Layer system for thin-film solar cells

Abstract

The present invention relates to a layer system (1) for thin-film solar cells (100) and solar modules, comprising an absorber layer (4), which includes a chalcogenide compound semiconductor, and a buffer layer (5), which is arranged on the absorber layer (4) and includes halogen-enriched Zn.sub.xIn.sub.1-xS.sub.y with 0.01x0.9 and 1y2,
wherein the buffer layer (5) consists of a first layer region (5.1) adjoining the absorber layer (4) with a halogen mole fraction A.sub.1 and a second layer region (5.2) adjoining the first layer region (5.1) with a halogen mole fraction A.sub.2 and the ratio A.sub.1/A.sub.2 is 2 and the layer thickness (d.sub.1) of the first layer region (5.1) is 50% of the layer thickness (d) of the buffer layer (5).

Claims

1. Layer system for thin-film solar cells, comprising: an absorber layer, which includes a chalcogenide compound semiconductor, and a buffer layer, which is arranged on the absorber layer and includes halogen-enriched Zn.sub.xIn.sub.1-xS.sub.y with 0.01x0.9 and 1y2, wherein the buffer layer consists of a first layer region adjoining the absorber layer with a halogen mole fraction A.sub.1 and a second layer region adjoining the first layer region with a halogen mole fraction A.sub.2 and a ratio A.sub.1/A.sub.2 is 2 and a layer thickness of the first layer region is 20% of a layer thickness of the buffer layer.

2. Layer system according to claim 1, wherein the ratio A.sub.1/A.sub.2 is from 2 to 1000.

3. Layer system according to claim 1, wherein an amount of the halogen in the first layer region amounts to an area concentration of 1.Math.10.sup.13 atoms/cm.sup.2 to 1.Math.10.sup.17 atoms/cm.sup.2.

4. Layer system according to claim 1, wherein y is from x+(1x)*1.3 to x+(1x)*1.5.

5. Layer system according to claim 1, wherein the halogen mole fraction in the buffer layer has a gradient that decreases from a surface facing the absorber layer to an interior of the buffer layer.

6. Layer system according to claim 1, wherein the layer thickness of the buffer layer is from 5 nm to 150 nm.

7. Layer system according to claim 1, wherein the halogen is chlorine, bromine, or iodine.

8. Layer system according to claim 1, wherein the chalcogenide compound semiconductor includes Cu(In,Ga,Al)(S,Se).sub.2.

9. Thin-film solar cell, comprising: a substrate, a rear electrode that is arranged on the substrate, a layer system according to claim 1 that is arranged on the rear electrode, and a front electrode that is arranged on the layer system.

10. Method for producing a layer system for thin-film solar cells, comprising: a) preparing an absorber layer, and b) arranging a buffer layer, which contains halogen-enriched Zn.sub.xIn.sub.1-xS.sub.y with 0.01x0.9 and 1y2 on the absorber layer, wherein the buffer layer consists of a first layer region adjoining the absorber layer with a halogen mole fraction A.sub.1 and a second layer region adjoining the first layer region with a halogen mole fraction A.sub.2 and a ratio A.sub.1/A.sub.2 is 2, and a layer thickness of the first layer region is 20% of a layer thickness of the buffer layer.

11. Method according to claim 10, comprising: in the step b), applying a metal-halide compound on the absorber layer and applying zinc indium sulfide on the metal-halide compound.

12. Method according to claim 10, comprising: in the step b), applying a metal-halide compound and zinc indium sulfide on the absorber layer.

13. Method according to claim 12, comprising: conveying the absorber layer past at least one steam beam of the metal-halide compound and at least one steam beam of indium sulfide and zinc sulfide.

14. Method according to claim 12, comprising: applying the metal-halide compound with chlorine, bromine, and/or iodine as halogen and sodium, potassium, aluminum, gallium, indium, zinc, cadmium, and/or mercury as metal.

15. Layer system according to claim 1, wherein the ratio A.sub.1/A.sub.2 is from 10 to 100.

16. Layer system according to claim 1, wherein the amount of the halogen in the first layer region amounts to an area concentration of 2.Math.10.sup.14 atoms/cm.sup.2 to 2.Math.10.sup.16 atoms/cm.sup.2.

17. Layer system according to claim 1, wherein the layer thickness of the buffer layer is from 15 nm to 50 nm.

18. Layer system according to claim 1, wherein the chalcogenide compound semiconductor includes CuInSe.sub.2, CuInS.sub.2, Cu(In,Ga)Se.sub.2, Cu(In,Ga)(S,Se).sub.2, or Cu.sub.2ZnSn(S,Se).sub.4.

Description

(1) The invention is explained in detail in the following with reference to drawings and an example. The drawings are not completely true to scale. The invention is in no way restricted by the drawings. They depict:

(2) FIG. 1 a schematic cross-sectional view of a solar cell with a layer system according to the invention,

(3) FIG. 2 a diagram of the transmission of a Zn.sub.xIn.sub.1-xS.sub.y buffer layer on glass for various ratios x,

(4) FIG. 3 a diagram of the short-circuit current strength of the solar cell with Zn.sub.xIn.sub.1-xS.sub.y buffer layer,

(5) FIG. 4 a diagram of the depth profile of chlorine of a layer structure according to the invention with a comparative example,

(6) FIG. 5 a diagram of the depth profile of the chlorine, copper, indium, sulfur, and selenium concentration of a layer structure according to the invention,

(7) FIG. 6 a diagram of the photoluminescence lifetime of a layer structure according to the invention with a comparative example,

(8) FIG. 7 a diagram of the efficiency of a layer structure according to the invention with comparative examples,

(9) FIG. 8 a diagram of the efficiency of another layer structure according to the invention with a comparative example,

(10) FIG. 9A a diagram of the efficiency of a layer structure according to the invention with a comparative example,

(11) FIG. 9B a diagram of the open circuit voltage of an alternative layer structure according to the invention with a comparative example,

(12) FIG. 10 an exemplary embodiment of the process steps according to the invention with reference to a flow diagram,

(13) FIG. 11 a schematic depiction of an in-line method for producing a buffer layer according to the invention,

(14) FIG. 12 a schematic depiction of an alternative in-line method for producing the buffer layer according to the invention.

(15) FIG. 1 depicts purely schematically a preferred exemplary embodiment of a thin-film solar cell 100 according to the invention with a layer system 1 according to the invention in a cross-sectional view. The thin-film solar cell 100 includes a substrate 2 and a rear electrode 3. A layer system 1 according to the invention is arranged on the rear electrode 3. The layer system 1 according to the invention comprises an absorber layer 4 and a buffer layer 5. A second buffer layer 6 and a front electrode 7 are arranged on the layer system 1.

(16) The substrate 2 is made here, for example, of inorganic glass, with it equally possible to use other insulating materials with sufficient stability as well as inert behavior relative to the process steps performed during production of the thin-film solar cell 100, for example, plastics, in particular polymers or metals, in particular metal alloys.

(17) Depending on the layer thickness and the specific material properties, the substrate 2 can be implemented as a rigid plate or flexible film. In the present exemplary embodiment, the layer thickness of the substrate 2 is, for example, from 1 mm to 5 mm.

(18) A rear electrode 3 is arranged on the light-entry side surface of the substrate 2. The rear electrode 3 is made, for example, from an opaque metal. It can, for example, be deposited on the substrate 2 by vapor deposition or magnetic field-assisted cathode sputtering. The rear electrode 3 is made, for example, of molybdenum (Mo), aluminum (Al), copper (Cu), titanium (Ti), or of a multilayer system with such a metal, for example, molybdenum (Mo). The layer thickness of the rear electrode 3 is, in this case, less than 1 m, preferably in the range from 300 nm to 600 nm, and is, for example, roughly 500 nm. The rear electrode 3 serves as a back-side contact of the solar cell 100. An alkali barrier, made, for example, of Si.sub.3N.sub.4, SiON, or SiCN, can be arranged between the substrate 2 and the rear electrode 3. This is not shown in detail in FIG. 1.

(19) A layer system 1 according to the invention is arranged on the rear electrode 3. The layer system 1 includes an absorber layer 4, made, for example, of Cu(In,Ga)(S,Se).sub.2, which is applied directly on the rear electrode 3. The absorber layer 4 has, for example, a thickness of 1.5 m.

(20) A buffer layer 5 is arranged on the absorber layer 4. The buffer layer 5 includes Zn.sub.xIn.sub.1-xS.sub.y with 0.01x0.9, and 1y2 and, for example, Zn.sub.0.1In.sub.0.9S.sub.1.45. The layer thickness d of the buffer layer 5 is from 5 nm to 150 nm and, for example, 35 nm. The buffer layer 5 consists of a first layer region 5.1 that adjoins the absorber layer 4 and is connected over its entire area to the absorber layer 4. Moreover, the buffer layer 5 includes a second layer region 5.2, which is arranged on the first layer region 5.1. The first layer region has a thickness d.sub.1 that is 50% of the layer thickness d of the entire buffer layer 5. The thickness d.sub.1 of the first layer region 5.1 is, for example, 10 nm. The first layer region 5.1 contains a halogen mole fraction A.sub.1 and the second layer region 5.2, a halogen mole fraction A.sub.2. The ratio of the halogen mole fractions A.sub.1/A.sub.2 is 2 and, for example, 10. For clarification, an exemplary curve of the halogen mole fraction A.sub.Halogen is depicted in FIG. 1 as a function of the layer depth s. The data of the halogen mole fraction A.sub.Halogen are plotted in atom-%, with the halogen mole fraction also including halogen ions and halogen in compounds.

(21) A second buffer layer 6 can be arranged above the buffer layer 5. The buffer layer 6 includes, for example, non-doped zinc oxide. A front electrode 7 that serves as a front-side contact and is transparent to radiation in the visible spectral range (window layer) is arranged above the second buffer layer 6. Usually, a doped metal oxide (TCO=transparent conductive oxide), for example, n-conductive, aluminum (Al)-doped zinc oxide (ZnO), boron (B)-doped zinc oxide (ZnO), or gallium (Ga)-doped zinc oxide, is used for the front electrode 7. The layer thickness of the front electrode 7 is, for example, roughly 300 to 1500 nm. For protection against environmental influences, a plastic layer (encapsulation film) made, for example, of polyvinyl butyral (PVB), ethylene vinyl acetate (EVA), or DNP can be applied to the front electrode 7. In addition, a cover plate transparent to sunlight that is made, for example, from extra white glass (front glass) with a low iron content and can have a thickness of, for example, 1 to 4 mm, can be provided.

(22) The described structure of a thin-film solar cell or a thin-film solar module is well known to the person skilled in the art, for example, from commercially available thin-film solar cells or thin-film solar modules and has also already been described in detail in numerous printed documents in the patent literature, for example, DE 19956735 B4.

(23) In the substrate configuration depicted in FIG. 1, the back electrode 3 adjoins the substrate 2. It is understood that the layer structure 1 can also have a superstrate configuration, in which the substrate 2 is transparent and the front electrode 7 is arranged on a surface of the substrate 2 facing away from the light-entry side.

(24) The layer system 1 can serve for production of integrated serially connected thin-film solar cells, with the layer system 1, the front electrode 7, and the rear electrode 3 patterned in a manner known per se by various patterning lines (P1 for rear electrode, P2 for front electrode/back electrode contact, and P3 for separation of the front electrode).

(25) FIG. 2 depicts the transmission of Zn.sub.xIn.sub.1-xS.sub.y layers on glass for various x as a function of wavelength. A buffer layer according to the prior art with x=0%, i.e., an In.sub.2S.sub.3 layer without a zinc fraction, serves as a comparative example. An increase in transmission over the entire wavelength range from 300 nm to 400 nm for increasing ratios of x is noted. The transmission increase is surprising for these small admixtures of zinc since the band gap changes only slightly for ratios x<10%. In the solar cell, this makes itself felt, as expected, in an increase in the short-circuit current, as shown in FIG. 3.

(26) FIG. 3 shows a diagram of the short-circuit current J.sub.SC of solar cells with Zn.sub.xIn.sub.1-xS.sub.y buffer layers 5 as a function of x. The short-circuit current J.sub.SC is normalized to the short-circuit current of a buffer layer according to the prior art with x=0, i.e., an In.sub.2S.sub.3 buffer layer without a zinc fraction. The normalized short-circuit current has a maximum for x=0.15 with J.sub.SC=1.026.

(27) FIG. 4 depicts the depth profile of chlorine measured by high-resolution time-of-flight secondary ion mass spectrometry (ToF-SIMS) of three differently prepared buffer layers 5 on an absorber layer 4. The chlorine fraction describes the fraction of all chlorine atoms present in the buffer layer 5 regardless of their oxidation state.

(28) The reference measurement was made using a comparative example on an indium sulfide buffer layer according to the prior art without a first layer region 5.1 with halogen enrichment. In the cases Cl-amount 1 and Cl-amount 2, two different amounts of sodium chloride were applied on the absorber layer 4 before the actual buffer deposition 5 made of indium sulfide. The NaCl-area concentration of Cl-amount 1 was roughly 210.sup.14 Cl-atoms/cm.sup.2. The NaCl-area concentration of Cl-amount 2 was roughly 310.sup.14 Cl-atoms/cm.sup.2 and was greater by 50% than with Cl-amount 1.

(29) In FIG. 5, the normalized sputtering time is plotted on the x-axis. The sputtering time corresponds to the sputter depth with the analysis beginning on the surface of the buffer layer 5 facing away from the absorber layer 4. The buffer layer 5 corresponds to the region 0 to 1 on the x-axis and the absorber layer 4 corresponds to the region with values >1. FIG. 4 shows that after somewhat more than half of the buffer layer with a normalized sputtering time of 0.6, a chlorine signal rises in a delta shape and drops again to lower values on obtaining the absorber at a normalized sputtering time of roughly 1. The intensities outside the maximum are attributable in part to diffusion and in part to smearing of the profile due to the relatively rough surface topography of the absorber layer 4. The comparative example also shows smaller amounts of chloride with normalized sputtering times between 0.6 and 0.8, which are attributable to impurities in the starting materials for the deposition. However, the low coating thickness of chloride in the comparative example is insufficient to exert a noteworthy positive effect on efficiency.

(30) FIG. 5 additionally depicts the depth profile of the chlorine enrichment with regard to the interface of absorber layer 4 and buffer layer 5 using Example Cl-amount 1 from FIG. 4 with the other elements copper, indium, sulfur, and selenium. As anticipated, in the region of the indium sulfide buffer layer 5, large fractions of sulfur and indium were measured and in the region of the absorber layer 4, large fractions of copper, selenium, indium, and a smaller fraction of sulfur. It was shown, in particular, that the intensity of the copper signal drops sharply from the interface between absorber layer 4 and buffer layer 5 in the direction of buffer layer 5 and overlaps only a little with the region of high chlorine intensities.

(31) The present invention is based, consequently, on the finding of the inventors that the relatively large halogen atoms or halogen ions localized at the interface reduce, as a diffusion barrier, the inward diffusion of impurities such as copper out of the absorber layer 4 into the buffer layer 5. The halogen atoms or halogen ions localized at the interface in the first layer region 5.1 alter the electronic properties of the buffer layer 5 itself in a positive manner. An inward diffusion of copper into In.sub.2S.sub.3 buffer layers is described, for example, in Barreau et al, Solar Energy 83, 363, 2009 and leads, via a reduction of the band gap, to increased optical losses. This occurs primarily through a shift of the valence band maximum, which could, in turn, have a disrupting effect on the formation of the energetically optimum band structure at the p-n-junction. Furthermore, halogen enrichment according to the invention at the interface between the absorber layer and the buffer layer can electronically mask and neutralize defects possibly occurring such that these are no longer active as recombination centers and, thus increase the efficiency of the solar cell overall. In order to detect the reduction of recombination centers, measurements of the photoluminescence lifetime were performed with different buffers on absorber/buffer heterojunctions.

(32) FIG. 6 depicts a diagram of the photoluminescence lifetime for various buffer layers. The lifetime was measured with a photoluminescence measuring station and a 633 nm-laser. Here, the typical penetration depth of the laser light is roughly 70 nm. This means that the photoluminescence of the surface and the region near the surface (in particular the space charge zone at the pn-junction) is measured. The photoluminescence decay time measured corresponds to an effective lifetime that is determined both by the recombination of the charge carriers in the volume and by the recombination at the interface to the buffer layer 5. FIG. 6 presents the typical lifetime of pn-junctions of a buffer layer 5 made of indium sulfide with chloride enrichment and two comparative examples according to the prior art.

(33) Comparative example 1 is a CdS buffer layer according to the prior art. The good lifetime of roughly 40 ns in pn-junctions with CdS buffer layers can be attributed to the very good reduction of interface defects due to the wet chemical processing. Comparative example 2 is a buffer layer according to the prior art made of indium sulfide without halogen enrichment. The pure In.sub.2S.sub.3 buffer layer results in a clear reduction of the lifetime to roughly 10 ns, which is attributable to an increased recombination of the charge carriers on the surface and in layers near the surface.

(34) The example shows a NaCl+In.sub.2S.sub.3/In.sub.2S.sub.3 buffer layer 5, wherein, before deposition of the buffer layer 5, sodium chloride was applied on the absorber layer 4. This results in the formation of a first layer region 5.1 with a halogen mole fraction A.sub.1. The buffer layer 5 presents a significantly increased lifetime of roughly 41 ns. The lifetime in the buffer layer 5 of the example falls within the range of the lifetime of Comparative Example 1, a CdS buffer layer. It can be concluded that despite the dry method of deposition of the buffer layer 5 of the example, a significant reduction of interface defects is achieved. The introduction of the halogen-rich first layer region 5.1 thus actually results in an improvement of the electronic properties of the absorber-buffer interface, in comparison to an indium sulfide buffer layer without halogen enrichment.

(35) The efficiency of buffer layers 5 with halide enrichment at the interface to the absorber layer 4 is improved relative to pure indium sulfide buffer layers over a wide range relative to the halogen mole fraction A.sub.1 in the first layer region 5.1.

(36) FIG. 7 depicts the efficiency of solar cells 100 with a buffer layer 5 with a NaCl pre-coating and two comparative examples according to the prior art. In Comparative Example 1, the buffer layer is a wet-chemically deposited CdS layer. In Comparative Example 2, the buffer layer is an indium sulfide layer without halogen enrichment deposited from the gas phase. The buffer layer 5 was produced by a NaCl pre-coating by pre-deposition of sodium chloride onto the absorber layer 4. Then, indium sulfide was deposited from the gas phase onto the NaCl pre-coating. As can be discerned from FIG. 7, the highest efficiencies, averaging 1.04 normalized to the efficiency of Comparative Example 2, i.e., an In.sub.2S.sub.3 buffer layer according to the prior art without halogen enrichment, are obtained with the buffer layer 5 with NaCl pre-coating.

(37) FIG. 8 presents the measurement of the efficiency for an In.sub.2S.sub.3 buffer layer according to the prior art without InCl.sub.3 pre-coating compared to the efficiency of an alternative buffer layer 5 with halogen enrichment through pre-coating with indium chloride (InCl3). InCl.sub.3 pre-coating also results in halogen enrichment in the first layer region 5.1 at the interface to the absorber layer 4, which dopes the interface and, moreover, reduces the surface defects. FIG. 8 clearly shows that halogen enrichment of the buffer layer 5 according to the invention increases the efficiency of solar cells 100. The increase in efficiency with InCl.sub.3 pre-coating is, in this case, roughly 4% and is in the same order of magnitude as with a NaCl pre-coating in FIG. 7.

(38) Copper, oxygen, and selenium can also be found in the buffer layer 5 in addition to the elements indium, sulfur, and chlorine. Indium sulfide has a relatively open lattice structure into which other elements such as sodium and copper can be incorporated quite well. The deposition of the buffer layer 5 can occur at relatively high temperatures, in particular at temperatures from room temperature to roughly 200 C. The subsequent transparent front electrode 7 is also deposited preferably at temperatures up to 200 C. At these temperatures, sodium, copper, and selenium can diffuse out of the absorber layer 4 or the front electrode 7 into the buffer layer 5. In this case, these elements can also be enriched at the interface through pre-coating with a metal-halogen compound in addition to halogen. Depending on the selection of the metal-halogen compound, the accompanying metal in the first layer region 5.1 of the buffer layer 5 will also be enriched. Due to the hygroscopic properties of the starting materials, enrichment by water from the ambient air is also conceivable.

(39) FIG. 9 A depicts the efficiency and FIG. 9 B the open circuit voltage as a function of the distribution of NaCl enrichment in the buffer layer 5. Solar cells 100 with layer systems 1, wherein sodium chloride was applied after deposition of the buffer layer (after), before deposition of the buffer layer (before), and within the buffer layer 5 (between), are compared. The solar cell 100 corresponds to the arrangement as it was described under FIG. 1. The layer system 1 before, wherein sodium chloride was deposited before the deposition of the buffer layer, presents the highest efficiency, averaging 14%, and the highest average open circuit voltage of 560 mV. FIG. 9 A and FIG. 9 B document the particular effectiveness of halogen enrichment in a first layer region 5.1 at the interface to the absorber layer 4 through pre-coating of the absorber layer 4 with sodium chloride (before) in contrast to positioning in the middle of the buffer layer 5 (between) or post-coating (after) at the interface to the second buffer layer 6.

(40) The essential idea for this invention is to combine the gain in current due to a fraction of zinc, as was shown in FIG. 3, with an increase of the open circuit voltage due to halogen enrichment of the buffer layer 5 at the interface to the absorber layer 4, as was shown in FIG. 9 B. Particularly advantageous layer systems 1 were successfully obtained in this manner.

(41) Experiments of the inventors demonstrated that doping of the entire buffer layer 5 with sodium and, in particular, with sodium chloride also increases the short-circuit current strength of a solar cell 100. This can be explained by a widening of the band gap in the buffer layer 5 by sodium doping. Since sodium is relatively mobile, it is not possible to rule out outward diffusion of sodium from the buffer layer into the absorber layer or the front electrode with the addition of sodium to an In.sub.2S.sub.3 buffer layer according to the prior art. However, the increase of sodium in the absorber layer can lead to an increase in the absorber conductivity, to the development of short circuits, and to the reduction in the width of the depletion zones. In the present invention, the widening of the band gap is obtained through the addition of zinc into the buffer layer 5 and separated from the positive effect of halogen enrichment in a first buffer layer 5.1 at the interface to the absorber layer 4. As experiments of the inventors showed, zinc is stably bound in the buffer layer 5. In this manner, particularly advantageous solar cells with particularly high efficiencies and particularly high stability can be obtained.

(42) FIG. 10 depicts a flow diagram of a method according to the invention. In a first step, an absorber layer 4 is prepared, for example, from a Cu(In,Ga)(S,Se).sub.2 semiconductor material. In a second step, the buffer layer 5 of sodium and chlorine, for example, of sodium chloride, and zinc indium sulfide is deposited. The zinc indium sulfide is deposited, for example, by simultaneous deposition of zinc sulfide and indium sulfide. The ratio of the individual components in the buffer layer 5 is regulated, for example, by control of the deposition rate, for example, by an aperture or by temperature control.

(43) In further process steps, a second buffer layer 6 and a front electrode 7 can be deposited on the buffer layer 5. In addition, connecting and contacting of the layer structure 1 to a thin-film solar cell 100 or to a solar module can take place.

(44) FIG. 11 presents a schematic depiction of an in-line method for producing a buffer layer 5 according to the invention. In an in-line method, the substrate 2 with rear electrode 3 and absorber layer 4 is conveyed past the steam beam 11 of a halogen-containing evaporator source and, for example, past a sodium chloride source 8. The transport direction is indicated by an arrow with the reference character 10.

(45) The amount of sodium chloride deposited is adjusted, for example, by opening and closing an aperture such that an NaCl amount of more than 110.sup.13 atoms/cm.sup.2 and less than 110.sup.17 atoms/cm.sup.2 is deposited on the surface.

(46) Then, the absorber layer 4 pre-coated with sodium chloride is conveyed past at least one indium sulfide source 9 and one zinc sulfide source 15. This occurs preferably without vacuum interruption. The layer thickness d of the buffer layer 5, the ratio x/(1x) of zinc to indium, and the halogen enrichment profile over the buffer layer 5 can be controlled by the deposition rates, transport speeds, and number of halogen sources 8, of indium sulfide sources 9, and of zinc sulfide sources 15.

(47) The source for the deposition of the metal-halogen compound, of indium sulfide or zinc sulfide is, for example, an effusion cell, a boat or crucible of a thermal evaporator, of a resistance heater, of an electron beam evaporator, or of a linear evaporator.

(48) FIG. 12 presents a schematic depiction of an alternative in-line method for producing a buffer layer 5 according to the invention. The indium sulfide source 9 and the zinc sulfide source 15 are arranged such that their steam beams 12,16 overlap virtually completely. The sodium chloride source 8 is arranged such that the steam beam 11 of the sodium chloride sources 8 and the steam beams 12,16 of the indium sulfide source 9 and the zinc sulfide source 15 overlap partially in an overlap region 14. The sodium chloride source 8, the indium sulfide source 9, and the zinc sulfide source 9 [sic] are, for example, effusion cells out of which sodium chloride or indium sulfide is thermally evaporated. Alternatively, any other form of generation of steam beams 11,12,16 is suitable for the deposition of the buffer layer 5 so long as the ratio of the mole fractions of chlorine, zinc, indium, and sulfur can be controlled.

(49) In this manner, for example, a gradient with a continuous decrease in halogen concentration can be formed in the buffer layer 5. As experiments of the inventors have demonstrated, such a gradient is particularly advantageous for the properties of the solar cell 100 according to the invention.

(50) The introduction of chlorine from sodium chloride, indium chloride, or zinc chloride into the indium sulfide buffer layer 5 has multiple special advantages. Sodium chloride is non-toxic and economical and can, as already mentioned, be readily applied using thermal methods. During thermal deposition, sodium chloride evaporates as NaCl molecules and does not dissociate to sodium and chlorine. This has the particular advantage that during evaporation, no toxic and corrosive chlorine develops.

(51) The introduction of chlorine from sodium chloride offers additional advantages from a production technology standpoint. Only one substance has to be evaporated, greatly simplifying the process compared to possible mixtures of substances such as NaCl/ZnS/In.sub.2S.sub.3. The vapor pressure curve of NaCl is known, for example, from C. T. Ewing, K. H. Stern, Equilibrium Vaporization Rates and Vapor Pressures of Solid and Liquid Sodium Chloride, Potassium Chloride, Potassium Bromide, Cesium Iodide, and Lithium Fluoride, J. Phys. Chem., 1974, 78, 20, 1998-2005, and a thermal vapor deposition process can be readily controlled by temperature. Moreover, an arrangement for vapor deposition of sodium chloride can be readily integrated into existing thermal indium sulfide coating equipment.

(52) Moreover, halide enrichment can be controlled and measured simply. Thus, for process control during the vapor deposition, a quartz resonator can be used for direct measurement of the rate. An optical control of the sodium amount and, consequently, the chloride amount can be used by means of emission spectroscopy. Alternatively, sodium chloride can be deposited on silicon and this can be investigated with x-ray fluorescence analysis (XRF), with an ellipsometer or a photospectrometer in-line or after the process. The introduction of chlorine from indium chloride or zinc chloride into the zinc indium sulfide buffer layer 5 has the advantage that indium and zinc are components of the buffer layer 5 and, consequently, no further function-impairing foreign metals get into the buffer layer 5.

(53) From the above assertions, it has become clear that by means of the present invention the disadvantages of previously used CdS buffer layers or the alternative buffer layers were overcome in thin-film solar cells, with the efficiency and the stability of the solar cells produced therewith also very good or better. At the same time, the production method is economical, effective, and environmentally safe. This was unexpected and surprising for the person skilled in the art.

REFERENCE CHARACTERS

(54) 1 layer system

(55) 2 substrate

(56) 3 rear electrode

(57) 4 absorber layer

(58) 5 buffer layer

(59) 5.1 first layer region

(60) 5.2 second layer region

(61) 6 second buffer layer

(62) 7 front electrode

(63) 8 sodium chloride source

(64) 9 indium sulfide source

(65) 10 transport direction

(66) 11 sodium chloride steam beam

(67) 12 indium sulfide steam beam

(68) 14 overlapping region

(69) 15 zinc sulfide source

(70) 16 zinc sulfide steam beam

(71) 100 thin-film solar cell, solar cell

(72) d layer thickness of the buffer layer 5

(73) d.sub.1 layer thickness of the first layer region 5.1

(74) s layer depth

(75) A.sub.1 halogen mole fraction in the first layer region 5.1

(76) A.sub.2 halogen mole fraction in the second layer region 5.2

(77) A.sub.Halogen halogen mole fraction