MEMS transducers for interacting with a volume flow of a fluid

Description

[0001] The present invention relates to a MEMS transducer for interacting with a volume flow of a fluid, such as a MEMS speaker, a MEMS microphone or a MEMS pump.

[0002] A key focus of MEMS technology (MEMS = microelectromechanical system), in addition to miniaturization, lies in the potential for cost-effective manufacturing of components in medium and high volumes. Electroacoustic MEMS loudspeakers are currently commercialized to a negligible extent. With few exceptions, MEMS loudspeakers consist of a membrane that is deflected quasi-statically or resonantly by a selected physical principle. The deflection depends linearly or non-linearly on the applied electrical signal (current or voltage). The signal exhibits a temporal variation, which is translated into a temporal variation of the membrane deflection. The back-and-forth movement of the membrane is transmitted in the form of sound into the surrounding fluid, which, for the sake of simplicity but not limitation, can be assumed to be air in the following.

[0003] In some cases, the diaphragm is actuated in only one direction. The restoring force is then provided by the mechanical spring action upon diaphragm deflection. In other cases, the actuation occurs in both directions, so the diaphragm can have very low stiffness.

[0004] The use of electrostatic, piezoelectric, electromagnetic, electrodynamic, and magnetostrictive principles has been described for actuating the membrane. An overview of MEMS sound transducers based on these principles can be found, for example, in [1]. DE 10 2012 223605 A1 discloses a MEMS component with beam-shaped displacement elements, each of which can be moved translationally within a trench structure. These displacement elements can be excited to a rotational "in-plane" movement. The displacement elements are suspended only on one side, allowing them to pivot about their suspension point within the respective trench structure. This can be used, for example, to generate a sequence of ultrasonic pulses whose superposition lies in the audible range.

[0005] Electrostatically operated transducers are based on the force that arises between two flat electrodes with different electrical potentials. In the simplest case, the arrangement corresponds to a plate capacitor, with one of the two plates being movably suspended. In practice, the movable electrode is designed as a membrane to prevent an acoustic short circuit. When a voltage is applied, the membrane warps towards the counter electrode. In a special embodiment, the membrane is operated in what is known as touch mode. In this case, the membrane touches the lower electrode, to which a thin insulating layer is applied to prevent a short circuit, as described, for example, in [2]. The contact surface is determined by the magnitude of the applied electrical voltage and thus varies over time according to the temporal progression of this voltage.The resulting oscillation serves to generate sound. In principle, with a conventional electrostatic setup, the membrane can only be attracted in the direction of the electrode. The restoring force can be determined, at least in part, by the stiffness of the membrane and must be sufficiently high to transmit even the higher frequencies in the audible sound range.

[0006] On the other hand, for a given electrical voltage, the deflection of the diaphragm can decrease with increasing stiffness. To overcome this problem, an approach was developed using a very soft diaphragm, which can then be controlled by an upper and lower electrode and thus deflected in both directions, as described in [3]. This loudspeaker uses a total of two such diaphragms suspended inside a cavity that, like a micropump, has an inlet and an outlet and is otherwise closed.

[0007] Piezoelectrically operated transducers utilize the inverse piezoelectric effect. An applied electrical voltage leads to a mechanical stress in a solid. Materials such as PZT (lead zirconate titanate), AIN (aluminum nitride), or ZnO (zinc oxide) are typically used in MEMS technology. These materials are usually applied as a functional layer to a membrane and structured in such a way that the membrane can be deflected or excited to vibrate depending on the electrical voltage applied to the functional layer. A disadvantage of piezoelectric functional layers is that they cannot be operated without hysteresis.In addition, the integration of the ceramic functional layers is complex and, due to the lack of CMOS compatibility (CMOS = complementary metal oxide semiconductor) with PZT and ZnO, is only possible under strict contamination control or in a separate clean room area.

[0008] Electromagnetically operated transducers are based on the force exerted by a soft magnetic material in a variable magnetic field (gradient). Implementing this principle requires not only the soft magnetic material but also a permanent magnet and a coil, which can be used to temporally control the spatial gradient of the magnetic field via current flow. The soft magnetic material is integrated, for example, into the membrane. All other components are provided in the assembly, as described, for example, in [4]. The structure is bulky, complex, and does not appear to be scalable for large quantities.

[0009] Electrodynamically driven transducers utilize the Lorentz force. This process, widely used in macroscopic loudspeakers, has also been used in some MEMS loudspeakers. The magnetic field is generated by a permanent magnet. A current-carrying coil is placed within the magnetic field. Typically, the coil is integrated into the membrane by depositing and structuring a metal layer, and the permanent magnet is added as an external component during assembly. The complexity and limitations of integrating all components in MEMS technology pose a similarly significant disadvantage as with electromagnetically driven transducers.

[0010] Magnetostrictively operated transducers are based on the contraction or expansion of a functional layer in the presence of a magnetic field. For example, vanadium permadur is positively magnetostrictive, meaning it expands when a magnetic field is applied. This contraction can be used to generate membrane vibration with a suitable design. In [1], vanadium permendur (Fe 49 Co 49 V 2 ) was used as the magnetostrictive functional layer, which was deposited on SiO 2 (silicon dioxide) via a chromium adhesion layer. The external magnetic field is provided by a micro-flat coil, which was realized using electroplated copper. With regard to complexity and limitations in integration, similar disadvantages apply as with the two aforementioned operating principles.

[0011] The classical and most widely used variants described above, which have as a common feature the use of a membrane that can be excited to oscillate, are supplemented below by certain modifications that were investigated due to specific disadvantages of the classical membrane principle.

[0012] Flexible membranes can also exhibit higher modes in the audible sound range and thus lead to parasitic vibrations, which reduces the acoustic quality (total harmonic distortion), see [1]. To avoid or reduce this effect, plates with significantly higher stiffness are used. Such a plate is connected to the chip via a very soft suspension, which also prevents acoustic short circuits, see [5].

[0013] A further modification is a segmented membrane, which was used in the magnetostrictive transducers described above. This represents a special topographical solution to the problem of the functional layer contracting or expanding in two directions. Specifically, the structure consists of several deflectable bending beams. According to [1], the arrangement can be considered acoustically closed for beam spacings of less than or equal to 3 µm. By appropriately dimensioning the individual beams with respect to a resonance frequency and the distances between the beams, a comparatively high acoustic bandwidth can be achieved, and the sound level curve can be adapted or optimized as a function of the vibration frequency.

[0014] In [6], Neumann et al. pursue the approach of using a multitude of small sub-diaphragms instead of a single, large membrane. Each sub-diaphragm has such a high resonance frequency that a quasi-static deflection can occur in the audible sound range. This enables, in particular, digital operation of the loudspeaker.

[0015] In summary, it can be concluded that, with regard to integration, known electrostatically driven diaphragm loudspeakers exhibit relatively small excursions, assuming moderate drive voltages. The electrostatic diaphragm loudspeaker by Kim et al. according to [3] can serve as a reference, for example. Each of the two diaphragms has an area of ​​2 x 2 mm 2< . The upper and lower electrodes are arranged at a distance of 7.5 µm. Depending on the geometry of the diaphragm and the increase in diaphragm stiffness with increasing excursion, the excursion is limited to typically 1 / 3 to 1 / 2 of the electrode spacing due to the so-called pull-in effect. If the higher value of 1 / 2 is assumed, the excursion is 7.5 µm / 2, namely once in one direction and once in the other.The displaced volume can be estimated by assuming that it corresponds to the volume of a deflected rigid plate with a displacement equal to half the maximum displacement of the membrane. For example, this yields: . ΔV ≈ 2 × 2 mm 2 × 50 % ∗ 2 × 7.5 μm / 2 = 15 × 10 − 3 mm 3 or ΔV / aktive Fläche = ΔV / A = ΔV / 4 mm 2 = 3.75 × 10 − 3 mm

[0016] A general problem in the manufacture of miniaturized membrane loudspeakers is to achieve a flat sound pressure curve as a function of frequency. The achievable sound pressure is proportional to the radiation impedance and the speed of the membrane. In the macroscopic sense, the membrane diameter is comparable to the acoustic wavelength. Here, the radiation impedance is proportional to the frequency, cf. [6]. High-quality loudspeakers are often designed so that the resonance f 0 lies below the audible sound range (for multi-way loudspeakers, the respective resonance frequency lies below the lower cut-off frequency of the corresponding electrical filter). For f >> f 0, the speed of the membrane is thus proportional to 1 / f. Overall, the expression p ∝ 1 for the frequency dependence of the sound pressure p results. This (simplified) consideration therefore results in a completely flat sound pressure curve.

[0017] As soon as the diameter of the sound source / membrane is much smaller than the sound wavelength to be generated, a quadratic dependence of the radiation impedance on the frequency can be assumed, as described in [7]. This is the case for MEMS loudspeakers with membranes on the order of millimeters. If f >> f 0 is assumed as above, the sound pressure curve follows the dependence p ∝ f. Low frequencies are reproduced with too low a sound pressure compared to the high frequencies. In quasi-static operation, the membrane speed is proportional to f. The sound pressure curve then follows the dependence p ∝ f 3< , which is even more unfavorable for low frequencies.

[0018] A concept for improved MEMS converters with high efficiency would therefore be desirable.

[0019] The object of the present invention is therefore to provide a MEMS converter and a method for producing the same which can influence a volume flow of a fluid with a high degree of efficiency and / or can be influenced by the volume flow with a high degree of efficiency.

[0020] This problem is solved by the subject matter of the independent patent claim.

[0021] The core idea of ​​the present invention is the recognition that the above object can be achieved by influencing a volume flow of a fluid particularly efficiently using an element that is deformable along a lateral direction of movement, or by allowing the volume flow to deflect such an element particularly efficiently. The lateral direction of movement, possibly perpendicular to a direction of the fluid flow, enables large surfaces of the deformable element that can interact with the volume flow, while simultaneously maintaining small dimensions of a chip surface, thus resulting in an efficient MEMS converter device with a high degree of efficiency.

[0022] The claimed invention relates to a MEMS system comprising: a MEMS transducer for interacting with a volume flow of a fluid, comprising: a substrate having a cavity; wherein the MEMS transducer comprises a plurality of electromechanical transducers, and each of these electromechanical transducers is connected to the substrate in the cavity and has an element deformable along a lateral direction of movement, wherein an in-plane deformation of the deformable element along the lateral direction of movement and the volume flow of the fluid are causally related; wherein each electromechanical transducer is connected to the substrate in a force-locking or form-locking manner or is formed integrally with the substrate; wherein the MEMS system further comprises: a control device configured to control the deformation of the deformable elements or to detect the deformation of the deformable elements;wherein the control device is designed to control the plurality of electromechanical transducers such that a first and an adjacent second electromechanical transducer move at least locally towards one another during a first time interval, and wherein the control device is designed to control the plurality of electromechanical transducers such that the first electromechanical transducer and a third electromechanical transducer arranged adjacent to the first electromechanical transducer, the first electromechanical transducer being arranged between the second and the third electromechanical transducer, move towards one another during a second time interval.

[0023] According to another embodiment, a MEMS loudspeaker comprises such a MEMS transducer and is configured to emit an acoustic sound wave or an ultrasonic wave. According to another embodiment, a MEMS pump comprises a MEMS transducer, so that the fluid can be transported based on the volume flow. According to another embodiment, a MEMS microphone comprises a MEMS transducer with a deformable element that is deformable along the lateral direction of movement. An advantage of these embodiments is that high efficiency can be achieved while utilizing a small chip surface area.

[0024] Advantageous embodiments are the subject of the dependent patent claims.

[0025] They show: Fig. 1 is a schematic perspective view of a MEMS converter according to an embodiment; Fig. 2a is a schematic perspective view of a MEMS converter comprising a plurality of electromechanical converters according to an embodiment; Fig. 2b is a schematic top view of the MEMS converter from Fig. 2a according to an embodiment. Fig. 2c a schematic perspective view of the MEMS converter from Fig. 2a , in which the electromechanical transducers have a deformed state of a deformable element, according to an embodiment; Fig. 3 is a schematic perspective view of a deformable element that is designed as a bimorph, according to an embodiment; Fig. 4a is a schematic perspective view of a deformable element that has three bimorph structures, according to an embodiment; Fig. 4b is a schematic perspective view of the deformable element according to. Fig. 4a in a deflected state according to an embodiment; Fig. 4c a schematic plan view of an arrangement of two deformable elements arranged adjacent to each other according to an embodiment; Fig. 5 a schematic plan view of a MEMS converter in which the electromechanical converters are compared with the MEMS converter from Fig. 2a have a modified configuration, according to one embodiment; Fig. 6a shows a schematic plan view of an electromechanical transducer in which straight spring elements are arranged between plate elements and deformable elements, according to one embodiment; Fig. 6b shows a schematic plan view of an electromechanical transducer in which spring elements of deflectable ends of the deformable elements are arranged at an angle of less than 90°, according to one embodiment; Fig. 6c shows a schematic plan view of an electromechanical transducer in which the spring elements are arranged at an angle of more than 90°, according to one embodiment; Fig. 6d shows a schematic plan view of an electromechanical transducer in which the substrate has a spring element adjacent to a deformable element, according to one embodiment;6e is a schematic plan view of an electromechanical transducer in which plate elements have cutouts, according to an embodiment; Fig. 7a is a schematic plan view of a deformable element connected to the plate element, according to an embodiment; Fig. 7b is a schematic plan view of a configuration in which the deformable element is firmly clamped and formed between the substrate, according to an embodiment; Fig. 7c is a schematic plan view of a configuration of the electromechanical transducer in which the deformable elements have cutouts in a central region, according to an embodiment; Fig. 7d is a schematic plan view of a configuration of the electromechanical transducer in which a first deformable element and a second deformable element are arranged parallel to one another; Fig.8a shows a schematic perspective view of a MEMS converter in which the deformable elements are alternately connected to the substrate and to an anchor element, respectively, according to an embodiment; Fig. 8b shows a schematic top view of the MEMS converter from . Fig. 8a according to an embodiment; Fig. 8c a schematic perspective view of the MEMS converter from Fig. 8a in a deflected state according to an embodiment; Fig. 8 shows a schematic plan view of the MEMS converter from Fig. 8b in the deflected state according to an embodiment; Fig. 9 shows a schematic perspective view of a stack having three MEMS transducers, according to an embodiment; Fig. 10 shows a schematic perspective view of a section of a MEMS transducer in which deformable elements are arranged between sides of the substrate, according to an embodiment; Fig. 11a shows a schematic view of a section of a MEMS transducer in which the electromechanical transducers are arranged obliquely with respect to a lateral direction of the substrate, according to an embodiment; Fig. 11b shows a schematic view of a section of a MEMS transducer that can be used as a pump, according to an embodiment; Fig. 12a shows a schematic view of a section of a MEMS transducer that can be used, for example, as a MEMS pump, in a first state; Fig. 12b shows the MEMS transducer from Fig. 12a in a second state; Fig. 13 a schematic view of two deformable elements that are connected to each other along a lateral extension direction, according to an embodiment; Fig. 14 a schematic view of a stack comprising two MEMS transducers that are connected to each other and have a common layer, according to an embodiment; Fig. 15 a schematic side sectional view of a deformable element that has two layers that are spaced apart and connected to each other via connecting elements, according to an embodiment; Fig. 16 a schematic plan view of a deformable element that is arranged adjacent to an electrode, according to an embodiment; Fig. 17 a schematic block diagram of a MEMS system according to an embodiment; Fig.Fig. 18 is a schematic top view of a MEMS transducer according to an embodiment having a plurality of electromechanical transducers with cantilevered beam elements; and Fig. 19 is a schematic top view of a MEMS transducer according to an embodiment having a plurality of electromechanical transducers with cantilevered beam elements. None of these figures illustrate a MEMS system with all features of the claimed invention. However, the features of the invention are clarified with reference to the figures and the accompanying text.

[0026] Before exemplary embodiments of the present invention are explained in more detail below with reference to the drawings, it is pointed out that identical, functionally equivalent or equivalent elements, objects and / or structures in the different figures are provided with the same reference numerals, so that the description of these elements shown in different exemplary embodiments is interchangeable or can be applied to one another.

[0027] The following refers to MEMS transducers (MEMS = microelectromechanical system). A MEMS transducer can have one or more electroactive components that cause a change in a mechanical component based on an applied electrical quantity (current, voltage, charge, or the like). This change can relate, for example, to a deformation, heating, or strain of the mechanical component. Alternatively or additionally, a mechanical influence on the component, such as a deformation, heating, or strain, can lead to an electrical signal or electrical information (voltage, current, charge, or the like) that can be detected at the component's electrical terminals. Some materials or components exhibit reciprocity, meaning that the effects are mutually interchangeable.For example, piezo materials can exhibit the inverse piezoelectric effect (deformation based on an applied electrical signal) and the piezoelectric effect (providing an electrical charge based on deformation).

[0028] Some of the embodiments described below relate to a deformable element of an electromechanical transducer being configured to interact with a volume flow of a fluid. An interaction can, for example, comprise a deformation of the deformable element, brought about by an electrical control signal, which leads to a movement, displacement, compression, or decompression of the fluid. Alternatively or additionally, the volume flow of the fluid can deform the deformable element, so that, based on the interaction between the volume flow and the deformable element, a presence, a characteristic (pressure, flow velocity, or the like), or other information regarding the fluid (such as a temperature) can be obtained. This means that a deformation of the deformable element along the lateral direction of movement and the volume flow of the fluid are causally related.MEMS can be manufactured using silicon technology, for example. The electromechanical transducer can comprise the deformable element and further elements, such as electrodes and / or electrical connections. The deformable element can be designed to deform (macroscopically) along a lateral direction of movement, i.e., an element or region can be movable along the lateral direction of movement. The element or region can be, for example, a beam end or a central region of a beam structure. Viewed microscopically, a deformation of the deformable element along the lateral direction of movement can result in a deformation of the deformable element perpendicular to the lateral direction of movement. The exemplary embodiments described below refer to the macroscopic view.

[0029] Embodiments may provide miniaturized, silicon-made loudspeakers, microphones and / or pumps that can generate the highest possible sound level, the highest possible sensitivity and / or the highest possible flow rate of the fluid, relative to their respective size.

[0030] Embodiments of the present invention can be used to generate airborne sound, particularly in the audible sound range. Embodiments thus relate to loudspeakers, particularly miniaturized loudspeakers, for example, for hearing aids, headphones, headsets, mobile phones, or the like. The reciprocal causal relationship between the volume flow and the deformation of the deformable element also enables application in loudspeakers. Embodiments thus relate to electroacoustic transducers.

[0031] Fig. 1 shows a schematic perspective view of a MEMS converter 10. The MEMS converter 10 is configured to interact with a volume flow 12 of a fluid. The fluid can be a gas (such as air) and / or a liquid. For example, the liquid can be a medical solution, a medication, a chemical for a technical process, or the like.

[0032] The MEMS converter 10 has a substrate 14. The substrate 14 can comprise any desired material. For example, the substrate 14 can comprise a wood material, a metal material, and / or a semiconductor material, such as a silicon material. The substrate 14 comprises a cavity 16. The cavity 16 can be understood, for example, as a recess or as an at least partially enclosed volume of the substrate 14. The fluid of the volume flow 12 can be arranged at least partially in the cavity 16.

[0033] The MEMS transducer 10 comprises an electromechanical transducer 18. The electromechanical transducer 18 is connected to the substrate 14. The electromechanical transducer 18 comprises a deformable element 22 that is deformable along a lateral movement direction 24. For example, applying an electrical signal to the electromechanical transducer 18 can lead to the deformation of the deformable element 22 along the lateral movement direction 24. Alternatively or additionally, the volume flow 12, when it impacts the deformable element 22, can cause the deformable element 22 to perform the deformation, so that an electrical signal based on the volume flow 12 can be obtained from the electromechanical transducer 18. This means that the deformation of the deformable element 22 and the volume flow 12 are causally related.For example, the electromechanical transducer 18 can comprise or consist of at least one, perhaps two, piezoelectric layers. Both layers can be deformed by electrical voltage. The electromechanical transducer can comprise additional elements, such as electrodes.

[0034] The substrate 14 can comprise one or more openings 26a-d through which the volume flow 12 can pass from an environment of the MEMS converter 10 into the cavity 16 and / or from the cavity 16 into an environment of the MEMS converter 10. A movement that the deformable element 22 performs during deformation can be understood as in-plane with respect to the substrate 14. The volume flow 12 can exit from the cavity 16 or enter it at least partially perpendicular to the direction of movement 24, as shown, for example, for the volume flow 12 through the opening 26c or 26d. Put simply, an in-plane movement of the deformable element 22 can lead to an out-of-plane movement of the volume flow 12 and vice versa. This means that the lateral direction of movement and / or the curvature of the deformable element can occur in-plane with respect to the substrate.

[0035] The openings 26c and 26d are arranged perpendicular to the lateral movement direction 24 in the substrate 14. The deformation of the deformable element 22 along the lateral movement direction 24 can result in a movement of at least a portion of the deformable element 22 toward the opening 26a, so that a partial cavity 28 is reduced in size based on the deformation. The pressure of the fluid located in the partial cavity 28 can be increased based on this. Simply put, the fluid can be compressed. This can enable the fluid to flow out of the partial cavity 28 or the cavity 16. The volume flow 12 perpendicular to the lateral movement direction 24 can be obtained through the openings 26d and 26c.

[0036] A base area of ​​the MEMS transducer 10 can, for example, be arranged in an x / y plane. A large dimension of the MEMS transducer 10 along a z-direction that is arranged perpendicular to the x-direction and / or the y-direction in space, or a large dimension of the deformable element 22 along the z-direction, can lead to an increase in the volume flow 12, while the base area of ​​the MEMS transducer 10 remains unchanged. An enlargement of the partial cavity 28 can lead to a negative pressure of the fluid in the partial cavity 28, so that the volume flow flows into the cavity 28 or 16 based on the deformation of the deformable element 22 perpendicular to the lateral movement direction 24.

[0037] The deformable element can have an axial extension, for example along the y-direction, which has a value in a range of at least 1 µm and at most 100 mm, preferably of at least 100 µm and at most 10 mm, and particularly preferably in a range of at least 500 µm and at most 5 mm. The deformable element 22 can have an extension along the lateral movement direction 24 which has a value in a range of at least 0.1 µm and at most 1000 µm, preferably of at least 1 µm and at most 100 µm, and particularly preferably in a range of at least 5 µm and at most 30 µm.The deformable element may have an extension along a lateral direction arranged perpendicular to the lateral movement direction, for example along the z-direction, which has a value in a range of at least 0.1 µm and at most 1000 µm, preferably of at least 1 µm and at most 300 µm and particularly preferably in a range of at least 10 µm and at most 100 µm.

[0038] Fig. 2a shows a schematic perspective view of a MEMS transducer 20 comprising a plurality of electromechanical transducers 18a-f. The electromechanical transducers 18a-f are connected to the substrate 14 and can each have an element deformable along the lateral movement direction 24, as described in connection with the Fig. 1 described.

[0039] The substrate 14 comprises, for example, a first layer 32a, a first spacer layer 34a, an intermediate layer 36, a second spacer layer 34b, and a second layer 32b, which are arranged one upon another in the stated order. According to further embodiments, one or more further layers can be arranged between two of the successively arranged layers. According to further embodiments, at least one of the layers 32a, 32b, 34a, 34b, and / or 36 has a multilayer structure.

[0040] The electromechanical converters 18a-f are designed and / or controllable so that they move partially towards each other and partially away from each other based on the volume flow 12 and / or based on a control.

[0041] For example, the electromechanical transducers 18a and 18b are configured to move away from each other, while the electromechanical transducers 18b and 18c move toward each other. Partial cavities 38a-c are arranged between the electromechanical transducers 18a and 18b, 18c and 18d, and 18e and 18f, wherein the partial cavities 38a-c can expand based on the deformation of the electromechanical transducers 18a-f. Partial cavities 42a and 42b are arranged between the electromechanical transducers 18b and 18c, and 18d and 18e, respectively, and can simultaneously shrink based on the movement or deformation. In a subsequent time interval, the deformation or movement of the electromechanical transducers or the deformable elements can be reversible, so that the volumes of the partial cavities 38a, 38b and 38c decrease while the volumes of the partial cavities 42a and 42b increase.

[0042] In other words, a structured layer, the spacer layer 34a, can be arranged on the lower cover (first layer 32a), which partially or completely closes off the chip on one side (for example, but without limitation, a bottom side). This structured layer can be used, for example, as a spacer between the lower cover and the intermediate layer 36 arranged on the structured layer 34a. A structured spacer layer 34b can in turn be arranged on the structured layer 36, which in its function as a spacer corresponds entirely or partially to the spacer layer 34a and can have an identical or similar shape. The MEMS converter 20 or its cavity can be partially or completely closed off along the z-direction by the upper cover, the second layer 32b. Fig. 2a 3 shows layer 32b as a partially cutaway view to illustrate elements arranged in the cavity region. Electromechanical transducers 18b and 18c, or 18d and 18e, can be arranged in pairs in an x / y plane of the intermediate layer 36, whereby such an arrangement can be repeated multiple times along a spatial direction, for example, along the x-direction.

[0043] The substrate can have a plurality of openings 26 connected to a plurality of partial cavities 38a-c or 42a-b, wherein, for example, each opening 26 can be connected to a partial cavity 38a-c or 42a-b. A volume of each partial cavity 38a-c or 42a-b can be influenced by a deflection state of at least one element 22 deformable along the lateral movement direction 24. Adjacent partial volumes can be complementarily enlarged or reduced during a first or second time interval. In simple terms, a partial volume of a partial cavity 38a-c or 42a-b can be reduced, while an adjacent partial volume of a partial cavity 42a-b or 38a-c is enlarged.

[0044] Rod structures 44 can be arranged in a region of one or more openings 26. The rod structures 44 can be arranged such that passage of the volume flow 12 in one or two directions is enabled, whereas penetration of particles into or out of the cavity is reduced or prevented. A shape of the layers 32a, 32b, 34a, 34b and / or 36 can be influenced, for example, during a manufacturing process by selectively removing and / or selectively arranging or growing layers. For example, the rod structures 44 can be formed from the layers 34a, 36 and / or 34b based on a selective etching process. Furthermore, a shape of the cavities 38a-c and 42a-b can be influenced during the manufacturing process.For example, walls of one or more layers 32a, 32b, 34a, 34b and / or 36 can be adapted to a movement of the deformable elements of the electromechanical transducers 18a-f, for example to enable an at least approximately constant and / or small distance between the deformable elements and the substrate 14.

[0045] A cover 43 can be arranged adjacent to or on the rod structures or rod elements. The cover 43 can be arranged adjacent to the cavity 16 and / or separated therefrom by the rod elements 44. The cover can comprise, for example, a nonwoven material (mesh material), a foam material, and / or a paper material. The cover can allow particles to penetrate into the cavity 16 or exit the cavity 16 with a diameter smaller than a distance between rod structures. Alternatively, the cover 43 can also be arranged adjacent to or on an opening 26 that does not have the rod elements 44.

[0046] If a free end of the movable elements moves, for example in a curved path and / or a circular path, the substrate 14 can have a parallel or similar shape in a region in which the movable end moves.

[0047] Fig. 2b shows a schematic top view of the MEMS converter 20 from Fig. 2a The electromechanical transducers 18a-f can, for example, be connected to the substrate 14 at elements 46a-c in a force-fitting or form-fitting manner. For example, one or more deformable elements of the electromechanical transducers 18a-f can be formed integrally with the elements 46a-c. The elements 46a-c can be arranged in a plane of the layer 36 or can be parts of the layer 36. An extension of the deformable elements 22 of the electromechanical transducers 18a-f can, for example, be less than or equal to an extension of the layers 34a, 36, and 34b along the z-direction. This means that the deformable elements 22 of the electromechanical transducers 18a-f can be arranged and movable without contact with the layer 32a and / or 32b. Alternatively, at least one deformable element can also be deformed with contact. For example, between the at least one deformable element and an adjacent layer, such as layer 32a and / or 32b, a low-friction, ie, having a low coefficient of friction. The low-friction layer can enable fluidic separation between partial cavities, as described, for example, for the wall structure 49. A coefficient of friction can, for example, be 10%, 20%, or 50% lower than a coefficient of friction of the layer 32a and / or 32b or the layer 34a and / or 34b. A frictional force between the deformable element 22 and adjacent layers can be lower than a force required to deform the deformable element 22. Based on a reduced frictional force, for example, a force to be provided by an actuator can be lower, so that the actuator can be designed with less power. Alternatively or additionally, a sensitivity of the deformable element 22 to the volume flow 12 can be increased.

[0048] The electromechanical transducers 18b and 18c form, for example, side walls of the partial cavity 42a (chamber). The movable elements 22 of the electromechanical transducers 18a-f can be positively attached to the elements 46a-c. A distance from the substrate 14 or from elements 48a-d of the substrate 14 can be arranged between a deflectable or movable end 52 of the deformable elements 22. The end 52 of the deformable element 52 can thus be arranged to be freely movable. One or more deformable elements 22 can be deflectable particularly far along the lateral direction 24 due to dimensional relationships, such as an extension along the x-direction in a ratio to an extension along the y-direction, simplified as a ratio of beam width to beam height. If the electromechanical transducers 18a-f are designed as actuators, for example, these actuators can be deflected when a corresponding signal is applied, iebe bent so that, for example, the end 52 of the deformable element 22 moves along a curved path. Depending on the course of this path, at least one of the elements 48a-d can be designed such that a distance between it and the end 52 remains approximately constant and / or small even when the deformable element 22 is deflected.

[0049] The MEMS converter 20 can have at least one wall structure 49. A movement of the actuators, electromechanical converters 18a-e, or deformable elements can, for example, result in a fluid-mechanical coupling to the adjacent chambers for a chamber 42a-b due to fluid flows triggered by the movement to fill the chamber 38a-c. Based on the fluid-mechanical coupling, a fluid flow 57 can occur between the partial cavities 42a and 38b. To reduce or avoid this direct coupling or the fluid flow 57, one or more partition walls (wall structures 49), which can optionally be designed to be immovable, can be arranged to separate adjacent chamber pairs 38 and 42. The wall structures can be implemented simply, for example, at the corresponding locations as an element formed continuously from the layers 34a, 36, and 34b.For example, such structures can remain in place during a selective etching process. The wall structure 49 can also increase the mechanical stability of the MEMS converter 20 and can facilitate a bonding process between the individual layers. The at least one wall structure 49 can have openings or be designed to be completely continuous, which makes it possible to specifically modify the damping resulting from the inflow / outflow of the fluid from the chambers 38a-c and 42a-b, in particular for adjusting the width of the resonance curve or generally for adjusting the dynamic properties of the actuator chamber systems.

[0050] Will the Fig. 2b together with the Fig. 1 From this perspective, a volume of the cavity 16 and / or the plurality of sub-cavities 38a-c and 42a-b can be influenced or determined by the layers 32a and 32b and side regions 53a and 53b of the substrate 14. The side regions 53a and 53b can be arranged between the layers 32a and 32b. The deformable elements of the electromechanical transducers 18a-c can be configured to perform a movement parallel to the first layer 32a and / or 32b at least in a section 55 of the lateral movement direction 24. This means that the deformable element can deform or move between the layers 32a and 32b.

[0051] A resonance frequency of a cavity or partial cavity can be influenced by the geometry of the volume, by a control frequency of the electromechanical transducers, and / or by a mechanical resonance frequency of the deformable element(s). (Partial) cavities that are at least partially fluidically separated from one another, for example by means of a wall structure 49, an arrangement of a low-friction layer, or based on an arrangement in different MEMS transducers, can have different resonance frequencies and / or be controlled at different frequencies, for example by means of a control device. A multi-way loudspeaker can be obtained based on different control frequencies and / or different resonance frequencies. Resonance frequencies of cavities are used, for example, in the field of cavity resonators or Helmholtz resonators.

[0052] Fig. 2c shows a schematic perspective view of the MEMS converter 20, in which the electromechanical converters 18a-f have a deformed state of the deformable element. For example, the deformable elements are deflected to a maximum deflection. Compared to the representation of the Fig. 2a a volume of the partial cavity 42a is reduced based on the deformation (bending) of the deformable elements (beams). For example, if a thickness (dimension along the z-direction or thickness direction) of the layers 34a and 34b (spacers) is small, a flow around the electromechanical transducers 18a-f or the deformable elements can be negligible during a movement of the electromechanical transducers 18a-f. This can also apply to a distance between the electromechanical transducer 18a-f and the substrate, for example the element 48. Based on the deformation of the deformable element, a volume of the fluid, for example an air volume, which corresponds to the volume difference of the partial cavities 42a in the Fig. 2a and 2c can correspond to, be delivered to an environment of the MEMS converter 20, for example in the form of the fluid flow (volume flow) 12.

[0053] A dimension of the spacer layer 34a or 34b along the z-direction, along which the first and second spacer layers 34a and 34b are arranged on the intermediate layer 36, may have a value in a range of at least 1 nm and at most 1 mm, preferably in a range of at least 20 nm and at most 100 µm, or particularly preferably in a range of at least 50 nm and at most 1 µm. For example, if the dimension of the spacer layers 34a and 34b is small compared to a dimension of the electromechanical transducers 18a-f along the z-direction, an extent of the fluid flow 57 flowing around the electromechanical transducer 18a-f from a first side to a second side (for example, from a positive x-direction to a negative x-direction or vice versa) while the deformable element deforms may be smaller than an extent of the volume flow 12 in the cavity.

[0054] The flow around or fluid flow 57 can, for example, be based on at least partial removal of the spacer layers 34a and / or 34b in an area in which the electromechanical transducer 18a-f moves. Simply put, based on the distance between the electromechanical transducer and adjacent layers, a fluid flow around moving elements (fluid losses) can result. These can be small compared to the fluid flow 12. For example, they can be smaller than the volume flow divided by the value 10, divided by the value 15, or divided by the value 20.

[0055] The electromechanical transducers can move towards and away from each other in pairs. For example, the electromechanical transducers 18a and 18b can move in relation to the state in Fig. 2b move away from each other in pairs and, in a subsequent time interval, move toward each other in pairs. At the same time, for example, the electromechanical transducers 18b and 18c can move toward or away from each other in pairs. Such a complementary movement of electromechanical transducers in pairs, which is also possible when the transducers are not arranged adjacent to each other, can result in at least partial but also complete compensation of inertial forces, so that a low degree of vibration or no vibration is retained in the MEMS transducer or is transmitted from the MEMS transducer to the environment.

[0056] In other words, a special feature of the chamber approach described so far is that the actuators always move toward or away from each other in pairs in opposite directions. Thus, (with appropriately careful design of the two active bending actuators bordering each chamber wall) no vibrations arise that would be disruptive, for example, when used as a hearing aid or in-ear headphones.

[0057] The fluid flow 12 can, for example, pass through the opening 26a and / or 26b. The openings 26a and 26b can be of identical design or can be adapted to a geometry of the adjacent partial cavity 38a or 42a. The opening 26a can, for example, have a variable cross-section along an axial direction (such as the y-direction), for example a dimension along the x-direction. The dimension of the opening 26b along the x-direction can decrease in a direction toward an interior of the MEMS transducer 20, i.e., toward the cavity or partial cavity 42a. Alternatively or additionally, the opening 26 can have a variable dimension or a variable cross-section along a further direction, such as a z-direction (thickness direction) perpendicular to the axial direction y. The variable cross-section can decrease from an outside of the MEMS transducer 20 in a direction toward the cavity 16.A tapered cross-section or a decreasing dimension of the opening 26 from the outside of the MEMS transducer 20 towards the cavity 16 along one or more directions x and / or z may be referred to as a funnel-shaped opening.

[0058] The possibly funnel-shaped opening 26b can be used as a device for impedance matching. Impedance matching can be advantageous, for example, when using the MEMS converter 20 as a loudspeaker. The design or geometry of the opening 26b can be analogous to macroscopic loudspeakers with dimensions of several centimeters. A shape of the opening 26b can enable the actual sound radiation to be defined by the outer surface of the funnel. The opening 26b can, for example, be formed continuously in the structured layers 34a, 36, and 34b. A rod grid 54, which comprises at least one rod element 44, can have openings or gaps between rod elements 44 and / or between rod elements 44 and the adjacent substrate. The gaps can be formed such that the fluid can flow through them.

[0059] The rod grid 54 can provide protection against particles penetrating the cavity of the MEMS converter 20. The width of the openings of the rod grid 54, i.e., the spacing between rod elements 44, can be configured such that the fluid flow 12 is fluidically influenced or uninfluenced to a desired degree. For example, or ideally, the spacing between the rod elements 44 can be smaller than the smallest gap spacing in the MEMS converter 20, so that the rod grid can filter a large number of, or even all, relevant particles. A gap spacing can, for example, describe the distance between a deformable element 18a-c and a layer 32a or 32b. The spacing between the rod elements 44 can, for example, be less than 5 µm, 1 µm, 0.1 µm, or 0.05 µm.

[0060] Dimensions of the rod elements 44 along the spatial directions can be implemented such that the rod elements 44 do not exhibit any resonances in the audible sound range, i.e., in a frequency range of at least 16 Hz and at most 22 kHz. Although the rod elements 44 are illustrated as being arranged on an outer side of the MEMS transducer 20, for example, in a region where the opening 26a or 26b has a maximum dimension along the x-direction, one or more rod elements can also be arranged at a different location of the opening 26a or 26b, for example, in a tapered region of the opening 26a or 26b.

[0061] By deforming the deformable elements, the volume of a partial cavity 42a can be reduced. During the same time interval, a volume of the chamber (partial cavity) 38a can increase. The partial cavity 38a can be connected to the environment of the MEMS transducer 20 in the same or similar manner as the partial cavity 42a via a funnel-shaped opening 26b and / or a rod grid 54 comprising one or more rod elements 44. Electromechanical transducers 18a-f can be configured to be driven at a different frequency or have a different resonant frequency. A volume of each partial cavity can change at a different frequency or at at least partially identical frequencies.

[0062] The opening 26a and the opening 26b can be arranged on or in sides of the MEMS converter 20 that are arranged opposite one another in space. For example, on a respective side having the opening 26a or 26b, the fluid flow can be expelled 12 or sucked in by means of the partial cavities 42a or 38a or a plurality of such partial cavities. This means that the fluid flow 12 can be generated in opposite directions. For example, in a first time interval, the volume flow 12 can be expelled in a negative y-direction from the opening 26a and sucked into the partial cavity 38a. In a second time interval, these directions can be reversed. A flow short circuit along the MEMS converter 20 can thus be prevented or excluded.

[0063] The deformable elements (beams) of the electromechanical transducers 18a-f may be configured to bend in accordance with an externally supplied signal.

[0064] The frequency at which the curvature occurs can influence or determine the frequency at which the volume flow 12 is generated and / or oscillates, thus influencing or determining a sound frequency. An amplitude of the oscillation determined via the supplied signal can influence or determine the amplitude of the volume flow 12 at one or more (resonance frequencies) and thus affect the sound level.

[0065] Likewise, at least one chamber (cavity or partial cavity) can function as a sensor element, and another chamber as an actuator element. This means that the MEMS transducer can comprise at least one sensor and at least one actuator-type deformable element. The movement of the beams is detected and evaluated. For example, the electromechanical transducers 18a and 18b can be controlled as actuators, while the electromechanical transducers 18c and / or 18d can be used as sensors for detection within the fluid. Electrostatic (capacitive), piezoelectric, or piezoresistive sensor elements can be integrated for detection. Such an element can be used as a microphone or pressure sensor. Such an integrated microphone or pressure sensor can also be used to monitor and control the properties of the loudspeaker chambers (actuators), the ultrasonic transmitter chamber, or the pump chamber.For this purpose, appropriate electronics must be used as a control circuit.

[0066] Further embodiments of the electromechanical transducers or actuators are explained below. Although the MEMS transducer 20 has been described as having an undeflected or non-actuated state with undeflected deformable elements, the states can also be interchangeable. This means that in a first, non-actuated state, the deformable elements can be deformed or curved and can deform into a less curved, more curved, or straight state based on a control signal.

[0067] Although the above explanations explain that an electrical signal is supplied to the MEMS converter 20, for example from a control device, the volume flow 12 can also lead to a deformation of the deformable elements, wherein the deformation can be obtained by means of an electrical signal at the MEMS converter 20, ie the MEMS converter 20 can also be configured as a sensor.

[0068] Reference is made below to advantageous developments of the deformable element. One or more electromechanical transducers can have deformable elements according to the developments described below.

[0069] Fig. 3 shows a schematic perspective view of a deformable element 30 designed as a bimorph. The deformable element 30 has a first layer 56 and a second layer 58, which are firmly connected to one another at least in places, advantageously over the entire surface. The first layer 56 and the second layer 58 are designed to deform, for example expand or contract, to different extents based on a mechanical, physical, or chemical influence. For example, the layers 56 and 58 can have different thermal expansion coefficients. Alternatively or additionally, the layer 56 or the layer 58 can be designed to expand or contract based on an electrical signal that is fed to the corresponding layer. For example, this layer can comprise piezo materials.

[0070] Different contractions or expansions of the layers 56 and 58 can lead to a deformation of the deformable element 30 along an actuation direction 59 or 59'. The actuation direction can be arranged parallel to the lateral movement direction 24. The actuation direction can be a direction along which the deformable element 30 can be deflected by applying a positive electrical voltage.

[0071] Alternatively or additionally, deformation along a further lateral movement direction 24' can also be used, which is based, for example, on a transverse contraction or transverse expansion of the deformable element 30 or the contraction or expansion of one of the layers. This means that the deformable element 30 can be configured to bend with its beam structure along an axial direction (such as the y-direction or in-plane) of the beam structure. This can be achieved based on a back-and-forth movement, i.e., along the lateral movement direction 24 and along an opposite direction.

[0072] In other words, the bimorph can correspond to a bar consisting of two layers. The layers are arranged, for example, in one direction (e.g., vertically) relative to one another. A passive layer (e.g., layer 56) can be firmly connected to an active layer (e.g., layer 58). By applying a suitable signal, a mechanical stress can be generated in the active layer 58, which leads to the contraction or expansion of the layer 58. The direction of the length change of the layer 58 can be selected such that the bimorph bends laterally in one direction (contraction) or the other (expansion).

[0073] Fig. 4a shows a schematic perspective view of a deformable element 40 having three bimorph structures 30a-c, as described in connection with Fig. 3 A schematic arrangement of the deformable element 40 in space along the x, y, and z directions is shown by way of example (but not limitation) in such a way that the deformable element 40 can be arranged, for example, in the MEMS converter 10 or 20. The deformable (sub-)elements 30a-c can have different dimensions from one another, for example along the x, y, or z direction. For example, the deformable elements 30a and 30c can have the same extent along the y direction. The actuation directions 59a-c of the deformable elements 30a-c can, for example, be arranged alternatingly or with a mutual orientation, for example in the positive / negative / positive x direction. In simplified terms, this can be understood to mean that the deformable elements 30a and 30c have the same length. The deformable element 30b may have a different length.For example, a length of the deformable element 30b can be twice as long as the comparable length of the element 30a or 30c. According to further embodiments, additional elements, such as spring elements, can also be arranged between the deformable elements 30a-c.

[0074] A direction along which the deformable elements 30a-c deflect upon application of an identical or comparable magnitude (such as a sign of an electrical voltage) can alternate along the length of the deformable element 40. This allows for an alternating curvature. Although the deformable element 40 is illustrated as comprising three deformable elements 30a-c, two deformable elements or more than three deformable elements 30 can be arranged.

[0075] Fig. 4b shows a schematic perspective view of the deformable element 40 in a deflected state. For example, the layers 58a-c are contracted, resulting in multiple curvatures along an axial path (y-direction).

[0076] In other words, three in Fig. 3 The bars shown in the drawing are arranged so that they are placed next to one another in the direction of their extension. This can be done in such a way that a first bar and a third bar (30a and 30c) have a curvature in a first direction and the second bar (30b) has a curvature in the other direction when a corresponding signal is received. In this way, an actuator can be obtained which, starting from its extended form, as shown in Fig. 4a is shown, without an applied signal deforms with a corresponding signal S-like, as shown in Fig. 4b is shown. The configuration with and without a signal is interchangeable. For example, the deformable elements 30 can have a pre-deflection or prestress that, based on the applied signal, leads to a reduced curvature or straight extension of the deformable element 30 and / or 40. For example, it can be assumed that the curvatures of the individual beams 30a-c are identical except for the sign, and that a length of the first and third beams 30a and 30c each corresponds approximately to a quarter of a total length of the deformable element, and that a length of the middle beam 30b corresponds approximately to half the length of the deformable element 40.

[0077] Fig. 4c shows a schematic plan view of an arrangement of two deformable elements 40a and 40b clamped on both sides, which are arranged adjacent to each other so that the partial cavity 38 is arranged between the deformable elements. The solid lines show, for example, an actuated state of the deformable elements 40a and 40b, while the dashed lines show an unactuated state. This description of the deformable elements is interchangeable, since the unactuated state can take on any desired shape during production.

[0078] The deformable elements 40a and 40b can be formed such that they have a curvature in the unactuated state. Furthermore, the deformable elements 40a and 40b can be formed from three segments 30a-1 to 30c-1 and 30a-2 to 30c-2, respectively, which execute a mutual curvature during actuation. Each segment, such as the middle segment 30b-a or 30b-2, can also be formed from two or more segments. Compared to the representations of the Fig. 4a und 4b The segments 30a-1, 30b-1, and 30c-1 can have different lengths relative to one another and to every other segment. The length can be adapted to the desired shape to be achieved during actuation. The S-shaped actuators have the very great advantage that they not only allow a large planar fill factor to be achieved, but they can also be clamped on both sides. The two-sided clamping significantly reduces the pre-deflection of the beams due to layer stress gradients, a technologically unavoidable occurrence. This allows the distances to the lower and upper covers of the substrate to be kept very small, which disproportionately reduces flow / pressure losses and thus not only significantly increases the efficiency of loudspeakers, ultrasonic transducers, microphones, and pumps, but may also enable their correct functioning in the first place.According to further embodiments, only one of the deformable elements 40 can be arranged, for example in the MEMS converter 10.

[0079] Fig. 5 shows a schematic top view of a MEMS transducer 50 in which the electromechanical transducers 18a-c have a modified configuration compared to the MEMS transducer 20. The electromechanical transducers 18a-c each comprise a first and a second deformable element 22a and 22b, 22c and 22d, and 22e and 22f, respectively. The deformable elements are arranged opposite one another. Deflectable ends of the beam elements are arranged facing one another. Regions where the deformable elements 22a-f are connected to the substrate are arranged facing away from one another.

[0080] The electromechanical transducers 18a-c each include a plate element 62a-c connected to the respective deformable elements 22a and 22b, 22c and 22d, and 22e and 22f, respectively. The respective plate element 62a-c may be connected to the deflectable ends of the respective deformable elements 22a-f.

[0081] The deformable elements 22a-f can be designed entirely or partially as deformable elements 30 or 40, or can have a different configuration. Different hatching of the deformable elements 22a and 22b, 22c and 22d, or 22e and 22f indicates that the deformation of the respective deformable element differs from one another. The deformable elements of an electromechanical transducer 18a-c can be arranged such that they perform a deflection of the deflectable ends along the same spatial direction, regardless of the respective design of the deformable element 22a-f.

[0082] For example, the Fig. 5 In the undeflected state shown, an actuation can result in the deflectable ends of the deformable elements 22a and 22b being moved along a positive x-direction. Furthermore, an actuation of the deformable elements 22c and 22d can result in a deflection of the respective deflectable ends being carried out along a negative x-direction. This enables the plate elements 62a and 62b to move towards one another during this actuation, so that the partial cavity 42a is reduced in size based on the movement of the plate elements. Alternatively or additionally, a negative pressure in the cavity 42a can result in the plate elements 62a and 62b moving towards one another, so that a deformation of the deformable elements 22a-d is obtained. Alternatively or additionally, it is also conceivable for one or more deformable elements 22a-d to be designed to be electrically passive.For example, an electrical potential can be applied to one or more plate elements 62a-c, so that, based on an electrical potential of the plate elements 62a and 62b, an attractive or repulsive force can be obtained between the plate elements 62a and 62b, which causes a movement of the plate elements 62a and 62b and thus a deformation of the deformable elements 22a-d. Alternatively or additionally, the deformable elements 22c-f and / or the plate elements 62b and 62c can be controlled simultaneously or with a time delay to obtain a deformation of the deformable elements 22c-f and a change in the volume of the partial cavity 38a.

[0083] In other words, Fig. 5 a variant of the Fig. 2a-c shown configuration, in which four bending beams 22a-d and 22c-f are used to narrow or widen each chamber (cavities 42a and 38a). In connection with the Fig. 2a-c This is described based on two bending beams (deformable elements). Fig. 5 shows a non-actuated state. The actuated and non-actuated states are interchangeable. Thus, in general, any controllable deformable element can be deformed when no signal is applied and can change its deformation depending on the signal, which also includes reaching a stretched (undeflected) state as a special case.

[0084] Vertically (e.g., along the y-direction) opposing bending beams, such as the deformable elements 22a and 22b or 22c and 22d, can each be connected to one another via a flexible web comprising the elements 64a and 64b. In a central region of the web thus obtained, a relatively rigid extension, the element 66, can be arranged. The plate element 62b, which is designed to be rigid or as rigid as possible, can in turn be arranged on this. Upon application of a corresponding signal, the plate elements 62a-c can move parallel towards or away from one another in order to reduce or increase the volumes of partial cavities. The parallel movement of the plate elements can enable the volume of the partial cavity 42a to be zero in the limiting case, which means that the plate elements 62a and 62b touch one another. Compared to a configuration as described in connection with the Figuren 2a-c As described in connection with the MEMS converter 20, such an arrangement can provide a fluid volume flow that is significantly higher than the volume flow of the MEMS converter 20. If the volume of the partial cavity 42a is reduced, the volume of the partial cavity 38b can be increased accordingly or at least based thereon. The fluid can be supplied, as described in connection with the MEMS converter 20, through an opening 26a, 26b, or 26c, respectively. The elements 64a and 64b can also be referred to as spring elements.

[0085] The deformable elements (bending beams) 22a and 22b can be designed to bend to the right (positive x-direction) when a signal is applied. The deformable elements 22c and 22d can be designed to bend to the left (negative x-direction) when a signal is applied. Both beam types (hatching of the deformable elements) can be designed to bend, for example, when a first signal is applied, as in connection with the Fig. 3 or 4 and in the opposite direction when a second signal is received. In this case, both the narrowing and the expansion of the chamber (partial cavity) to the original size can be achieved independently of the mechanical restoring force due to the bending of the beams. The first and second signals can, for example, be a positive and a negative electrical voltage. For example, if the Fig. 3 considered, the layers 56 and 58 can each be active layers or a further active layer can be arranged on the layer 56 on a side facing away from the layer 58, wherein the two active layers can be addressed separately from one another in order to obtain a deflection in one or the other direction.

[0086] A volume between two opposing deformable elements, such as the deformable elements 22c and 22d and the plate element 62b connected to them, can also change upon movement or deformation of the bending beams. The plate element 62 can, for example, be rigid. To enable improved pressure equalization, the deformable elements 22c and / or 22d and / or connecting elements 64 and 66, which connect the plate element 62b to the deformable elements 22c and 22d, can be locally thinned or thinned to provide a local flow channel. This can be achieved, for example, by additional structuring or etching. The connecting elements 64a, 64b, and 66 can be arranged in a T-arrangement. The connecting element 66 can have a high rigidity compared to the elements 64a and 64b.During deformation of the deformable elements 22c and 22d, the elements 64a and 64b can thus preferably deform in order to enable a rectilinear movement of the respective plate element.

[0087] The following are based on the Fig. 6a-e advantageous embodiments are described in which the plate element 62a or 62b is connected to opposite deformable elements 22a and 22b or 22c and 22d.

[0088] Although the following statements refer to a connection of the plate elements to the deformable elements, which are each configured identically, different electromechanical converters and / or connections of individual deformable elements to a plate element can be designed differently. The details described below do not describe exhaustive advantageous developments and can be implemented alone or in combination with one another or in other advantageous embodiments.

[0089] Fig. 6a shows a schematic top view of a configuration in which straight spring elements 68 are arranged between the plate element 62a or 62b and the deformable elements 22a and 22b, or 22c and 22d. The spring elements 68 can be formed from a material of the deformable elements 22a-d or a material of the plate elements 62a or 62b and / or can be formed integrally with one or more of these elements. For example, the spring elements 68 can be at a right angle to the plate elements 62a or 62b.

[0090] Fig. 6b shows an alternative configuration in which spring elements 68' of deflectable ends of the deformable elements are arranged at an angle α of less than 90°, for example 30 or 40°. This allows a Fig. 6a increased distance between the contact points on the plate element 62a, which can lead to reduced deflection of the plate element 62a during movement.

[0091] Fig. 6c shows a configuration in which the spring elements 62a are arranged at an angle a of more than 90°. This can, for example, lead to reduced restoring forces of the spring elements 68 if the configuration as shown in Fig. 6a is shown, is used for comparison.

[0092] Fig. 6d shows a configuration where the configuration consists of Fig. 6a is modified in such a way that a spring element 72a or 72b is arranged in regions of the substrate 14 adjacent to which the electromechanical transducer 18a is arranged or the respective deformable element is connected to the substrate 14.

[0093] The spring element 72a and / or 72b can, for example, be at least partially defined by a recess (cavity) 74a or 74b in the substrate 14. This means that, for example, the stiffness of the substrate 14 can be locally reduced by the recesses 74a or 74b, so that the spring elements 72a and 72b are formed. Although the recesses 74a and 74b are illustrated as extending beyond adjacent deformable elements 22a and 22c, or 22b and 22d, in the substrate 14, the recess 74a or 74b can also be arranged adjacent to only one deformable element or adjacent to multiple deformable elements. Alternatively, the substrate 14 can also have multiple recesses or spring elements.

[0094] In other words, Fig. 6d A configuration in which a further structure in the form of a bending spring (spring elements 72a and 72b) to which the deformable elements (beams) are attached can lead to a further reduction of the tensile stress. Such bending spring elements can, for example, also be integrated into the rigid plate, as in the configuration of the Fig. 6e shown and described in connection with the recesses 76a-d. These elements can deform into an S-shape in the event of beam deflection and reduce the tensile load on the rigid plate.

[0095] Fig. 6e shows a configuration of electromechanical transducers 18a and 18b in which the plate elements 62a and 62b are compared with the configuration as described in connection with the Fig. 6d described, have recesses 76a-d adjacent to a region where the plate elements 62a and 62b, respectively, are connected to the deformable elements via the spring elements 68. A distance between the recesses 76a-d and a side of the plate elements 62a and 62b facing the deformable elements can influence a stiffness of the plate element 62a and 62b, respectively, in this region. The recesses 76a-d enable reduced restoring forces acting on the deformable elements 22a-d.

[0096] In other words, the Fig. 6a-e Variants for the design of the moving elements or the electromechanical converters. These differ from a design as used in connection with the Fig. 5 described, for example or in particular by the fact that the Fig. 5 shown elements 64a or 64b were fused with the stiffener 66 to form the spring elements 68. The configuration according to Fig. 6a can have a higher stiffness against parasitic tilting of the plate elements 62a and 62b about an axis perpendicular to the plane of the drawing (x / y plane). Similar can be achieved for the configurations according to the Fig. 6b und 6c All three configurations also allow for larger deflections of the bending beams compared to the configuration from Fig. 5 . There, the element 64a or 64b (flexible web) can be subjected to tensile stress during a deflection of the beams, which can result in an increasing mechanical resistance to the beam deflection of the deformable elements with increasing deflection. In the variants according to the Fig. 6a-c the mechanical connection of the two deformable elements can be made significantly softer (less stiff), since the connecting spring elements 68 can react with a bending, which can represent a significantly lower mechanical resistance if these elements are designed accordingly.

[0097] The connecting elements / springs 68 and / or those associated with the Fig. 5 The elements / springs 64a-b described may also have a curved or meandering shape. This allows for increased flexibility in a preferred direction. The configurations described in connection with the Fig. 6d und 6e allow a reduction of the tensile load, which would lead to an effective stiffening of the deformable element. The Fig. 6a-e The configurations described neglect inlet and outlet openings 26. If these openings are arranged, recesses or spring elements in the substrate in regions where the opening is arranged can be omitted. Alternatively or additionally, one, several, or each of the spring elements 72a, 72b obtained by at least one recess and / or in the plate elements 62a or 62b can be realized based on two or more separate and independent spring elements.

[0098] The following Fig. 7a-c describe possible arrangements of deformable elements and plate elements as examples.

[0099] Fig. 7a shows the deformable element 40 connected to the plate element 62. The plate element 62 can, for example, be arranged directly on the deformable element 40.

[0100] Fig. 7b shows a configuration in which the deformable element 40a is firmly clamped between the substrate 14 and is designed to deform along the lateral direction 24. Two further deformable elements 40b and 40c are arranged between the deformable element 40 and the plate element 62, the ends of which can be connected to one another. Starting from the connections, the deformable elements 40b and 40c can be aligned with one another such that a curvature of the respective deformable element 40b or 40c points away from the other deformable element. The deformable elements 40a-c can, for example, be controlled jointly or react jointly to the volume flow of the fluid, wherein, for example, a joint control of the deformable elements 40a-c leads to an increase in the travel, i.e., to an increase in the distance by which the plate element 62 is deflected.This means that at least one further deformable element can be arranged between the deformable element and the plate element, which is designed to increase a travel range of the deformable element when actuated jointly with the deformable element.

[0101] Fig. 7c shows a configuration of the electromechanical transducer 18, in which the deformable elements 40a-c have recesses 70a or 70b in a central region, which enable a fluidic coupling of a volume 82 between the deformable elements 40b and 40c with another partial cavity, for example, the partial cavity 38a. The deformable elements 40a, 40b, and / or 40c can each be designed in two parts to provide the recesses 78a and 78b. Alternatively or additionally, the recesses 78a and 78b can be designed as recesses that are enclosed along a thickness direction (z-direction) by additional material of the deformable elements 40a, 40b, and 40c, respectively.

[0102] In other words, Fig. 7a a configuration with actuated S-shaped bending beams according to Fig. 4 , in which a connection to the bending beam is arranged in the middle of the rigid plate. To increase the deflection, the bending actuators can be arranged several times in series. Fig. 7b und 7c schematically show an arrangement of three S-actuators connected in series. According to further embodiments, two S-actuators (deformable elements 40) or more than three actuators can be connected in series. The hatching of the deformable elements in the Fig. 7a-c are, for example, in accordance with the hatchings as shown in Fig. 4 was selected. Different hatchings may indicate different curvature directions of the respective sections. Fig. 7c shows a configuration that has an opening in the middle of the S-shaped actuators (recesses 78a and 78b), which allows improved ventilation of the space (cavity 82).

[0103] Fig. 7d shows a configuration of the electromechanical transducer in which a first deformable element 40a and a second deformable element 40b are arranged parallel to each other along the y-direction. This allows for an increase in the force with which the plate element 62 is deflected. Ends of the deformable elements can be connected to each other or arranged together on the substrate. Alternatively, two or more deformable elements 40a and 40b can be arranged parallel along another direction, for example along the z-direction (thickness direction). Alternatively or additionally, a series connection and a parallel connection of deformable elements can also be combined.

[0104] Moving elements can collide with another moving element or a fixed element if they are subjected to a high or excessive deflection. This can lead to sticking. The moving elements or the fixed elements can preferably be equipped with spacers (bollards), which significantly reduce the contact area and thus reduce or prevent sticking. Instead of so-called bollards, small structures designed as spring elements can also be arranged. In addition to preventing sticking, this can reverse the momentum when two elements collide, reducing or preventing energy losses and improving the dynamic behavior of the actuators.

[0105] Fig. 8a shows a schematic perspective view of a MEMS converter 80, in which the deformable elements are alternately connected to the substrate or the intermediate layer 36 or to an anchor element 84 that is connected to the substrate. For example, the deformable element 22a is firmly connected to the substrate at the ends in the regions 46 and 48 of the intermediate layer 36 and is designed to execute an S-shaped movement, as explained by way of example in connection with the deformable element 40. The adjacently arranged deformable element 22b is connected to the anchor element 84. The anchor element 84 is arranged in a central region of the deformable element 22b and can be connected to the spacer layer 34a or the layer 32a. This means that the substrate can have an anchor element.

[0106] Side walls of the intermediate layer 36, which are arranged adjacent to movable ends of the deformable elements 22a or 22b, can be shaped based on a movement shape of the deformable elements 22a or 22b, respectively.

[0107] Fig. 8b shows a schematic top view of the MEMS converter 80, wherein the spacer layer 34b and the layer 32b are not shown for example. The MEMS 80 includes the rod elements 44 in regions of the openings 26. The regions 48 can include the spring elements 72a-c. The regions 48 are shown as a top view of the intermediate layer 36 for example.

[0108] The anchor element 84 may be integrally formed with the deformable element 22b and / or a layer of the substrate. As shown in Fig. 8 However, as shown, the anchor element 84 can extend beyond the deformable element 22b along the z-direction to connect the layers 32a and 32b to one another. This enables a reduced susceptibility of the layers 32a and 32b to vibration. Alternatively, the anchor element 84 can also be formed from a different piece and / or from a different material than the mechanically deformable element 22b. The deformable element 22a arranged adjacent thereto is, for example, firmly connected on both sides to the substrate in the regions 48 or 46, for example in a form-fitting or force-fitting manner.

[0109] A distance 85 between rod elements 44 can, for example, be less than 1 µm, than 0.1 µm or 0.05 µm.

[0110] The anchor element 84 can be arranged in a central region of the deformable element 22b. The central region can, for example, comprise a geometric center of gravity of the deformable element. The central region can, for example, be the beam segment 30b of the deformable element 40.

[0111] Fig. 8c shows a schematic perspective view of the MEMS transducer 80 in a deflected state. Outer regions of the deformable element 22b may have moved in a direction toward the deformable element 22a, with locations of the outer ends of the deformable element 22a remaining substantially unchanged. A central region of the deformable element 22a may have moved in a direction toward the deformable element 22b, with a location of the central region of the deformable element 22b based on the anchor element 84 remaining substantially unchanged.

[0112] Fig. 8d shows a schematic plan view of the MEMS converter 80 in the deflected state as shown in Fig. 8c The volume of the cavity 42 is compared with the view of the Fig. 8b reduced, whereas a volume of the partial cavity 38 is increased. The spring element 72a can lead to a reduced force introduction into the deformable element 22a, but can also be omitted. A first partial cavity 42 can be arranged between the beam structures of the first electromechanical transducer and the second electromechanical transducer, or between the actuators 22a and 22b, which adjoins an opening 26 of the substrate.

[0113] In other words, show Fig. 8a and 8b A schematic 3D representation or a top view of a variant in which a chip area of ​​the MEMS converter can be used very efficiently. As in the basic configuration, as described in connection with the Fig. 2a-c As described, exclusively or predominantly bending actuators can be used, ie the additional rigid plate element can be omitted. The chamber 42 is, as in Fig. 8a illustrated, limited by two non-deflected S-actuators 22a and 22b. The S-actuator 22a, which limits the left (negative x-direction), can be connected to the rest of the component with its two ends at the top or bottom in the drawing (i.e., along the positive or negative y-direction). The S-actuator 22b, which limits the right, can be attached to a post (anchor element) 84. The two ends of this S-actuator can be freely movable. The post 84 can be firmly connected to the upper and lower covers 32a and 32b, respectively. When a signal is applied, both actuators bend in an S-shape. Fig. 8a The spring element 72a, which is shown hidden and influenced by a recess, can serve as strain relief. The spring element is in the plane of the drawing of the Fig. 8b along the lateral movement direction 24 in the element 48, so that the spring element 72a is firmly clamped along the lateral movement direction 24. The spring element 72a can, as shown for example in Fig. 8 As shown, the spring element 72a has a fixed connection to the spacer layers 34a and 34b and can also be clamped. Alternatively, the layers 34a and 34b can also be structured such that the spring element 72a has no contact with the spacer layer 34a and / or 34b and can thus have a higher degree of compliance.

[0114] As it is in the Fig. 8c and 8dAs shown, the bulbous bulges of the S-actuator 22a can be moved toward the post 84 so that the center of the S-actuator 22a almost touches the center of the S-actuator 22b. At the same time, the free ends of the S-actuator 22b have moved toward the fixed clamping of the S-actuator 22a so that they also almost touch. The actuated shape of the two S-actuators can be approximately the same or identical, so that the chamber 42 can close practically or almost completely with sufficient deflection of the actuators. The original volume of the chamber 42 can therefore be used entirely for generating the volume flow or for detecting it.To the extent that chamber 42 loses volume, chamber 38 can gain volume. With appropriate dimensioning of the elements influencing the flow, it can be prevented that an excessive pressure difference between chambers 38 and 42, caused by dynamic effects, impairs the movement of the actuators. Elements 46 and 48 can be designed such that the distance to the free ends of actuators 22b remains small and / or approximately constant, regardless of the deflection of the ends. Spiral spring elements 72a can be arranged to provide strain relief for actuators 22a, as described above.

[0115] The above-described embodiments may include additional actuators arranged in the resulting flow channels. These additional actuators may, for example, not serve to directly generate sound, as the electromechanical transducers 18 can enable, but may instead be used to variably adjust the flow characteristics. This allows, for example, the damping and consequently the width of the resonance curve to be individually adjusted for each chamber in a requirement-specific and flexible manner during operation of the component (MEMS transducer).

[0116] In the estimate given above, the volume change per active area (ΔV / A) for a membrane loudspeaker according to the state of the art was estimated at 3.75 µm. This can be achieved, as explained below, for a Fig. 8a-c The MEMS converters shown were re-estimated using dimensions that are reasonable for microtechnology to obtain an estimate for an active area ΔV / A. For this purpose, a width of the actuators (in Abbildung 8a in the x-direction) a value of 5 µm can be assumed. The width of the post 84 can also have a value of 5 µm. For the distance between the actuators, which form the side walls of the chamber 38, (approximately in the Fig. 8a and 8b in the undeflected state) can be assumed to be 10 µm. For a distance of the actuators, which form the side walls of the chamber 42, ( Fig. 8a and 8b In the undeflected state) 100 µm can be assumed. A planar fill factor F p , which can indicate which portion of the active area is usable for the generation of a volume flow, can then be F p = 100 / 5 + 100 + 5 + 10 = 83 % ΔV / A can be expressed as: ΔV / A = A x F ph / A = F ph

[0117] In the above expression, h can be the height of the chamber (e.g. the z-direction in Fig. 8a ). For simplicity, only the actuator height can be assumed for this purpose. The thickness of the spacer layers 34a and 34b can be neglected. In comparison with the above 3.75 µm for the diaphragm loudspeakers, it becomes clear that an actuator height of just 3.75 µm / F p (i.e., 4.5 µm) is sufficient to provide the same volume flow per active area. With an actuator thickness h of approximately 50 µm, which can be manufactured using micromechanical technology without increased effort, the value can already be more than a factor of 10 higher than that of the MEMS diaphragm loudspeaker.

[0118] In embodiments according to the MEMS converter 80, which are designed without rigid plates, parasitic vibrations can be controlled or reduced much more easily due to the significantly reduced number of mechanical elements and mechanical connections than in variants that have the plate elements and possibly further deformable elements between the deformable element and the plate element. A serial connection of actuators, as for example in the Fig. 7b und 7c shown, can be used to achieve larger strokes or greater forces.

[0119] Fig. 9 shows a schematic perspective view of a stack 90. ​​The stack 90 comprises a MEMS transducer 80a, which is connected to further MEMS transducers 80b and 80c to form the stack 90 and is arranged in the stack 90. ​​The electromechanical transducers of the MEMS transducer 80a and a further MEMS transducer 80b and / or 80c can be jointly controlled. This means that, with a constant chip area, a volume flow that can be generated or detected is increased. Although the stack 90 is described as comprising the MEMS transducers 80a, 80b, and 80c, other MEMS transducers 10, 20, and / or 50 can alternatively or additionally be arranged. Although the stack 90 is described as including three MEMS transducers, the stack 90 may also include a different number of MEMS transducers, such as two, four, five, six, or more. The cavities or subcavities of the MEMS transducers, respectively.Adjacent MEMS transducers arranged in stack 90 can be connected to one another. The cavities or sub-cavities can be connected, for example, by openings in layers between individual MEMS transducers.

[0120] In other words, based on silicon technology, wafers or chips (MEMS transducers) can be stacked, for example, using bonding processes, which can further increase the volume flow, unlike traditional membrane loudspeakers. By using technologies to thin the individual wafers or chips before stacking, the stack height can be kept low. Such a technology can include, for example, an etching process and / or a grinding process.

[0121] A reduction in the thickness of layers 32a and / or 32b, which are arranged adjacent to one another, can be carried out to the extent that one or even both of these layers are removed. Alternatively or additionally, to reduce the stack height, a manufacturing process can be carried out such that certain lower or upper covers (layers 32a and 32b, respectively) are omitted. For example, the stack 90 could be formed such that the MEMS converter 80b and / or 80c are each designed without layer 32b.

[0122] Fig. 10 shows a schematic perspective view of a section of a MEMS converter 100, in which deformable elements 22a-d are arranged between sides of the substrate 14. The deformable elements 22a and 22b are indirectly connected via the anchor element 84a. This means that ends of the deformable elements 22a and 22b can be firmly connected to the substrate, optionally to the anchor element 84a, and thus (firmly) clamped. This means that the deformable elements 22a-d or other deformable elements according to further embodiments can have a beam structure. The beam structure can be firmly clamped at a first and a second end. Clamping the ends of a deformable element 22a-d or a beam structure makes it possible to reduce or significantly reduce pre-deflection of the deformable elements (for example due to layer stress gradients).This allows the gaps between the covers and the actuators to be much smaller, which has significant efficiency advantages for some applications.

[0123] The deformable elements 22a-d are, for example, each firmly clamped on both sides. A firm clamping can be achieved by arranging or creating the deformable elements 22a and / or 22b on the substrate 14 and / or on an anchor element 84a or 84b. Dashed lines 88 indicate an undeflected state, whereas solid bars 92 indicate a deflected shape of the deformable elements 22a-d. Formations or elements 94a and 94b of the substrate 14 can enable positioning of the deformable elements 22a-d along the y-direction. A paired position of electromechanical transducers 18a-c can be shifted based on the elements 94a and 94b. Electromechanical transducers 18a and 18b arranged adjacent to one another and / or in pairs can be deformable in opposite directions to one another.

[0124] The deformable element 22a and optionally an opposing deformable element 22c can be designed to influence a partial cavity section 96a based on the deformation, i.e., to enlarge or reduce it, or to carry out a deformation based on the volume flow in the partial cavity section 96a. The deformable element 22b and optionally the oppositely arranged deformable element 22d can be designed to influence a partial cavity section 96b. The partial cavity sections 96a and 96b can be connected to one another, for example, in a region of the anchor elements 84a and 84b. The deformation of the deformable elements 22a-d can be obtained such that the deformable elements 22a and 22c, or 22b and 22d, deform at a different frequency from one another, i.e.,A volume change in the partial cavity section 96a can occur at a frequency that differs from a frequency at which a volume of the partial cavity section 96b changes. If the MEMS transducer is used, for example, as a loudspeaker, different frequencies can be obtained in the partial cavity sections based on the frequency-different volume change. If the MEMS transducer 100 is used, for example, as a microphone, the partial cavity sections 96a and 96b can, for example, have different resonance frequencies from one another. Alternatively, further partial cavity sections and further deformable elements can be arranged along the y-direction, so that the MEMS transducer 100 can, for example, generate further frequencies or have further resonance frequencies.

[0125] Alternatively, the deformable elements 22a and 22b or the deformable elements 22c and 22d can also be directly connected to one another. For example, anchor elements can be arranged in a central region of one or more deformable elements 22a-d to influence the deformation of the deformable elements 22a-d. This means that the deformable elements 22a and 22b can be directly connected to one another. Alternatively, a spring element or another element can also be arranged between the deformable elements 22a and 22b.

[0126] The MEMS converter 100 can be designed such that in a first time interval the volume flow 12 is obtained in a positive y-direction from openings 26 and subsequently, in a second time interval the volume flow 12 is obtained in a negative y-direction from openings 26.

[0127] In other words, Fig. 10 a configuration in which, in turn, possibly exclusively, S-shaped actuators are arranged. To clarify the principle, the S-shaped actuators can be shown in the figure both actuated (solid lines 92) and non-actuated (dashed lines 88). Actuated and non-actuated states can also be interchangeable through appropriate design. The S-shaped actuators (deformable elements 22a-d) can be clamped at one (upper) or at their other (lower) end. The anchor elements 84a-b can be used for this purpose. The anchor elements 84a-b can be formed from layers 34a, 36, and 34b and connected to a layer 32a and / or 32b. Distances between the free ends of the S-shaped actuators and elements 94a or 94b can be eliminated based on this configuration. This can enable lower flow losses.A starting substrate can be processed in such a way that the actuators can be produced from it, wherein the starting substrate can have layer stress gradients or layer stress gradients can be introduced during the production of the actuators. A deflection of the deformable elements induced thereby can be reduced or prevented based on the arrangement of the anchor elements 84a and / or 84b. In particular, the suspension of the deformable elements on both sides can lead to a reduction or prevention of a deflection thereof in the direction of one of the layers 32a or 32b. The spacer layers 34a and 34b can therefore be thinner, which in turn can reduce the flow losses. Each chamber (partial cavity section 96a or 96b) can be delimited by two S-shaped actuators. In the example of the . Fig. 10 Two chambers can be connected in series. The number of chambers connected in series can be selected based on the available area on the chip, taking into account the acoustic properties, in particular the resonance frequency of the S-shaped actuators or the actuator-chamber system, and can vary between 1 and a high number, for example, more than 3, more than 5, or more than 10.

[0128] Elements 94a and 94b can be arranged optionally, meaning that MEMS converter 100 can also be designed without these elements. If, for example, due to a special design or control of the electromechanical converters and / or the deformable elements, a corresponding part of the actuator is not deflected, spacing from substrate 14 by means of elements 94a or 94b can be omitted. A multiple S-actuator (wave-shaped actuator) can be implemented. In particular, this arrangement enables low resonance frequencies to be achieved, since a resonance frequency of the beam (deformable element) can decrease with increasing length.

[0129] Fig. 11a shows a schematic plan view of a section of a MEMS converter 110, in which the electromechanical converters 18a-b are arranged in a different configuration compared to the configuration of the Fig. 10 are arranged obliquely relative to a lateral direction of the substrate 14, for example, the x-direction. With an extension along the y-direction that is the same as that of the MEMS transducer 100, the electromechanical transducers 18ab can have a longer axial extension. This can enable larger partial cavity sections 96a and / or 96b and / or a higher number of partial cavity sections or deformable elements connected in series.

[0130] An outer beam segment 30a of a deformable element can be indirectly connected to an outer beam segment 30c of another deformable element via the anchor element 84. Alternatively, the beam segments 30a and 30c can also be directly connected to one another.

[0131] In other words, Fig. 11a another embodiment in which the active area is larger than in the embodiments of the Fig. 10 rotated by 45°, whereby the available chip area can be utilized to a greater extent if necessary. Funnel-shaped openings 26 can be designed so that the sound can be emitted preferably perpendicular to the chip edge surface, ie, along the y-direction in the positive or negative direction thereof.

[0132] Each of the deformable elements described above can also be formed as a plurality of interconnected deformable elements.

[0133] Fig. 11b shows a schematic plan view of a section of a MEMS converter 110', which can be used, for example, as a pump. Compared to the MEMS converter 110 from Fig. 11a The partial cavity sections 96a and 96b can be connected to the surroundings of the MEMS converter 110' via two openings 26a and 26b. The partial cavity sections 96a and 96b can be connected to a first side 97a of the MEMS converter 110' via the opening 26a and to a second side 97a of the MEMS converter 110' via the opening 26b. The first side 97a and the second side 97b can, for example, be arranged opposite one another. Alternatively, the sides 97a and 97b can also be at an angle to one another. For example, one of the sides 97a or 97b may comprise a side surface of the MEMS transducer 110' and the other side 97b or 97a may comprise a main side (e.g., a top or bottom side) of the MEMS transducer 110'.

[0134] Based on a deformation of the deformable elements 22a-d, the fluid flow can be generated from the first side 97a to the second side 97b or vice versa through the MEMS converter 110'. For example, the deformable elements 22a and 22c can be deformed in a first time interval, and the volume of the partial cavity section 96a can be reduced. In a second time interval, the volume of the partial cavity section 96b can be reduced. Based on an order of reducing or increasing the volumes, a direction of the volume flow 12 can be influenced. Alternatively, several partial cavity sections can be arranged one behind the other, or only one partial cavity section can be arranged.

[0135] Put simply, the function of a pump can be achieved by generating the volume flow 12 through the MEMS transducer according to a flow principle instead of back and forth, analogous to a loudspeaker. An inlet and an outlet side of the MEMS transducer can be arranged opposite one another, but can alternatively also be at an angle to one another or be spatially or fluidically spaced from one another on the same side. The cavity comprising the partial cavity sections 96a and 96b can have the openings 26a and 26b in the substrate. At least one of the electromechanical transducers 18a or 18b can be designed to provide the volume flow 12 based on the fluid. For example,At least one of the electromechanical transducers 18a or 18b may be configured to convey the fluid through the first opening 26a in a direction of the cavity based on an actuation of the electromechanical transducer, or to convey the fluid through the second opening 26b in a direction away from the cavity based on the actuation, or vice versa.

[0136] Although a pump function is described in connection with the MEMS converter 110', other embodiments described here may also be usable as a pump or micropump, for example by adapting an arrangement of openings of the cavity, partial cavity or at least a partial cavity section.

[0137] If the deformable elements 22a and 22e are simultaneously deflected, a negative pressure (alternatively, positive pressure) can result in an intermediate volume that counteracts the deformation or deflection. This volume can have an opening, for example, in layer 32a and / or 32b, allowing pressure equalization within this volume. This enables efficient operation of the MEMS converter 110'.

[0138] Fig. 12a shows a schematic view of a MEMS converter 120, which can be used, for example, as a MEMS pump, in a first state. The MEMS converter 120 has, for example, two deformable elements 22a and 22b, which have a beam structure and are clamped or fixedly clamped on both sides to the substrate 14. Alternatively, the MEMS converter 120 can also be designed with one deformable element or with more than two deformable elements.

[0139] Fig. 12b shows the MEMS converter 120 in a second state. Based on a deformation of at least one deformable element 22a and / or 22b, starting from the first state, as shown in Fig. 12a is shown, the second state can be obtained. Starting from the second state, the first state can be obtained based on a re-deformation of the deformable element or elements. In the second state, the partial cavity 38 between the deformable elements 22a and 22b is, for example, enlarged compared to the first state. During a transition from the first to the second state, a negative pressure can arise in the partial cavity 38. During a transition from the second state to the first state, a negative pressure can arise in the partial cavity 38.

[0140] Between a deformable element 22a or 22b and the substrate 14, a partial cavity 42a or 42b is arranged, the volumes of which can be reduced or enlarged complementarily to the volume of the partial cavity 38, wherein an overpressure or underpressure based on the deformation of the deformable elements can also be obtained complementarily to the partial cavity 38.

[0141] A valve structure 85a-f can be arranged in a region of a respective opening 26. One or more valve structures 85a-f can be formed, for example, from a material of the substrate 14. The valve structures can be formed integrally with one or more layers of the substrate 14 and can be produced, for example, by means of an etching process.

[0142] The valve structures can be configured to impede, i.e., reduce or prevent, a flow of the volume flow 12 through the opening 26 at least along one direction. For example, the valve structures 85b, 85d, and 85f can be configured to reduce or prevent the fluid from escaping from the respective partial cavity. Alternatively or additionally, the valve structures 85a, 85c, and 85e can be configured to reduce or prevent the fluid from entering the respective partial cavity. One or more valve structures 85a-f can be passively configured, for example, as a cantilevered bending beam structure or tongue structure. Alternatively or additionally, one or more valve structures 85a-f can be actively configured, for example, as an electromechanical transducer or deformable element. Simply put, the valve structures 85a-f can be actuated like the other actuators (electromechanical transducer) of the MEMS transducer.

[0143] The valve structure 85d can, for example, be designed to allow the volume flow 12 to flow into the partial cavity 38 based on a negative pressure in the partial cavity 38, while the valve structure 85c simultaneously reduces or prevents the volume flow 12 from entering the partial cavity 38. Occurs, as in Fig. 12b shown, an overpressure occurs in the partial cavity 38, the valve structure 85c can be designed to allow the volume flow 12 to flow out of the partial cavity 38 based on the overpressure, while the valve structure 85d simultaneously reduces or prevents the volume flow 12 from escaping from the partial cavity 38.

[0144] The function of the valve structures 85a, 85b, or 85e and 85f can be the same or comparable with respect to the partial cavities 42a and 42b, respectively. The valve structures 85a-f can also be referred to as check valves and enable, for example, the adjustment of a preferred direction of the volume flow 12.

[0145] Although the MEMS converter is described such that, for example, the volume flow from the partial cavities 38, 42a and 42b flows along the same direction (positive y-direction) and during different time intervals during which a transition between the first and the second state occurs, the valve structures can also be arranged such that the volume flow from at least one partial cavity 38, 42a or 42b flows along a different direction, such as the negative y-direction.

[0146] Although the MEMS converter is described as having valve structures 85a-f disposed at each aperture 26, alternatively, valve structures may be disposed at none or only some of the apertures 26.

[0147] Although the valve structures can be passively designed for a function as a check valve, the valve structures can also be actively formed, meaning they can be controllable and, in the sense of actuators, provide an open or closed state of the valve based on the control. In particular, two valve structures 85a and 85b, 85c and 85d, or 85e and 85f, each assigned to a partial cavity, can be controlled such that pressure pulses arise in the fluid stream 12, for example, by a control device connected to the MEMS converter. For example, the electromechanical converters 18 can be actuated such that an overpressure or underpressure is built up in the fluid within the partial cavities 42a, 42b, and only then is an opening of the valve structures 85a-f controlled.

[0148] In other words, such pressure pulses can also be used to approximate a low-frequency sound wave using short pressure pulses. This can be achieved in a nearly continuous manner by using several chambers arranged in series. A similar approach is also possible with chambers arranged parallel to one another. Fig. 12a shows an example in the non-actuated state, in which each chamber is provided with a valve at the top and bottom, which can be actively formed. Each valve can be opened or closed individually. Partial opening / closing is also conceivable. The valve beams can be designed and operated in the same way as the movable side walls, i.e. the deformable elements. They can therefore be based on the same or identical actuator principle. These valve bending beams can also be designed so that they can be moved in both directions or close the opening (by means of a corresponding counterforce applied by the bending actuator valve) when fluid flows (apart from a very small gap that is required for the movement). This design provides full flexibility for controlling the fluid flow in terms of direction and negative / positive pressure, individually for each chamber.If the direction of fluid flow is determined, stops for the valve beams can also be used ("check valve").

[0149] In other words, in the first state, the middle chamber (partial cavity 38) can be expanded by the two dark actuators (deformable elements 22a and 22b), while the two outer chambers (partial cavities 42a and 42b) are compressed. The former chamber fills with the fluid from the lower region via the check valve 85d. The latter pushes fluid into the upper region through the check valves 85a and 85e, respectively. In the second state, the middle chamber is compressed. Fluid is pushed into the upper region. The outer chambers fill with the fluid from the lower region. The examples of Figuren 12A und 12B are not part of the claimed invention.

[0150] Fig. 13 shows a schematic view of a first deformable element 22a and a second deformable element 22b, which are connected to one another along a lateral extension direction 98 of the deformable element 22a and / or 22b. A spring element 102 is arranged between the deformable element 22a and the deformable element 22b. The spring element 102 can cause reduced mechanically induced restoring forces in the deformable elements 22a and 22b. For example, the spring element 102 can have a low stiffness in a direction 98', which is arranged perpendicular to the direction 98, and a high stiffness along a direction 98", which can be arranged perpendicular to the directions 98 and 98' in space. The deformable elements 22a and 22b and the spring element 102 can be arranged, for example, as the deformable element 22a in the MEMS converter 110.

[0151] In other words, to provide strain relief for the S-shaped actuators 22a-d clamped on both sides, suitable spring elements 102 can be arranged at clamping locations or, for example, in an area between clamping locations, approximately in the center of the actuators. The spring element 102 is inserted, for example, in the center of the actuators and is particularly flexible in the desired direction (98') and rigid in both directions (98 and 98"), i.e., it has a high or higher rigidity. The spring element 102 can be arranged between deflectable ends of the deformable elements 22a and 22b. The spring element 102 can have a lower rigidity along the lateral movement direction 24 than in a direction perpendicular to the lateral movement direction 24.

[0152] Fig. 14 shows a schematic view of a stack 140 comprising a MEMS transducer 80'a and a MEMS transducer 80'b, which are connected to each other and have a common layer 32 compared to the MEMS transducer 80, that is, a layer 32a or 32b of the MEMS transducer 80 is removed.

[0153] Furthermore, the MEMS converter 80'a has the openings 26 in the layer 32b, which means that a radiation direction of the volume flow 12 or a penetration direction of the volume flow 12 is tilted vertically compared to the MEMS converter 80. This means that a cover surface of the MEMS converter can form an outer side of the stack, wherein the MEMS converter can have an opening in the cover surface that is arranged facing away from a side facing the second MEMS converter, wherein the volume flow 12 of the MEMS converter 80'a enters or enters the cavity perpendicularly or oppositely to the volume flow of the MEMS converter 80'b.

[0154] A membrane element 104 can be arranged on the MEMS converter 80'a. The membrane element 104 can be arranged such that an exit of the volume flow 12 from the cavity and through the membrane element 104 or an entry of the volume flow 12 into the cavity 16 is at least partially prevented. The cavity can extend to regions that are arranged outside the MEMS converter 80'a and between the MEMS converter 80'a and the membrane element 104. Based on the volume flow 12, a deflection of the membrane element 104 can be effected. The membrane element 104 can be arranged on the MEMS converter 80'a, for example, by means of a frame structure 106. The frame structure 106 can be arranged on one side of the MEMS converter 80'a, for example on a main side of the layer 32b.

[0155] Alternatively, a tilt at an angle other than 90° can also be implemented. The MEMS converter 80'b can have openings on or in the layer 32b, so that the volume flow 12 can enter and / or exit cavities on two sides of the stack 140, with the sides being arranged opposite one another.

[0156] Alternatively or additionally, stack 140 may also include another or different MEMS transducer, such as MEMS transducer 20 or 80. For example, MEMS transducer 20 may be arranged between MEMS transducers 80'a and 80'b. This allows volume flow 12 to enter or exit cavities along a direction perpendicular to a corresponding direction of MEMS transducer 80'a.

[0157] In other words, sound outlet openings 26 can also be arranged in the lower cover 32a and / or the upper cover 32b instead of on the chip side surfaces. Fig. 14 shows a corresponding simplified representation. The openings 26 in the upper cover 32b are visible. Similar openings may be located in the lower cover 32b, but are not visible based on the perspective view. The layer 32 may also have openings, meaning that cavities, sub-cavities, and / or sub-cavity sections of the MEMS transducers 80'a and 80'b may be connected to one another. Chambers lying vertically (along the z-direction) one above the other may be connected to one another via the openings in the layer 32.

[0158] A grid comprising one or more rod elements (grid webs) 44, which can be designed to adjust the damping and in particular as protection against particles, can also be used in the Fig. 14 The variant described can be easily implemented. For example, the openings 26 in the upper cover 32b or the lower cover 32a can be formed using a wet or dry chemical etching process. Before etching, the desired grating can be structured in an additionally applied thin layer that has a suitably high selectivity with respect to the etching of the openings. For the etching of the openings 26, an etching process with a suitably high isotropy or lateral undercut can be selected, so that the grating webs 44 can be undercut. For example, the grating can be manufactured in a silicon oxide or nitride layer and the covers made of silicon, which can then be structured using deep reactive ion etching (DRIE). This process can be adjusted so that undercuts on the order of micrometers can be achieved.Alternatively, a wet chemical etching can be carried out using tetramethylammonium hydroxide (TMAH) and / or potassium hydroxide (KOH) or nitric acid (HNA).

[0159] With a corresponding funnel-shaped design of the openings in the lower cover 32a and the upper cover 32b, the sound exit surface can thus comprise a larger proportion of the chip area and, if necessary, can be designed larger compared to MEMS transducers that have an exit on a side surface, such as the MEMS transducer 80. This option offers further design flexibility with regard to acoustic properties and damping. A combination of sound exit openings in the covers 32a and 32b and the side surfaces between the cover surfaces 32a and 32b is a feature of further embodiments. A preferred variant for highly integrated systems can comprise the provision of openings in the cover 32b to emit the sound upwards and the provision of the pressure equalization openings on the side in order to be able to easily mount the component, for example, on a printed circuit board.

[0160] In general, the sound inlet openings or sound outlet openings 26 can be designed such that the acoustic properties and / or the damping properties can be specifically adjusted. The lower and / or upper layers 32a and 32b can, in principle, also be capable of vibration. The vibration of these elements can be suppressed or reduced by suitable additional connecting elements in the intermediate layers 34a and 34b or 36, for example by the anchor elements 84. The suppression or reduction can comprise shifting the vibration to a frequency range that lies outside the audible sound. Alternatively or additionally, the vibration of the layers 32a and / or 32b can also be specifically implemented to optimize the acoustic radiation, wherein targeted connections can also be used in the layers and additionally the rigidity orthe acoustic properties of layers 32a and 32b can be adjusted by appropriate structuring (through openings or blind holes).

[0161] It is also possible to apply a membrane to the upper cover 32b, which is then excited to oscillate by the volume flow 12 of the chambers. This is schematically indicated by the dashed line 104. In a simple case, a spacer in the form of a frame or spacer 106 can be arranged on the upper cover 32b, on which the membrane 104 can be arranged or clamped. The production of such a membrane 104 can be carried out using known micromechanical processes. Alternatively, the membrane 104 can also be arranged inside the cavity or partial cavity and / or cover only one or a portion of the openings 26.

[0162] For some of the previously described embodiments of the MEMS transducers (such as MEMS loudspeaker components), it may be the case that there are chambers that can generate a partial volume flow independently of some, several, or all other chambers, for example in partial cavities or partial cavity sections. Chambers can be realized that consist of sub-chambers connected in a lateral and / or vertical direction (lateral, see, for example, Fig. 10 and 11 ) (vertical see for example Fig. 14 ), although embodiments also show a combination thereof. Such connected sub-chambers (such as the partial cavity sections 94a and 94b) can be used to generate a partial volume flow that is independent of or dependent on other chambers or sub-chambers. A case in which one chamber (partial cavity) can generate a volume flow independently of one another can be referred to as a mono-chamber. A chamber that can generate a volume flow based on multiple sub-chambers (partial cavity sections) can be referred to as a composite chamber.

[0163] The above-described embodiments can be modified so that both types of chambers can be combined as desired. Thus, embodiments are possible in which exclusively mono-chambers or exclusively composite chambers are arranged. Alternatively, embodiments in which both types of chambers are arranged are feasible.

[0164] In other words, when using only mono-chambers, the resonance frequencies of all actuator / chamber systems can be identical or even designed differently. For example, certain frequency ranges in the sound radiation can be emphasized by increasing the number of corresponding mono-chambers. In particular, the frequency response (sound pressure level as a function of frequency) can be shaped by appropriate distribution of the resonance frequencies and the width of the resonance curves via damping, for example, by dimensioning the grille openings or, more generally, the sound outlet openings or flow channels. Smoothing the frequency response plays a particularly important role in this process.

[0165] Partial cavities and / or partial cavity sections can emit the volume flow at different frequencies and / or be optimized for the detection of specific frequencies of the volume flow based on spatial extents of the volumes, a geometry of the electromechanical transducers and / or a frequency with which the electromechanical transducers are operated.

[0166] In a further embodiment, only mono-chambers are used. The sound outlet openings can be arranged exclusively laterally. Three chips / disks (MEMS transducers) can be stacked one on top of the other. The upper chip can be optimized for sound radiation in a first (e.g., high frequency range). A second, approximately middle, MEMS transducer can be adapted to a second frequency range (e.g., medium frequencies). A third MEMS transducer can be adapted to a third frequency range, for example, for low frequencies. This can result in a three-way loudspeaker. The arrangement of the three channels (three MEMS transducers) could also be done in one chip by using a first number N 1 of chambers laterally for the high frequencies, a second number N 2 of chambers for medium frequencies, and a third number N 3 for low frequencies.This principle can be easily extended to an N-way system in the lateral direction and, when stacked, also in the vertical direction. In another embodiment, an N-way system is designed so that the sound is generated via Fourier synthesis of the corresponding harmonics with frequencies N*f 1 , where f 1 represents the lowest frequency.

[0167] This means that a MEMS transducer can be arranged with at least one further MEMS transducer to form a stack, wherein a stack can be obtained, for example, by arranging at least two MEMS transducers along a lateral direction (such as the x-direction) and / or a thickness direction (such as the z-direction). Alternatively, the MEMS transducers can also be arranged at a distance from one another. The cavity of the MEMS transducer and the cavity of the at least one further (second) MEMS transducer can have different resonant frequencies.

[0168] In actuator mode, i.e., the deformable elements are actively deformed, an N-way loudspeaker can be obtained, where N denotes a number of MEMS transducers with mutually different resonance frequencies. In sensor mode, for example, different frequency ranges of the volume flow can be detected with different MEMS transducers. This enables, for example, a Fourier synthesis of the volume flow. For example, the control device 128 can be designed to detect the deformation of the deformable elements of one or more of the electromechanical transducers of the MEMS transducer and the further MEMS transducer. The control device can be designed to calculate a Fourier synthesis (Fourier analysis) based on the electrical signals and output a result.

[0169] The examples just presented using mono-chambers can also be realized using composite chambers, whereby the individual sub-chambers of a composite chamber have identical resonance frequencies.

[0170] When using composite chambers, the connected sub-chambers can also support different frequencies by positioning the resonance maxima accordingly. For example, three sub-chambers could form a three-way system. For example, the airflow modulated at low frequency in the rear sub-chamber (first section along an axial extension) would additionally experience a mid-frequency modulation in the middle sub-chamber (second section along an axial extension) and a high-frequency modulation in the front part of the chamber (third section along an axial extension).

[0171] At high frequencies, the required stroke, i.e., the deflection of the electromechanical transducers, is less than at low frequencies to generate the same sound pressure. The chambers or sub-chambers used for high frequencies can therefore be designed with a smaller chamber volume or a smaller distance between the actuator side walls that define the chamber.

[0172] During operation, a phase shift can be introduced between chambers of the same frequency via the control so that the wavefront is tilted and does not exit perpendicular to the surface (phased array).

[0173] In all the variants presented so far and in the following, each chamber is surrounded by at least one second chamber, into which air flows to equalize pressure when air flows into the first chamber, or vice versa. This is especially evident when there are no partition walls between these chambers, since an actuator, when moving, increases the volume of one chamber while simultaneously decreasing the volume of the other chamber, or vice versa.

[0174] For use as a loudspeaker in hearing aids or in-ear headphones, for example, the external air (i.e., the air outside the ear) is often not moved through the loudspeaker. Instead, the volume in the ear canal is periodically varied by the vibration of, for example, a membrane. This can be achieved in all the variants shown and those presented below by keeping the corresponding openings closed. These openings are located either on the top side, bottom side, or side of the chip. To achieve this, the bar grids need not be structured at these locations.

[0175] Generally speaking, and applicable to all loudspeaker applications, bar grilles can be replaced in specific locations or completely with a closed membrane. This minimizes particle sensitivity and enables operation, especially in contaminating or corrosive gases and liquids.

[0176] In the following, measures in the design and operation of the bending actuators are presented, which aim to represent the desired frequency response as well as possible.

[0177] By incorporating several additional spring elements, which divide the bending actuator into individual elements, the effective stiffness of the actuators and thus the resonance frequency can be reduced. For example, Abbildung 15 where a single spring element was used to divide the bending actuator into two elements. The division into two or more elements is important for achieving a resonant frequency in the low frequency range of audible sound, since without such a measure the bending actuators have natural frequencies in the kHz range with typical dimensions of the bending actuators (e.g. width 5 µm, length 2 mm, material silicon). Alternatively or additionally, an additional mass element can be specifically provided on the bending actuator or on the rigid plate, if present, in order to reduce the resonant frequency. Such an element can be easily provided when structuring layer 36. The mode of operation of an additional mass Δm can be explained using a model of the harmonic oscillator.

[0178] The oscillation amplitude A(ω) of an element of mass m suspended by a spring of stiffness k is given by sinusoidal excitation with a force of amplitude F 0 as: A ω = F 0 m 1 ω − ω 0 2 2 + c m ω 2

[0179] Where ω is the angular frequency of the excitation and c is the damping constant. If the resonator is operated in the quasi-static range, the amplitude is independent of the mass. The following applies for ω << ω 0 : A ω ≈ F 0 / k

[0180] An additional mass Δm thus changes the natural frequency ω 0 to the lower value ω 0- , but the amplitude of the oscillation remains unchanged. The situation is different when the bending actuator is operated in the range of its natural frequency. For ω ≈ ω 0 , the first term in the root of Equation 3 can be neglected in comparison to the second term, and the following applies: A ω ≈ F 0 / c ω 0 −

[0181] Since ω 0- is inversely proportional to the square root of the oscillator's mass, an increase in mass causes a corresponding decrease in ω 0- and thus an increase in amplitude. The increased amplitude is achieved under the condition c ω 0- < k. The possibility of constructing the cantilevers so that they can bend in one direction or the other depending on the address or signal was already described above. This means that the restoring force no longer has to be applied via the mechanical spring effect when the cantilever bends. The lower the stiffness of such a cantilever, the greater the deflection for a fixed input energy.

[0182] While all considerations have focused on the audible sound range, it is conceivable to design the component for ultrasound generation as well. In principle, it is also conceivable to equip beams with position-sensing elements (e.g., piezoresistive, piezoelectric, capacitive, etc.) instead of actuators, thus providing a component that functions as a microphone.

[0183] For the core manufacturing of MEMS speakers using silicon technology, well-known wafer bonding processes and deep reactive ion etching can be used. The production of the actuators depends on the selected operating principle and will be omitted for now. This part can be integrated modularly into the following example process. The following illustration refers to a component with only lateral openings for airflow.

[0184] BSOI (Bonded Silicon on Insulator) wafers are used as the starting material. The carrier wafer (handle wafer) forms the lower cover 32a of the MEMS loudspeaker component. The buried oxide layer of the BSOI wafer can later function as a spacer layer 34a. The active layer of the BSOI wafer can correspond to layer 36. The carrier wafer can have a thickness of 500 to 700 µm and can be further thinned if necessary—possibly at the end of the process. The buried oxide layer can have a thickness of 50 nm to 1 µm. The active layer of the BSOI wafer can have a thickness of 1 to 300 µm. Layer 36 is preferably structured, for example, using deep reactive ion etching (DRIE). After this structuring, the buried oxide layer (34a) can be removed or at least thinned, at least locally in the movement area of ​​the actuators. This can be done wet-chemically, e.g.with BOE (buffered oxide etch - buffered HF solution) or dry chemically, e.g., using gaseous HF (hydrofluoric acid). After at least partial removal of the spacer layer 34a in the movement area of ​​the actuators, a low-friction layer can be deposited, e.g., by vapor deposition (chemical vapor deposition - CVD or atomic layer deposition - ALD), which closes or significantly reduces the gap between the layer 34a and the actuators (deformable elements). Alternatively, areas in which no bonding occurs can be defined during the bonding of the wafers for the production of the BSOI wafers by depositing and structuring suitable layers, as described, for example, in US Pat. No. 7,803,281 B2. Such a process can be used for the upper and lower covers. The layer 34b is preferably structured by means of reactive ion etching (RIE), for example.With these two structures, all elements in layers 36 and 34b are produced, as shown in the corresponding figures. This also includes the rod-shaped lattice structure.

[0185] The previously described deposition of a low-friction layer can also be used for the upper cover (layer 32b). This is then applied, for example, to the cover before bonding. The spacer layer 34b can then be omitted. For example, a low-friction layer can be obtained by depositing a material. A friction value can be, for example, 10%, 20%, or 50% lower than for a material of layers 32a, 34a, 34b, or 32b.

[0186] Layer 36 can also be used as an electrical conductor with appropriate doping. Vertical electrical isolation in layer 36 is particularly advantageous when actuators are to be excited at different frequencies. This can be achieved, for example, by so-called filled trenches, as described in [8]. The use of open trenches for electrical isolation is also a possibility.

[0187] A layer is applied and structured onto a second wafer, which can be formed as a silicon wafer with a typical or possible thickness of 500 to 700 µm and will, for example, form the upper cover 32b. This layer corresponds to the spacer layer 34b. The thickness of this layer preferably corresponds to that of the buried oxide layer. All materials that enable the subsequent bonding of the second wafer to the BSOI wafer are available as material for the spacer layer. An example here is silicon oxide, preferably thermal oxide for the direct bonding of silicon oxide to silicon. Alternatively, polysilicon can also be used for the direct bonding. Another alternative is to etch suitable depressions into the second wafer so that the function of both the upper cover 32b and the function of the spacer layer 34b is mapped onto the wafer.These recesses can be omitted, at least in the area of ​​actuator movement, if the wafer is coated with a suitable low-friction layer at these locations, thus eliminating the need for a gap between the actuator (moving element) and the cover (layers 32a and / or 32b). A further layer on the second wafer—apart from auxiliary layers (masking) for structuring—can then be omitted. This also enables direct bonding of silicon to silicon.

[0188] In addition to direct bonding, it is also possible to use adhesive bonding processes, so that the spacer layer 34b is then made of a polymer material (e.g., BCB). Au-Si eutectic bonding processes or anodic bonding processes (layers containing sodium ions) are also conceivable, but not preferred due to the lack of CMOS compatibility.

[0189] After bonding the two wafers, the core of the wafer assembly manufacturing process is complete. The fabrication of the electrical wiring and contacts, as well as any required electrical insulation structures, was not performed. These elements can be provided using known standard processes according to the state of the art: fabrication of conductive paths, e.g., by sputtering and patterning AlSiCu; vertical insulation by deposition and patterning of oxides; lateral insulation by open or filled insulation trenches that completely penetrate layer 36.

[0190] The separation of components with side openings requires, in particular, the protection of the bar grids. This can be achieved, for example, by connecting the component within a frame to the frame, for example, via four thin webs. For this purpose, the lower cover 32a and upper cover 32b, as well as layers 34a, 36, and 36b, must be structured accordingly. Anisotropic etching processes such as TMAH, KOH, and DRIE are particularly suitable for this structuring. DRIE structuring of layer 36 is the preferred variant, especially for structuring along the bar grids. To remove the components from the wafer composite, the webs are destroyed. This can be done, for example, mechanically or by laser processing.

[0191] It is also conceivable not to structure the lower cover 32a for the separation process, but only layers 34a, 36, 34b, and 32b. Layer 36, in particular, can be structured using DRIE to create the vertical alignment of the bar grids. A trench then emerges from the chip surface, terminating at the lower cover 32b. This trench can then be filled with a polymer material (e.g., photoresist). The polymer serves as protection against contamination during the subsequent sawing and separation process. After sawing, the components are rinsed and cleaned to remove the sawing sludge. The polymer is then removed using suitable solvents or in an oxygen plasma.

[0192] If openings in the lower and upper lid are used instead of the side openings, the production must be expanded, as already explained in the context of Abbildung 16 described. For the separation process, the lower and upper openings can be protected, for example, by a film, allowing for sawing or laser cutting. Alternatively, the openings can be sealed with a polymer material, such as photoresist, for the separation process and subsequently removed using a solvent or oxygen plasma.

[0193] Components are preferably stacked in a wafer assembly using bonding processes. Electrical contact can then be established either through electrical contacts (bond pads) in the respective layer 36 or, when using TSVs (through-silicon vias), via so-called bumps on the chip underside. TSVs can also be used to electrically connect the stacked individual chips. TSVs and bumps can also be used for unstacked chips.

[0194] In order to achieve greater stability of the bar grids 54, the spacer layers 34a and 34b can remain unstructured in the area of ​​the bar grids.

[0195] Preferred design variants for the production of the lateral bending actuators are described below.

[0196] In principle, known electrostatic, piezoelectric, thermomechanical and electrodynamic principles can be used to actuate the bending beams.

[0197] A simple electrostatic operating principle can also be implemented without active bending beams for some of the component variants shown above. The MEMS converter 50 can be designed such that rigid plate elements 62a and 62b are designed as capacitor plates or have capacitor plates that move toward each other due to an electrical potential difference until the elements 64, which then act as a bending spring, exhibit a corresponding mechanical counterforce.

[0198] Alternatively, the cantilevers can be deflected directly via an additional, fixed counter electrode. The use of comb electrodes to increase the forces or deflection is also conceivable.

[0199] Another electrostatic principle is based on the use of a cantilevered beam that is clamped at one end and has a very short distance from an electrode at its clamping point, and this electrode distance increases with increasing distance from the clamping point. The distance at the clamping point can be zero. If an electrical voltage is applied between the cantilever beam and the electrode, a portion of the cantilever beam, determined by the magnitude of the electrical voltage and the stiffness of the beam, clings to the electrode. Based on the principle described here, the space between the beam and the electrode forms the chamber 42a, the volume of which can be varied as described.

[0200] A basic principle of such actuators is described in the literature. In [9], for example, vertically deflecting actuators are presented. The variation of the electrode spacing is achieved by the targeted introduction of layer stresses during the manufacture of the cantilever beams. For the component described in this application, actuators based on this principle could easily be realized by appropriately structuring layer 36. In addition to the structuring of layer 36, which is required anyway, an insulating layer must be applied between the electrode and the cantilever beam, which is easy to implement using known microsystem technology methods. The introduction of layer stress is not necessary, since the cantilever beams already receive the desired shape through the structuring. In the manner described here, the actuators can be deflected laterally and can therefore be used for the component principle described above.

[0201] The electrostatic operating principle offers a number of advantages regarding integration and scalability for high volume production. No external components such as magnets or coils are required, and no materials that are contamination-critical for cleanrooms, especially CMOS-compatible cleanrooms, are required. However, the membrane approach pursued to date has several disadvantages. These include the insufficient coverage of the entire audible sound range with a single vibrating membrane or plate. However, the approach of operating the membrane or membranes quasi-statically solves this problem due to the lack of resonance enhancement at the expense of the deflection and thus at the expense of the achievable volume flow or the achievable sound level. The latter are related as follows for a fixed volume, such as for in-ear headphones

[11] : SPL = 20 log 10 − 1.4 P 0 Δ V P ref V 0

[0202] SPL stands for "Sound Pressure Level", P 0 is the normal pressure, ΔV is the achievable volume change through the loudspeaker, P ref is the reference pressure which indicates a measure of the hearing threshold, it is 20 µPa, V 0 is the volume of the ear cavity in the case of in-ear headphones or hearing aids and corresponds to approximately 2 cm 3< .

[0203] With MEMS loudspeakers, it is therefore desirable to achieve the highest possible volume flow per chip area or per volume of the entire loudspeaker. Electrodynamic transducers, for example, can achieve very high diaphragm excursions and thus a high volume flow. However, the volume of the overall structure is very large due to the required permanent magnets. For loudspeakers in mobile phones, which are increasingly limited in one dimension, this approach generally appears limiting.

[0204] Piezoelectric bending actuators require the deposition of a piezoelectric layer on a substrate. This piezoelectric layer could, for example, be layer 58 of Fig. 3 which is then arranged laterally to the layer 56, which, for example, comprises or consists of silicon. The production of such actuators is possible using surface micromechanical processes.

[0205] Lateral thermomechanical actuators in the form of a cold and a warm arm, as described in

[10] , can be easily integrated by taking the corresponding geometries into account during the DRIE structuring of layer 36 as described above.

[0206] Another variant for thermomechanical actuators involves the use of bimorphs, which are heated by an electric current. To produce such a bimorph, for example, an oxide layer could be conformally deposited after structuring layer 36, so that all sidewalls are also coated. This oxide layer could then be removed from all but one sidewall of the flexure element using masking and etching processes.

[0207] The use of an electrodynamic operating principle is easily implemented for the bending beams clamped on both sides. When current flows through the beams or through a separately applied conductor structure, the beams experience a force in a magnetic field, which leads to deflection. The direction of current flow can be selected for the individual beams according to the desired deflection direction. The optional production of the conductor tracks is carried out using standard surface micromechanics processes. In this case, the additional topography must be taken into account when selecting the thickness of the spacer layer 34b.

[0208] The preferred design for the bending actuator is a lateral electrostatic actuator, which utilizes very small electrode spacing and can therefore operate at low voltages. Such lateral actuators are described, for example, in EP 2 664 058 B1. This technology allows the production of all the bending actuator and component variants described above and can be easily integrated modularly into the core part of the component manufacturing process described above.

[0209] In the following, reference is made to the flow losses during the movement of the side walls, i.e., the deformable elements. Assuming laminar flow, it can be shown in a simple model that the flow losses, such as volume flows from chamber 42a to chamber 38a in Fig. 2a Compared to the useful volume flow rates, i.e., the volume flow that penetrates outwards or penetrates from the outside to the inside, can be kept suitably low if the spacer layers 34a and 34b are small compared to the thickness of layer 36. The same applies to the distance at the possibly free end of a bending beam to the laterally limiting structure. The latter can be omitted for bending actuators clamped at both sides. If the flow losses are calculated for this configuration in the model of a laminar flow through rectangular pipes, a loss due to flow around the useful volume flow of approximately 3% can result if the following dimensions are assumed: Bending actuator: Length: 1 mm, height: 30 µm, width: 10 µm. Chamber: For the calculation of the outward flow resistance, an average width of 50 µm was assumed. This underestimates the flow resistance at large deflections of the bending actuators. Layer thicknesses of spacers 34a and 34b: 0.5 µm each

[0210] The assumed dimensions are merely examples and can be easily realized using micromechanical technologies. The assumption of laminar flow could be incorrect due to the narrow width of the actuators (above: 10 µm), which corresponds to the tube length. However, this assumption is a worst-case scenario, since flow resistance increases when turbulence occurs. To encourage such turbulence, the bending actuators in layer 36 can be provided with suitable laterally formed elements. Arrangements that form vortices when flowing around them are considered suitable. Alternatively or additionally, deliberate roughening of the surface of the cover 32a and 32b facing the chamber can promote the formation of turbulent flow.

[0211] Fig. 15 shows a schematic side sectional view of a deformable element 150 having a first layer 112 and a second layer 114, which are spaced apart and connected to one another via connecting elements 116, wherein the connecting elements 116a-c are arranged at an angle of ≠ 90° to the layer 114 and the layer 112. For example, the layers 112 and 114 can have an electrode. Alternatively, an electrode can be arranged on each of the layers 112 and / or 114. Based on the application of an electrical potential, a repulsive or attractive force can be generated between the layers 112 and 114. The attractive or repulsive force can lead to a deformation of the elements 116a-c, so that a deflectable end 122 of the deformable element 144 facing away from a clamped end 118 can be deflected along the lateral movement direction 24.

[0212] This means that the deformable element 150 can have a first layer 114 and a second layer 116, wherein spacers 116a-c can be arranged between the first layer 114 and the second layer 116. The spacers 116a-c can be arranged in an inclination direction 124 oblique to a course of the layers 112 and 114. An attractive force between the layers 112 and 114 can cause the deformable element 150 to bend.

[0213] The deformable element 150 can be flat or simply curved along the inclination direction. Alternatively, the deformable element or the layers 112 and / or 114 can also have at least two discontinuously arranged sections, for example, following a sawtooth pattern.

[0214] Fig. 16 shows a schematic plan view of a deformable element 160 arranged adjacent to an electrode 126. The deformable element 160 may have a further electrode 127 or may be the further electrode 127. Based on an applied electrical potential between the electrode 126 and the further electrode 127 of the deformable element 160, an electrostatic or electrodynamic force F can be generated. Based on the electrostatic or electrodynamic force F, a deformation of the deformable element 160 can be effected.

[0215] In a state of the deformable element 160 unaffected by the volume flow or the electrical potential, i.e., the force F, a distance between the deformable element 160 and the electrode 126 can be variable along the axial extension direction 98 of the deformable element. In a region where the mechanical transducer or the deformable element 160 is connected to the substrate 14, the distance can be minimal. This enables a high degree of controllability of the deformation of the deformable element 160. Alternatively, the distance between the electrode 126 and the deformable element 160 can be arbitrarily variable or constant along the extension direction 98.

[0216] According to embodiments, electromechanical transducers can be formed as electrostatic transducers, as piezoelectric transducers, as electromagnetic transducers, as electrodynamic transducers, as thermomechanical transducers or as magnetostrictive transducers.

[0217] Based on a force that can be generated, a deformation of the deformable element can be caused or a deformation of the deformable element can be detected or determined.

[0218] Fig. 17 shows a schematic block diagram of a MEMS system 170 having the MEMS transducer 80 connected to a control device 128 configured to control the electrodynamic transducers of the MEMS device 80 and / or to receive electrical signals from the electrodynamic transducers of the MEMS device 80.

[0219] If the MEMS converter 80 has, for example, a plurality of electromechanical transducers 18, the control device 128 can be configured to control the plurality of electromechanical transducers such that a first and an adjacent second electromechanical transducer move at least locally toward one another during a first time interval. The control device 128 can be configured to control the plurality of electromechanical transducers such that the first electromechanical transducer and a third electromechanical transducer arranged adjacent to the first electromechanical transducer move toward one another during a second interval. The first electromechanical transducer can be arranged between the second and the third electromechanical transducer. For example, these can be the electromechanical transducers 18a-c, wherein the electromechanical transducer 18b can be the first electromechanical transducer.

[0220] Alternatively or additionally, the control device 128 can be configured to receive and evaluate an electrical signal based on a deformation of the deformable element. For example, the control device 128 can be configured to determine a frequency or an amplitude of the deformation. This means that the system 170 can be operated as a sensor and / or actuator.

[0221] The system 170 can be operated, for example, as a MEMS loudspeaker, wherein the volume flow 12 can be an acoustic sound wave or an ultrasonic wave.

[0222] Alternatively, the system 170 can be implemented as a MEMS pump. A cavity of the substrate can have a first opening 26 and a second opening 26 in the substrate 14. The electromechanical transducer 18 can be configured to provide the volume flow 12 based on the fluid. The electromechanical transducer can be configured to convey the fluid through the first opening 26 in a direction toward the cavity based on an actuation of the electromechanical transducer 18, or to convey the fluid through the second opening in a direction away from the cavity based on the actuation.

[0223] Alternatively, system 170 can be operated as a MEMS microphone, wherein an electrical signal can be obtained at a terminal of electromechanical transducer 80 or another connected electromechanical transducer based on the deformation of the deformable element. The deformation of the deformable element can be effected based on the volume flow 12.

[0224] Although the system 170 is described with the control device 128 connected to the MEMS converter 80, another MEMS converter may also be arranged, such as the MEMS converter 10, 20, 50, 100, or 110. Alternatively or additionally, a plurality of MEMS converters may be arranged according to previously described embodiments. Alternatively or additionally, a stack of MEMS converters may be arranged, such as the stack 90 or 140. Alternatively or additionally, at least two MEMS converters may be arranged. At least a first MEMS converter and a second MEMS converter may have cavities or partial cavities and / or electromechanical transducers with different resonance frequencies, such as a chamber with 500 Hz actuators, another chamber or another (partial) cavity with 2 kHz actuators, etc.).

[0225] Fig. 18 shows a schematic plan view of a MEMS transducer 180 having a plurality of electromechanical transducers 18a to 18i, wherein the electromechanical transducers 18a to 18f are arranged laterally offset from one another in a first cavity 16a, and the electromechanical transducers 18g to 18i are arranged laterally offset from one another in a second cavity 16b. The cavities 16a and 16b can have an opening in a bottom and / or cover surface (not shown) of the substrate 14. The MEMS transducer 180 can be used as a loudspeaker and / or microphone, which applies both to individual electromechanical transducers 18a to 18i and to the electromechanical transducers 18a to 18f or 18g to 18i of a respective cavity 16a and 16b. The speakers and / or microphones can also be designed to be optimized for the transmission or reception of sound waves via vibrations. For example,It can be placed on the human body, ideally close to a bone, to transmit or record information using structure-borne sound. In this case, a preferred variant is one in which all actuators move in the same direction. This means, regardless of the approach, that a chamber has two movable walls. The electromechanical transducers 18a to 18i comprise cantilevered beam elements.

[0226] In other words, the left chamber, cavity 16a, contains laterally or vertically movable bending actuators that preferably oscillate in phase, thus causing the chip to vibrate and transmit sound. The right chamber, cavity 16b, contains three lateral or vertical bending actuators that also preferably oscillate in phase, but whose dimensions (thickness, length, or width) represent a different frequency range than the left chamber.

[0227] Fig. 19 shows a schematic plan view of a MEMS transducer 190 comprising a plurality of electromechanical transducers 18a to 18i. The electromechanical transducers 18a to 18f are arranged laterally offset from one another and space adjacent cavities 16a to 16k or subcavities from one another. The electromechanical transducers 18a to 18i comprise beam elements clamped on both sides.

[0228] Although the examples of implementation of the Fig. 18 und 18 are shown such that the MEMS converter 180 has only beam elements clamped on one side and the MEMS converter 190 has only beam elements clamped on both sides, the embodiments can also be combined with one another as desired, so that similar electromechanical converters can be arranged independently of one another in each cavity 16a or 16b, or different electromechanical converters can be arranged within a cavity.

[0229] In other words, Fig. 19 the same principle as in Fig. 18 , but this time bending actuators clamped on both sides are used.

[0230] Further embodiments relate to a method for producing a MEMS transducer. The method comprises providing a substrate having a cavity. Furthermore, the method comprises producing an electromechanical transducer, which has an element deformable along a lateral movement direction, with the substrate. The producing step is carried out such that a deformation of the deformable element along the lateral movement direction and a volume flow interacting with the MEMS transducer are causally related. The electromechanical transducer can be produced, for example, by molding it from the substrate, for example by an etching process and / or by a deposition process for depositing additional layers.

[0231] Although the previously described embodiments refer to the volume flow being generated by two electromechanical transducers moving toward each other, the volume flow can also be obtained based on, or in causal interaction with, a movement of an electromechanical transducer relative to a rigid structure, for example, the substrate. This means that a volume of a partial cavity or a partial cavity section can be influenced by a single electromechanical transducer.

[0232] Previously described embodiments comprising a deformable element configured to perform multiple curvature and / or connected to a plate element can be compared to the configuration described in connection with the Fig. 1 described, can be used to generate a significantly higher volume flow or to react to a volume flow much more sensitively.

[0233] Examples of embodiments allow the frequency-dependent curve of the sound pressure to be flexibly adjusted, in particular to enable the frequently desired case of a frequency curve that is as flat as possible.

[0234] In order to make a frequency-dependent sound pressure curve as flat as possible with as few chambers in the MEMS transducer, it is advantageous if the quality factor of the oscillating cantilevers is low, i.e. the cantilevers have a broad resonance curve. For this purpose, the beams can be clamped in such a way that the beam vibration is additionally damped by means of a damping material. The beam clamp is preferably made of a non-crystalline material. This includes silicon oxide, polymers such as SU8 or other resists. Damping of the beam vibration can also be achieved electrically. For example, during the free beam oscillation of an electrostatic or piezoelectric actuator when a voltage is applied, a periodically changing current flows due to the change in capacitance. Suitable electrical resistors result in power loss that damps the vibration.A complete electrical oscillating circuit (i.e., an additional integrated or external coil is provided) is also possible. Damping can also be achieved by implementing additional structures on the cantilever beams, which represent a significant flow resistance for the fluid as it flows into or out of the chamber.

[0235] Especially for the representation of low resonance frequencies – for the generation or detection of low frequencies – it can be advantageous to increase the mass of the bending beams. To avoid significantly increasing stiffness, additional structures are preferably installed in the area of ​​the largest vibration amplitudes. In the case of a beam clamped at one end, the optimal location or area of ​​the largest vibration amplitudes is the end of the bending beam. In the case of a beam clamped at two ends, this is the center of the beam.

[0236] In other words, one finding of the present invention is that a volume flow is generated or can be detected by compression or expansion of chambers, i.e. partial cavities or partial cavity sections that can be formed in a silicon chip. Each chamber can be provided with an inlet or outlet through which a fluid, such as air, can flow in or out. The chambers can be closed by a fixed cover along a direction perpendicular to the lateral direction of movement (for example, top and bottom). At least one of the lateral walls of each chamber is movable or deformable and can be displaced by an actuator such that the volume of this chamber is reduced or increased.

[0237] Embodiments of MEMS converters described above may have electrical connections, bond pads or the like that are not shown in the figures for the sake of clarity.

[0238] The embodiments described above relate to multi-way loudspeakers or N-way loudspeakers that can be obtained based on different resonance frequencies of at least two cavities or sub-cavities. The electromechanical transducers and the cavities or sub-cavities can be coordinated with one another such that a sound pressure level (SPL) is, at least in sections, a function of the resonance frequency, i.e., several actuator chambers can have different frequency curves (SPL = f(frequency)). This means that values ​​of sound pressure levels obtained based on the deformation of the deformable elements and based on the sub-cavities have a relationship with a frequency of the volume flow flowing out of or into the respective sub-cavity. The relationship can be represented as a function, whereby the function can be linear, for example, such as SPL = x*frequency + b, where x and b are variables.Alternatively, the function can also be nonlinear, such as quadratic, exponential, or based on a root function. This functional relationship can easily be transferred to different subcavities or cavities arranged in different MEMS transducers. Thus, the frequency of the volume flow can describe a frequency-dependent pressure curve in the fluid.

[0239] The silicon chips of MEMS converters can be designed and extracted from the wafer-level composite during wafer-level manufacturing to achieve a customized shape for the respective application. For example, the chip can be designed round for use as a speaker in hearing aids or in-ear headphones, or hexagonal, which is more suitable for consuming silicon area on the wafer.

[0240] Although some aspects have been described in connection with a device, it is understood that these aspects also represent a description of the corresponding method, so that a block or component of a device can also be understood as a corresponding method step or as a feature of a method step. Similarly, aspects described in connection with or as a method step also represent a description of a corresponding block, detail, or feature of a corresponding device.

[0241] The above-described embodiments are merely illustrative of the principles of the present invention. It is understood that modifications and variations of the arrangements and details described herein will be apparent to others skilled in the art. Therefore, it is intended that the invention be limited only by the scope of the following claims and not by the specific details presented in the description and explanation of the embodiments herein. Literatur

[0242] [1] Albach, Thorsten Sven: Magnetostriktive Mikroaktoren und deren Anwendung als Mikrolautsprecher, Dissertation, Universität Erlangen-Nürnberg (2013). [2] Roberts, Robert C. et al.: Electrostatically Driven Touch-Mode Poly-SiC Microspeaker, Sensors, IEEE 2007 (2007), p. 284-287. [3] Kim, H. et al.: Bi-directional Electrostatic Microspeaker with Two Large-Deflection Flexible Membranes Actuated by Single / Dual Electrodes, Sensors, IEEE 2005 (2005), p. 89-92. [4] Rehder, J.; Rombach, P.; Hansen, O.: Magnetic flux generator for balanced membrane loudspeaker. In: Sensors and Actuators A: Physical 97 (2002), Nr. 8, p. 61-67. [5] Neri, F.; Di Fazio, F.; Crescenzi, R.; Balucani, M.: A novel micromachined loudspeaker topology. In: 61 st Conf. on Electronic Components and Technology, ECTC, IEEE 2011 (2011), p. 1221-1227. [6] Neumann, J. J., Gabriel, K. J.: CMOS-MEMS Acoustic Devices, in: Advanced Micro and Nanosystems, Vol. 2. CMOS-MEMS. Edited by H. Baltes et al., Wiley-VCH Verlag, Weinheim (2005). [7] Lerch R.; Sessler, G.; Wolf, D.: Technische Akustik, Springer Verlag (2009). [8] Schenk, H. et al.: A resonantly excited 2D-micro-scanning-mirror with large deflection, Sensors and Actuators A 89 (2001), p. 104-111. [9] Rosa, M. A. et al.: A novel external electrode configuration for the electrostatic actuation of MEMS based devices, J. Micromech. Microeng.(2004), p. 446-451.

[10] Kumar, V.; Sharma, N. N.: Design and Validation of Silicon-on-Insulator Based U Shaped thermal Microactuator, Int. J. Materials, Mechanics and Manufacturing, Vol. 2, No. 1 (2014), p. 86-91.

[11] Cheng, Ming-Cheng et al.: A lilicon microspeaker for hearing instruments, J. Micromech. Microeng. 14 (2004), p. 859-866.

Claims

1. MEMS system (170), comprising: an MEMS transducer (10; 20; 50; 80; 80'; 100; 110) for interacting with a volume flow (12) of a fluid, comprising: a substrate (14) comprising a cavity (16); and wherein the MEMS transducer (10; 20; 50; 80; 80'; 100; 110) comprises a multitude of electromechanical transducers (18; 18a-f), and each of these electromechanical transducers (18; 18a-f) is connected to the substrate (14) in the cavity (16) and comprises an element (22; 22a-f; 30; 40; 150; 160) deformable along a lateral movement direction (24), wherein a deformation of the deformable element (22; 22a-f; 30; 40; 150; 160) in-plane along the lateral movement direction (24) and the volume flow (12) of the fluid are causally related; wherein each electromechanical transducer (18; 18a-f) is connected to the substrate (14) in a force-fitted or in a form-fitted manner; or is formed integrally with the substrate (14); wherein the MEMS system further comprises: a control device (128) configured to drive the deformation of the deformable elements (22; 22a-f; 30; 40; 150; 160) or to detect the deformation of the deformable elements (22; 22a-f; 30; 40; 150; 160); wherein the control device (128) is configured to drive the multitude of electromechanical transducers (18; 18a-f) such that a first (18b, 18d) and a neighbouring second electromechanical transducer (18c, 18e) at least locally move towards each other during a first time interval, and wherein the control device (128) is configured to drive the multitude of electromechanical transducers (18; 18a-f) such that the first electromechanical transducer (18b, 18d) and a third electromechanical transducer (18a, 18c) arranged adjacent to the first electromechanical transducer (18b, 18d), wherein the first electromechanical transducer (18b, 18d) is arranged between the second (18c, 18e) and the third electromechanical transducer (18a, 18c), move towards each other during a second time interval.

2. MEMS system according to claim 1, wherein the electromechanical transducers (18; 18a-f) of the MEMS transducer are configured to, in response to an electrical drive (129a), causally cause a movement of the fluid in the cavity (16) and / or, in response to the movement of the fluid in the cavity (16), to causally provide an electrical signal (129b).

3. MEMS system according to claim 1 or 2, wherein a first subcavity (42a, 42b) is arranged between the first electromechanical transducer (18b, 18d) and the second electromechanical transducer (18c, 18e) and a second subcavity (38a, 38b) is arranged between the second electromechanical transducer (18b, 18d) and a third electromechanical transducer (18a, 18c); and wherein the first, second and third electromechanical transducers (18a, 18b, 18c) are configured to causally cause a movement of the fluid in the cavity (16) in response to an electrical drive (129a); and wherein the first and the second electromechanical transducer (18b, 18d) are configured to change a volume of the first subcavity with a first frequency, wherein the first (18b, 18d) and the third electromechanical transducer (18a, 18c) are configured to change a volume of the second subcavity with a second frequency.

4. MEMS system according to claim 3, wherein the first subcavity and the second subcavity of the MEMS transducer comprise resonance frequencies different from each other.

5. MEMS system according to claim 4, wherein the first and the second electromechanical transducer (18b, 18d) of the MEMS transducer are configured to change a volume of the first subcavity with a first frequency, wherein the first (18b, 18d) and the third electromechanical transducer (18a, 18c) are configured to change a volume of the second subcavity with a second frequency.

6. MEMS system according to one of claims 3 to 5, wherein the volume flow (12) and the deformation of the deformable elements (22; 22a-f; 30; 40; 150; 160) of the MEMS transducer are causally related with the change of the volumes of the first subcavity (42a, 42b) and the second subcavity (38a, 38b).

7. MEMS system according to one of the preceding claims, wherein a first subcavity (42a, 42b) is arranged between a first electromechanical transducer (18b, 18d) and a second electromechanical transducer (18c, 18e) and a second subcavity (38a, 38b) is arranged between the second electromechanical transducer (18b, 18d) and a third electromechanical transducer (18a, 18c). wherein the first and the second electromechanical transducer (18b, 18d) are configured to change a volume of the first subcavity with a first frequency, wherein the first (18b, 18d) and the third electromechanical transducer (18a, 18c) are configured to change a volume of the second subcavity with a second frequency.

8. MEMS system according to one of the preceding claims, wherein a first deformable element (22; 22a-f; 30; 40; 150; 160) of the first electromechanical transducer of the MEMS transducer and a second deformable element of a second electromechanical transducer of the MEMS transducer include a bar structure (30) configured to be curved in-plane with respect to the substrate (14).

9. MEMS system according to one of the preceding claims, wherein the electromechanical transducer (18; 18a-f) of the MEMS transducer includes a plurality of deformable elements (22; 22a-f; 30; 40; 150; 160) at least indirectly connected in an axial direction (y) of the electromechanical transducer (18; 18a-f), which are configured to each affect a volume of a first and of a second subcavity portion (96a, 96b).

10. MEMS system according to claim 9, wherein the electromechanical transducer (18; 18a-f) of the MEMS transducer is configured to, in response to an electrical drive (129a), causally cause a movement of the fluid in the first (96a) and the second subcavity portion (96b), wherein the deformable elements (22; 22a-f; 30; 40; 150; 160) are configured to change the volumes of the first (96a) and the second subcavity portion (96b) with a frequency different from each other.

11. MEMS system according to one of the preceding claims, wherein the substrate (14) of the MEMS transducer comprises an anchor element (84); wherein one of the deformable elements (22; 22a-f; 30; 40; 150; 160) is connected to the anchor element (84) in a center region (30b) of an axial extension direction (y) of the deformable element (22; 22a-f; 30; 40; 150; 160); or wherein the deformable element (22; 22a-f; 30; 40; 150; 160) is connected to a further deformable element at an outer bar segment (30a, 30c) via the anchor element (84).

12. MEMS system according to one of the preceding claims, wherein the electromechanical transducer (18; 18a-f) of the MEMS transducer is formed as electrostatic transducer, piezoelectric transducer, electromagnetic transducer, electrodynamic transducer, thermomechanical transducer or magnetostrictive transducer.

13. MEMS system according to one of the preceding claims, including at least one further MEMS transducer (10; 20; 50; 80; 80'; 100; 110), wherein the cavity (16) of the further MEMS transducer (10, 20, 50, 80, 80', 100', 110) comprises a resonance frequency different from a resonance frequency of the cavity (16) of the MEMS transducer (10; 20; 50; 80; 80'; 100; 110), wherein the control device is configured to detect the deformation of the deformable element (22; 22a-f; 30; 40; 150; 160) of the MEMS transducer and of the further MEMS transducer and to compute a Fourier synthesis based on the electrical signals.

14. MEMS system according to one of the preceding claims, including at least one further MEMS transducer (10; 20; 50; 80; 80'; 100; 110), wherein the cavity (16) of the further MEMS transducer (10; 20; 50; 80; 80'; 100; 110) comprises a resonance frequency different from a resonance frequency of the cavity (16) of the MEMS transducer (10; 20; 50; 80; 80'; 100; 110), wherein the control device is configured to drive the deformation of the deformable element (22; 22a-f; 30; 40; 150; 160) of the MEMS transducer and of the further MEMS transducer with frequencies different from each other.