DE102006037652A1 - Expanded silicon substrate for the production of sensors/solar cell or for the production of further substrates, consists of monocrystalline regions having a uniform crystal orientation to its respective surface-normal - Google Patents

Abstract

The expanded silicon substrate (10) for the production of sensors or solar cell, or for the production of further substrates with a thickness d of 0.1-100 mu m and width and length of 1 cm, consists of monocrystalline regions (1a, 1b, 1c) with a length- and a width expansion of 1 cm. The regions have a uniform crystal orientation to its respective surface-normal (13) with a deviation of 5[deg]. The expansion of the planar substrate is 0.5 m in a longitudinal direction (12) and its mechanical characteristics permit a reversible winding on a cylinder surface (14) with a diameter of 5 m. The expanded silicon substrate (10) for the production of sensors or solar cell, or for the production of further substrates with a thickness d of 0.1-100 mu m and width and length of 1 cm, consists of monocrystalline regions (1a, 1b, 1c) with a length- and a width expansion of 1 cm. The regions have a uniform crystal orientation to its respective surface-normal (13) with a deviation of 5[deg]. The expansion of the planar substrate is 0.5 m in a longitudinal direction (12) and its mechanical characteristics permit a reversible winding on a cylinder surface (14) with a diameter of 5 m. The monocrystalline regions have an expansion of 3 cm. The length of the foil substrate is 3 m. The foil is wound reversibly on the cylinder surface with a diameter of 2 m. The substrate has an unrolled length of 10 m and width of 0.1-10 m. The crystal orientation is predominantly {100} toward the normal of the substrate surface in the respective place with a disorientation of 5[deg]. The grain boundaries are found on a regular lattice or are located in a surface of 1 cm around the grid lines of the lattice. The grid line distance is 3 cm, and other substrate areas are free of grain boundaries. Within the ranges mentioned around the grid lines, a crystal defect density is increased or deviation of the plane parallelism of the surfaces is increased or dispersion of the substrate thickness is increased. The crystal orientations in longitudinal direction of the individual pieces of substrate have same values (within a deviation of 5[deg]). The foil substrate is part of a solar module in cut form. The foil contained in the module on both surfaces, has a random pyramidal structure with an expansion of the pyramid of 0.1-10 mu m. The foil is attached on a glass-, ceramic- or metal plate and the attachment is interfered between the silicon foil and the plate by aluminum or silver containing material. An independent claim is included for a procedure for the production of an expanded silicon substrate.

Description

field of use

The The invention relates to a device according to the preamble of the claim 1, and a method for its production according to the preamble of Claim 13.

State of the art and their disadvantage

Monocrystalline Silicon substrates are used to produce a variety of semiconductor devices. These include also solar cells produced from monocrystalline silicon wafers, the opposite Silicon solar cells produced by layer deposition on other substrate materials be prepared, such. Cells made of amorphous silicon, clearly higher Achieve efficiencies. Efficiency benefits also arise across from Solar cells made on the basis of multicrystalline silicon become.

According to the state Technically, these substrates are predominantly sawed and sawed Polishing solid monocrystals made. The maximum extent the substrate surface is by the dimensions limited to the solid crystal used.

Out The maximum expansion results in restrictions on efficient production because instead of less big Parts many individual substrate discs must be treated. Also for the production large-area components, such as. large photovoltaic modules would be a Substrate that needed area the application is advantageous.

The Sawing and Polishing the hard silicon is very expensive. The consumption high-purity (and therefore valuable) semiconductor silicon is high because the sawn ones Due to the process, disks must have a certain minimum thickness (about 100-200 μm). For many Components would be In contrast, a significantly lower substrate thickness sufficient or even advantageous. In addition, the sawing process Material in the area of approximately 200 to 300 μm wide kerfs destroyed.

In order to avoid some of the disadvantages mentioned, different tape drawing methods have been developed for special applications, such as photovoltaics, such as SSP or EFG [cf. A. Götzberger, B. Voss, J. Knobloch, Solar Energy: Photovoltaics, Teubner Stuttgart (1994), page: 146-148 ]. In contrast to older melt crystal growth processes, such as the Czochralski process, the resulting growth body has the advantageous band shape for use, so that sawing and polishing can be dispensed with. Disadvantage of the tape drawing method, however, is that it does not succeed to breed monocrystals. Accordingly, the density of other structural crystal defects is comparatively high. Therefore, the achievable efficiency of solar cells made of this material is lower than monocrystals. Furthermore, the flexibility of the bands is low. It is not enough to wrap on a roll. Long pieces break easily during handling, therefore the strips are cut into handy pieces (eg 0.15m x 0.15m) immediately after manufacture and are thus comparable to the size of sawn slices. With regard to maximum substrate size, the previous tape drawing methods therefore have the disadvantages already mentioned. Also, the band thickness is typically at least 200 microns, which is significantly thicker than is absolutely necessary for a solar cell, so that these methods still require relatively much silicon.

To avoid some of the disadvantages mentioned, layer transfer processes (LTP) have been developed. In contrast to the strip drawing methods already discussed, the layers are obtained from a single crystal, so that the layers produced are likewise monocrystalline and contain fewer structural crystal defects. One such LTP process is the PSI (Porous Silicon) process [cf. R. Brendel, Thin-film crystalline silicon solar cells, Wiley-VCH, Weinheim (2003), pages 108 to 115 ]. For this purpose, a porous double layer is produced electrochemically on the surface of a monocrystalline silicon wafer. On the upper of these layers, a nonporous monocrystalline layer is deposited by means of epitaxy. The lower of the porous layers is used in the further process sequence finally as a predetermined breaking point for detachment of the thin (eg 5 microns) epitaxially produced silicon layer. The detachment takes place under application of force. The detached silicon layer is processed further to the component, the disk, however, repeatedly used again, resulting in economic benefits. Since the layers are obtained from a silicon wafer ge, however, the disadvantages already described in terms of limited surface area of the substrate remain. Furthermore, no self-supporting silicon foils are produced, but a glass substrate is already adhered to the epitaxial layer before detachment, ie the layer produced is transferred from the original substrate to a new substrate. The glued glass substrate, however, means significant limitations for further processing. For example, all subsequent process steps must be carried out at a temperature compatible with the adhesive and the glass substrate. Furthermore, heating and cooling rates are limited by the thermal shock resistance of the glass. On the other hand, the production and further processing of self-supporting layers with this method would be very difficult, because the tearing takes place under application of force and, in particular, cracks in the edge region are generated, which reduce the mechanical strength drastically. In another process (quasi-monocrystalline silicon QMS), self-supporting porous layers are produced [cf. PJ Rostan, C. Berge, U. Rau, JH Werner, Proceedings 15th PVSEC Shanghai (2005), pp. 628-629 ]. In contrast to the PSI method, no epitaxy is initially applied to the porous layer system, ie the detached layer is also completely porous. The detachment of the layer is carried out by wet chemical etching without external mechanical force application. However, due to their porosity, the layers have a low mechanical strength, which restricts further processing. The disadvantages regarding the limited extent remain.

Object of the invention

task The invention is an apparatus in the form of an extended To provide substrate for the production of semiconductor devices, the self-economical makes, the usual size restrictions in terms of surface area overcomes and allows the use of highly efficient role to role production process, Furthermore to provide a method of manufacturing said substrate.

Solution of the task

These The object is achieved by a device having features of claim 1 or solved by a method according to claim 13.

Advantages of the invention

By the invention will be compared the prior art disadvantages avoided while achieving significant advantages.

at the method can Slides of the desired Substrate thickness are produced, so that only a little of the high purity (and thus expensive) semiconductor silicon is consumed. costly Sawing and polishing eliminated. Due to the small thickness and the flexibility of the films, there are advantages when embedding in a solar module. Besides, can be flexible electronic Components such as e.g. Sensors or solar cells are produced. These components also work well in non-planar devices surfaces integrate. According to the invention can after the production of a first generation film, in the still slices made of a solid crystal are needed Further substrate films produced directly from this first generation become. The substrate films can in any length and width are produced.

by virtue of their flexibility and sufficient mechanical strength, the film can be rolled up. For the Production and finishing of the films can roll to roll procedures be applied, which in terms of Materialzu- and discharge to individual slices a significant increase in efficiency means. Despite high material throughput, the processes can be hermetic closed chambers performed In this case, complete film rolls are inserted or removed. The litigation in closed chambers is advantageous because it impurities in air, moisture or dust can be prevented. The Avoidance of dust and particle entry is in all semiconductor processes very critical and otherwise usually only by production in elaborate clean rooms manageable. Some of the processes also have toxic or explosive effects Gases are used which must not escape uncontrolled. at the role to role production can the individual processes are carried out in independent machines. Opposite one Line production, in which all involved processes work synchronously and with the same throughput have to, that means a significant simplification. Despite the procedure achievable substrate extent, the films have all the essential Advantages of Monocrystals. Due to the low defect density can be used to produce solar cells with high efficiency. Due to the defined crystal orientation of the surfaces let surface-structuring etching processes very good, given a target low film thickness essential for the achievement of a high efficiency is. A few grain boundaries, the process caused at the joints on the other hand, they are of little importance because they are only predefined Places arise and in further processing, e.g. as part of structuring and interconnection, and thus neutralized can be.

As a result the according to the procedure special process management at transfer process Microcracks are avoided especially in the edge area, which a high breaking strength of the film is ensured. This is an essential prerequisite for the production and processing large-area substrate films.

The Second generation films are produced by epitaxy. That means inhomogeneities the dopant distribution, which in the produced by Schmelzklechtung Massive crystals are formed as a result of macrosegregation and microsegregation, can be avoided. In epitaxy, however, the dopant insertion can be controlled very precisely become.

Another advantage of the process is that both sides of the film are ready for further processing are accessible. This allows a variety of device optimizations, such. B. a pyramid-structured surface on both sides, which is optimal for the quantum efficiency of a thin-film solar cell.

Description of exemplary embodiments

embodiments The invention are illustrated in the drawings and are in Following closer described. Show it:

1 Perspective drawing of a substrate film produced by joining individual pieces of film partially rolled up.

2 Substrate film produced by joining individual pieces of film.

3 An assembly for joining immediately before melting.

4 An assembly for joining during melting.

5 An arrangement accordingly 4 from another perspective.

6 An assembly for joining before melting.

7 An arrangement for porosifying a substrate wafer.

8th An arrangement for porosifying a substrate film in a continuous process.

9 Detail of an arrangement accordingly 8th for the purpose of separating the electrolyte in the area between the moving foil and the fixed partition wall.

10 Arrangement for epitaxially coating a rolled-up silicon foil in the roll to roll method.

11 Layer structure in the edge area immediately before the beginning of the detachment.

12 Arrangement for detaching the daughter foil from the parent foil in a continuous process.

1. film substrate

In 1 and 2 is an example of a silicon foil produced by joining 10 shown. The foil 10 is made of monocrystalline foil pieces 1a . 1b . 1c etc. composed. The thickness d is 0.1 .mu.m to 100 .mu.m, preferably 1 to 30 .mu.m, so that the film, as in 1 indicated, flexible and thus on a cylinder 14 with diameter r can be rolled up. The roll-up is an essential feature of the film 10 because it is the prerequisite for efficient production and processing of long films. The diameter r is less than 1/3 of the film length z, but not more than 5 m, preferably not more than 2 m.

The width b is 1 cm to 10 m, preferably 0.1 m to 3 m and the length z is at least 0.5 m and can due to the Aufrollbarkeit very large values, such as 10 km, assume. Preferably, the joints are in the transverse direction 11 and longitudinal direction 12 in the planar arranged film in a regular grid, with grid spacing p and q, for example 3 cm to 50 cm, arranged. The surface normal 13 has the same orientation to the crystal lattice at every point, for example the <100> direction. The area immediately around the joint (ie joining area with a distance f up to 1 cm from the joint) is characterized by the occurrence of at least one grain boundary (which always means sub-grain boundaries) that runs along the joint, while outside this area there are no grain boundaries occur. An increased density of crystal defects may be localized especially in the joining region. In addition, due to the melting process in the joining area, there will often be an increased deviation from the plane parallelism and from the average film thickness in this area.

The Crystal orientation is parallel to the surface of the individual pieces of film preferably aligned regularly. This orientation can be used for all film pieces be similar, leaving the planar rolled-up film, aside from the grain boundaries in the region of the joints, a monocrystal equivalent. Also in this case, the grain boundaries mentioned, due of technical inaccuracies in the alignment of the individual pieces of film when joining, practically unavoidable.

Alternatively, the orientation of the film pieces can also change between several orientations. In the application example of 2 alternates the crystal orientation in the longitudinal direction 16 of the pieces of foil between checkered <010> and <011>. The crystal orientation in the direction of the surface normal, however, is everywhere <100>. An alternating arrangement can have advantages for mechanical reasons because then the crystallographic cleavage planes preferred in silicon ( 111 ) in the composite film do not form uniform levels, so that, for example due to a production disruption cracks in the film have less dramatic consequences.

2. Add

3 . 4 and 5 show an example of an arrangement 20 for joining, or extending the already described substrate film 10 , According to 3 becomes first a silicon foil pieces 1 on impact or overlap on the already manufactured substrate film 10 created. The process steps that are suitable for the film piece 1 will be described later.

Plates of quartz glass 21 (above) and 22 (below) with a thickness of approx. 0.5 to 3 mm serve to fix the position of the foil piece 1 and the foil 10 , According to 4 Then the silicon is melted in the region of the impact or overlap. For this purpose, by means of a laser 24 a radiation cone 25 generated with a wavelength in the range of 0.2 to 1.2 microns and on the spot 26 directed. The plate 22 is for the radiation 25 transparent and permeates them. The melting zone 26 can, but does not have, the plates 21 . 22 touch. The width w of the melting zone must be greater than the overlap area and is typically 3 μm to 1 cm, preferably 30 μm to 1000 μm.

5 shows the same arrangement from above, but for clarity, without a laser 24 and without covering quartz glass plate 21 , The arrow 27 shows the direction of movement of the melting zone 26 by uniformly moving the radiation cone 25 is effected. In the direction of movement before the melting zone 26 there is still a foil edge 23 before, which is no longer present behind the melting zone. When not too fast movement (ie up to 3000 mm / min, preferably up to 100 mm / min) takes place at the rear end of the molten zone 26 an ordered crystal growth. At least one grain boundary occurs 28 which spreads in the same direction. In order to avoid edge-induced problems at the beginning of the process, it may be appropriate, as in the 5 shown the described process with 2 lasers 24 symmetrical to the plane 29 to expire. For a higher throughput, it may be useful to have additional lasers parallel to the edge 23 to be arranged, whereby only one section is processed by each laser.

In the crystallized area, various defects such as grain boundaries or dislocations may occur, especially on wall contact (in this case contact with the plates 21 or 22 ) are due to the phase boundary during crystal growth. Therefore, a two-step process can be useful. In the first step, a joint is produced as described, wherein additionally by pressing the plates 21 versus 22 a reliable contact of the films 1 and 10 in the overlap area during melting is effected. Wall contact is tolerated. In the second step is by appropriate measures, such as reduction of the contact pressure or inflow of gas into the column 31 and 32 which prevents wall contact in a re-melting of the joint area. The melt width w in the second step must be slightly larger than in the 1st step.

After completion of a joining, the laser operation is interrupted, the plate 21 raised the slide 10 moved horizontally to the left, so that the next edge comes into the radiation range of the laser. Then a new silicon foil piece 1 hung up. After lowering the plate 21 the process starts again.

Another arrangement for joining 30 is in 6 shown. First, the silicon films 10 and 1 by clamping between jaws 32 respectively. 31 (eg of quartz glass or coated steel) according to 6a fixed. The foil edges 34 and 33 must be parallel to each other and have the same distance s (eg 3 mm) to the respective clamping jaw. The tolerances are preferably of the order of magnitude, the film thickness. If the edges do not meet this requirement, then it is expedient to bring the films to dimensional stability after clamping, for example by laser cutting. By means of a positioning arrangement, not shown, then the pair of jaws 31 in the x and y directions according to 6b shifted so that an overlap with the 1 to 30 times the film thickness takes place and also an elastic deflection of the films is generated, so that the resulting bending stress is sufficient to reliably press the film ends against each other. The further processing can be carried out by melting by laser as in arrangement 20 be performed.

at the described joining processes It may vary by atmosphere come to the formation of silicon oxide. To prevent the oxide formation or to keep it within controlled limits it is convenient to Process in an inert or reducing atmosphere or controlled To carry oxygen content, ie e.g. to use a mixture of nitrogen and hydrogen. For the same reason, the use of vacuum may be appropriate. In these cases it is advantageous if the process is in a not shown Closed chamber is realized. Due to the rollability of the resulting film are compact dimensions of the chamber of e.g. 2 to 4 m in each spatial direction feasible.

3. porosity

For the production of the expanded silicon foil film pieces are required, which can be produced with a transfer process described below. Alternative to already be As a result of the subsequent joining process, foils of a later generation can also be produced directly from the foil of an earlier generation, which serves as a substrate foil, by means of the now following detachment process.

For detachment, a porous double layer must first be created. For this purpose, according to application example in 7 the silicon wafer 52 or according to 8th the silicon foil 53 into a vessel 51 with electrolyte-containing liquid (liquid surface 54 ), for example, a mixture of HF, water and propanol, immersed. The silicon substrate 52 respectively. 53 effected in connection with the partition 55 a spatial separation of the solution in two sections, the anode compartment 56 where the silicon acts as an anode and the cathode compartment 57 where the silicon acts as a cathode. In the anode compartment, the silicon is porosified, here the electrolyte must contain hydrogen fluoride. In the respective subarea is also in each case the associated counter electrode 58 and 59 , which can be realized for example by platinum networks. The counter electrodes are coupled via an external circuit, not shown.

In the case of porosification of a film 53 It is expedient to pass the film in continuous feed, according to the direction 66 to edit and the film by means of rotating rollers 61 respectively. In this case, it may be difficult to create a hermetically sealed connection between the bulkhead 55 and the foil 53 to achieve. Remedy can be done here according to 9 the flushing of an inert medium, represented by the directional field 65 , such as nitrogen gas or deionized water in the space 62 the partition 55 create. For example, a gap 62 of less than 1 mm, by means of the support 64 , adjusted, then it can prevent penetration of the electrolyte 63 and thus a contact of the electrolyte between the two subspaces are largely prevented. Finally, the imbibed medium mixes with the electrolyte or rises up in the form of gas bubbles.

To achieve a porosity, it is advantageous if the silicon is doped. If porosity is carried out without illumination of the silicon, then p-doped silicon having a dopant concentration of more than 1 16 cm ^-3 , preferably more than 3 e 18 cm ^-3, must be used. Alternatively, n-doped silicon may be used, in which case it is necessary to generate the p-type carrier required for porosity by light irradiation on the silicon electrode. Since in this case the porosification takes place only where light is irradiated, plus a region which is given by the diffusion length of the p-type charge carriers, in this way the porosity can be restricted to an advantageous spatial range.

After applying an external current, the pore growth occurs at the silicon anode. The penetration depth of the pores into the silicon is controlled substantially over the duration of the process. The ratio of the pore volume to the total volume, ie the porosity within the porous layer is controlled by the external current flow and possibly the illuminance, wherein a high porosity is generated by a high current flow. Typical currents per anode area of the silicon are 2 to 300 mA / cm ² . Advantageously, it is first an approximately 0.1 to 1 micron thick low-porous layer, ie porosity, for example 10-30% and below about 0.1 to 1 micron thick highly porous layer to produce with a porosity of eg 40 to 70% ,

At the electrodes 58 . 59 , the silicon anode and silicon cathode generate gases, in the application examples O ₂ at the electrode 59 , and H ₂ on the other electrodes. The associated rising bubbles of the various gases are fed separately by means not shown walls of a suction. To the gas discharge at the silicon cathode, ie the underside of the substrate 52 respectively. 53 To promote, it is appropriate to tilt the substrate slightly to the horizontal.

4. Annealing and epitaxy

For the epitaxy process, it is advantageous if the pores are closed at the surface beforehand. For this purpose, the porosified silicon substrate is annealed under H ₂ atmosphere at about 900 to 1200 ° C. In this case, a rearrangement of the silicon takes place, in which the porosity is obtained, but the surface is closed. Thereafter, a non-porous layer can be applied by Epi taxie. Advantageous for this purpose is a large-scale deposition process.

In 10 is an embodiment of a roll to roll epitaxy plant 70 shown. The two roller chambers 71 and 73 and the process chamber 72 together form a closed chamber, only for installation or removal of the rollers 67 including substrate wound thereon 75 is opened. The roller chambers are each a slot 74 , with a width of eg 0.1 mm to 10 mm connected to the process chamber. Inert gas is flowed in the region of the slot, so that no process gas flows into the roller chambers and also no gas from the roller chambers enters the process chambers. The latter is important because there may be residual amounts of water vapor in the roller chambers which would adversely affect the epitaxy process. The substrate film 75 is according to the feed direction 76 continuously added and removed. infrared lamps 68 shine through the quartz glass plate 69 through to the substrate and heat it in the area within the process chamber 72 at about 1100 ° C.

The coating gas consisting of H ₂ and HSiCl ₃ flows, represented by the arrows 77 , perpendicular to the plane through the inlet 78 down into the process chamber, rises on the opposite wall, not shown, and flows along the bottom of the substrate 75 where it causes an epitaxial silicon deposition. The reacted gas escapes through the outlet 79 ,

5. Peel off

After epitaxial application of a nonporous layer 83 , according to 11 , this layer can be detached, the highly porous layer 82 serves as a predetermined breaking point, which has only a low breaking strength due to the porosity. As in 11 shown it is advantageous for initializing the detachment process when at the end of the film 80 on both sides eg Kunsthof films 84 glued on. Mechanical then becomes in the area of the porosized layers 85 and 82 generates a notch stress. 12 shows an embodiment for a continuous feed, corresponding to the arrows 96 , performed peeling process. The detachment can additionally by dipping the film 80 into a vessel 91 with a silicon etching solution (with surface 94 ), such as a mixture of HF, HNO ₃ and water. Since the etching rate of porous silicon is much higher than that of nonporous silicon, the porous layers are in the range 95 selectively removes the notch created during peeling. The thus low mechanical force and the etching of open pores prevents microcracks from forming on the edges of the film, which is crucial for sufficient strength for further processing of the substrate film. In order to avoid cracks, in particular at the edges of the film, it is expedient if the epitaxial layer 83 according to 11 not to the edge of the film 81 is enough, but if everywhere in the border area an uncovered area 86 the porous layers 85 and 82 can be attacked immediately by the etching solution.

The detached layer (ie daughter foil) 93 has similar properties as the original substrate film (ie parent film) 92 except for a slightly smaller extent.

6. Overall process

Of the joining process is only for the first generation of films required. Other generations can too produced by a combination of porosity, peeling and epitaxy become. The mother foil loses due to the porosity and detachment a part of their film thickness. Epitaxy can cause this thinning out Need undone become.

7. Further processing

For further processing, it is advantageous to edit the film also in roll to roll processes. Depending on requirements, both sides of the film can be processed. For solar cells, for example, the following process steps are possible:
The diffusion of an acceptor or donor dopant to create a pn junction, the wet chemical structure etching to produce random pyramids on the surfaces to increase the optical absorption of the light, especially in the long wavelength range of the irradiated spectrum.

In addition, can also the application of amorphous silicon layers, the different Can have functions such as. the creation of a heteroemitters, the production of a Surface passivation or in the case of several amorphous layers for producing a Tandem solar cell, done this way.

8. Apply to a solid substrate

Then, if the applied structures a strong bimetallic bending would result in the film it is necessary to apply the foil to a solid substrate beforehand. This counts e.g. the application of contact fingers for electrical contacting at a solar cell. The film is cut for this purpose and e.g. by means of aluminum or an aluminum-containing paste on a massive substrate, such as e.g. a glass plate, attached, which then also serve for module encapsulation in the case of solar modules can. When heated to about 600 ° C The aluminum provides a firm connection between glass and silicon and at the same time to the back serve electrical contact. Structuring and frontal Contacting is preferably done after this step.

Claims (30)

Device in the form of a substrate ( 10 ) of silicon for the production of semiconductor components or for the production of further such substrates having a thickness d of 0.1 μm to 100 μm and a width and length of at least 1 cm, predominantly consisting of monocrystalline regions ( 1a ) 1b ) 1c ), etc., with a length and width extent of at least 1 cm each, all of which have a uniform crystal orientation to their respective surface normal ( 13 ) with a deviation of at most 5 °, characterized in that the exp z of the planar substrate ( 10 ) in the longitudinal direction is at least 0.5 m and its mechanical properties are a reversible rolling on a cylindrical surface ( 14 ) with a diameter r of at most one third of the length z, but not more than r = 5 m. Device according to Claim 1, characterized in that the monocrystalline regions ( 1a ) 1b ), etc. have at least an extension of 3 cm, the length z of the film ( 10 ) is at least 3 m and the film is reversible on a cylindrical surface ( 14 ) with a diameter of at most 2 m. Device according to one of the preceding claims, characterized in that the substrate ( 10 ) has a rolled-out length z of at least 10 m and a width b of 0.1 m to 10 m. Device according to one of the preceding claims, characterized in that the crystal orientation predominantly <100> in the direction of the normal ( 13 ) of the substrate surface at the respective location is with a misorientation of at most 5 °. Device according to one of the preceding claims, characterized characterized in that the grain boundaries are on a regular grid or in an area up to 1 cm wide around the lines are located of said grating, wherein the grid line spacing p and q is at least 3 cm, and the other substrate areas are free of grain boundaries. Device according to claim 5, characterized in that that within the said areas around the grid line respectively exactly one grain boundary is located. Apparatus according to claim 5 or 6, characterized that within the said areas around the grid lines, opposite to the Areas outside, increases the crystal defect density or the deviation from the plane parallelism of the surfaces is increased or the scattering of the substrate thickness is increased. Apparatus according to claim 5, characterized in that the crystal orientation in the longitudinal direction ( 16 ) of the individual substrate pieces ( 1a ) 1b ), etc. assumes alternating values, so that, for example, the <010> and <011> orientations alternate in a checkerboard pattern. Apparatus according to claim 5, characterized in that the crystal orientation in the longitudinal direction ( 16 ) of the individual substrate pieces ( 1a ) 1b ), etc. all have the same values (within a maximum deviation of 5 °). Device according to one of the preceding claims, characterized in that the film substrate ( 10 ) is in cut form part of a solar module. Apparatus according to claim 10, characterized in that the film contained in the module ( 10 ) on both surfaces has a random pyramidal structure with an extension of the pyramids of 0.1 .mu.m to 10 .mu.m. Device according to one of the preceding claims, characterized in that the film ( 10 ) is mounted on a solid substrate such as a glass, ceramic or metal plate and the attachment between the silicon foil and the solid substrate is mediated by aluminum or silver containing material. Method for producing a device according to one of the preceding claims, characterized in that a first generation of films by joining of either monocrystalline foil pieces or already one with joined to this method Film is made by two films or pieces of film as described juxtaposed and in the joining area be melted and crystallized again, with the individual foil pieces in terms of their surface normal have the same crystal orientation, more generations of films optionally also by means of a combination of epitaxy and detachment processes, using an earlier Generation as a substrate. Method according to claim 13, characterized in that that at the coincidence shifted the phase boundary at a speed of less than 3 m / min and an orderly crystal growth takes place, in the joining area predominantly exactly a grain boundary is created, which runs along the join. Method according to claim 13 or 14, characterized that melting in the joining area by laser radiation and the laser beam along the joint is moved at a speed of less than 3 m / min. Method according to claim 13 or 14, characterized that melting in the joining area by infrared radiation at a color temperature of the radiator of 2000 K to 3500 K takes place by the radiation either on a Spot or a line along the joint with a line width from at most 1 cm is focused. Method according to one of the preceding claims, characterized in that for joining the pieces of film in the region of the joint between quartz glass parts, for example in the form of plates or rolls stabilized above and below and any overlapping areas are pressed against each other. Method according to one of the preceding claims, characterized characterized in that the addition in a two-step process, with the first step made by melting and crystallizing a compound and in the second step, the previously crystallized material again melted and crystallized, the melt in the second step except to the silicon foil has no contact with a solid or to glass. Method according to one of the preceding claims, characterized characterized in that cut immediately before joining the foil edges the slides remain clamped afterwards and with an overlap width, which corresponds to 1 to 30 times the film thickness, positioned there be melted and crystallized. Method according to one of the preceding claims, characterized characterized in that during the joining process the one to be joined Place with a gas mixture consisting of inert gas or hydrogen is applied or the joining process in a vacuum. Method according to one of the preceding claims, characterized characterized in that before peeling Another generation of foil or for the production of the foil pieces for the first Generation a porous one Layer is produced on the film or a silicon wafer, the as breaking point for the detachment process serves. Method according to claim 21, characterized that the detachment process, according to the description, by admitting a corrosive solution in the peeling supporting notch becomes, the porous one Silicon opposite nonporous Silicon preferably dissipates. Method according to claim 21 or 22, characterized that, the one to be replaced nonporous Layer is applied by epitaxy after porosification. Method according to claim 23, characterized that according to the description on the edges the film the epitaxial layer does not completely cover the porous layer, so that along a line along the edges the porous one Layer to the surface enough. Method according to claim 21, characterized that a mother foil after peeling one or more daughter films as needed by epitaxy again thickened. Method according to claim 23 or 25, characterized that the epitaxy process as described from roll to roll, with the roll chambers along with the process chamber form an overall closed container and the substrate feedthrough between roller and process chamber through narrow slots. Method according to claim 21, characterized that the generation of the porous layer in an electrically conductive solution (Electrolyte) takes place by the silicon substrate at least part a limitation is that the solvent in two subspaces separates, each having an electrode connected via an external circuit as a result, on one side of the delimiting substrate that Silicon porous becomes. Method according to Claim 27, characterized that a silicon foil, according to the description, with constant feed rate in the range of 0.3 mm / s to Continuously tracked at 5 m / s so that eventually becomes the entire film, except for edge areas, on one side a porous layer Has. Method according to claim 27 or 28, characterized that the separation of subspaces through a moving substrate and a fixed wall, wherein substrate and wall are separated by a narrow gap and into the gap an electrically insulating liquid, e.g. deionized Water, or a gas flushed is what causes penetration of the electrolyte into the gap and thus a contact of the electrolyte between the subspaces largely to prevent. Method according to Claim 27, characterized that n-type silicon is porosified by the to be porosized Place in addition with Light (i.e., wavelength in the range of 0.2 μm up to 1.2 μm) is irradiated, creating a sharp spatial demarcation of the moment porosified site takes place.