AU652998B2 - Solar cell - Google Patents

Description

OPI DATF 07/09/q2 AOJP DATE 15/10l/92 APPLN. TD 11PS3 92 PCT MLIMSFR PCT/Cq2/lflf?? I OBER DIE SENS (PCT) (51) Internationale Paten tklassifi kation 5 H01L 31/052, 31/0236 Al (11I) Internationale Ver~ffentlichungsnummer: (43) Internationales Ver~iffentlicliungsdatum:- 20.

WO 92/14270 August 1992 (20.08.92)1 (21) Internationales Aktenzeichen. PCT/CH92/00022 (22) Internationales Anrneldedatum: 3. Februar 1992 (03.02.92) Prioritaitsdatcn: 334/91-0 4. Februar 1991 (04.02.91) CH (71) Anmelder (ffir alle Bestimmrungsstaaten ausser US): G-t- -S EL±SCHA-F--R-FOR-DERI-hG--E ER-I N-DIJ4 STR1 EGRI EN.TIE-R-TE-N-FO RS-H4NG--A-N4---EN- -SeHwVEI-ER-ISC-HEN-HOC HSCHU LENl-UN D-WEi-- 4ER N-S-T-TA-T1 NN-E-TH -ZE NTRUM41- FW)- -[G-H4CH]-HadeneggsteigeCH8092-'ztrrich-(eH)T- (72) Erfinder, and Erfinder/Anmelder (nur fiir US) MORF, Rudolf, Hans ECH/CHj; Ziegeiwies, CH-8314 Kyburg (CH).

(74) Gemneinsamer Vertreter: G ESE LLSC H AFT Z UR FO R DE- RUNG DER INDUSTRIEGRIENTIERTEN FOR- SCHUNG AN DEN SCHWEIZERISCHEN HOCH- SCHULEN UND WEITEREN INSTITUTIONEN ETH ZENTRUM (IFWV); Haldeneggsteig 4, D-9092 Z~rich (CH).

(81) Bestimmungsstaaten: AT (europaiisches Patent), AU, BE (europ~iisches Patent). CA. CH (europdiisches Patent), DE (europ~iisches Patent), DK (europdiisches Patent). ES (europaiisches Patent), FR (europ~isches Patent), GB (curoptiisches Patent). GR (europ~iisches Patent), HU, IT (europhiisches Patent), JP, LU (europgiisches Patent), MC (europ~iisches Patent), NL (europ~iisches Patent), SE (europaiisches Patent), US.

Ver~ffentlicht Alit internanionalein Reclierchenbericht.

A~

C

C c U1, (54)Title: SOLAR CELL (54) Bezeichnung: SOLARZELLE (57) Abstract In order to increase the efficiency of a (poly)crystalline solar cell, the invention proposes that the back of the cell be fitted with an asymmetric optically active incident-light relief grating designed to intensify absorption of red light and infra-red radiation in the vicinity of the band gap near I.1 micrometres. Linear, asymmetric onedimensional diffraction gratings mounted on the back of the solar cell are used to increase significantly the light captured in both directions of' polarization and at any angle of incidence. Also proposed is a simple process for the large-scale manufacture of solar cells with this type of grating structure, the main step in the process being a stamping operation.

(57) Zusammenfassung Zur Verbesserung des Wirkungsgrades einer (po- "L NJ ly)-kristallinen Solarzelle wird vorgeschlagen, dieselbe 16' an ihrer Rtickseite mit einem asymmetrischen, optisch wirksamen Auflicht-Reliefgitter zu versehen, um die Absorption von rotem und infrarotem Licht in der Mite des "Band Gaps", nahe bei 1,1 micro m zu verstairken. Lineare, asymmetrische, eindimensionale, auf der Rickseite der Solarzelle angebrachte Beugungsgitter werden verwendet, um den Lichteinfang in beiden Polarisationsrichtungen und bei beliebigen Einfallswinkeln wirkungsvoll zu erh~hen. Emn Verfahren zur einfachen, grosstechnischen Herstellung von Solarzellen mit derartigen Gitterstrukturen wird beschrieben, bei welchem als wesentlicher Verfahrensschritt emn Pragevorgang verwendet wird.

PCT/CH92/00022 Description: Solar Cell The invention relates to a photo-voltaic cell, in particular an Si solar cell, comprising a thin layer structure with a base and emitter, which is produced by doping or, in the case of the emitter also as an inversion layer, where the light enters the front of the layer structure and electrical energy can be removed at suitably arranged cell contacts of the layer structure.

It is known that many attempts to gather light effectively have been made in the course of constructing solar cells with a high degree of efficiency. According to theory it is possible to attain degrees of efficiency up to 30% in monocrystalline solar cells in connection with solar radiation in accordance with the Standard AM 1.5 Green "High Efficiency Solar Cells", Trans.

Tech. Publ., 1987, pp. 69 82). Degrees of efficiency of up to 24% have been attained with prototype cells of monocrystalline silicon and of 17% to 18% with polycrystalline silicon. However, commercially available cells for use on earth typically have degrees of efficiency of only 13 to 16% for monocrystalline silicon, for polycrystalline silicon 13%.

It is known that the degree of efficiency of Si solar cells can be improved by the use of thin Si wafers of 10 to 100 micro m thickness if the diffusion length of the minority charge carriers is great in comparison with the cell thickness, the recombination rate at the surface can be kept sufficiently small and the entering sunlight in the wavelength range between 400 nm and 1200 nm can be quantitatively absorbed by the Si solar cell in sufficient amounts.

For improved coupling-in of the solar light, grids of square pyramids of arbitrarily distributed sizes are formed, often in combination with AR layers, by anisotropic etching of (100)oriented silicon wafers. These pyramids simultaneously act in a reflection reducing manner, as means for increasing the path length of weakly absorbed light in the material and for capturing light (Rittner et al. in JAP 47(1976), p. 2999; Yablonovitch and Cody in IEEE Transact. on Electron Dev., Ed. 29 (1982), p. 300).

It has also been proposed as an effective means for capturing of light to dispose micro-grooves with a perpendicular orientation to each other on both sides of the cell. These and similar measures are based on geometric-optical considerations. Two disadvantages in particular result from this: on the one hand, the reflectionreducing properties to not attain the quality of a good antireflection coating with dielectric layers, on the other, the silicon surface is increased by these measures, because of which the recombination rate at the surface is unduly increased. Also, a surface is created as a rule by structuring the surface which is not the (100)-surface and therefore cannot be rendered passive as easily. Thus, with the known measures, two of the mentioned prerequisites for an efficient functioning of a this Si solar cell have been violated.

Furthermore, a solar cell is known from US 4436608 which contains on the back of the layer structure a two-dimensional hexagonal diffraction grid and which has on the front an antireflection coating, because of which incident light is totally reflected on the front following diffraction at the grid in the layer structure and in this way is partially captured (light trapping). The disadvantage in this connection lies in that the capture of light is poorly performed and that the cited production of the solar cell appears to be uneconomic.

-2- 3 It is the object of the invention to disclose a photovoltaic solar cell which optimally converts the energy of the available sunlight, has a low recombination rate at the surface and the efficiency of which does not require any large diffusion lengths of the minority charge carriers in the silicon material, as well as a process for its economical production.

According to one aspect of the present invention there is provided a photovoltaic solar cell, comprising a thin layer structure with a base and an emitter; an anti-reflection layer, where incident light enters a front surface of the layer structure and electrical energy can be removed at suitably arranged cell contacts of the layer structure; and an asymmetrical, optically acting surface relief diffraction grating is disposed on a back surface of the layer structure, said grating having geometric and optical structural parameters selected such that the light entering on the back is mainly diffracted in modes which only propagate within the layer structure and which undergo total reflection on one of the layers of said layer structure.

According to another aspect of the present 2 invention there is provided a process for making a photovoltaic solar cell, said solar cell comprising a thin layer structure with a base and an emitter, an anti- *reflection layer, where incident light enters a front surface of the layer structure and electrical energy can be removed at suitably arranged cell contacts of the layer 30 structure, and an asymmetrical, optically acting surface 7relief diffraction grating disposed on a back surface of the layer structure, said grating having geometric and optical structural parameters selected such that the light entering on the back is mainly diffracted in modes which 35 can only propagate within the layer structure and which undergo total reflection on one of the layers of said layer 3A structure, said process comprising the step of applying said surface relief diffraction grating, produced by an embossing process, to a back surface of said layer structure.

Exemplary embodiments of the invention will be described below by means of the drawing figures.

Fig. 1 is a sectional view of an Si solar cell, Fig. 2a is a sectional view of an embodiment of a symmetrical incident light relief grid on an Si solar cell, Fig. 2b is a sectional view of a first embodiment of an asymmetrical incident light relief grid of the invention on an Si solar cell, Fig. 3 is a sectional view of a second embodiment of an asymmetrical incident light relief grid of the invention on an Si solar cell, i* o o

'"N

Fig. 4a is a sectional view of a first embodiment of an embossing process in accordance with the invention for producing a grid structure on an Si solar cell prior to the embossing step, Fig. 4b is a sectional view of a first embodiment of an embossing process in accordance with the invention for producing a grid structure on an Si solar cell following the embossing step, Fig. 5 is a sectional view of a second embodiment of an embossing process in accordance with the invention for producing a grid structure on an Si solar cell, Fig. 6 is a sectional view of a third embodiment of an embossing process in accordance with the invention for producing a grid structure on an Si solar cell.

The layer structure of an Si solar cell illustrated in Fig.

I comprises in a manner known per se a thicker layer 1 of doped silicon, a thinner layer 2 of n-doped silicon, and their boundary layer 3 in the form of a p-n transition layer. The layer 2 is used as an emitter, the layer 1 is designated the base. A polycrystalline silicon layer of a thickness of 10 to 100 micro m can also take the place of the layers 1 to 3. The front 4 of the emitter layer 2 and the back 5 of the base layer 1 support cell contacts 6 and 7, from which the electrical energy converted by the cell is removed. The front 4 has a protective layer 8 of SiO(2), the thickness of which should not appreciably exceed 10 nm and which has further functions, to be described below. The light to be converted enters the front 11 of the solar cell from above in the direction of the arrow 10. The layer 12 is a layer of glass which is relatively thick and lends mechanical sturdiness to the entire solar cell. A possible embodiment of the coupling -4layers 13 and 15 consists in making them of TiO(2) and the layer 14 of Sio(2). The purpose of this coupling layer is a reflectionpoor coupling-in of light to the layer 8. TiO(2) can also be replaced by SiN(X) (silicon nitrite) with a refraction index of 2.2 to Light with a penetration depth of more than one half the cell thickness is only partially absorbed in the Si material. But light of lesser penetration depth is almost completely absorbed.

With a presumed cell thickness of 30 micro m, it follows from the known connection between penetration depth and light wavelength that measures for capturing light only need apply to wavelengths larger than approximately 800 nm. In this case the penetration depth is greater than 10 micro m. Furthermore, an upper wavelength limit of 1200 nm results from the fact that the photoelectrically effective absorption of silicon becomes zero at 1200 nm. Thus, only light between wavelengths of 800 to 1200 nm, which reaches the back of the cell, needs to be captured. The solar cell embodied in accordance with Fig. 1 provides this wavelength-specific light capture.

In the known embodiment in accordance with Fig. 1, the sunlight, after passing through the coupling layers 13 to enters into the front 4 of the 30 micro m thick cell. Thus, only light of a wavelength of more than approximately 800 nm impinges on the back 5 of the cell. If a simple flat mirror were disposed here, this light would pass through the Si material a second time and would then leave the cell following partial absorption. This leads to a doubling of the optically effective cell thickness. To increase the length of the light path in the silicon considerably beyond this, an optical reflection grid (16 in Fig. 2a) is placed on the back 5, instead of the flat mirror. Based on the special embodiment of this reflection grid, to be further described below, the light is bent back into the Si material, as indicated by arrows 17 and 17'. In the course of this, light is first primarily diffracted into higher order modes, for example first and/or second, and secondly the angle of diffraction is so large, that these modes undergo total reflection on the front of the cell and for this reason the path of the bent-back light in the Si is increased by at least a factor of four, i.e. in respect to the structure with the flat mirror by a further factor of two.

The electrical cell contacts 6, 7 at the front 4 and back must be embodied as line or point contacts, those on the back also as area contacts. The dielectric layer 8 of SiO(2) or Si(3)N(4) of a thickness of less than approximately 10 nm thus fulfills the following important functions: It is used to attain a sufficiently low recombination rate at the surface, and it permits the effective coupling-in of light thanks to is narrow thickness and its not too low refraction index.

An anti-reflection coating, consisting of three dielectric layers and shown in Fig. 1, is used for reducing the reflection losses on the front 11 of the Si solar cell. This coating can be applied either directly to the Si layer 2 or to the SiO(2) layer 8 of less than approximately 10 nm thickness, which is used to protect the Si surface of the emitter layer 2. As shown, the upper layer 13 of the three layers consists of TiO(2) and is approximately 15 nm thick. It lies on the center layer 14 of SiO(2), and is approximately 30 nm thick. This, in turn, lies on the third layer 15 of TiO(2) of a thickness of approximately nm, which in turn lies on the protective layer 8. The antireflection coating 13 to 15 has been optimized in relation to the spectrum of the solar radiation in accordance with the Standard AM -6as well as in relation to the variable angle of incidence of the sunlight.

The incident light relief grid 16 illustrated in Fig. 2a for capturing light at the back 5 of the Si solar cell has a coating of a gold, silver or aluminum film. This grid 16 bends the light at a shallow angle back into the cell (arrows 17) in such a way, that the path length is increased. It was shown by numerically exact solutions of the Maxwell equations and from first tests with monocrystalline Si and one-dimensional grids that an increase in the path length by a factor of at least four is attainable, if a grid 16 with a rectangular profile is used, which has a period of 310 nm and a depth (step height) of 120 nm to 150 nm. Because linear (one-dimensional) grids only diffract light in one polarization direction very effectively, such a grid will bend back into the Si only approximately one half of the non-polarized light falling on the grid, with the mentioned gain in path length.

Light of the other polarization is diffracted less effectively.

It is possible to use crossed grids, which bend back light of all polarizations into the si material to make the improvement fully effective even for non-polarized light. In accordance with the above recited structural principles for preventing any mutual interference between electrical and optical properties, the grid 16 has been applied on the support layer 18 which rests on the base layer 1 of the cell shown in Fig. 2a. The thickness of the support layer 18 advantageously is below 0.5 micro m.

The operation of light capture will now be explained by means of Fig. 2a.

After passing through the solar cell, in particular the layers 8 and 1, the incident wave 10 impinges on the grid 16. For -7the sake of simplicity the layers 12 to 14 of Fig. 1 are no longer shown. The wave 10 is diffracted at the symmetrical grid 16, where the excited modes for right- and left-propagating waves 17 and 17' are necessarily equally strong. The grid has been optimized such that as little light of the zero order as possible is reflected, which has been indicated by the small arrow 23. The waves 17 and 17' undergo total reflection at the cell surface 11 at the latest and impinge a second time on the grid 16 as reflected waves 24 and 24', where they are again diffracted and leave the solar cell mainly as waves 25 and 25', so that they are no longer available for further light capture. Only a small portion of the waves 24 and 24' is reflected in the direction of the arrows 26 and 26'. This has to do with the symmetry in regard to time reversal. The operation of time reversal means the reversal of all path directions of all modes and, if the grie _S loss-free, the time-reversed situation is also a solution of the Maxwell equations. The result of this is that the diffraction efficiency for the beam 25', generated by the beam 24' is as good as that for the beam 17 and 17', generated by the beam 10. The fact that a main portion of the light energy is coupled out with the light waves 25 and 25' represents a limitation of symmetrically arranged grids.

Asymmetrical grid structures in accordance with the invention behave completely different, as will be shown by means of Fig. 2b. It is the purpose of these assymetrical grids to suppress the coupling out of the guided modes 24 into the mode as much as possible. This is possible if the diffraction efficiency for perpendicular light incidence shows characteristic preferred directions and is very small in the directions opposite to those. Fig. 2b shows a possible embodiment of the solar cell in section. On the back 5 the base layer 1 of p-silicon has the -8protective layer 9 of SiO(2), on which an incident light relief grid 16', embedded in a support layer 18, has been applied. As low as possible a surface recombination is assured by the protective layer 9. Thus, this embodiment permits the improvement of the electrical surface properties of the Si solar cell, which is very important for thin cells.

With the embodiment of the solar cell shown in Fig. 2b, the electrical properties of the silicon and of the p-n transition therefore remain unaffected by the added grid 16' for light capture. For meeting the previously mentioned prerequisite (2) for a low recombination rate at the surface, it is already known to coat the Si surface with SiO(2). However, if the SiO(2) layers are not to disturb the effect of the optical arrangement for light capture, their thickness must be kept below approximately 10 nm.

But if the grid 16' is intended to bend back in a useful way the incident light at a shallow angle into the Si material, the refraction index ci the support layer 18 must be greater than approximately 2.6. Also, their absorption coefficient in the wavelength range above 800 nm must be sufficiently low, so that light is not absorbed in this layer 18. Anorphous silicon, SIPOS (SiO(x), i.e. semi-insulating polysilicon) or non-stoichiometric silicon nitride SiN(x) can be worked into a support layer material in such a way that the mentioned conditions are met. A particularly welcome property of amorphous silicon with a band gap of 1.7 eV is the fact that it does not absorb in the wavelength range of 800 to 1200 nm of interest here. Refraction indices larger than 2.6 can also be attained with composite material.

Composite materials can also be produced with plastic deformable polymers. In this case the grid 16' can be usably produced by an embossing process.

In place of the described anti-reflection layer 13 to other reflection-reducing means can be used. Phase grids with a suitable step height or phase retardWion are suitable for this, particularly for the broad wavelength range of 400 nm to 1200 nm.

If the solar cell is employed in combination with preceding optical collector systems, the range of the angle of incidence of the light to be processed can be kept small. By means of a correctly selected grid geometry, only a right-propagating wave 17 of higher order, mainly of the first and/or second order, is excited by the incident wave 10. As small as possible a portion is reflected in the zero order, which is indicated by the small arrow 23. The wave 17' of the order of minus 1 or minus 2 is practically riot excited, which is a result of the asymmetrical grid arrangement. The wave 17 is totally reflected at the SiO(2) layer 8 and is again diffracted as wave 24 at the grid 16', but is for all practical purposes not reflected into a wave 25 of the zero order, which would be coupled out, but into higher modes.

These are the cause of repeated reflections in the silicon layer, which has been indicated by the arrow 26. Thus, the assymetrical grid choice makes it possible to effectively prevent the coupling out of the wave 25. This constitutes an essential prereq "site for the effective light capture in the active Si material.

The assymetrically embodied grid structure illustrated in Fig. 3 has step-like indentations cut into a metallic 16', consisting of Al, Ag or Au, as will be described further down by rnans o. the process. The metallic layer 16' borders the silicon layer 18. The grid structure has a periodicity of 660 nm, where each step length 19 is 220 nm and each associated step depth 20 is nm.

This grid structure was optimiz( y the exact solution of the Maxwell equations for a cell thicknuss of 30 micro m. In the course of this, an additional excellent property resulted in the light capture in both polarization directions being equally good (E-vectors and H-vectors parallel to the grid lines). The embodiment of a two-dimensional grid can therefore be omitted.

Diffuse light can be captured just as well. The efficiency of light capture can be quantified by the specification of an effective equivalent cell thickness. The effective equivalent cell thickness for the grid shown in Fig. 3 is approximately 6 to 7 times the physical thickness, i.e. a 200 nm thick solar cell with a planar reflector on the back would correspond to a 30 micro m thick solar cell with such a Blaze grid on the back. The grid geometry must be adapted for other cell thicknesses, because the spectrum of the incident light at the back of the cell depends on the cell thickness. Based on the absorption characteristics of (poly-)crystalline silicon or other semiconductors with an indirect band gap, this spectrum is displaced towards longer wavelengths with increasing cell thickness. Therefore the grid must have its greatest efficiency in another wavelength range.

Optimization has the result that for cell thicknesses above micro m the optically effective cell thickness for diffuse, unpolarized light (standard spectrum AM 1.5) always lies at least at a factor of six. For collimated light, factors of up to twenty are possible.

The manufacturing process for Blaze grids will be addressed in what follows.

The commonly customary manufacturing processes for submicron grid structures comprise holographic illumination of a substrate coated with a photoresist, development of the -11photoresist, and subsequent substrate etching with use of the structured photoresist as etching mask.

A related process consists in that additionally chromium, for example, is vacuum-evaporated at an angle to the structured photoresist. This is performed prior to the etching process to achieve increased selectivity of the etching step. This process is complicated and impairs the quality of the passivation layer.

An embossing process with the process variations listed below is suggested for producing the grid structure in accordance with the invention: 1. A portion of the back of an Si solar cell is shown in Fig. 4a.

A highly reflecting dielectric 18, capable of being embossed, is applied to the silicon material 1, which has been rendered passive by means of an SiO(2) layer 9 on the back. This layer consists for example of a thin TiO(2) layer 18 (or of another dielectric with an even greater dielectric constant). For generating the grid structure, the layer 18 is now embossed with a metal die 27, which can be made of nickel. The direction of embossing is indicated by the arrow 28. Use of a sol-gel process represents a possible realization of the application of the layer 18.

Subsequently (possibly after thermal hardening (pre-baking) and a tempering step at 400 to 500°C), the grid (Fig. 4b) is coated with a metal layer 16', which can consist of Al, Ag or Au.

This can be done by vacuum evaporation or sputtering. This ends the formation of the grid.

-12- The dielectric constant of the applied material 18 must be as close as possible to that of the silicon. Otherwise the waves of higher order which must be excited in the photo-electric layer cannot propagate in the dielectric. Therefore strong reflections of these modes occur at the phase boundary between the dielectric and the silicon. This leads more to light capture within the metallic reflector instead of within the silicon, as desired.

2. The grid is produced independently of the solar cell.

In this case two ways of production are defined: a. The grid is embossed into a suitable film (Fig. preferably in a very thin film, the thickness of which is less than 10 micro m. Mylar, polyester, polycarbonate or polyvinyl chloride for example can be used as the material for the film.

The embossed film is subsequently metallized with Al, Ag or Au 16', and then coated with a dielectric or semiconducting material 18, the dielectric properties of which are as close as possible to those of silicon. Flattening of the profile is desired. A small thickness is advantageous, if possible in the micron or sub-micron range. Amorphous silicon, which is practically absorption-free between 800 and 1200 nm, is particularly suitable as material 18.

b. The grid structure is embossed into a metal-coated plastic film. Fig. 6 shows a plastic film 29 with an applied metal layer 16' following the embossing step.

This method is known to be used today for producing holograms, such as are used on credit cards or other identification cards for reasons of security or for decorative purposes. As in process this metal grid is then coated with a thin layer 18 of a dielectric (thickness in the micron or sub- -13micron range), again preferably with amorphous silicon, which is sputtered on, for example.

c. The grid structure is embossed directly into a metal film.

As in process this metal grid is then coated with a thin layer 18 of a dielectric, again preferably with amorphous silicon, which is sputtered on, for example.

In the three cases b. and the produced "grid film" is applied to the back of an already produced solar cell, as is indicated in Fig. 6 by the Si layer i, the SiO(2) layer 9 and the cell contacts 7.

Application of the "grid film" is preferably performed in a vacuum, which is indicated by the arrows 30. The connection (bonding) to the back can be aided by the application of an electric field. The small thickness of the "grid film" and the use of a flexible plastic material assure good adhesion properties. Use of a polar film substrate can aid with the adhesion properties.

Although the exemplary embodiments described relate to silicon solar cells, the embodiments of the invention can also be used for other photo-voltaic solar cells, for example GaAs solar cells, CuInSE(2) solar cells, CdS solar cells, Ge solar cells and Se solar cells.

-14-

Claims (21)

1. A photovoltaic solar cell, comprising a thin layer structure with a base and an emitter; an anti-reflection layer, where incident light enters a front surface of the layer structure and electrical energy can be removed at suitably arranged cell contacts of the layer structure; and an asymmetrical, optically acting surface relief diffraction grating is disposed on a back surface of the layer structure, said grating having geometric and optical structural parameters selected such that the light entering on the back is mainly diffracted in modes which only propagate within the layer structure and which undergo total reflection on one of the layers of said layer structure.

2. A photovoltaic solar cell in accordance with claim i, wherein, in the surface relief diffraction grating, light energy flow in higher order modes displays a preferred direction even when incident light i7 20 perpendicular and is as large as possible, while the light flow is small in the opposite direction, and a diffraction 'efficiency in this preferred direction does not have mirror-sjmnmetry.

3. A photovoltaic solar cell in accordance with claim 1 or 2, wherein the surface relief diffraction "o grating is applied to a support material which, in the oO wavelength range of the light employed, has a refraction index of more than 2.6, said support material being located between the relief grating and the layer structure.

4. A phntovoltaic solar cell in accordance with any one of claims 1 to 3, wherein the solar cell is an Si solar cell including a passivation layer, which is located between one of the surface relief diffraction grating or [I 16 the support material, and the layer structure, said passivation layer rendering the surface passive.

A photovoltaic solar cell in accordance with any one of claims 1 to 4, wherein the surface relief diffraction grating is one-dimensional and the diffraction efficiency of which does not have mirror-symmetry.

6. A photovoltaic solar cell in accordance with any one of claims 1 to 5, wherein the surface relief diffraction grating has periodical, step-like indentations.

7. A photovoltaic solar cell in accordance with any one of claims 1 to 5, wherein the surface relief diffraction grating has a periodic sawtooth structure.

8. A photovoltaic solar cell in accordance with any one of claims 1 to 6, wherein the surface relief diffraction grating has a periodicity of 600 to 700nm and two steps each of a depth of 50 to 60nm and 100 to 120nm.

9. A photovoltaic solar cell in accordance with any one of claims 1 to 8, wherein said surface relief diffraction grating is metallised for a wavelength range between 800nm and 120 0 nm, in one of a monocrystalline and polycrystalline solar cell.

A photovoltaic solar cell in accordance with any one of claims 1 to 9, wherein the layer structure includes n-doped silicon and p-doped silic-' and has a thickness 25 between 10gm and 100pm.

11. A photovoltaic solar cell in accordance with any ,one of claims Ito 9, wherein the layer structure includes an inversion border layer and p-doped silicon or n-doped silicon and has a thickness between 10pm and 100gm. 17

12. A photovoltaic solar cell in accordance with any one of claims 3 to 11, wherein the support material of the surface relief diffraction grating has a thickness of less than 500nm and includes one of amorphous silicon, semiconducting polysilicon, non-stoichiometric silicon nitride, and a composite of polymers with metallic and/or semiconducting particles.

13. A process for making a photovoltaic solar cell, said solar cell comprising a thin layer structure with a base and an emitter, an anti-reflection layer, where incident light enters a front surface of the layer structure and electrical energy can be removed at suitably arranged cell contacts of the layer structure, and an asymmetrical, optically acting surface relief diffraction grating disposed on a back surface of the layer structure, said grating having geometric and optical structural parameters selected such that the light entering on the back is mainly diffracted in modes which can only propagate within the layer structure and which undergo total reflection on one of the layers of said layer structure, said process comprising the step of applying said surface relief diffraction grating, produced by an embossing process, to a back surface of said layer structure.

14. A process in accordance with claim 13, wherein 25 the embossing process is performed on a material which can be embossed and the embossed material is deposited on the :"back of the cell. aom

15. A process in accordance with claim 13 or 14, :wherein, after the embossing process is completed, a metal 30 layer including one of Al, Ag and Au, is vacuum-evaporated or sputtered on the back of the cell.

16. A process in accordance with claim 13, wherein the embossing process acts on a plastic film, the thickness 3' i! 1$ 18 of which is less than 10pm and which comprises one of mylar, polyester, polycarbonate and polyvinyl chloride, at the end of the embossing process the film is metallised in a first step with Al, Ag, or Au, in a second step is coated with a dielectric, where the embossing profile is flattened out and the applied layer is less than lm, ang the grating film created in this manner is applied to the back surface of the solar cell.

17. A process in accordance with any one of claims 13 to 16, wherein the embossing process acts on a plastic film provided with an applied metal layer and is subsequently coated with a dielectric, which has a layer thickness of less than ltm.

18. A process according to claim 17 wherein said dielectric consists of sputtered silicon.

19. A process in accordance with claim 17 or 18, wherein the embossing process is made directly into the metal layer without substrate.

A photovoltaic cell substantially as herein described with reference to as illustrated by any one or more of Figures 2b to 6 of the accompanying drawings.

21. A process for making a photovoltaic solar cell substantially as herein described with reference to and as illustrated by one cr more of Figures 2b to 6 of the 25 accompanying drawings. 0* Dated this 2nd day of February, 1994. GESELLSCHAFT ZUR FORDERULNG DER INDUSTRIE- ORIENTIERTEN FORSCHUNG AN DEN SCHWEIZERISCHEN HOCHSCHULEN UND WEITEREN INSTITUTIONEN ETH ZENTRUM (IFW) By its Patent Attorneys: GRIFFITH HACK CO. Fellows Institute of Patent Attorneys of Australia. Abstract: For the improvement of the degree of efficiency of a (poly) -crystalline solar cell it is proposed to provide the same on its back with an asymmetrical, optically effective incident light relief grid, so as to increase the absorption of red and infrared light in the vicinity of the "band gaps" near 1.1 micro m. Linear, asymmetrical, one-dimensional diffraction grids, applied to the back of the solar cell, are employed to increase light capture in both polarization directions and at arbitrary angles of incidence. A method for the simple, mass production of solar cells with such grid structures is described, wherein an embossing step is employed as the essential process step. (Fig. 2b) -19-