Sound transducer in a microelectromechanical system comprising a sectionally activatable membrane and a method for manufacturing the membrane

Abstract

The present invention relates to a sound transducer in a microelectromechanical system (1) comprising a membrane (4) attached to a frame structure (2) with vertical (24, 24', 24") and horizontal (30, 30', 30") sections, wherein the membrane (4) comprises a continuous support layer (10) and a piezoelectric layer (12, 14) formed on the horizontal sections (30, 30', 30") and at least partially on the vertical (24, 24', 24") sections. The invention further relates to a method for manufacturing the membrane (4).

Description

Technical field

[0001] The present invention relates to a sound transducer in a microelectromechanical system comprising a membrane attached to a frame structure with vertical and horizontal sections, wherein the membrane comprises a continuous support layer and a piezoelectric layer formed on the horizontal sections and at least partially on the vertical sections. The invention further relates to a method for manufacturing the membrane. State of the art

[0002] Electromechanical devices are in increasing demand at ever smaller scales, which is why their manufacture is a central area of ​​microsystems technology. Particularly in the field of communications technology, micro-scale loudspeakers and microphones have gained increasing popularity in recent years. Transducers in microelectromechanical systems (also known as MEMS transducers) are known from the prior art.

[0003] Early MEMS transducers often featured a planar diaphragm that was excited to vertical vibrations piezoelectrically, electromagnetically, or electrostatically. The disadvantage of such planar diaphragm systems, namely reduced sound power, was to be overcome by using meandering diaphragms. These meandering diaphragms exhibit a 3D folded-sheet structure, meaning that the diaphragm displays a waveform in which peaks and troughs continuously transition into each other.

[0004] Such membranes are usually multi-layered, with a piezoelectric layer continuously applied to a support layer, meaning over the entire length of the membrane.

[0005] The production of such multilayered meandering membranes is achieved via a multi-stage deposition process, either by chemical vapor deposition (CVD), in particular atomic layer deposition (ALD), or by physical vapor deposition (PVD), in particular sputtering. Both processes are known from the prior art.

[0006] The ALD process offers the advantage of being easily repeatable for specific layer thicknesses without damaging the previous layer. A disadvantage is that the meandering membranes, due to their inherent height differences (wave crests and troughs), require very long deposition times to achieve the necessary layer thicknesses. Furthermore, the deposited layers often contain oxygen impurities, which is particularly detrimental to the piezoelectric layer. Moreover, given the required layer thicknesses and the associated long deposition times, implementing the ALD process as a batch process is difficult, which would, of course, be desirable.

[0007] Sputtering, on the other hand, is known for producing precisely thin piezoelectric layers; however, due to the physics of deposition, there is a risk that these layers will not form uniformly. This can lead to gaps in the layers, particularly in the areas of the wave depressions. Furthermore, in the very thin regions of the layers, there is a risk of electrical breakdowns, which means that the applicable voltage is severely limited.

[0008] Accordingly, there is immense demand for membranes with non-planar geometry that can be manufactured simply, quickly and cost-effectively, with individual layers free of gaps or impurities. Disclosure of the invention

[0009] According to the invention, a sound transducer in a microelectromechanical system is proposed, comprising a frame structure and a diaphragm attached to the frame structure. The diaphragm comprises vertical sections, each with an upper end and a lower end. The upper ends of the vertical sections preferably form the top surface of the diaphragm, and the lower ends preferably form the bottom surface of the diaphragm. The upper ends of the vertical sections are connected to each other via horizontal sections. The diaphragm comprises a support layer and a piezoelectric layer, the support layer being continuous over the entire length of the diaphragm. The piezoelectric layer is fully formed on the horizontal sections of the diaphragm and at least partially on the vertical sections.The piezoelectric layer can be a first piezoelectric layer or a second piezoelectric layer; in the case of only one piezoelectric layer, this always refers to the first piezoelectric layer. In the case of multiple piezoelectric layers, the invention always refers to the last layer applied.

[0010] The frame structure according to the present invention can, in principle, have any geometry, with the membrane preferably being attached to two opposite sides of the frame structure. The attachment can be direct or via additional retaining devices. The frame structure can comprise, but is not limited to, silicon compounds, nitrides, germanium compounds, gallium compounds, indium compounds, carbon, glass, or mixtures thereof.

[0011] The vertical sections of the membrane according to the present invention run almost parallel to the side walls of the frame structure and thus in the direction of flow of the fluid, wherein the fluid can be liquid or gaseous, but preferably gaseous.

[0012] The horizontal sections connect the upper ends of the vertical sections, thus running orthogonally to the vertical sections, starting at the respective side wall of the frame structure. The first horizontal sections attached to the frame structure may also be slotted for improved flexibility.

[0013] According to the invention, the membrane can be a membrane with a 3D folded structure, as described above. In this case, the lower ends of the vertical sections are connected to each other via bridge sections. In such a membrane, the frame structure is thus followed by a horizontal section, a vertical section, a bridge section, another vertical section, and so on. Alternatively, the membrane according to the invention can be designed as a membrane with a lamellar structure. In this variant, the lower ends of the vertical sections are designed as exposed ends; the bridging sections according to the 3D folded structure are therefore not present.

[0014] According to the invention, the piezoelectric layer is therefore not continuous across the entire membrane, particularly not at the bridge sections in the case of a membrane with a 3D folded structure. In the case of more than one piezoelectric layer, this applies especially to the last layer applied, i.e., the uppermost piezoelectric layer, but can also apply to all subsequent piezoelectric layers. The piezoelectric layer is thus only present on the support layer in sections, i.e., with interruptions. The coverage of the vertical sections with the piezoelectric layer is greater than 0% and less than or equal to 100%.

[0015] In suitable embodiments, the support layer is made of microcrystalline silicon or polysilicon. When the membrane is connected to a voltage source, the support layer, together with the voltage source, functions as the bottom electrode, i.e., the lowest conductive layer. Preferably, the first piezoelectric layer is an AlN (aluminum nitride) layer or an AlScN (aluminum scandium nitride) layer.

[0016] In a first embodiment of the transducer according to the invention, the piezoelectric layer is formed as a taper at the vertical sections. This means that the piezoelectric layer is present at its full thickness at the upper end of the vertical section, i.e., directly at the transition from the horizontal to the vertical section. This thickness is in the range of 100 nm to 1000 nm, preferably in the range of 300 nm to 500 nm. Along the vertical section, i.e., from the upper end to the lower end, the thickness of the piezoelectric layer decreases continuously, so that the layer thickness in cross-section is wedge-shaped.In a further alternative embodiment of the sound transducer according to the invention, the piezoelectric layer is partially formed on the vertical section, meaning that the thickness of the piezoelectric layer is uniform from the upper end of the vertical section to approximately halfway along the vertical section, and that no piezoelectric layer is present on the second half, i.e., from halfway to the lower end of the vertical section. Preferably, at least one-third, and particularly preferably at least one-quarter, of the vertical section, starting from the upper end, is covered with a piezoelectric layer.

[0017] In a further alternative embodiment of the transducer according to the invention, the piezoelectric layer is almost completely formed on the vertical sections, that is, precisely up to the transition to the bridge sections in the case of the 3D folded structure or the exposed ends of the lamellar structure. However, the transition to the bridge sections themselves is not covered by the piezoelectric layer in order to avoid a kink in the piezoelectric layer. Preferably, no piezoelectric layer is formed on any of the bridge sections, which leads to better actuation of the vertical sections. In a less preferred embodiment, the piezoelectric layer is at least not formed on adjacent bridge sections, so that the intervening vertical section is subject to freer movement.If adjacent bridge sections are equipped with a piezoelectric layer, a so-called "buckling" can occur, which means that the piezoelectric layer on the adjacent bridge sections works against the movement of the vertical sections.

[0018] In a preferred embodiment of the present invention, the membrane comprises a first piezoelectric layer and additionally a first conductive layer, wherein the first conductive layer is formed on the horizontal sections and at least partially on the vertical sections of the membrane, preferably in the areas where a first piezoelectric layer is also formed. However, the first conductive layer preferably does not extend completely to the respective ends of the first piezoelectric layer, but begins and ends with a certain offset, meaning that the respective end sections of the first piezoelectric layer are not covered by the first conductive layer in order to avoid short circuits. Preferably, the first conductive layer functions as the top electrode when applied to a voltage source.The first conductive layer can comprise platinum, tungsten, tin oxide, monocrystalline silicon, polysilicon, molybdenum, titanium, tantalum, metal silicide, aluminum, graphite, copper or mixtures thereof.

[0019] In suitable embodiments, the membrane comprises at least one insulating layer. This layer can be formed between the support layer and the first piezoelectric layer. Alternatively or additionally, it can be formed between the first piezoelectric layer and the first conductive layer. In any case, the insulating layer is formed on the horizontal sections and at least partially on the vertical sections of the membrane. The insulating layer can also be continuous on the vertical sections or continuous over the entire length of the membrane on the support layer. The at least one insulating layer prevents electrical short circuits between the individual layers, particularly when a voltage is applied. Preferably, the at least one insulating layer comprises Si3N4 (silicon nitride).Furthermore, it is possible that the insulation layer, especially if it is formed on the first piezoelectric layer, is also formed as a tapered section.

[0020] In one embodiment of the transducer according to the invention, the diaphragm forms multilayer stacks in certain sections. In this embodiment, the diaphragm additionally comprises a second piezoelectric layer and a second conductive layer, wherein the second piezoelectric layer and the second conductive layer are formed on the horizontal sections of the diaphragm and at least partially on the vertical sections of the diaphragm. By applying different voltages to the support layer / first conductive layer or the first conductive layer / second conductive layer by means of a voltage source, the diaphragm can be excited differently, in particular, asymmetrical horizontal vibrations can also occur.Alternatively, one of the two conductive layers can be used as a sensor layer, meaning that the deflection of the membrane (vibration) is detected and converted into an electrical signal.

[0021] Preferably, the following multilayer stack, in particular in this order, borders the continuous support layer of the membrane on its horizontal sections and at least partially on its vertical sections: i) optionally an insulation layer ii) first piezoelectric layer iii) first conductive layer iv) second piezoelectric layer v) second conductive layer

[0022] According to the invention, the membrane, in particular the vertical sections of the membrane, is designed and configured to perform vibrations. In the case of the membrane with a 3D folded-sheet structure, these can be symmetrical horizontal vibrations or asymmetrical horizontal vibrations. In the case of the membrane with a lamellar structure and exposed ends, these are unsupported horizontal vibrations, which can also be symmetrical or asymmetrical.

[0023] According to the invention, a method for producing the membrane, in particular the membrane according to the invention, is further proposed, comprising the steps, in particular in this order: a) Provision of a converter structure with depressions b) Application of a continuous support layer c) Optional application of a continuous insulation layer d) Application of a first piezoelectric layer to the horizontal sections of the membrane and at least partial application to the vertical sections e) Application of a first conductive layer to the horizontal sections of the membrane and at least section by section to the vertical sections f) Removal of the converter structure while maintaining a frame structure

[0024] The recesses in step a) are obtained by reactive ionic deep etching or KOH etching using a suitable method. The continuous support layer in step b) is preferably obtained using a CVD process, as this provides fast deposition times. The CVD process can be selected from metal-organic vapor deposition (MOCVD), plasma-enhanced chemical vapor deposition (PECVD), or low-pressure chemical vapor deposition (LPCVD), but is not limited to these methods. The optional continuous insulating layer in step c) is applied using an ALD or CVD process in suitable embodiments. The piezoelectric layer in step d), preferably the section-by-section application on the vertical sections, can be obtained particularly well using a PVD process.Here, the process parameters are set such that the converter structure with the already applied layers b) and optionally c) tilts or rotates during the process, whereby angle values ​​of up to 90° are theoretically possible, but preferably these are less than 45° and particularly preferably between 2 and 15°. The application of the first conductive layer in step e) is also preferably achieved via such a process.

[0025] In an alternative version, the procedure can additionally include the following steps: h) Application of a second piezoelectric layer to the horizontal sections of the membrane and at least partial application to the vertical sections i) Applying a second conductive layer to the horizontal sections of the membrane and at least sectionally to the vertical sections, taking place after step e) and before step f).

[0026] The method according to the invention can have all the features that have already been discussed for the sound transducer according to the invention. Advantages of the invention

[0027] The sound transducer according to the invention in a microelectromechanical system with a membrane attached to a frame structure having vertical and horizontal sections, wherein the membrane comprises a continuously formed support layer and a piezoelectric layer formed on the horizontal sections and at least partially on the vertical sections, can be obtained as a sound transducer in a microelectromechanical system which has high sound conductivity and can be manufactured simply and cost-effectively.

[0028] By applying a voltage source to the vertical sections of the membrane, these can be selectively excited; that is, a stress gradient is created relative to the supporting layer, causing the vertical sections to bulge. Since the principle is not based on electrostatic attraction, adhesion of the vertical sections of the membrane can be prevented.

[0029] Because the first piezoelectric layer, the first conductive layer, and the potentially second piezoelectric and second conductive layers are not necessarily continuous across the entire membrane, and especially not at the bridging sections, deposition times in the ALD process can be significantly reduced, potentially to about 1 / 60th of the original deposition time. Oxygen impurities in the manufacturing process can also be significantly reduced.

[0030] In sputtering, the membrane structure according to the invention prevents gaps from forming in the individual layers, since in particular the treatment of the bridge sections is omitted.

[0031] Likewise, the risk of electrical breakdowns or breakthroughs is minimized.

[0032] Since the membrane is only formed section by section across all layers, only smaller capacitances need to be driven, resulting in lower power consumption. Brief description of the drawings

[0033] Embodiments of the invention are explained in more detail with reference to the drawings and the following description.

[0034] They show: Fig. 1 a schematic cross-sectional view of a first sound transducer according to the invention in a microelectromechanical system; Fig. 2 a schematic cross-sectional view of an alternative sound transducer according to the invention in a microelectromechanical system; Fig. 3 a schematic cross-sectional view of another sound transducer according to the invention in a microelectromechanical system; Fig. 4 a schematic cross-sectional view of a further alternative sound transducer according to the invention in a microelectromechanical system; Fig. 5 a schematic cross-sectional view of an additional alternative sound transducer according to the invention in a microelectromechanical system and Fig. 6 a method according to the invention for manufacturing a sound transducer in a microelectromechanical system. Embodiments of the invention

[0035] In the following description of embodiments of the invention, identical or similar elements are designated by the same reference numerals, and repeated descriptions of these elements are omitted in individual cases. The figures represent the subject matter of the invention only schematically.

[0036] Fig. Figure 1 shows a sound transducer 1 in a microelectromechanical system comprising a frame structure 2 and a membrane 4 attached to the frame structure 2. Fig. Figure 1a) shows a membrane 4 with a 3D folding structure 6, the membrane 4 according to Fig. 1b) has a lamellar structure 8. In both variants a) and b), the membrane comprises 4 vertical sections 24, 24', 24", each with an upper end 26, 26', 26" and a lower end 28, 28', 28". The upper ends 26, 26', 26" of the vertical sections 24, 24', 24" are connected to each other via horizontal sections 30, 30', 30". The two variants a) and b) differ from each other in that the membrane 4 with the 3D folding structure 6 according to Fig. 1a The lower ends 28, 28', 28" of the vertical sections 24, 24', 24" are connected to each other via bridging sections 32, 32', 32", whereas in the membrane 4 with lamellar structure 8, the lower ends 28, 28', 28" of the vertical sections 24, 24', 24" terminate in exposed ends 44, 44', 44" (i.e., no bridging sections 32, 32', 32" are present). Both membranes 4 comprise a support layer 10 and a first piezoelectric layer 12, the support layer 10 being continuous over the entire membrane 4. The first piezoelectric layer 12, on the other hand, is only formed on the horizontal sections 30, 30', 30" of the membrane 4 and at least partially on the vertical sections 24, 24', 24". In the Fig. 1a) The membrane 4 shown has the first piezoelectric layer 12 on the vertical sections 24, 24', 24" on the left side of the Fig. 1a) partially formed 34b, on the right side of the Fig. 1a) almost fully developed 34c. In Fig. 1b) The first piezoelectric layer 12 is partially formed on the vertical sections 24, 24', 24" 34b. The membrane 4 according to Fig. 1a), in particular their vertical sections 24, 24', 24", are designed and configured to perform symmetrical horizontal vibrations 36. In Fig. 1b) the membrane 4, in particular its vertical sections 24, 24', 24", is designed and configured to carry out unsupported horizontal vibrations 38.

[0037] Fig. Figure 2 shows an alternative embodiment of a sound transducer 1 according to the invention. In contrast to the sound transducer 1 according to Fig. In both variants a) for the diaphragm 4 with 3D folded structure 6 and b) for the diaphragm 4 with lamellar structure 8, the sound transducer 1 shown here comprises an additional insulating layer 16, which in the embodiment shown here is applied between the support layer 10 and the first piezoelectric layer 12 to prevent short circuits (not shown here). In both variants a) and b), the insulating layer 16 is applied continuously over the entire diaphragm 4, in particular continuously over the entire support layer 10.

[0038] Fig. Figure 3 shows an alternative embodiment of the sound transducer 1 according to the invention in a microelectromechanical system, wherein the diaphragm 4 has a 3D folded structure 6. In addition to the one already described in Fig. 1 a) The membrane 4 shown additionally comprises an insulating layer 16, which, in contrast to Fig. 2a) is applied between the first piezoelectric layer 12 and a first conductive layer 18. In the embodiment shown here, the first piezoelectric layer 12 is applied to the horizontal sections 30, 30', 30" and to the vertical sections 24, 24', 24" as a taper 34a, which means that the thickness of the first piezoelectric layer 12 on the vertical sections 24, 24', 24" decreases continuously from the upper end 26, 26', 26" to the lower end 28, 28', 28" so that a wedge shape is formed in cross-section. The insulating layer 16 and the first conductive layer 18 are formed with a constant thickness throughout the vertical sections 24, 24', 24".

[0039] Fig. Figure 4 shows a further alternative embodiment of a sound transducer 1 according to the invention for a membrane 4 with a 3D folded structure 6 according to Fig. 4a) and for a membrane 4 with lamellar structure 8 according to Fig. 4b). In contrast to the ones already in place for Fig. In the membranes 4 discussed here, the first piezoelectric layer 12 is partially formed in both variants a) and b) on the vertical sections 24, 24', 24" of the membrane 4, 34b, here only to about a quarter of the length of the vertical sections 24, 24', 24". In contrast to the one in Fig. In the embodiment shown in Figure 3, the membrane 4 shown here does not have an insulating layer 16. Furthermore, the first conductive layer 18 is not continuous over the entire length of the membrane 4, but only over the sections of the first piezoelectric layer 12. Furthermore, the membranes 4 of the Fig. 4 a second piezoelectric layer 14 and a second conductive layer 20 are formed, these also only being formed in the sections where the first piezoelectric layer 12 is formed. Thus, multilayer stacks are formed section by section.

[0040] Fig. Figure 5 shows an additional alternative embodiment of a sound transducer 1 according to the invention for a membrane 4 with a 3D folded structure 6 according to Fig. 5a) and for a membrane 4 with lamellar structure 8 according to Fig. 5b). In contrast to the ones in Fig. In the embodiments shown in Figure 4, the first piezoelectric layer 12 is designed as a continuous layer, i.e., over the entire length of the membrane 4. The first conductive layer 18 is also designed as a continuous layer. In the embodiment shown here, the second piezoelectric layer 14 and the second conductive layer 20 extend only to one-quarter of the length of the vertical sections 24, 24', 24". When a voltage is applied to the support layer 10 / first conductive layer 18 and to the first conductive layer 18 / second conductive layer 20 (not shown here), depending on the voltage difference, either symmetrical horizontal vibrations 36 of the vertical sections 24, 24', 24" can occur, as shown in Figure 1, or asymmetrical horizontal vibrations 40 can occur, as shown in Figure 2.

[0041] Fig.Figure 6 shows a method according to the invention for manufacturing a sound transducer 1, wherein the sound transducer 1 is shown in a cross-sectional view, in particular for manufacturing a sound transducer 1 according to the invention. In a first step, a transducer structure 22 with recesses 42, 42', 42" is provided, wherein the recesses 42, 42', 42" are etched as described above. In a second step, a support layer 10 is applied continuously, this being done by CVD as described above. In the next step, a first piezoelectric layer 12 is applied to the horizontal sections 30, 30', 30" of the diaphragm 4 and at least partially to the vertical sections 24, 24', 24", such that these are covered to one-third in the embodiment shown here, which is preferably achieved by PVD as described above.A first conductive layer 18 is then applied to the horizontal sections 30, 30', 30" of the membrane 4 and at least partially to the vertical sections 24, 24', 24". The first conductive layer 18 is not applied completely to the areas of the first piezoelectric layer 12, but with an offset, meaning that the ends of the first piezoelectric layer 12 remain free, thus preventing an electrical short circuit (not shown here). Finally, the transducer structure 22 is removed while retaining the frame structure 2.

[0042] The invention is not limited to the embodiments described here and the aspects highlighted therein. Rather, within the scope specified by the claims, a multitude of modifications are possible that fall within the bounds of what is considered skilled in the art.

Claims

[1] Sound transducer in a microelectromechanical system (1) comprising a framework structure (2) and a membrane (4) attached to the frame structure (2), wherein the membrane (4) comprises vertical sections (24, 24', 24") each comprising an upper end (26, 26', 26") and a lower end (28, 28', 28"), wherein the upper ends (26, 26', 26") of the vertical sections (24, 24', 24") are connected to each other via horizontal sections (30, 30', 30"), wherein the membrane (4) comprises a support layer (10) and a piezoelectric layer (12, 14), wherein the support layer (10) is continuous over the entire membrane (4), wherein the piezoelectric layer (12, 14) is formed on the horizontal sections (30, 30', 30") of the membrane (4) and at least partially on the vertical sections (24, 24', 24"). [2] Sound transducer (1) according to claim 1, wherein the membrane (4) has a 3D folded structure (6), wherein the lower ends (28, 28', 28") of the vertical sections (24, 24', 24") are connected to each other via bridge sections (32, 32', 32") or wherein the membrane (4) has a lamellar structure (8) wherein the lower ends (28, 28', 28") of the vertical sections (24, 24', 24") terminate in exposed ends (44, 44', 44"). [3] Sound transducer (1) according to claim 1 or 2, wherein the piezoelectric layer (12, 14) is not formed continuously over the entire membrane (4), in particular not formed at the bridge sections (32, 32', 32"). [4] Sound transducer (1) according to at least one of the preceding claims, wherein the piezoelectric layer (12, 14) on the vertical sections (24, 24', 24") is formed as a taper (34a), partially (34b) or almost completely (34c). [5] Sound transducer (1) according to at least one of the preceding claims, wherein the diaphragm (4) comprises a first piezoelectric layer (12) and additionally a first conductive layer (18), wherein the first conductive layer (18) is formed on the horizontal sections (30, 30', 30") of the diaphragm (4) and at least partially on the vertical sections (24, 24', 24") of the diaphragm (4). [6] Sound transducer (1) according to at least one of the preceding claims, wherein the membrane (4) comprises at least one insulating layer (16), wherein the insulating layer (16) is formed on the horizontal sections (30, 30', 30") of the membrane (4) and at least sectionally on the vertical sections (24, 24', 24") of the membrane (4), in particular wherein the insulating layer (16) is formed between the support layer (10) and the first piezoelectric layer (12) or wherein the insulating layer (16) is formed between the first piezoelectric layer (12) and the first conductive layer (18). [7] Sound transducer (1) according to at least one of the preceding claims, wherein the membrane (4) additionally comprises a second piezoelectric layer (14) and a second conductive layer (20), wherein the second piezoelectric layer (14) and the second conductive layer (20) are formed on the horizontal sections (30, 30', 30") of the membrane (4) and at least partially on the vertical sections (24, 24', 24") of the membrane (4). [8] Sound transducer (1) according to claim 7, wherein the following multilayer stack is adjacent to the support layer (10) of the membrane (4), in particular in this order: i) optionally an insulation layer (16) ii) first piezoelectric layer (12) iii) first conductive layer (18) iv) second piezoelectric layer (14) v) second conductive layer (20) [9] Sound transducer (1) according to at least one of the preceding claims, wherein the diaphragm (4), in particular the vertical sections (24, 24', 24") of the diaphragm (4), are designed and configured to perform symmetrical horizontal vibrations (36) and / or unsupported horizontal vibrations (38) and / or asymmetrical horizontal vibrations (40). [10] Method for producing a membrane (4), in particular a membrane (4) according to at least one of the preceding claims, comprising the steps, in particular in this order: a) Provision of a converter structure (22) with recesses (42, 42', 42") b) Application of a continuous support layer (10) c) Optional application of a continuous insulation layer (16) d) Application of a first piezoelectric layer (12) to the horizontal sections (30, 30', 30") of the membrane (4) and at least section-by-section application to the vertical sections (24, 24', 24") e) Application of a first conductive layer (18) to the horizontal sections (30, 30', 30") of the membrane (4) and at least section-by-section application to the vertical sections (24, 24', 24") f) Removal of the converter structure (22) while maintaining a frame structure (2)

Images