The present invention relates to a method for manufacturing solar cells, and more specifically thin-film solar cells.

The technique used to produce thin-film solar cells is the PVD-tech- nique (Physical Vapor Deposition) , in which copper, indium and selenium are vapourized and deposited on a heated substrate. The substrate normally includes a glass plate on which a layer of molybdenum (Mb) is sputtered. Ihe <_"c ^* rπpound c»pper-indαum-diselenide (CuInSe2) is formed on the substrate. A layer of cadmium sulphide (CdS) and a transparent electric contact in the form of doped zinc oxide (ZnO) are laid over a layer of cadmium sulphide (CdS) .

Vapourization of the various substances takes place in a chamber. When the glass plate has large dimensions, the plate is positioned station¬ ary in relation to the sources of the aforesaid substances.

Three problems are primarily encountered when producing thin-film solar cells based on CuInSe2 by means of vacuum deposition (PVD) . This also applies with layer ccttπpositions other than just OιInSe2.

Ihe first of these problems resides in the ability to produce layers on large surfaces in a manner which will provide a uniform layer. Thin-film solar cells normally have a size of 1 x 0.4 metres. The CuInSe2-layer is particularly sensitive and should be uniform in order to achieve good furtctioning of the solar cell.

Ihe second of the aforesaid problems is that the composition of the CuInSe2-layer should vary over its thickness, so that the composition of that part of the layer distal away from the glass plate will contain more indium than is given by the stoichiσmetric composition. Correspondingly, that part of the layer which lies nearest the glass plate should contain a relatively higher percentage of copper. This is because of the desire to suppress the formation of so-called copper nodules on the upper boundary surface of the layer, since such nodules

reduce the efficiency of the solar cell.

Ihe known technique solves this problem by first applying a layer rich in copper and then applying to this first layer in a second process stage a second layer which is rich in indium. Application of a large quantity of indium .suppresses the formation of a copper-rich surface, which in turn suppresses the formation of said nodules. This known technique thus requires two different process stages, thereby rendering the process complicated and expensive.

Ωie third of the aforesaid problems resides in a relatively low yield of vapourized substances cn the substances deposited on the substrate.

These problems are solved by means of the presort invention.

Ihe present invention thus relates to a method of anufacturing solar cells, more specifically tliin-film solar cells, in which a multilayer -structure is built-up on a substrate, said method comprising the steps of depositing a layer consisting of a cαmpσund of several basic substances, such as c»p ^* er-indiuπirdiselenide (CXιIhSe2) or a closely related compound in which copper (Cu) and indium (In) can be replaced totally or partially with silver (Ag) or gallium (Ga) respectively, and where Se is replaced either totally or partially with sulphur (S) or tellurium (Tte), and in which the ccnceπtration of the basic sub- stances in the layer varies, said method being characterized in that the substrate is placed on the inside of a rotatable, generally circular carrier device; in that the substrate is heated; in that sources of the basic substances to be deposited on the substrate are present; in that one or more of the basic substances is/are deposited on the substrate by passing a respective source axially throug the cylindrical carrier device; and in that the cylindrical carrier device and/or respective sources can be moved while rotating the cylindrical carrier device.

The invention will now be described in more detail with reference to an exeπplifying eo-todimeπt thereof illustrated in the accαπpanying drawing, in which

Figure 1 is a schematic sectional view of a solar cell;

Figure 2 is a schematic, axial view of an arrangement for carrying out the invention; and

Figure 3 is a schematic side view of the arrangement.

Figure 1 is a schematic sectional view of a tiin-fil solar cell. Thin-film solar cells are constructed on substrates of relatively large sizes, for instance a size of 1 x 0.4 m. Such structures include a very large number of cells above the substrate surface, these cells being mutually connected electrically. Figure 1 shows only a part of one such cell. The technique of manufacturing a large number of -mutually separate, but electrically connected cells on a substrate surface is well known and will not be described here.

The substrate 1 is normally a glass plate of suitable thickness, generally 3 ram. A layer 2 of molybdenum (Mo) is first sputtered onto the glass surface. Ihe molybdenum layer forms an electric back- contact and positive pole, or terminal, in the finished cell. The molybdenum layer may have a thickness of, e.g., 0.5 micrometer. There is then applied to the molybdenum layer a layer 3 of CuInSe2, having a thickness of 2 micrometers, for instance. There is then applied onto the layer 3 a first layer 4 of cadmium sulphide (CdS) to a thickness of from 5 to 500 A for instance, and then an electrical contact in the form of a transparent, doped zinc oxide layer (ZnO) 6 which is applied to a thickness of from 1 to 2 irt croπieters, for instance. A metallic contact 5 made of aluminium for instance may also be used.

When sunlight falls on the solar cell, an electric voltage will occur between the contact 6 (ZnO), which is a minus pole, and the back contact 2.

As before mentioned, the present invention relates to a method for manufacturing solar cells, more specifically thin-film solar cells, in which a multi-layer structure is built-up on a substrate, said method ∞r-prising the steps of depositing a layer consisting of a compound of several basic substances, such as copper-indiurrt-diselenide (CuInSe2) or a closely related cαrπpound in which copper (Cu) and

indium (In) can be replaced totally or partially with silver (Ag) and gallium (Ga) respectively, and in which Se can be replaced totally or partially with sulphur (S) or tellurium (Te), where the concentration of the basic substances in the layer varies.

More specifically, the present invention relates to a method for manufacturing thin-filia solar cells in -which a layer consisting of copper-indiiirnr-diselenide (CuInSe2) is deposited and in which the concentration of copper and indium -respectively vary in the layer for the reasons given in the introduction.

The present invention is described belcw with reference to an ex¬ emplifying eπ-x iinent thereof in * ^* hic ^* h a layer of CuInSe2 is deposited, althouc^i it will be understood that the invention is not restricted to such a layer but can be applied for depositing layers which comprise other basic substances.

According to the invention, the substrate 1 is placed on an inner surface 11 of a rotatable, generally cylindrical carrier device 10, as illustrated sciianatically in Figure 2. The whole of the carrier device 10 is placed in a vacuum chamber 12 (see Figure 3) of a suitable known kind. The -rølybdeπum layer 2 has been applied to the substrate previous to placing the substrate on the carrier device 10.

The carrier device may have any suitable cxaistruσtion. For instance, it may comprise a framework of steel beams provided with glass-plate or substrate attachment devices.

Ωie arrangearteπt includes a selenium source 13 from which selenium is deposited onto the substrate. The position of the selenium source is not critical. Nevertheless, selenium will be present in sufficient concentrations in the chamber atmosphere.

The CuInSe2-layer 3 is deposited, by passing a copper source and an indium source axially through the cylindrical carrier device 10 while the device is rotated about its own axis. The sources and the carrier device preferably move continuously in relation to one another.

with the substance sources moving from one open end of the carrier device to its other end.

Movement of the sources 14, 15 through the carrier device may be effected by moving the cylindrical carrier device 10 and/or the sources 14, 15.

When depositing the basic substances, the sources 13, 14, 15 are heated so as to generate in the chamber an atmosphere of predetermined composition, in a known manner. The substrate is heated to a tempera¬ ture of 300-600°C.

Figure 3 illustrates schematically only one conceivable embodiment of the inechanical part of the arrangement. The carrier device may be provided with circular beams 15, 16 which extend around the periphery of said device and rest on rollers 17, 18. The rollers 17, 18 may be driven, for rotation of the carrier device 10. The sources 13, 14, 15 may be mounted on a rod 19 -which can be moved in relation to a carrier arm 21 in the directions indicated by the arrow 20. For instance, a ball screw may be fixedly mounted in relation to the carrier arm 21, which coaσts with a drive unit 22 fixedly connected to the rod 19. The rod is moveable between the rear position shown in full lines in Figure 3 and the forward position shown in broken lines 23.

The chamber 12 has a door 24.

The sources may be stationarily mounted instead of being moveable, and the carrier device mounted for movement in the direction of its geometric axis.

Because, in accordance with the invention, the CuInSe2-layer 3 is deposited fcy moving a copper source and an indium source at a uniform speed axially through the carrier device 10 while rotating the carrier device 10 about its own axis, a uniform layer will be deposited despite the sources delivering their respective basic substances, in accordance with a given space distribution.

According to a much preferred embodiment, the copper source and the indium source are spaced apart at a predetermined distance in their movement directions. Furthermore, the copper source 14 is brought into position first and then the indium source 15.

Because the copper source and the indium source are placed at prede¬ termined distance apart and the copper source 14 is brought into position first and then the indium source 15, the amount of copper deposited on the substrate structure will initially be greater than the amount of indium that is deposited, i.e. a larger amount of copper nearest the substrate, whereafter the amount of indium deposited will be greater than the amount of copper deposited. This means that the CuEhSe2-layer 3 will be richer in copper nearest the substrate and richer in indium nearest the CdS-layer 4.

According to one preferred ernbodimeπt of the invention, the distance between the copper source and the indium source is in the order of 10% to 25% of the radius of the carrier device.

.according to a much preferred e_πbod__meπt, deposition is effected by moving the copper source and the indium source relative to the carrier device solely once and in one direction. In the case of the aforesaid dimensions, this means that the sources will move continuously "t±rough the carrier device over a time period of 30 minutes.

Alternatively, the sources may be moved backwards and forwards several times and the teπperatures of respective sources controlled so as to obtain a copper-rich layer nearest the substrate and an indium-rich layer on the opposing surface. It is extremely difficult, however, to achieve sufficient reproducibility with this particular technique.

In that case in which the sources are moved only once t ^* hrough the carrier device and then returned to their aforesaid wi1 ^** hdrawn position, the temperature of the sources is decreased so no substance will be deposited when the sources are withdrawn through the carrier device, back to their original positions.

According to one preferred embodiment, the cylindrical carrier device is rotated at a speed greater than about 1 to 5 r.p.m.

According to another preferred embodiment, the circumferential inner surface of the cylindrical carrier device is essentially covered with substrate, preferably planar substrate, such as to form a polygonal cross-section; see Figure 2 in which the substrates form an eight- sided figure in cross-section. This greatly increases the yield of the quantity of substances delivered by the sources to the quantity of substances deposited on the substrate in comparison with the known technique. The yield is troublesomely low in the case of the known technique.

According to a further preferred eπ ^* bcx ^* liment, the length of the cylin- drical carrier device corresponds to the length of a solar cell panel, i.e. about 1 metre, and a diameter such that a substrate having a width of 0.4 metre will form a polygon when seen in cross- section, preferably a hexagonal or decagonal figure.

It will be obvious from the aforegoing that the present invention solves the three problems mentioned in the introduction. Consequently, the present invention provides an inexpensive and quick method of manufacturing thin-film solar cells of very high quality.

The invention, however, is not restricted to the deposition of CuInSe2, but may also be used in conjunction with other related compounds where copper is replaced totally or partially with silver and in which indium is replaced solely or partially with gallium and where selenium is replaced totally or partially with sulphur or tellurium may be used.

Although the invention has been described with reference to a number of exemplifying ertibodiments thereof, it will be obvious that the carrier device and also the means for support-ing the substance sources may be cxsnstructed in other ways suitable for carrying out the inven¬ tive method. Furthermore, the sources may be positioned other than immediately behind one another. The above-mentioned dimensions and speeds may also be adapted to suit prevailing conditions and are

therefore not limited to the aforegiven values.

The present invention is therefore not restriced to the aforedescribed and illustrated exemplifying emtodiments thereof, since these embodi- ments can be modified v thin the scope of the following Claims.