2.

A silicon layer, preferably an amorphous silicon layer 7 is further deposited on the transparent conductive oxide layer 3, preferably using atmospheric pressure chemical vapor deposition (APCVD) process with silane (SiH ₄ ) as the silicon precursor. It should be obvious to a person skilled in the art that as the technology advantages, the inventive concept can be implemented in variety of ways. The invention and its embodiments are not restricted to the above examples, but may vary within the scope of the claims.