MEMS transducers with improved performance

Abstract

The present invention relates to a MEMS transducer having a vibrating diaphragm 1 for generating or capturing pressure waves in a fluid in the vertical direction, the vibrating diaphragm 1 being held by a support 4 and having two or more vertical sections 2 formed parallel to the vertical direction and having at least one layer of actuator material 11. The ends of the vibrating diaphragm 1 are in contact with electrodes 13, which means that by actuation of at least one electrode 13, the two or more vertical sections 2 can be excited to generate horizontal vibrations, or this means that excitation of the two or more vertical sections 2 to generate horizontal vibrations can result in an electrical signal being generated in at least one electrode 13.

Description

[Technical field]

[0001] The present invention relates to a MEMS transducer having a vibratable membrane for generating or receiving fluid pressure waves in a vertical direction, the vibratable membrane being supported by a carrier and the vibratable membrane exhibiting two or more vertical sections formed parallel in the vertical direction and having at least one layer of actuator material. At least one end of the vibratable membrane is preferably connected to an electrode, so that by driving the at least one electrode, the two or more vertical sections can be induced to generate horizontal vibrations or so that when the two or more vertical sections are induced to vibrate in a horizontal direction, an electrical signal can be generated in the at least one electrode. [Background technology]

[0002] Today, microsystems technology is used in many fields of application for the manufacture of small mechanical and electronic devices. The microsystems (microelectromechanical systems, MEMS for short) that can be manufactured in this way are very small (in the micrometer range), have excellent functionality and are much less expensive to manufacture.

[0003] MEMS transducers, such as MEMS loudspeakers or MEMS microphones, are also known in the prior art. Current MEMS loudspeakers are mostly designed as planar membrane systems, driving a vibrable membrane perpendicular to the emission direction. The vibrations are induced, for example, using piezoelectric, electromagnetic or electrostatic actuators.

[0004] An electromagnetic MEMS speaker for portable devices is described by Shahosseini et al. in 2015. The MEMS speaker features a hardened silicon microstructure that is the sound radiator, and the moving part is suspended from a carrier via silicon drive springs, allowing large out-of-plane displacements using an electromagnetic motor.

[0005] Stoppeletal et al. (2017) presented a bidirectional loudspeaker concept based on concentric piezoelectric actuators. A special feature is that the diaphragm is not a closed design but has eight piezoelectric unimorph actuators, each consisting of a piezoelectric layer and a passive layer. The outer woofer is made of four one-sided clamp actuators with a trapezoidal shape, while the inner tweeter is formed by four triangular actuators connected to a rigid frame by springs. By separating the membranes, improved sound at higher power should be possible.

[0006] A drawback of such planar MEMS speakers is their limitations, especially with regard to sound output at low frequencies. One of the reasons for this is that the sound pressure level that can be generated is proportional to the square of the frequency of a given displacement. Therefore, to obtain a sufficient sound output, a membrane displacement of at least 100 μm or a large area membrane in the range of square centimeters is required. Both conditions are difficult to achieve with MEMS technology.

[0007] Therefore, in the prior art it has been proposed to design MEMS speakers that do not have a closed membrane vibrating in the vertical emission direction, but have a number of moving elements that can be driven to generate lateral or horizontal vibrations. The advantage of this is that an increased volume flow can be moved over a smaller surface, thus providing increased sound output.

[0008] For example, a MEMS speaker based on this principle is disclosed in U.S. Patent Application Publication No. 2018 / 0179048A1 or by Kaiseretal et al. in 2019.

[0009] The MEMS speaker has several electrostatic bending actuators, which are arranged between the top and bottom wafers as vertical lamellae and can be driven to generate lateral vibrations by appropriate control. Here, the inner lamella forms the actuator electrode facing the two outer lamellae. There is an air gap between the three bending lamellae, except for the connection node of the electrodes, which is also electrically isolated. When an electric potential is applied inwards with respect to the outwards, it creates a pulling force on both sides due to the designed curvature in the direction that is the preferred direction predefined by the anchors. The bulge of the outer lamellae is used for mobility. The restoring force is given by the force of a mechanical spring. Pulling and pushing movements are therefore not possible.

[0010] Another drawback is that a gap between the flex actuator and the lid / bottom wafer is necessary for the actuator's mobility, resulting in air venting between the two chambers, which limits the lower cutoff frequency. Furthermore, to avoid the drag effect and acoustic damage, the lateral movement of the flex actuator is limited, thus limiting the sound output.

[0011] An alternative MEMS-based pneumatic pulse or sound generation system is described in US Patent Application Publication No. 2019 / 0116417A1. The device has a front and rear chamber and a number of valves, the front and rear chambers being separated from each other by a folded membrane. In one embodiment, the folded membrane has a rectangular cross-section serpentine structure with horizontal and vertical sections. Piezo actuators are arranged in each horizontal section, and synchronized expansion and contraction of the horizontal sections induces lateral movement of the vertical sections. The proposed principle allows an increasing volumetric flow, and thus an increasing sound output, to be generated on a still small chip surface.

[0012] However, one drawback is the increased effort required for synchronous actuation of the piezo actuators. There is also potential for improvement with regard to the volumes displaced by lateral vibrations, but this is limited by the geometric arrangement of the horizontal sections that are driven in one direction.

[0013] From US patent application 2002 / 006208A1 and JP patent 3919695B2, a piezoelectric speaker is known in which two piezoelectric films are formed in an accordion-shaped membrane. In the folded form, the membrane is fixed, for example by a screw connection, and is laterally clamped respectively by a pair of corrugated plates that stabilize the diaphragm as a composite side frame. A number of electrodes are attached in a structured form to the peaks and valleys of the membrane and are insulated from each other by strips of non-conductive material. To drive the electrodes, electrode cables are arranged on the pair of plates or on the side frames. Alternatively, the pair of plates can also be at least partially formed by conductive material.

[0014] The macroscopic piezoelectric loudspeakers of US Patent Application No. 2002 / 006208A1 and JP Patent No. 3919695B2 are obtained with an assembly process that cannot be miniaturized in a common way to obtain MEMS loudspeakers. In particular, the intended clamping of the membrane with the two-part side frame, the structured attachment of the electrodes to the crests and troughs of the membrane, or the connection of the electrodes with the electrode cables of the side frame cannot be transferred to the MEMS process.

[0015] Therefore, in light of the shortcomings of the prior art, there is a need for alternative or improved solutions for MEMS-based speakers. [Prior art documents] [Patent documents]

[0016] [Patent Document 1] US Patent Application Publication No. 2018 / 0179048A1 [Patent Document 2] U.S. Patent Application Publication No. 2019 / 0116417 A1 [Patent Document 3] U.S. Patent Application No. 2002 / 006208 A1 [Patent Document 4] Japanese Patent No. 3919695 B2 [Non-Patent Document]

[0017] [Non-Patent Document 1] F. Stoppel, C. Eisermann, S. Gu-Stoppel, D. Kaden, T. Giese, and B. Wagner, "NOVEL MEMBRANE-LESS TWO-WAY MEMS LOUDSPEAKER BASED ON PIEZOELECTRIC DUAL-CONCENTRIC ACTUATORS", Transducers, 2017, Kaohsiung City, Taiwan, June 18 - 22, 2017 [Non-Patent Document 2] Iman Shahosseini, Elie LEFEUVRE, Johan Moulin, Marion Woytasik, Emile Martincic, etc., "Electromagnetic MEMS Microspeaker for Portable Electronic Devices", Microsystem Technologies, Springer Verlag (Germany), 2013, page 10 <hal-01103612> [Non-Patent Document 3] Bert Kaiser, Sergiu Langa, Lutz Ehrig, Michael Stolz, Hermann Schenk, Holger Conrad, Harald Schenk, Klaus Schimmanz, and David Schuffenhauer, "Concept and proof for an all-silicon MEMS microspeaker utilizing air chambers," Microsystems & Nanoengineering, Volume 5, Article number 43 (2019) [Non-Patent Document 4] Kazuo Sato, Mitsuhiro Shikida, Yoshihiro Matsushima, Takashi Yamashiro, Kazuo Asaumi, Yasuroh Iriye, and Masaharu Yamamoto, "Characterization of orientation-dependent etching properties of single-crystal silicon: effects of KOH concentration", Sensors and Actuators A, Vol. 64 (1988), pp. 87-93. [Non-Patent Document 5] Seidel, H., Csepregi, L., Heuberger, A., and Baumgartel, H., (1990), "Anisotropic etching of crystalline silicon in alkaline solutions", Journal of the Electrochemical Society, Vol. 137, 10.1149 / 1.2086277 Summary of the Invention [Problem to be solved by the invention]

[0018] The object of the present invention is to provide a MEMS transducer, in particular a MEMS speaker or a MEMS microphone, that does not present the drawbacks of the prior art, as well as a method for manufacturing said MEMS transducer. One object of the present invention is in particular to provide a high-performance MEMS speaker or a MEMS microphone with high sound quality or audio quality, which at the same time are characterized by a simple, low-cost and compact design. [Means for solving the problem]

[0019] This object is solved by the features of the independent claims. Preferred embodiments of the invention are set forth in the dependent claims.

[0020] The present invention preferably relates to a MEMS transducer for interacting with a volumetric flow of a fluid, the MEMS transducer comprising: - Career and - a vibratable membrane for generating or receiving vertical fluid pressure waves, the vibratable membrane being supported by a carrier; the vibratable membrane has two or more vertical sections formed substantially parallel to a vertical direction and having at least one layer of actuator material, at least one end of the vibratable membrane being connected to at least one electrode; As a result, two or more vertical sections can be induced to generate a substantially horizontal vibration by driving at least one electrode, or as a result, when two or more vertical sections are induced to vibrate in a substantially horizontal direction, an electrical signal can be generated in at least one electrode.

[0021] Particularly preferably, the MEMS transducer may be a MEMS speaker. In a particularly preferred embodiment, the invention relates to a MEMS speaker, the MEMS speaker comprising: - Career and - a vibratable membrane for generating sound waves in a vertical emission direction, the vibratable membrane being supported by a carrier; wherein the vibratable membrane has two or more vertical sections formed substantially parallel to the emission direction and having at least one layer of actuator material, and at least one end of the vibratable membrane is preferably connected to at least one electrode such that by actuating the at least one electrode the two or more vertical sections can be induced to generate substantially horizontal vibrations.

[0022] This MEMS speaker design can achieve a MEMS speaker with high audio output and simplified control.

[0023] Unlike known planar MEMS speakers, the vibratable membrane itself does not need to operate over a large area of ​​several square centimeters or with a displacement of more than 100 μm to generate sufficient sound pressure. Instead, multiple vertical sections of the vibratable membrane can move an entire extended volume in the vertical emission direction with small horizontal or lateral movements of a few micrometers.

[0024] Compared to the solution of U.S. Patent Application Publication No. 2018 / 0179048A1 or Kaiser et al. of 2019, the claimed MEMS speaker is characterized by a simplified structure, control, and manufacturing process.

[0025] In particular, the creation of vertical lamellae or bending actuators for MEMS speakers by Kaiser et al. in 2019 is complicated. In addition, sufficiently accurate vertical etching is only possible if the height of the lamellae is limited, which limits the sound output.

[0026] Instead, with the solution according to the invention, the vertical section of the vibrable membrane can be realized in a MEMS design using a simple manufacturing process, as will be explained in detail below. Furthermore, the principle of the actuator according to the invention avoids the pulling or sticking of the vertical section. In contrast to the solution of Kaiser et al. in 2019, the one-sided electrodes do not introduce a potential difference in the gap between the vertical sections. In addition to avoiding overvoltages or pulling, this also reduces dust accumulation, since for example the external electrode can be placed at ground potential.

[0027] Another unique advantage of the described MEMS speaker is the simplified driving. Whereas US2019 / 0116417A1 requires multiple piezoelectric actuators to be connected to the horizontal section, the proposed MEMS speaker can be driven using at least one end electrode. This reduces manufacturing costs, minimizes error sources, and also inherently provides synchronous control of the vertical section to generate horizontal vibrations.

[0028] In this way, the air volume present between the vertical sections can be moved very precisely by horizontal vibration along the vertical emission direction, resulting in improved sound, even at high sound output levels.

[0029] "MEMS speaker" preferably means a speaker based on MEMS technology and in which the sound generating structure has dimensions at least partially in the micrometer range (1 μm to 1000 μm). The vertical section of the vibratable membrane may preferably have dimensions in the range of less than 1000 μm, for example width, height and / or thickness. Here, it may be preferred that, for example, only the height of the vertical section is dimensioned in the micrometer range, while, for example, the length may be a larger dimension and / or the thickness may be of a smaller size.

[0030] Advantageously, the vibratable membrane design can not only be used to create MEMS speakers with large sound output and simplified control, but also allows for the creation of particularly powerful MEMS microphones with high audio quality, for example.

[0031] Therefore, in a preferred embodiment, the invention further relates to a MEMS microphone, comprising: - Career and - a vibratable membrane for receiving sound waves in a vertical direction, the vibratable membrane being supported by a carrier; wherein the vibratable membrane has two or more vertical sections formed parallel in a vertical direction and having at least one layer of actuator material, and at least one end of the vibratable membrane is preferably connected to at least one electrode such that when the two or more vertical sections are induced to vibrate in a horizontal direction an electrical signal can be generated at the at least one electrode.

[0032] The design of a MEMS microphone is structurally similar to that of a MEMS speaker, especially with regard to the design of the vibrable membrane. Instead of driving electrodes to generate horizontal vibrations, i.e. sound pressure waves, the MEMS microphone is designed to receive the same vertical sound pressure waves. Preferably, there is an air volume between the vertical sections, which moves along the vertical detection direction when sound waves are received. The sound pressure waves induce horizontal vibrations in the vertical sections, which in turn causes the actuator material to generate a corresponding periodic electrical signal.

[0033] "MEMS microphone" preferably means a microphone based on MEMS technology and in which the sound receiving structure has dimensions at least partially in the micrometer range (1 μm to 1000 μm). The vertical section of the vibratable membrane may preferably have dimensions in the range of less than 1000 μm, for example width, height and / or thickness. Here, it may be preferred that, for example, only the height of the vertical section is dimensioned in the micrometer range, while, for example, the length may be a larger dimension and / or the thickness may be of a smaller size.

[0034] The term MEMS transducer therefore refers to both MEMS microphones and MEMS speakers. In general, MEMS transducers refer to transducers that are based on MEMS technology and that interact with a volumetric flow of a fluid, the structures that interact with the volumetric flow, i.e. receive or generate pressure waves in the fluid, have dimensions in the micrometer range (1 μm to 1000 μm). The fluid can be a gaseous fluid, but also a liquid fluid. The structures of the MEMS transducer, in particular the vibratable membrane, are designed to generate or receive pressure waves in the fluid.

[0035] For example, the MEMS transducer may be related to sound pressure waves, as in the case of a MEMS speaker or a MEMS microphone. However, the MEMS transducer may be equally suitable as an actuator or sensor of other pressure waves. Thus, the MEMS transducer is preferably a device that converts pressure waves (e.g., acoustic signals that are sound pressure waves) into electrical signals or vice versa (conversion of electrical signals into pressure waves, e.g., acoustic signals).

[0036] Applications of MEMS transducers as energy harvesters using alternating air or water pressure are also possible, where the electrical signal can be dissipated as harvested electrical energy, stored or fed to other (consumer) devices.

[0037] End side preferably means the arrangement of at least one electrode at one end of the vibratable membrane such that a connection with an electronic system, for example in the case of MEMS loudspeakers a connection to a current or voltage source, preferably at the end where the membrane is suspended from the carrier, can be established. Electrode preferably means an area made of a conductive material, preferably metal, adapted for such establishment of a connection with an electronic system, for example in the case of MEMS loudspeakers a connection to a current and / or voltage source. The member may preferably be an electrode pad. Particularly preferably, the electrode pad is used to establish a connection with the electronic system, the electrode pad itself being connected to a conductive metal layer, which may extend over the entire surface of the vibratable membrane. The conductive layer, which is partly integral with the electrode pad, is called electrode in the following, for example top electrode or bottom electrode.

[0038] Particularly preferably, the layer of conductive material, preferably metal, in the sense of a top or bottom electrode, is present as a continuous, full-surface or consistent layer of the vibratable membrane, forming a substantially homogeneous surface and is not particularly structured, instead preferably two or more vertical sections are connected to end electrodes or electrode pads using an unstructured layer of conductive material, preferably metal.

[0039] Advantageously, it is not particularly necessary to create separate connection areas for different vertical sections of the vibratable membrane. In contrast to the approach of macroscopic piezoelectric speakers according to US patent application 2002 / 006208 A1 and JP patent 3919695 B2, there is no need to attach structured top or bottom electrodes. Instead, the top or bottom electrodes can in each case be applied as a continuous layer of conductive material and are connected by at least one end electrode or electrode pad. The manufacturing process is therefore significantly simplified and miniaturized MEMS transducers can be produced in large quantities by batch processing.

[0040] In a preferred embodiment, the MEMS transducer has two end electrodes, preferably such that a connection to an electronic system, e.g. a current or voltage source, can be established between the electrodes at either end of the vibratable membrane, between which there are two or more vertical sections, so that the actuator layers of the vertical sections can be driven using the end electrodes.

[0041] Therefore, the provision of end-side electrodes is preferably distinct from the connection means, which actuates or, in the case of a MEMS microphone, receives the electrical signal generated by each vertical section, each with a separate electrode. Thus, the MEMS transducer preferably has just one or just two electrodes for end-side connection and no separate electrode / electrode pad for connecting the central vertical section.

[0042] Preferably, the layer of actuator material in the vertical section acts as a component of a mechanical bimorph, and lateral bending of the vertical section is induced by actuating the actuator layer via electrodes, or a corresponding electrical signal is generated by the induced lateral bending.

[0043] In a preferred embodiment, the two or more vertical sections show at least two layers, one layer having an actuator material and a second layer having a mechanical support material, at least the layer having the actuator material being connected to end electrodes such that a change in shape of the actuator material relative to the mechanical support material can generate horizontal vibration. In one embodiment, a mechanical bimorph is formed by a layer of actuator material (e.g., piezoelectric material) and a passive layer acting as a mechanical support layer. Both lateral and longitudinal piezoelectric effects can be used for bending.

[0044] When the actuator layer is actuated, it can expand or compress, for example, laterally or vertically. This generates a stress gradient on the mechanical support layer, causing lateral bending or vibration. By alternating polarity on the electrodes, as shown in Figure 1, preferably a pushing and pulling action can be obtained, which can cause substantially the entire air volume between the vertical sections to be moved alternately in the vertical ejection direction.

[0045] The advantage of the actuator principle is therefore that it converts horizontal vibration of the vertical section very efficiently into vertical volume movement or sound generation.

[0046] Since the actuator principle is not based on electrostatic attraction but on a relative change in the shape of the actuator layer with respect to the support layer (e.g. compression, extension, shear), it is possible to keep the membrane sections from sticking, instead the vertical sections can be in contact with each other, so that their displacement is not restricted.

[0047] In a further preferred embodiment, the two or more vertical sections have at least two layers, both layers having actuator material and connected to respective end electrodes, and the horizontal vibration can be generated by a change in shape of one layer relative to the other layer. Thus, in one embodiment, the horizontal vibration of the vertical sections is generated by a relative change in shape of the two active actuator layers, rather than by a stress gradient between the active actuator layer and the passive support layer.

[0048] The actuator layers may be made of the same actuator material and driven differently. The actuator layers may also be made of different actuator materials, for example piezoelectric materials with different deformation coefficients.

[0049] In the sense of the present invention, a "layer with actuator material" is preferably also called an actuator layer. By actuator material is preferably meant a material which undergoes a change in shape, for example elongation, compression or shear, when a voltage is applied to it, or vice versa, which generates a voltage when it changes shape.

[0050] Preferred materials are those that have electric dipoles that undergo a change in shape when a voltage is applied, and the orientation of the dipole and / or the electric field can determine the preferred direction of the shape change.

[0051] The actuator material may preferably be a piezoelectric material, a polymeric piezoelectric material, and / or an electroactive polymer (EAP).

[0052] Particularly preferably, the piezoelectric material is selected from the group comprising lead zirconate titanate (PZT), aluminium nitride (AlN), aluminium scandium nitride (AlScN) and zinc oxide (ZnO).

[0053] Polymeric piezoelectric materials preferably include polymers that exhibit internal dipoles, thus giving them piezoelectric properties. This means that when an external voltage is applied, the piezoelectric polymer material changes shape (e.g., compresses, stretches, or shears) (in a manner similar to the typical piezoelectric materials discussed above). An example of a preferred piezoelectric polymer material is polyvinylidene fluoride.

[0054] This allows for a macroscopic solution in which a polymeric piezoelectric material layer is provided on a mechanical support layer and wrapped over the upper and lower combs. Preferably, the polymeric piezoelectric material layer (including electrodes) is first provided on the support layer (possibly including opposing electrodes). Subsequently, the upper and lower combs (preferably of a MEMS structure) are moved in opposite directions to each other so that a folded membrane with actuable vertical sections is formed.

[0055] In the sense of the present invention, the "layer with mechanical support material" is preferably also called the support layer. The mechanical support material or support layer preferably functions as a passive layer that can resist changes in shape of the actuator layer. Unlike the actuator layer, the mechanical support material preferably does not change shape when a voltage is applied. The mechanical support material is preferably conductive and can therefore be used to directly contact the actuator layer. However, in some embodiments, the mechanical support material may also be non-conductive, for example coated with a conductive layer.

[0056] Particularly preferably, the mechanical support material is monocrystalline silicon, polysilicon or doped polysilicon.

[0057] When a voltage is applied, the shape of the actuator layer changes, while the layer of mechanical support material remains substantially unchanged. The resulting stress gradient between the two layers (mechanical bimorph) preferably causes a horizontal bending. For this, the thickness of the support layer compared to the thickness of the actuator layer is preferably selected such that a sufficiently large stress gradient is generated for bending. For example, in the case of doped polysilicon, which is the mechanical support material, and a piezoelectric material such as PZT or AlN, substantially equal thicknesses, preferably between 0.5 μm and 2 μm, have proven to be particularly suitable.

[0058] The terms substantially, approximately, about and the like preferably describe a tolerance range of less than ±20%, preferably less than ±10%, even more preferably less than ±5%, especially less than ±1%. The indications substantially, approximately, about and the like always also disclose and include the exact value referred to.

[0059] In this way, when the actuator layer is periodically driven, for example with an AC voltage, horizontal vibrations can be rapidly and accurately generated to emit sound.

[0060] To ensure horizontal vibration, the piezoelectric material may preferably have a C-axis orientation perpendicular to the surface of the vertical section, thereby utilizing the transverse piezoelectric effect. Other orientations and, for example, the use of the longitudinal piezoelectric effect to create horizontal bending or vibration (see FIG. 1) may also be preferred.

[0061] The electrical connection, i.e. the application of a voltage, of the actuator layer and / or the layer made of mechanical support material can be made directly via the end electrodes or can be assisted by a layer made of a conductive material.

[0062] Therefore, in a preferred embodiment, the vibratable membrane comprises at least one layer of electrically conductive material.

[0063] In a preferred embodiment, the conductive material is selected from the group consisting of platinum, tungsten, (doped) tin oxide, single crystal silicon, polysilicon, molybdenum, titanium, tantalum, titanium-tungsten alloys, metal silicates, aluminum, graphite, and copper.

[0064] The vertical and horizontal (or lateral) direction designations preferably refer to the preferred directions in which the vibratable membrane is oriented to generate or receive pressure waves of a fluid. The vibratable membrane is preferably suspended horizontally between at least two side regions of the carrier, while the vertical direction (direction of interaction with the fluid) for generating or receiving pressure waves is orthogonal to the suspended direction. In the case of a MEMS speaker, the (interacting) vertical direction corresponds to the vertical direction in which the MEMS speaker emits sound. In this case, vertical preferably means the direction of sound emission, while horizontal means the direction orthogonal to the direction of sound emission. In the case of a MEMS microphone, the (interacting) vertical direction corresponds to the vertical direction in which the MEMS microphone detects sound. In this case, vertical preferably means the direction of sound detection or recording, while horizontal means the direction orthogonal to the direction of sound detection or recording.

[0065] Thus, a vertical section of the vibratable membrane preferably denotes a section of the vibratable membrane that is substantially oriented in the emission direction of the MEMS speaker or the detection direction of the MEMS microphone. Those skilled in the art will appreciate that the vertical section of the vibratable membrane does not have to be precisely vertically aligned, but is preferably substantially aligned in the emission direction of the MEMS speaker or the detection direction of the MEMS microphone.

[0066] In a preferred embodiment, the vertical sections are oriented substantially parallel to the vertical direction, where substantially parallel means within a tolerance of ±30°, preferably ±20°, more preferably ±10° about the vertical direction.

[0067] The vibratable membrane can therefore preferably present not only a serpentine shape with a rectangular cross section, but also a curved or undulating shape or a sawtooth (zigzag) shape.

[0068] Preferably, the vertical and / or horizontal sections are at least partially or entirely straight, but the vertical and / or horizontal sections may also be curved, at least partially or entirely. In case the cross-section of the vibratable membrane is curved or undulating, alignment preferably refers to the tangents at the respective midpoints of the curved vertical and / or horizontal sections.

[0069] The vibratable membrane is preferably horizontal to the direction of sound emission or detection, while sound waves are generated or conversely detected by driving the vertical section.

[0070] In a preferred embodiment of the invention, the carrier has two side regions between which the vibratable membrane is horizontally arranged.

[0071] The carrier is preferably a frame structure, which is substantially formed by a continuous outer boundary in the form of a side wall of the area kept free. The frame structure is preferably stable and not easily bent. In the case of an angular frame shape (triangular, quadrilateral, hexagonal or generally polygonal contour), the individual side areas which preferably substantially form the frame structure are specifically called side walls.

[0072] The vibratable membrane is preferably held by at least two side walls of the carrier, two of which can be seen in cross section in the examples of figures 1 to 9.

[0073] Preferably, however, the preferred carrier has four side regions, the further end faces being generally parallel to the cross section depicted, and the further two side walls span the framing structure.

[0074] The vibratable membrane is preferably suspended planarly in the free area. The extension of the plane of the vibratable membrane points in the horizontal direction, while the vertical section is substantially orthogonal thereto. As for the end faces, the membrane may be glued to these side walls or, for greater mobility, may be fitted into slots in the side walls. The slots may advantageously represent, for example, dynamic high-pass filters that couple the front and rear volumes.

[0075] In a preferred embodiment of the present invention, the carrier is formed of a substrate, preferably selected from the group consisting of monocrystalline silicon, polysilicon, silicon dioxide, silicon carbide, silicon germanium, silicon nitride, nitride, germanium, carbon, gallium arsenide, gallium nitride, indium phosphide, and glass.

[0076] These materials are easy and cheap to process and suitable for large-scale production in semiconductor and / or microsystem manufacturing. The carrier structure can be flexibly manufactured based on the material and / or manufacturing method. In particular, it is possible to manufacture the MEMS transducer with the vibratable membrane together with the carrier, preferably in one (semiconductor) process, preferably on a wafer. This further simplifies and cheapens the manufacturing process, so that a small and robust MEMS transducer can be provided at low cost.

[0077] In a preferred embodiment, the vibratable membrane is formed with a lamellar structure (layered structure) or a serpentine structure, preferably the specification of the lamellar or serpentine structure refers to the shape of the cross section of the vibratable membrane.

[0078] A lamellar structure refers to an arrangement of preferably similar parallel layers, preferably forming a vertical section. The individual lamellae are preferably oriented such that the surfaces of the lamellae are substantially parallel in a vertical direction, preferably in the emission or detection direction. The lamellae are preferably stacked in layers to form a mechanical bimorph. For example, each lamella can have an actuator layer and a passive layer made of a support material, and / or two separately controllable actuator layers.

[0079] Those skilled in the art will appreciate that the lamellae do not need to be aligned precisely parallel to the vertical direction, but rather the lamellae are preferably substantially aligned with the emission direction of a MEMS speaker or the detection direction of a MEMS microphone.

[0080] In a preferred embodiment, the vertical sections or lamellae are oriented substantially parallel to the vertical direction, where substantially parallel means within a tolerance of ±30°, preferably ±20°, particularly preferably ±10° around the vertical direction.

[0081] It may be preferred that the lamellae are planar, which in particular means that the extension in each of the two planar dimensions of the lamellae (height, width) is greater than the extension in the dimension perpendicular to the plane of the lamellae (thickness). For example, a size ratio of at least 2:1, preferably at least 5:1, 10:1 or more may be preferred.

[0082] Preferably, the vibratable membrane has a plurality of lamellae forming a vertical section, for example 2, 3, 4, 5, 10, 15, 20, 30, 40, 50 or more lamellae may be preferred, which allows for high efficiency of the desired sound emission or sound detection in a limited space.

[0083] In one embodiment, the vibratable membrane is formed by lamellae, preferably of vertical section, which are connected to each other via conductive bridges or horizontal sections. Suitable bridges are for example metal bridges (see FIG. 10) or bridges made of other conductive materials. The conductive bridges ensure, on the one hand, the mechanical integrity of the vibratable membrane. On the other hand, they advantageously allow the connection of all lamellae by end electrodes. The lamellae can thus be driven synchronously and generate horizontal vibrations or detect horizontal vibrations, advantageously with minimal complexity in the control and manufacture of the lamellae.

[0084] The serpentine structure preferably refers to a structure formed by a series of sections, which are substantially mutually orthogonal in cross section. The mutually orthogonal sections are preferably vertical and horizontal sections of the vibratable membrane. The serpentine structure is particularly preferably rectangular in cross section. However, it may also be preferred that the serpentine structure is sawtooth (zigzag) or curved or undulating in cross section. This is particularly true when the vertical sections are not aligned exactly parallel to the vertical emission or detection direction, but for example make an angle of ±30°, preferably ±20°, particularly preferably ±10° with the vertical direction.

[0085] In preferred embodiments, the horizontal section may also not be exactly 90° orthogonal to the vertical emission or detection direction, but may for example form an angle between 60° and 120°, preferably between 70° and 110°, and particularly preferably between 80° and 100°, including the vertical direction.

[0086] In the case where the cross-section of the vertical and / or horizontal sections of the vibratable membrane is curved or undulating in shape, the alignment preferably refers to a tangent at the midpoint of the vertical and / or horizontal sections, respectively.

[0087] The serpentine structure therefore preferably corresponds to a membrane folded along its width. In the sense of the present invention, the vibratable membrane may therefore preferably also be called a bellows. The parallel folds of the bellows preferably form vertical sections. The connection sections between the folds preferably form horizontal sections. The vertical sections are preferably, for example, 1.5, 2, 3, 4 or more times longer than the horizontal sections.

[0088] With regard to the function of the vibratable membrane in serpentine form to generate or receive sound waves, the vertical sections are decisive in a similar manner as the lamellae described above. The vertical sections are preferably stacked in layers to form a mechanical bimorph. For example, each vertical section may have an actuator layer as well as a passive layer made of a support material and / or two separately controllable actuator layers. Preferably, the horizontal sections of the folded membrane may be constructed identically to the vertical sections (see in particular Figures 3 to 7). However, it may likewise be preferred that the horizontal sections, in contrast to the vertical sections, do not present actuator layers but only mechanical support layers and / or conductive layers.

[0089] In a preferred embodiment, at least one layer of actuator material of the vibratable membrane is a continuous layer. Continuous preferably means that there is no break in the cross-sectional profile. Thus, in this embodiment, there is preferably a continuous layer of actuator material in both the vertical and horizontal sections. Advantageously, no structuring is necessary. A continuous layer is particularly easy to manufacture and ensures synchronous actuation during operation of the MEMS speaker.

[0090] The performance of a MEMS transducer, in particular a MEMS speaker or a MEMS microphone, may be significantly determined by the number and / or size of the vertical sections.

[0091] In preferred embodiments, the vibratable membrane has 3, 4, 5, 10, 15, 20, 30, 40, 50, 100 or more vertical sections.

[0092] In preferred embodiments, the vibratable membrane has a vertical section of 10000, 5000, 2000 or 1000 or less.

[0093] The preferred number of vertical sections provides high sound output with minimal chip surface area without sacrificing voice or audio quality.

[0094] The vertical section is preferably planar, which in particular means that the extension in each of the two plane dimensions of the lamella (height, width) is greater than the extension in the dimension perpendicular to the plane of the lamella (thickness). For example, a size ratio of at least 2:1, preferably at least 5:1, 10:1 or even more may be preferred.

[0095] In the sense of the present invention, the height of a vertical section preferably corresponds to the dimension along the sound emission or sound detection direction, while the thickness of a vertical section preferably corresponds to the sum of the layer thicknesses of the layer or layers forming the vertical section. The length of a vertical section preferably corresponds to the dimension perpendicular to the height or thickness. In the cross-sectional views of the figures below, the height and thickness are shown diagrammatically (not necessarily to scale), while the length dimension corresponds to the depth of the (not visible) drawing of the figures.

[0096] In a preferred embodiment, the height of the vertical section is between 1 μm and 1000 μm, preferably between 10 μm and 500 μm. Intermediate ranges from the above ranges may also be preferred, such as 1 μm to 10 μm, 10 μm to 50 μm, 50 μm to 100 μm, 100 μm to 200 μm, 200 μm to 300 μm, 300 μm to 400 μm, 400 μm to 500 μm, 600 μm to 700 μm, 700 μm to 800 μm, 800 μm to 900 μm, or even 900 μm to 1000 μm. Those skilled in the art will recognize that the limits of the aforementioned ranges may be combined to obtain other preferred ranges, such as 10 μm to 200 μm, 50 μm to 300 μm, or even 100 μm to 600 μm.

[0097] In a preferred embodiment, the thickness of the vertical section is between 100 nm and 10 μm, preferably between 500 nm and 5 μm. Intermediate ranges from the above ranges may also be preferred, such as 100 nm to 500 nm, 500 nm to 1 μm, 1 μm to 1.5 μm, 1.5 μm to 2 μm, 2 μm to 3 μm, 3 μm to 4 μm, 4 μm to 5 μm, 5 μm to 6 μm, 6 μm to 7 μm, 7 μm to 8 μm, 8 μm to 9 μm, or even 9 μm to 10 μm. Those skilled in the art will recognize that the limits of the aforementioned ranges may be combined to obtain other preferred ranges, such as 500 nm to 3 μm, 1 μm to 5 μm, or even 1500 nm to 6 μm.

[0098] In a preferred embodiment, the length of the vertical section is between 10 μm and 10 mm, preferably between 100 μm and 1 mm. Intermediate ranges from the aforementioned ranges may also be preferred, such as 10 μm to 100 μm, 100 μm to 200 μm, 200 μm to 300 μm, 300 μm to 400 μm, 400 μm to 500 μm, 500 μm to 1000 μm, 1 mm to 2 mm, 3 mm to 4 mm, 4 mm to 5 mm, 5 mm to 8 mm, or even 8 mm to 10 mm. Those skilled in the art will recognize that the limits of the aforementioned ranges may also be combined to obtain other preferred ranges, such as 10 μm to 500 μm, 500 μm to 5 μm, or even 1 mm to 5 mm.

[0099] The above-mentioned preferred dimensioning of the vibratable membrane and the vertical section makes it possible to provide a particularly small MEMS transducer, in particular a MEMS speaker or a MEMS microphone, which combines high performance together with excellent sound or audio quality.

[0100] In a preferred embodiment of the invention, the vibratable membrane is formed by a serpentine structure with alternating vertical and horizontal sections, and a retaining structure is attached to at least two horizontal sections, said retaining structure being directly or indirectly connected to the carrier. For example, the retaining structure may be provided in the substrate material of the carrier, i.e., the retaining structure may be formed directly from the substrate of the bottom wafer. Alternatively, the retaining structure may be connected to the horizontal sections as a separate ridge or bump of the top wafer.

[0101] The retaining structures may preferably be present on one and / or both sides of the vibratable membrane, i.e. preferably on the upper horizontal section and / or on the lower horizontal section.

[0102] In particular, when suspending a larger vibratable membrane between the side walls of a carrier, the use of a support structure is advantageous as it allows stabilization without adversely affecting sound generation or sound capture.

[0103] Since the serpentine horizontal section is at least substantially mechanically neutral, anchoring the horizontal section using a retaining structure is advantageous as it does not cause any undesirable stresses between the membrane and the retaining structure or carrier.

[0104] To ensure the described driving and inducing of horizontal vibrations or detection of horizontal vibrations, various layers may be provided in the structure of the vibratable membrane.

[0105] The connection, i.e. application or detection of a voltage, of one or more actuator layers and / or one or more layers of mechanical support material can be achieved directly via end electrodes or can be assisted by a layer of conductive material.

[0106] Therefore, in a preferred embodiment, the vibratable membrane comprises at least one layer of electrically conductive material.

[0107] In a preferred embodiment, the conductive material is selected from the group consisting of platinum, tungsten, (doped) tin oxide, single crystal silicon, polysilicon, molybdenum, titanium, tantalum, titanium-tungsten alloys, metal silicates, aluminum, graphite, and copper.

[0108] In a preferred embodiment, the vibratable membrane has three layers: an upper layer made of a conductive material and connected to the upper electrode, a middle layer made of an actuator material, and a lower layer made of a conductive material.

[0109] Preferably, the conductive material of the upper and / or lower layer may be a mechanical support material, such that this layer has a dual function: on the one hand, it ensures the connection of the actuator layer to the potential that can be applied to the end electrodes, and on the other hand, it acts as a mechanical support layer in the manner described when the actuator layer is accordingly driven to generate a horizontal bending or vibration.

[0110] Such an embodiment can be realized by a simple manufacturing process, as shown by way of example in Figure 2. In a preferred embodiment shown in Figures 2G, 3 and 4, the vibratable membrane is a serpentine structure, comprising a continuous top layer of conductive material (metal), a continuous middle layer of actuator material, and a bottom layer of conductive mechanical support material. To improve the connection, it may also be possible to reverse the order of the layers or to add another conductive layer in contact with the mechanical support layer and / or the actuator layer.

[0111] In another preferred embodiment, the vibratable membrane has two layers of actuator material separated by an intermediate layer of conductive material, preferably metal, the intermediate layer being connected to a first electrode and at least one of the two layers of actuator material being connected to a second electrode via another layer of conductive material, preferably metal.

[0112] As explained above, in a preferred embodiment, two actuator layers can also be used to move the vertical section in horizontal vibration, for example with different drives. To transfer the change in potential from the end electrodes to the respective actuator layers, preferably two or more intermediate layers of conductive material can be provided. Preferably, the layer of conductive material, for example metal, in this case preferably serves only for connection and not as mechanical support layer. In the sense of a bimorph for a MEMs speaker, the stresses required to induce bending or vibration are themselves induced by the different control of the actuator layers themselves.

[0113] Therefore, preferably the layer of conductive material, such as metal, can be particularly thin (less than 500 nm, preferably less than 200 nm).

[0114] Figure 5 shows an example of such a preferred embodiment. It has a vibratable membrane that is a serpentine structure with two layers of actuator material separated by an intermediate layer of conductive material (metal). The intermediate layer is connected to a first end electrode pad, while the upper actuator layer is connected to a second end electrode pad via another layer of conductive material. The lower layer of conductive material is not connected to any of the electrodes. It may also be possible to reverse the layer order or omit the lower layer of conductive material that is not in contact with the electrodes.

[0115] In the embodiments described above, the actuator layer, and if necessary the mechanical support layer, are preferably continuous, i.e. extending in cross section from one end of the membrane (preferably where the first electrode is present) over several alternating horizontal and vertical sections to the second end of the membrane (preferably where the second electrode is present).

[0116] It has been realised by the inventors that for the operating principle of a MEMS transducer, preferably a MEMS speaker, it is sufficient to provide a mechanical bimorph in the vertical section.

[0117] In a preferred embodiment, at least one actuator layer is not continuous and is present only in the vertical sections and not in the horizontal sections. In this case, it may be preferred that the mechanical support layer, if present, extends continuously or that it does not extend continuously and is, for example, only provided in the vertical sections. It is preferred to use one or more continuous layers of a conductive material, preferably a metal, in order to be able to connect the vertical sections with end electrodes.

[0118] A preferred fabrication process for an embodiment with a non-continuous actuator layer is shown in Figure 7. Here, a selective spacer etch of the actuator layer can be performed on the horizontal sections, so that only the vertical sections of the membrane have a layer of actuator material. The continuous layers of mechanical support material can both be dielectric to avoid shorts between the upper and lower conductive layers (also called the top and bottom electrodes).

[0119] This embodiment features a particularly efficient drive and high performance, in that only the vertical section is induced to alternately bend or vibrate, while the horizontal section is kept mechanically neutral. Advantageously, the amount of displacement can be increased even further with each step of the drive.

[0120] In the above embodiment, the serpentine morphology of the vibratable membrane is preferably realized by suitable deposition or etching of functional layers.

[0121] Alternatively, the vibratable membrane can also be manufactured by providing vertical sections and connecting the vertical sections using a metal bridge.

[0122] In a preferred embodiment, the vertical section of the vibratable membrane has two layers, a first layer made of actuator material and a second layer made of flexible support material, the vertical sections being connected via a horizontal metal bridge.

[0123] Several individual piezoelectric ceramic elements, having layers of mechanical support material and layers of piezoelectric material as well as sacrificial layers, may preferably be produced for this purpose, as shown in Figure 10. By multiple processing steps, including interlayer connections and metal filling, as well as stacking and dicing of the piezoelectric ceramic elements, a highly efficient membrane can be advantageously achieved in a robust and process-efficient manner.

[0124] In this embodiment, a continuous, homogenous conductive layer is not necessary: ​​instead, the connection of the actuator layers in the vertical sections is ensured by metal bridges and conductive mechanical support material.

[0125] In a preferred embodiment, the vibrable membrane is coated with a layer of non-stick material. By non-stick (or anti-adhesive) material is meant in particular a material with low surface energy that is almost inert to the environment and thus prevents the accumulation of dust or other undesirable particles. The non-stick material can illustratively be formed by a carbon layer, for example a diamond-like carbon (DLC) layer, or also by a layer with a perfluorocarbon (PFC), such as polytetrafluoroethylene (PTFT).

[0126] In a preferred embodiment of the invention, the MEMS transducer, preferably a MEMS speaker, comprises a control unit configured to drive at least one electrode such that two or more vertical sections are induced to generate horizontal vibrations. The control unit is preferably configured to drive the electrodes to ensure that the frequency of the horizontal vibration is between 10 Hz and 20 kHz.

[0127] In a preferred embodiment of the invention, the MEMS transducer, preferably a MEMS microphone, comprises a control unit configured to detect an electrical signal provided by at least one electrode, the electrical signal being generated by horizontal vibration of two or more vertical sections. Preferably, the control unit of the MEMS microphone is configured to receive and process an electrical signal corresponding to a frequency of the horizontal vibration between 10 Hz and 20 kHz, and is thus adapted for sound detection in the audible range.

[0128] The control unit is therefore preferably configured and adapted to drive the vibratable membrane (or the actuator layer of the vertical section) with an electrical signal to generate horizontal vibrations and emit sounds in the audible frequency range, or to receive and process corresponding electrical signals when the vibratable membrane is driven.

[0129] Preferably, in the case of a MEMS speaker, the vertical section of the membrane is directly driven by an audio signal, thus greatly simplifying the actuation to generate sound, as opposed to the combination of separate membrane units and actuation of multiple valves according to US 2019 / 0116417 A1.

[0130] The control unit may preferably comprise a data processing unit for generating or receiving the electrical signals.

[0131] In the sense of the present invention, a data processing unit preferably refers to a unit adapted and configured to receive, transmit, store and / or process data relating to the driving of the electrodes or the reception of electrical signals generated by the electrodes. The data processing unit preferably also comprises an integrated circuit, e.g. an application specific integrated circuit, a processor, a processor chip, a microprocessor or a microcontroller for processing data, and optionally a data memory, a random access memory (RAM), a read-only memory (ROM) or a flash memory for storing data.

[0132] In a preferred embodiment, the control unit is integrated together with the other components of the MEMS transducer (carrier, vibratable membrane) on a printed circuit board or circuit board. This preferably means that the MEMS transducer and the electronic system required for driving or sensing are seamlessly integrated. In addition to the control unit, other electronic components such as communication interfaces (preferably wireless, e.g. Bluetooth), amplifiers, filters or sensor systems may also be mounted on one and the same printed circuit board.

[0133] Advantageously, a compact integrated solution is achieved in which a MEMS transducer, preferably a MEMS speaker or a MEMS microphone, can be produced integrally with a desired electronic system, in a limited space, preferably in a low-cost CMOS process suitable for mass production.

[0134] In a further preferred embodiment, the vibratable membrane carried by the carrier is placed at the front of the housing surrounding a rear resonant volume, so that the sound emission of such a MEMS speaker is preferably towards the open front side (sound port), whereby the sound is improved by the rear resonant volume, especially for the lower frequencies.

[0135] In a further preferred embodiment, there is a vent in the housing to prevent acoustic shorts and / or to assist with audio. The vent is preferably small compared to the audio port, for example less than 100 μm, preferably less than 50 μm in maximum dimension.

[0136] In another aspect, the present invention relates to a method for manufacturing the above-mentioned MEMS transducer, preferably a MEMS speaker or a MEMS microphone, comprising the steps of: - etching the substrate, preferably from the front side, to form structures, preferably serpentine structures; - optionally depositing an etch stop - depositing at least two layers, where at least a first layer comprises an actuator material and a second layer comprises a mechanical support material, or where at least two layers comprise an actuator material. - connecting the first layer and / or the second layer to an electrode. - etching, preferably from the rear face, and optionally removing the etch stop, Thereby, a vibratable membrane, preferably in the form of a serpentine structure, is held by a carrier (4) formed by a substrate (8), the vibratable membrane (1) having at least two or more vertical sections (2) for generating or receiving vertical fluid pressure waves, the vertical sections being formed parallel in the vertical direction, so that the two or more vertical sections can be induced to generate horizontal vibrations by driving at least one electrode, or As a result, when two or more vertical sections are induced to vibrate horizontally, an electrical signal can be generated in at least one electrode.

[0137] A person skilled in the art will recognize that the technical features, definitions and advantages of the described preferred embodiment of the MEMS transducer, preferably a MEMS speaker or a MEMS microphone, also apply to the described manufacturing process and vice versa. Preferably, the described manufacturing method serves to produce a MEMS transducer with a folded vibratable membrane in a serpentine structure. An example of a preferred manufacturing process is illustrated in Figures 2A-2G, 8A-8J or 9.

[0138] For example, one of the preferred materials mentioned above can be used as the substrate. During etching, a blank, for example a wafer, can be formed into the desired basic shape of the serpentine structure. In a next step, the layers for the vibrable membrane are preferably deposited.

[0139] The step of depositing at least one layer of electrically conductive material preferably comprises, in addition to depositing one layer, depositing several layers, in particular a layer system, which comprises at least two layers deposited in a planned manner relative to each other. The deposition of the layers or layer system preferably serves to define a vibratable membrane with a vertical section, which can be induced to generate horizontal vibrations.

[0140] Preferably, the deposition may be selected from the group comprising physical vapor deposition (PVD), in particular thermal evaporation, laser beam evaporation, arc evaporation, molecular beam epitaxy, sputtering, chemical vapor deposition (CVD) and / or atomic layer deposition (ALD). The deposition may include a plating process, in particular, for example in the case of a substrate made of polysilicon.

[0141] The etching and / or structuring may preferably be selected from the group comprising dry etching, wet chemical etching and / or plasma etching, in particular reactive ion etching, deep reactive ion deep etching (Bosch process).

[0142] In a preferred embodiment, the etching of the substrate, preferably from the front side, to form the structuring is characterized in that the substrate has a crystalline structure and that a plurality of trenches are produced by etching along the lattice vectors of the crystalline structure. The trenches are preferably defined as parallel slits from the front side of the substrate. After deposition of a functional layer and appropriate rear surface treatment, the vibratable membrane is formed as a bellows with a serpentine cross section (see in particular figures 2 and 8).

[0143] The preferred etching along the orientation of the crystalline substrate can advantageously result in smooth, quasi-crystalline trenches with high precision orientation and negligible roughness at substantial depths of 200 μm, 300 μm, 400 μm, 500 μm or more.

[0144] It is also advantageous for the surface normals of the trench sides to be aligned with the crystal structure, preferably with orthogonal lattice vectors.

[0145] When a layer of actuator material, preferably a piezoelectric material, is applied to such a structured substrate, the orientation of the actuator material may also be quasicrystalline. In particular, piezoelectric materials such as AlN, AlScN or PZT advantageously exhibit columnar growth on the sidewalls of trenches oriented in this way, which ensures that the piezoelectric layer has a particularly precise c-axis orientation perpendicular to the surface of the vertical sections of the resulting film.

[0146] Thus, creating horizontal vibrations through the lateral piezoelectric effect can be particularly effective and accurate, and can provide improved sound in the case of a MEMS speaker, or detection capabilities in the case of a MEMS microphone.

[0147] In a preferred embodiment, the etching of the substrate, preferably from the front side, to form a structuring, preferably a serpentine structure, is characterized in that the substrate has a crystalline structure and that a plurality of trenches aligned with lattice vectors is realized at least partially by wet-chemical etching, preferably an anisotropic etching dependent on the crystal orientation is performed.

[0148] Preferably, for this purpose, an etching agent is used that has significantly different etching rates for the two orthogonal crystallographic orientations of the substrate, for example, the etching rate may be 50, 10, 150, 200 or more times faster in a first crystallographic orientation of the selected substrate than in a second crystallographic orientation that is orthogonal to the first crystallographic orientation.

[0149] The substrate is preferably oriented such that the first crystal orientation in which the etching rate increases is aligned with the surface normal of the substrate surface. An etch mask can be used to define areas on the substrate surface that are not to be etched. The etch mask can preferably define a frame in which slots or strips are left free for forming the trenches. The areas remaining between the parallel trenches to be formed can serve as a substrate for the horizontal sections of the membrane.

[0150] The anisotropic wet chemical etch is followed by a selective etch perpendicular to the substrate surface to form deep vertical trenches. Thus, etching in the orthogonal (horizontal) direction is reduced. The larger the anisotropy factor of the etch, which depends on the crystal orientation, the less noticeable the undercut will be.

[0151] For example, particularly good results may be obtained using potassium hydroxide (KOH) as an etchant for silicon crystalline substrates. <110> Regarding directional etching, <111> They show a clear directional preference for orientation, as shown by Sato et al. (1988) <110> The etching rate of KOH for a silicon single crystal in the direction can be 1.455 μm / min, and the perpendicular <111> This is 291 times faster than the azimuth direction (etching rate is 0.005 μm / min).

[0152] FIG. 9 shows how proper alignment of silicon crystals can reliably produce nearly perfectly smooth, deep trenches with crystal-oriented sidewalls to ensure the growth of c-axis oriented piezoelectric material.

[0153] Those skilled in the art will appreciate that alternative crystal orientation dependent etchants such as tetramethylammonium hydroxide (TMAH) can be used as well (see, for example, Seidel et al., 1990).

[0154] Advantageously, this process is not only suitable for scale-up for mass production, but in addition the serpentine-shaped vibratable membranes that can be produced in this way are also characterized by a particularly precise alignment of the vertical sections, which results in an improved vibration behavior, i.e. sound production or detection.

[0155] If further structuring of the vibratable membrane is desired, this can be carried out, for example, by further etching processes. Likewise, further material can be deposited or doping can be carried out by conventional processing.

[0156] To connect the layers, suitable materials such as copper, gold and / or platinum may be further deposited by conventional processes. Preferably, physical vapor deposition (PVD), chemical vapor deposition (CVD) or electrochemical deposition may be used for this purpose.

[0157] These process steps can be used to produce finely structured vibratable membranes according to the desired definition of vertical and horizontal sections, preferably suspended between two lateral regions of a stable carrier, with dimensions in the micrometer range. The manufacturing steps are part of the standard process steps of semiconductor processing and are therefore proven and suitable for mass production.

[0158] Therefore, in a further aspect, the present invention also relates to a MEMS transducer producible by the above manufacturing process.

[0159] Those skilled in the art will recognize that special features of the manufacturing process, such as the crystal orientation dependent etching to form deep trenches with quasi-crystalline smooth surfaces, are directly transferred to the structural features of the MEMS transducer. If the trench sidewalls are quasi-crystalline smooth surfaces, a vibratable membrane with multiple vertical sections of serpentine shape, as explained above, can be formed in a particularly precise manner. Also, the c-axis orientation of the actuator material, preferably the piezoelectric material, can be directly followed by using the preferred manufacturing process.

[0160] In another aspect, the present invention relates to a method for manufacturing the above-mentioned MEMS transducer, the method comprising the steps of: - providing a plurality of individual piezoelectric ceramic elements having a sacrificial layer, a layer of conductive material, and a layer of piezoelectric material; - defining holes for interlayer connections and metal filling of the piezoelectric ceramic elements; - stacking and optionally dicing the piezoelectric ceramic elements so as to obtain a stack of piezoelectric ceramic elements connected by metal bridges. - removing the sacrificial layer and inserting the stack of piezoelectric ceramic elements into a carrier, whereby each piezoelectric ceramic element is connected to one electrode, Thereby, a vibratable membrane, preferably in the form of a lamellar structure, is held on a carrier formed by a substrate, the vibratable membrane having at least two or more vertical sections generating or receiving vertical fluid pressure waves, the vertical sections being formed parallel in the vertical direction, so that the two or more vertical sections can be induced to generate horizontal vibrations by driving at least one electrode, or so that when the two or more vertical sections are induced to vibrate in the horizontal direction, an electrical signal can be generated in at least one electrode.

[0161] A person skilled in the art will recognize that the technical features, definitions and advantages of the described preferred embodiment of the MEMS transducer, preferably a MEMS speaker or a MEMS microphone, also apply to the described manufacturing process and vice versa. The described manufacturing method is preferably for producing a MEMS transducer having a vibratable membrane with a lamella structure, the lamella being a mechanical bimorph and connected by a metallic bridge. An example of the preferred manufacturing process is shown in Figures 10A-10F and 11.

[0162] In an alternative manufacturing process, several individual piezoelectric ceramic elements can be advantageously used to obtain, by hole definition, metal filling, and stacking and dicing, a vibratable membrane with lamellae that are vertical sections connected using metal bridges.

[0163] Piezoelectric ceramics are preferably ceramic materials that exhibit separation of charge under the influence of deformation by an external force or undergo a shape change when a voltage is applied. The piezoelectric ceramic element preferably comprises not only a piezoelectric layer but also a layer of the mechanical support material described above, as well as a sacrificial layer.

[0164] The sacrificial layer is used to process and create the metal bridge, and the sacrificial layer itself will not become part of the vibratable membrane.

[0165] The sacrificial layer can preferably be, for example, a photoresist. Such a material changes its solubility when irradiated with light, in particular with UV light. In particular, such a material can be a so-called positive resist, the solubility of which increases as a result of UV irradiation. This allows the targeted removal of the sacrificial layer after metal filling to generate metal bridges.

[0166] In another aspect, the present invention relates to a method for manufacturing a MEMS transducer, the method comprising the steps of: - providing a plurality of individual piezoelectric ceramic elements having a layer of mechanical support material that is electrically conductive and a layer of piezoelectric material; - providing upper and lower frames with recesses for a plurality of individual piezoelectric ceramic elements; - fixing the piezoelectric ceramic elements in the recesses of the upper and lower frames, preferably with an adhesive; - depositing at least one continuous conductive layer for connecting the piezoelectric ceramic elements with at least one electrode, Thereby, a vibratable membrane, preferably in the form of a lamella structure, is held in a carrier formed by upper and lower frames and generates or receives vertical fluid pressure waves, the vibratable membrane having at least two or more vertical sections formed parallel in the vertical direction, so that the two or more vertical sections can be induced to generate horizontal vibrations by driving at least one electrode, or so that when the two or more vertical sections are induced to vibrate horizontally, an electrical signal can be generated in at least one electrode.

[0167] A preferred embodiment is shown in Figure 12. Advantageously, in this embodiment, no structured connection is required: instead, the connection is made using a continuous conductive surface from the front and / or rear of the MEMS transducer.

[0168] In a preferred embodiment, the upper and lower frames are formed from a non-conductive material, such as a polymer. Preferably, a 3D printing process can be used to form the frames.

[0169] To connect the individual lamellae or piezoelectric ceramic elements, a continuous layer of conductive material, preferably a metal, is preferably deposited from the front (front electrode) or from the rear (back electrode). The deposition can be carried out, for example, by means of a sputtering process.

[0170] A person skilled in the art will recognize that the technical features, definitions and advantages of the described preferred embodiments of a MEMS transducer, preferably a MEMS speaker or a MEMS microphone, also apply to the described manufacturing process and vice versa. Preferably, the described manufacturing method serves to produce a MEMS transducer having a vibratable membrane with a lamella structure, the lamella being a mechanical bimorph and connected by a layer of a continuous conductive material, preferably a metal.

[0171] The invention will be further explained below with reference to figures and examples which serve to illustrate preferred embodiments of the invention and are not intended to limit the same. [Brief description of the drawings]

[0172] [Figure 1] 1A is a cross-sectional view of a preferred embodiment of a MEMS speaker according to the present invention, in an idle state (A) and in an active state (B). [Diagram 2] 1A-1D are diagrams of a preferred method of manufacturing a MEMS speaker with a vibratable membrane exhibiting a serpentine shape in cross section. [Diagram 3] 1 is a diagram of a preferred embodiment of a MEMS speaker with a serpentine-shaped vibratable membrane, the horizontal section of which is supported by a retaining structure. [Figure 4] 1 is a diagram of a preferred embodiment of a MEMS speaker with a serpentine-shaped vibratable membrane, the horizontal section of which is supported by a retaining structure. [Diagram 5] FIG. 2 is a diagram of a preferred embodiment of a MEMS speaker comprising two actuator layers separated by an intermediate layer made of a conductive material. [Figure 6] FIG. 1 is a diagram of a preferred drive system for operating a MEMS speaker. [Figure 7] A MEMS speaker is preferably integrated into the front of the housing with a resonant volume at the rear. [Figure 8] 1 is a diagram of a preferred method of manufacturing a MEMS speaker comprising a vibratable membrane with a serpentine-shaped cross section, only the vertical sections of which carry a layer of actuator material. [Figure 9] FIG. 1 shows a preferred structuring of a crystalline morphology substrate for forming deep trenches using a crystal orientation dependent etching process. [Figure 10] FIG. 1 illustrates a preferred method of manufacturing a MEMS speaker with a vibratable membrane based on individual piezoelectric ceramic elements. [Figure 11] FIG. 1 is a diagram of preferred electrical connections for a MEMS speaker with a vibratable membrane based on individual piezoelectric ceramics. [Figure 12] FIG. 1 illustrates a preferred method of manufacturing a MEMS speaker with a vibratable membrane based on individual piezoelectric ceramic elements. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0173] Figure 1 shows a preferred embodiment of a MEMS speaker according to the invention, Figure 1A shows the idle state, while Figure 1B shows two phases during driving of the MEMS speaker.

[0174] The MEMS speaker comprises a vibratable membrane 1, which generates sound waves in a vertical emission direction, and which is held in a horizontal position by a carrier 4. In cross section, the vibratable membrane 1 has a serpentine structure with a horizontal section 3 and a vertical section 2. The vertical section is formed parallel to the emission direction and presents at least one actuator layer, for example a layer made of piezoelectric material. The connection between the vibratable membrane 1 and the actuator layer is preferably realised by means of electrodes at both ends. For this purpose, for example electrode pads (not shown) can be arranged on the carrier 4.

[0175] The vertical section is preferably a mechanical bimorph that can be induced to generate horizontal vibrations as a result of a suitable drive. The vertical section 2 can have for this purpose, for example, a first layer of actuator material and a second layer of mechanical support material. By driving the actuator layer, a stress gradient can be generated, which can result in curvature or vibration. It may also be preferred that the vertical section 2 has two actuator layers that are driven in opposite directions to cause curvature of the vertical section 2 as a result of a relative change in the corresponding shape.

[0176] Fig. 1B shows, by way of example, two stages during actuation. Advantageously, thanks to the multiple vertical sections 2 of the vibratable membrane 1, a small horizontal movement (bending) of a few micrometers allows the entire increased volume to be moved in the vertical emission direction and thus used for sound production. By actuation, during one stage, almost the entire air volume between the vertical sections can be moved up or down along the emission direction, allowing here a particularly efficient implementation.

[0177] 2 shows diagrammatically a preferred manufacturing method for producing a MEMS loudspeaker comprising a vibratable membrane 1 with a serpentine cross section, which may preferably also be called a folded membrane or bellows.

[0178] 2A shows the etching of a substrate 8 from the top or front side to form a structuring. In the process, parallel deep trenches are etched into the substrate 8. The structures formed represent bellows, or meanders in cross section.

[0179] Subsequently, a layer of etch stop 9 (FIG. 2B) is deposited, which may be, for example, TEOS or PECVD. A layer of mechanical support material 10 (FIG. 2C) and a layer of actuator material 11 are deposited on the etch stop 9. The mechanical support material 10 may be, for example, doped polysilicon, while a piezoelectric material may be used for the actuator material 11. A layer thickness of, for example, 1 μm may be preferred. The piezoelectric material may preferably have a C-axis orientation perpendicular to the surface, so that the transverse piezoelectric effect is used. Other orientations and, for example, the use of a longitudinal effect may be preferred.

[0180] Figure 2E shows the preferred deposition of a top electrode on the entire surface, which is a layer of conductive material 12. End side connections can be achieved, for example, using electrode pads 13 (Figure 2F).

[0181] 2F and 2G show etching of another substrate 8 from the back and bottom surfaces, respectively, and removal of the etch stop.

[0182] Thus, the manufacturing steps 2A-2G result in a vibratable membrane 1, the cross section of which exhibits a serpentine structure. Advantageously, by providing a continuous actuator layer 11 and end side connections 13, the vertical section 2 can be efficiently actuated to generate horizontal vibrations (see FIG. 1). As can be seen in FIG. 2G, actuation is preferably realized by means of two electrodes, such actuator layer 12 being preferably connected both from the front side (top electrode, conductive layer 12) and from the rear side (bottom electrode, via conductive mechanical support material 10) (see FIG. 6A).

[0183] A retaining structure 14 may be provided to stabilize the membrane 1 suspended between the side walls of the carrier 4. As shown in Figures 3 and 4, the retaining structure may preferably support a horizontal section 3 of the vibratable membrane 1. Advantageously, the horizontal section 3 is mechanically neutral (see Figure 1B) and therefore does not induce undesirable stresses between the membrane 1 and the retaining structure 14 or the carrier 4 during actuation.

[0184] 5 shows an alternative preferred embodiment of a MEMS speaker in which a vibratable membrane 1 has two actuator layers separated by an intermediate layer of conductive material 12, preferably metal. The intermediate layer is connected to a first end electrode pad 13, while in the embodiment shown the upper actuator layer 11 is connected to a second end electrode pad 13 via another layer of conductive material 12.

[0185] FIG. 6 shows a preferred drive system for operating the described MEMS speaker.

[0186] Fig. 6A shows a preferred drive system of the MEMS speaker, with an actuator layer 11 and a passive mechanical support layer 10. Drive is preferably performed using two end electrode pads 13, so that horizontal vibrations can be generated by a change in shape of the actuator material relative to the mechanical support material. The actuator layer 11 is preferably connected from both the front (top electrode 13, conductive layer 10) and rear (bottom electrode 13, conductive mechanical support material 10). For example, an AC voltage, which is an audio input signal, can be applied to the front electrode pad 13 (left), while the rear electrode pad 13 (right) is grounded.

[0187] FIG. 6B shows a preferred drive system for a MEMS speaker comprising two actuator layers 11 separated by an intermediate layer of conductive material 12, preferably metal.

[0188] The upper actuator layer 11 is preferably driven by the front surface (top electrode 13 and upper conductive layer 12) and the middle conductive layer 12. The lower actuator layer 11 is preferably driven by the rear surface (bottom electrode 13 and lower conductive layer 12) and the middle conductive layer 12. In the illustrated embodiment, an AC voltage, e.g. an audio input signal, may be applied to the top and bottom used electrode pads 13 (left), while the middle layer 12 is grounded via another electrode pad 13 (right).

[0189] 7 shows an example of a preferred integration of a MEMS speaker according to the invention in an enclosure 15. Preferably, the vibratable membrane 1 held by a carrier 4 is placed at the front (sound port) of the enclosure. The enclosure also encloses a rear resonant volume (back volume 16). Vents 17 may be provided to prevent acoustic short circuits or to aid sound.

[0190] 8 shows an alternative manufacturing method for providing a MEMS loudspeaker with a vibratable membrane 1 according to the invention. The process steps shown in FIGS. 8A to 8D are similar to those in FIG.

[0191] 8A shows the etching of a substrate 8 from the top or front side to form a structured, preferably serpentine, structure. In this process step, deep parallel trenches are etched into the substrate 8. The structures formed represent bellows, or in cross section, a serpentine.

[0192] Subsequently, a layer of etch stop 9 (FIG. 2B), which can be, for example, TEOS or PECVD, is deposited. A layer of mechanical support material 10 (FIG. 2C) and actuator material 11 are deposited on the etch stop 9. The mechanical support material 10 can be, for example, doped polysilicon, while preferably a piezoelectric material is used for the actuator material 12.

[0193] In contrast to the embodiment shown in Figure 2, the actuator layer 11 is not connected to the upper conductive layer as a continuous layer. Instead, a spacer etch of the actuator layer 11 (Figure 8F) is performed on the horizontal sections of the membrane, so that only the vertical sections of the membrane still have a layer of actuator material 11.

[0194] A continuous dielectric layer 18 is then preferably deposited to prevent shorts between the top and bottom electrodes to be deposited later (FIG. 8G). A continuous conductive layer 12, the top electrode, allows for a front connection (FIG. 8H).

[0195] 8I and 8J show etching of another substrate 8 from the back or underside and, optionally, depositing a continuous conductive layer 12, which is the back electrode.

[0196] Figure 9 shows a preferred way of producing a structured substrate 8. In a similar way to the process step shown in Figure 8a, parallel deep trenches are etched into the substrate 8. The structures formed represent bellows, or meanders in cross section, onto which a vibratable membrane can be deposited in a meandering manner.

[0197] A preferred production of the structured substrate 8 of FIG. 9 is characterized by taking advantage of the crystal structure of the substrate 8, with trenches formed along the lattice vectors of the crystal structure.

[0198] In this way, particularly smooth quasicrystalline trenches of considerable depth, up to 200 μm, 400 μm or more, with high precision of orientation can be obtained. It is also advantageous that the surface normal of the trench sidewall is aligned with a lattice vector orthogonal to the lattice vector in the direction in which the etching process is carried out.

[0199] For example, if silicon is used as the substrate, the silicon substrate 8 may be present as shown in FIG. 9, preferably with Miller indices <110> Therefore, preferably, the crystal structure <110> The lattice vector of is perpendicular to the surface of the as yet unstructured substrate. An etching mask 24, e.g., SiO 2 A hard mask can be used to define horizontal areas or stripes on the substrate surface that should not be etched.

[0200] Smooth, precisely oriented trenches create a <111> Relative to direction <110> This is obtained by anisotropic etching along a preferred direction. Advantageously, for this purpose, wet chemical processes can be used, which are suitable for mass production in batch processes. For example, potassium hydroxide is <111> Crystal orientation <110> As shown in Sato et al. (1988), <110> The etching rate of KOH for silicon single crystal is 1.455 μm / min, while <111> The azimuthal etch rate is only 0.005 μm / min. Due to the speed of the anisotropic etch, it is possible to obtain deep trenches with little underetching using wet chemical processes.

[0201] For example, to form a 400 μm deep trench, KOH is <110> It can be used for 275 minutes on silicon substrates with an orthogonal orientation. <111> The orientation reduces the etch rate by a factor of 291, so that only 1.37 μm of underetching will occur during this period. Even local variations in the intensity of the underetching process will result in orientation variations of much less than 1° over the considerable depth of the 400 μm trench. Instead, the process can achieve deep, nearly perfectly vertical trenches characterized by high precision and smooth quasicrystalline orientation.

[0202] As a further advantage, the sidewalls of the trenches thus obtained, in which the vertical sections of the membrane are formed, have a crystal orientation (here <111> ). This situation promotes columnar growth of piezoelectric materials, such as AlN or PZT. This can ensure in a particularly precise manner that the piezoelectric material has a c-axis orientation perpendicular to the surface of the vertical section, so that the transverse piezoelectric effect can be used to generate horizontal vibrations.

[0203] FIG. 10 illustrates a preferred manufacturing method for producing a MEMS speaker with a vibratable membrane based on an individual piezoelectric ceramic.

[0204] First, a plurality of individual piezoelectric ceramic elements 19 are produced, comprising a layer 10 of mechanical support material (e.g. doped polysilicon) and a layer 11 of piezoelectric material as well as a sacrificial layer 20 (see Figs. 10A and 10B). The sacrificial layer 20 can be, for example, photoresist. The layer 10 of mechanical support material can preferably be conductive to ensure the connection. It is also possible to apply to one layer of piezoelectric material 11 one or two layers of conductive material 12, which serve to make the electrical connection with the piezoelectric material.

[0205] Subsequently, holes for interlayer connections and metal filling 21 are defined (see FIG. 10C). The piezoelectric ceramic elements 19 are stacked (FIG. 10D) and cut (dicing 22 in FIG. 10E) so as to obtain two or more stacks of piezoelectric ceramic elements 19 connected by metal bridges 21 (see FIG. 10E).

[0206] After removing the sacrificial layer 20 (FIG. 10F), the stacked piezoelectric ceramic elements 19 are inserted into the carrier 4, preferably with the first and last piezoelectric ceramic element each being connected to an electrode 13 (FIG. 10E).

[0207] In this way, a vibratable membrane 1 is obtained, also in the middle of the carrier 4, having at least two or more vertical sections 2 formed parallel to the emission direction and which can be driven to vibrate horizontally, generating acoustic waves in a vertical emission direction.

[0208] The actuator principle is again preferably based on a relative change in the shape of the actuator layer 11 with respect to the mechanical support layer 10. This does not require a continuous actuator layer. The connection of all vertical sections 2 with end-side actuation is ensured by a metal bridge 23 in combination with the conductive layer 12.

[0209] FIG. 11 shows the preferred electrical connections of a MEMS speaker with a vibratable membrane based on individual piezoelectric ceramics.

[0210] Figure 11A shows a top view of the MEMS speaker, and Figure 11B shows a side view of the MEMS speaker. The individual lamellae or vertical sections are driven in parallel by electrode pads 13, with U-shaped spacers on either side of the lamellae creating the mechanical and electrical connection with the next lamella.

[0211] FIG. 12 shows an alternative manufacturing method for providing a MEMS speaker with a vibratable membrane based on individual piezoelectric ceramics.

[0212] Advantageously, in the illustrated embodiment, in contrast to the embodiment according to Figure 10 or Figure 11, a structured connection may not be necessary. Instead, the connection may be made by means of a continuous conductive surface from the front (front electrode) or rear (rear electrode), as will be explained below.

[0213] In a manner similar to the manufacturing method according to Figure 10, a plurality of individual piezoelectric ceramic elements 19 are produced, having a layer 10 of mechanical support material (e.g. doped polysilicon) and a layer 11 of piezoelectric material. Preferably, the layer 10 of mechanical support material may be electrically conductive.

[0214] Additionally, an upper frame 25 and a lower frame 26 are provided, each having a recess or groove 27 for receiving the piezoelectric ceramic element 19. Preferably, the upper and lower frames are made of a non-conductive material, such as a polymer. Preferably, a 3D printing process may be used to form the frames.

[0215] To secure the piezoelectric ceramic elements 19, it may be preferable to use an adhesive, which is preferably applied first to the recesses 27 (see FIG. 12A). After securing the piezoelectric ceramic elements 19 in their respective recesses 27 in the lower frame 26, an adhesive may be applied to the piezoelectric ceramic elements 19 such that the upper frame secures the piezoelectric ceramic elements 19 to its upper surface (see FIG. 12B).

[0216] To connect the individual lamellae or piezoelectric ceramic elements 19, a continuous layer of conductive material, preferably metal, is preferably deposited (invisibly) from the front (front electrode) or rear (back electrode), for example by means of a sputtering process.

[0217] In this way it is also possible to obtain a vibratable membrane 1 having at least two or more vertical sections 2 formed parallel to the emission direction and which can be induced to vibrate horizontally in order to generate sound waves in the vertical emission direction. A composite frame 25, 26 can act as a carrier for the vertical sections 2. [Explanation of symbols]

[0218] 1 vibratable membrane 2. Vertical section of the vibrating membrane 3 Horizontal section of the vibrating membrane 4. Career 5 Air volume between vertical sections 8 Substrate 9 Etch stop 10 Layer of mechanical support material, preferably doped polysilicon 11 A layer of actuator material (actuator layer), preferably a layer of piezoelectric material 12 A layer of conductive material, preferably metal 13 Electrode connection part, preferably an electrode pad 14 Holding structure 15. Cabinet 16 Rear resonance volume 17 Ventilation hole 18 Layers of Dielectric Material 19 Piezoelectric ceramic elements 20 Sacrificial Layer 21 Holes defined for metal-filled interconnects 22 Cutting (dicing) 23 Metal Bridge 24 Etching Mask 25 Upper Frame 26 Lower Frame

Claims

1. 1. A MEMS transducer that interacts with a volumetric flow of a fluid, comprising: - Career (4) and a vibratable membrane (1) supported by said carrier (4) for generating or receiving pressure waves in said fluid of vertical direction, said vibratable membrane (1) being manufactured together with said carrier (4) in a semiconductor process; having said vibratable membrane (1) exhibits two or more vertical sections (2) formed substantially parallel to said vertical direction, said vertical sections (2) having at least one layer of actuator material (11), at least one end of said vibratable membrane (1) being connected to at least one electrode (13), The MEMS transducer is characterized in that the two or more vertical sections (2) can be induced to vibrate horizontally by driving the at least one electrode (13), or when the two or more vertical sections (2) are induced to vibrate horizontally, an electrical signal can be generated in the at least one electrode.

2. 2. A MEMS transducer according to claim 1, characterized in that the MEMS transducer is a MEMS speaker, in which an air volume (5) is preferably present between the vertical sections (2) and is moved along a vertical emission direction as a result of the horizontal vibration to generate sound waves, or the MEMS transducer is a MEMS microphone, in which an air volume (5) is preferably present between the vertical sections (2) and is moved along a vertical detection direction when receiving sound waves.

3. said two or more vertical sections (2) having at least two layers, one layer (11) having an actuator material and a second layer (10) having a mechanical support material, at least said layer (11) having actuator material is connected to an electrode (13); whereby horizontal vibrations may be generated by changes in shape of the actuator material relative to the mechanical support material; or Horizontal vibration thereby causes a change in shape of the actuator material relative to the mechanical support material, generating an electrical signal. A MEMS transducer according to any one of claims 1 to 2, characterized in that

4. The two or more vertical sections (2) have at least two layers, both layers (11) having actuator material and each connected to an electrode (13), respectively; and The horizontal vibration may be produced by a change in shape of one layer relative to the other layer, or The horizontal vibration causes a change in shape of one layer relative to the other layer, generating an electrical signal. A MEMS transducer according to any one of claims 1 to 3, characterized in that

5. the carrier (4) has two side areas and the vibratable membrane (1) is arranged horizontally between the two side areas, and / or The carrier (4) is formed of a substrate (8), preferably selected from the group consisting of monocrystalline silicon, polysilicon, silicon dioxide, silicon carbide, silicon germanium, silicon nitride, nitride, germanium, carbon, gallium arsenide, gallium nitride, indium phosphide, and glass. A MEMS transducer according to any one of claims 1 to 4, characterized in that

6. A MEMS transducer according to any one of the preceding claims, characterized in that the vibratable membrane (1) is formed in a serpentine structure.

7. 7. A MEMS transducer as claimed in any one of claims 1 to 6, characterized in that the vibrable membrane (1) is formed in a serpentine structure with alternating vertical sections (2) and horizontal sections (3), at least two of the horizontal sections (3) having holding structures (14) attached thereto and connected directly or indirectly to the carrier (4).

8. 8. The MEMS transducer according to claim 1, wherein the actuator material comprises a piezoelectric material, a polymeric piezoelectric material, and / or an electroactive polymer (EAP), the piezoelectric material being preferably selected from the group comprising lead zirconate titanate (PZT), aluminium nitride (AlN), aluminium scandium nitride (AlScN), and zinc oxide (ZnO).

9. 9. A MEMS transducer according to any one of claims 1 to 8, characterized in that the vibrable membrane (1) has three layers, an upper layer (12) made of a conductive material, a middle layer (11) made of an actuator material and a lower layer (12) made of a conductive material, the conductive material of the upper and / or lower layer being preferably a mechanical support material.

10. 10. A MEMS transducer as claimed in any one of claims 1 to 9, characterized in that the vibrable membrane (1) has two layers (11) of actuator material separated by an intermediate layer (12) of conductive material, preferably of metal, said intermediate layer (12) being connected to a first electrode (13) and at least one of the two layers (11) of actuator material being connected to a second electrode (13) via another layer (12) of conductive material, preferably of metal.

11. 11. MEMS transducer according to any one of the preceding claims, characterized in that the vibratable membrane (1) is coated with a layer of anti-stick material.

12. A MEMS transducer according to any one of claims 1 to 11, characterized in that the vibrable membrane (1) supported by the carrier (4) is arranged at the front of a housing (15) enclosing a rear resonant volume (16), and ventilation holes (17) are preferably present in the housing (15) to avoid acoustic short circuits and / or to assist sound.

13. A method for manufacturing a MEMS transducer according to any one of claims 1 to 12, comprising the steps of: - etching the substrate (8), preferably from the front side, in order to form a structuring, preferably a serpentine structure; - optionally depositing an etch stop; - depositing at least two layers, where at least a first layer (11) comprises an actuator material and a second layer (10) comprises a mechanical support material, or where at least two layers (11) comprise an actuator material; - connecting said first and / or second layer to an electrode (13); - etching, preferably from the rear side, and optionally removing said etch stop; having Thereby, a vibratable membrane (1), preferably in the form of a serpentine structure, is supported by a carrier (4) formed by said substrate (8), said vibratable membrane (1) having at least two or more vertical sections (2) for generating or receiving vertical fluid pressure waves, said vertical sections being formed parallel to said vertical direction, whereby said two or more vertical sections (2) can be induced to vibrate in a horizontal direction by driving said at least one electrode (13), or Thereby, when the two or more vertical sections (2) are induced to vibrate in a horizontal direction, an electrical signal may be generated in the at least one electrode.

14. 1. A MEMS transducer that interacts with a volumetric flow of a fluid, comprising: - Career (4) and a vibratable membrane (1) supported by said carrier (4) for generating or receiving pressure waves in said fluid of vertical direction; having said vibratable membrane (1) exhibits two or more vertical sections (2) formed substantially parallel to said vertical direction, said vertical sections (2) having at least one layer of actuator material (11), at least one end of said vibratable membrane (1) being connected to at least one electrode (13), Whereby the two or more vertical sections (2) can be induced to vibrate horizontally by driving the at least one electrode (13), or when the two or more vertical sections (2) are induced to vibrate horizontally, an electrical signal can be generated in the at least one electrode; and 1. A MEMS transducer, characterized in that the vertical sections (2) of the vibrable membrane (1) have two layers, a first layer (11) made of actuator material and a second layer (10) made of conductive support material, the vertical sections (2) being connected via a horizontal metal bridge (23), and the vertical sections (2) being respectively connected to electrodes (13).

15. A method of manufacturing the MEMS transducer of claim 14, comprising the steps of: - providing a plurality of individual piezoceramic elements (19) comprising a sacrificial layer (20), a layer of conductive material (12) and a layer of piezoelectric material (11); - defining holes (21) for the interlayer connections of said piezoelectric ceramic elements and for metal filling; - stacking and optionally dicing (22) said piezoelectric ceramic elements (19) so as to obtain a stack of said piezoelectric ceramic elements (19) connected by metal bridges (23); - removing said sacrificial layer (29) and inserting said stack of piezoceramic elements (19) into a carrier (4), said piezoceramic elements (19) being respectively connected to an electrode (13); having Thereby, a vibratable membrane (1), preferably in the form of a lamellar structure, is supported by said carrier (4), said vibratable membrane (1) having at least two or more vertical sections (2) for generating or receiving vertical fluid pressure waves, said vertical sections being formed parallel to said vertical direction, whereby said two or more vertical sections (2) can be induced to vibrate in a horizontal direction by driving said at least one electrode (13), or Thereby, when the two or more vertical sections (2) are induced to vibrate in a horizontal direction, an electrical signal may be generated in the at least one electrode.

Images