EP4091340A1

EP4091340A1 - Mems transducer with increased performance

Info

Publication number: EP4091340A1
Application number: EP21700309.4A
Authority: EP
Inventors: Alfons Dehé; Achim Bittner; Lenny Castellanos
Original assignee: Hann-Schickard-Gesellschaft fuer Angewandte Forschung eV
Current assignee: Hann-Schickard-Gesellschaft fuer Angewandte Forschung eV
Priority date: 2020-01-17
Filing date: 2021-01-15
Publication date: 2022-11-23
Also published as: JP2023511538A; EP3852391A1; KR20220130720A; US20230047856A1; CN115280797A; EP3852391B1; US11800294B2; WO2021144400A1

Abstract

The invention relates to an MEMS transducer that comprises a vibrating diaphragm (1) for producing or picking up pressure waves in a fluid in a vertical direction, wherein the vibrating diaphragm (1) is held by a support (4) and the vibrating diaphragm (1) has two or more vertical sections (2) that are formed parallel to the vertical direction and comprise at least one layer of an actuator material (11). The ends of the vibrating diaphragm (1) are in contact with an electrode (13), which means that actuation of the at least one electrode (13) allows the two or more vertical sections (2) to be excited to produce horizontal vibrations or which means that excitation of the two or more vertical sections (2) to produce horizontal vibrations results in an electrical signal being able to be generated at the at least one electrode (13).

Description

MEMS CONVERTER WITH INCREASED PERFORMANCE

DESCRIPTION

The invention relates to a MEMS transducer, which comprises an oscillating membrane for generating or receiving pressure waves of a fluid in a vertical direction, the oscillating membrane being held by a support and the oscillating membrane having two or more vertical sections which are parallel to the vertical Direction are formed and comprise at least one layer of an actuator material. The vibratable membrane is preferably in contact with an electrode at the end, so that the two or more vertical sections can be excited to horizontal vibrations by controlling the at least one electrode or so that when the two or more vertical sections are excited to horizontal vibrations at the at least one electrode electrical signal can be generated.

Background and prior art

For the production of compact, mechanical-electronic devices, microsystem technology is used in many fields of application today. The microsystems (microelectromechanical system, MEMS for short) that can be produced in this way are very compact (micrometer range) with excellent functionality and ever lower production costs.

MEMS converters such as MEMS loudspeakers or MEMS microphones are also known from the prior art. Current MEMS loudspeakers are mostly designed as planar membrane systems with a vertical actuation of a vibratable membrane in the direction of emission. The excitation takes place, for example, by means of piezoelectric, electromagnetic or electrostatic actuators.

A MEMS electromagnetic speaker for mobile devices is described in Shahosseini et al. Described in 2015. The MEMS loudspeaker has a stiffening silicon microstructure as a sound emitter, with the movable part being suspended from a carrier via silicon mainsprings in order to enable large displacements out of the plane by means of an electromagnetic motor.

Stoppel et al. 2017 reveals a two-way loudspeaker whose concept is based on concentric piezoelectric actuators. As a special feature, the vibration diaphragm is not designed to be closed, but rather includes eight piezoelectric unimorph actuators, each of which consists of a piezoelectric and a passive layer. The outer woofers consist of four trapezoidal actuators clamped on one side, while the inner tweeters are formed by four triangular actuators, which are connected to a rigid frame by means of a spring. The separation of the membrane should allow an improved sound image with higher power. A disadvantage of such planar MEMS loudspeakers is their limitation with regard to the sound power, in particular at low frequencies. One reason for this is that the sound pressure level that can be generated is proportional to the square of the frequency for a given deflection. For sufficient sound power, either deflections for the vibration diaphragms of at least 100 μm or large-area diaphragms in the square centimeter range are necessary. Both conditions are difficult to implement using MEMS technology.

In the prior art, it has therefore been proposed to design MEMS loudspeakers which do not have a closed membrane for vibrations in the vertical emission direction, but a large number of movable elements that can be excited to lateral or horizontal vibrations. It is advantageous here that an increased volume flow can be moved over a small area and thus an increased sound power can be provided.

A MEMS loudspeaker based on this principle is used, for example, in US 2018 /

0179048 A1 or Kaiser et al. Revealed in 2019.

The MEMS loudspeaker comprises a plurality of electrostatic bending actuators, which are arranged between a top and bottom wafer as vertical lamellas and can be excited to lateral vibrations by appropriate control. Here, an inner lamella forms an actuator electrode opposite two outer lamellae. With the exception of a connection node between the electrodes, which are still galvanically separated, there is an air gap between the three curved lamellae. If there is a potential inside against outside, this leads to a mutual attraction due to the curvature of the design in the direction of a preferred direction, which is specified by an anchor. The bulges of the outer lamellae are used for mobility. The restoring force is given by a mechanical spring force. A pull-push operation is therefore not possible.

Another disadvantage is that gaps between the bending actuators and the top / bottom wafers, which are necessary for their mobility, lead to ventilation between the two chambers. This limits the lower limit frequency. Furthermore, the lateral movement of the bending actuators and therefore the sound power are restricted in order to avoid a pull-in effect and acoustic breakdown.

An alternative air pulse or sound generation system based on MEMS is described in US 2019/011 64 17 A1. The device comprises a front and rear chamber and a plurality of valves, the front and rear chambers being separated from one another by means of a folded membrane. In one embodiment, the folded membrane has a rectangular meandering structure with horizontal and vertical sections in cross section. Piezo actuators are positioned on the respective horizontal sections in order to bring about a lateral movement of the vertical sections by synchronized expansion or compression of the horizontal sections. With the proposed principle, an increased volume flow and thus sound power can also be generated on a small chip surface. The disadvantage, however, is the increased effort for the synchronized drive of the piezo actuators. There is also potential for improvement with regard to the volume displaced by the lateral vibrations, which is limited by the geometric arrangement of the horizontal sections actuated on one side.

From US 2002/006208 A1 and JP 3 919695 B2 a piezoelectric loudspeaker is known in which two piezoelectric films are formed into a diaphragm with an accordion shape. When folded, the diaphragm is clamped in at the sides by a wave-shaped pair of plates, which are fixed, for example, by means of screw connections and, as a composite side frame, stabilize the vibrating diaphragm. Several electrodes in a structured form are applied to the wave crests and troughs of the diaphragm and are isolated from one another by strips of non-conductive material. To control the electrodes, electrode lines are arranged in the plate pair or side frame. Alternatively, the pair of plates can also be formed at least partially from a conductive material.

The macroscopic piezoelectric loudspeaker US 2002/006208 A1 and JP 3 919695 B2 is obtained in an assembly process which cannot be miniaturized in an obvious manner in order to obtain a MEMS loudspeaker. In particular, the intended clamping of the diaphragm in a two-part side frame, the structured application of multiple electrodes on wave crests and valleys of the diaphragm or the connection of the electrodes to electrode lines in the side frame cannot be transferred to a MEMS process.

In light of the disadvantages of the prior art, there is thus a need for alternative or improved solutions for MEMS-based loudspeakers.

Object of the invention

The object of the invention is to provide a MEMS converter, in particular a MEMS loudspeaker or MEMS microphone, and a method for producing the MEMS converter, which do not have the disadvantages of the prior art. In particular, it was an object of the invention to provide a powerful MEMS loudspeaker or MEMS microphone with high sound quality or audio quality, which is characterized at the same time by a simple, inexpensive and compact structure.

Summary of the invention

The object is achieved by the features of the independent claims. Preferred embodiments of the invention are described in the dependent claims.

The invention preferably relates to a MEMS converter for interaction with a volume flow of a fluid

- a carrier,

an oscillatable membrane for generating or absorbing pressure waves of the fluid in a vertical direction, the oscillatable membrane being held by the carrier and wherein the vibratable membrane has two or more vertical sections which are formed essentially parallel to the vertical direction and comprise at least one layer of an actuator material, the vibratable membrane being in contact with at least one electrode at the end so that the at least one electrode is controlled two or more vertical sections can be excited to essentially horizontal oscillations or so that when the two or more vertical sections are excited to essentially horizontal oscillations at the at least one electrode, an electrical signal can be generated.

The MEMS converter can particularly preferably be a MEMS loudspeaker. In a particularly preferred embodiment, the invention relates to a MEMS loudspeaker comprising

- a carrier,

An oscillatable membrane for generating sound waves in a vertical emission direction, the oscillatable membrane being held by the carrier, the oscillatable membrane having two or more vertical sections which are essentially parallel to the emission direction and comprise at least one layer of an actuator material , wherein the vibratory membrane is preferably in contact at the end with at least one electrode, so that the two or more vertical sections can be excited to essentially horizontal vibrations by controlling the at least one electrode.

Due to the construction of the MEMS loudspeaker, a MEMS loudspeaker with high sound power and simplified control can be obtained.

In contrast to known planar MEMS loudspeakers, the vibratory membrane itself does not have to be operated over a large area of several square centimeters or with a deflection of more than 100 μm in order to generate sufficient sound pressure. Instead, the majority of the vertical sections of the vibratable membrane can move an increased total volume in the vertical emission direction with small horizontal or lateral movements of a few micrometers.

Compared to solutions according to US 2018/0179048 A1 or Kaiser et al. In 2019, the claimed MEMS loudspeaker is characterized by a simplified structure, control and manufacturing process.

In particular, the provision of the vertical lamellae or bending actuators for a MEMS loudspeaker according to Kaiser et al. 2019 is complex. In addition, sufficiently precise vertical etchings are only possible for limited lamella heights, which limits the sound power.

By means of the solution according to the invention, the vertical sections of the oscillatable membrane can instead be obtained in MEMS design by means of simple manufacturing steps, as will be explained in detail below. In addition, the actuator principle according to the invention avoids pull-in or sticking of the vertical sections. in the In contrast to the solution by Kaiser et al. In 2019, no potential differences are obtained in a gap between the vertical sections due to the one-sided electrodes. In addition to avoiding an overvoltage or a pull-in, this can also reduce the accumulation of dust, since, for example, an external electrode can be connected to a ground potential.

Another particular advantage of the MEMS loudspeaker described is the simplified control. While in US 2019/011 64 17 A1 a large number of piezoelectric actuators have to be contacted on the horizontal sections, the proposed MEMS loudspeaker can be operated by means of at least one end electrode. This reduces the manufacturing effort, minimizes sources of error and also inherently leads to a synchronous control of the vertical sections to horizontal oscillations.

In this way, the air volumes which are present between the vertical sections can be moved extremely precisely by the horizontal oscillations along the vertical emission direction. The result is an improved sound image, even with high sound power levels.

A “MEMS loudspeaker” preferably denotes a loudspeaker which is based on MEMS technology and whose sound-generating structures are at least partially dimensioned in the micrometer range (1 μm to 1000 gm). For example, the vertical sections of the oscillatable membrane can preferably have a dimension in the range of less than 1000 μm in terms of width, height and / or thickness. It can also be preferred here that, for example, only the height of the vertical sections are dimensioned in the micrometer range, while, for example, the length can have a larger dimension and / or the thickness can have a smaller size.

The design of the vibratory membrane can advantageously not only be used to form a MEMS loudspeaker with high sound power and simplified control. Likewise, for example, the provision of a particularly powerful MEMS microphone with high audio quality is made possible.

In a preferred embodiment, the invention thus also relates to a MEMS microphone comprising

- a carrier,

an oscillatable membrane for receiving sound waves in a vertical direction, wherein the oscillatable membrane is held by the carrier, and wherein the oscillatable membrane has two or more vertical sections which are formed parallel to the vertical direction and comprise at least one layer of an actuator material , wherein the vibratory membrane is preferably in contact at the end with at least one electrode, so that when the two or more vertical sections are excited into horizontal vibrations at the at least one electrode, an electrical signal can be generated. The structure of the MEMS microphone is structurally similar to that of the MEMS loudspeaker, particularly with regard to the design of the vibratable membrane. Instead of controlling the electrodes to generate horizontal vibrations and therefore sound pressure waves, the MEMS microphone is designed to pick up sound pressure waves in the same vertical direction. Thus, there are preferably air volumes between the vertical sections which are moved along a vertical detection direction when sound waves are picked up. The vertical sections are excited to vibrate horizontally by the sound pressure waves, so that the actuator material generates a corresponding periodic electrical signal.

A “MEMS microphone” preferably denotes a microphone which is based on MEMS technology and whose sound-picking structures are at least partially dimensioned in the micrometer range (1 pm to 1000 gm). For example, the vertical sections of the oscillatable membrane can preferably have a dimension in the range of less than 1000 μm in terms of width, height and / or thickness. It can also be preferred here that, for example, only the height of the vertical sections are dimensioned in the micrometer range, while, for example, the length can have a larger dimension and / or the thickness can have a smaller size.

The term MEMS converter is therefore to be understood as meaning both a MEMS microphone and a MEMS loudspeaker. In general, the MEMS converter denotes a converter for interaction with a volume flow of a fluid, which is based on MEMS technology and whose structures for interaction with the volume flow or for absorbing or generating pressure waves of the fluid are dimensioned in the micrometer range (1 pm to 1000 pm ) exhibit. The fluid can be either a gaseous or a liquid fluid. The structures of the MEMS transducer, in particular the vibratory membrane, are designed to generate or absorb pressure waves from the fluid.

For example, as in the case of a MEMS loudspeaker or MEMS microphone, they can be sound pressure waves. The MEMS converter can, however, also be suitable as an actuator or sensor for other pressure waves. The MEMS converter is therefore preferably a device that converts pressure waves (e.g. acoustic signals as sound pressure) into electrical signals or vice versa (conversion of electrical signals into pressure waves, e.g. acoustic signals).

The MEMS converter can also be used as an energy harvester, using alternating pneumatic or hydraulic pressures. In these cases, the electrical signal can be dissipated as generated electrical energy, stored or fed to other (consumer) devices.

At the end, preferably means a positioning of the at least one electrode so that contact with electronics, for example a current or voltage source in the case of a MEMS loudspeaker, can take place at one end of the vibratable membrane, preferably at one end at which the membrane is attached The carrier is suspended. Electrode preferably means an area made of a conductive material (preferably a metal) which is used for such contact with electronics, for example a current and / or voltage source in the Case of a MEMS speaker. It can preferably be an electrode pad. The electrode pad is particularly preferably used to make contact with electronics and is itself connected to a conductive metal layer which can extend over the entire surface of the vibratory membrane. In the following, the conductive layer is sometimes referred to together with an electrode pad as an electrode, for example as a top electrode or bottom electrode.

Particularly preferably, the layer made of a conductive material, preferably metal, in the sense of a top or bottom electrode, is a continuous or full-area or contiguous layer of the vibratable membrane, which forms an essentially homogeneous surface and is in particular not structured. Instead, the two or more vertical sections are preferably contacted with the end-side electrodes or the electrode pad by means of an unstructured layer made of a conductive material, preferably metal.

In particular, it is advantageously not necessary to create separate contact areas for different vertical sections of the vibratable membrane. In contrast to the approach for a macroscopic piezoelectric loudspeaker according to US 2002/006208 A1 and JP 3 919695 B2, the application of a structured top or bottom electrode is not necessary. Instead, a top or bottom electrode can be applied as a cohesive layer made of a conductive material, which is contacted by means of at least one end electrode or an electrode pad. This significantly simplifies the manufacturing process and allows miniaturized MEMS transducers to be made available in large numbers by means of a batch process.

In preferred embodiments, the MEMS transducer comprises two electrodes at the end. Preferably, the contact with electronics, e.g. a current or voltage source, can be made with the electrodes at opposite ends of the vibratory membrane, between which the two or more vertical sections are present, so that the actuator layer (s) in the vertical sections by means of the end electrodes can be controlled.

The end-side provision of the electrodes is thus preferably distinguished from a contact which controls the respective vertical sections with respective separate electrodes or, in the case of a MEMS microphone, picks up electrical signals generated. The MEMS converter thus preferably comprises exactly one or exactly two electrodes for contacting the end and no further electrodes (pads) for contacting central vertical sections.

The layer made of an actuator material in the vertical sections is preferably used as part of a mechanical biomorph, with a lateral curvature of the vertical sections being caused by controlling the actuator layer via the electrode or with a corresponding electrical signal being generated by an induced lateral curvature.

In a preferred embodiment, the two or more vertical sections have at least two layers, one layer comprising an actuator material and a second layer comprising a mechanical support material and at least the layer comprising the Actuator material and is present in contact with an end electrode, so that the horizontal vibrations can be generated by changing the shape of the actuator material compared to the mechanical support material. In the embodiment, the mechanical bimorph is formed by a layer of actuator material (for example a piezoelectric material) and a passive layer, which functions as a mechanical support layer. Both a transverse and a longitudinal piezo effect can be used for the bending.

When the actuator position is activated, it can experience, for example, a transverse or longitudinal stretching or compression. This creates a stress gradient in relation to the mechanical support layer, which leads to a lateral curvature or oscillation. As illustrated in FIG. 1, a push-pull operation can preferably take place by changing polarity on the electrodes, whereby almost the entire air volume between the vertical sections can alternately be moved in the vertical emission direction.

The advantage of the actuator principle is therefore a highly efficient translation of the horizontal vibrations of vertical sections into a vertical volume movement or sound generation.

Since the actuator principle is not based on electrostatic attraction, but on a relative change in shape (e.g. compression, stretching, shear) of the actuator layer in relation to a support layer, sticking of the membrane sections can be excluded. Instead, the vertical sections can finally touch and are therefore not restricted in their deflection.

In a further preferred embodiment, the two or more vertical sections comprise at least two layers, both layers comprising an actuator material and being in contact with electrodes at each end and the horizontal vibrations being able to be generated by changing the shape of one layer compared to the other layer. In the embodiment, the horizontal oscillation of the vertical sections is therefore not generated by a stress gradient between an active actuator layer and a passive support layer, but rather by a relative change in shape of two active actuator layers.

The actuator layers can consist of the same actuator material and can be controlled differently. The actuator layers can also consist of different actuator materials, for example of piezoelectric materials with different deformation coefficients.

In the context of the invention, the “layer comprising an actuator material” is preferably also referred to as an actuator layer. An actuator material preferably means a material which, when an electrical voltage is applied, undergoes a change in shape, for example an expansion, compression or shear, or, conversely, generates an electrical voltage with a change in shape.

Preference is given to materials with electric dipoles which experience a change in shape when an electric voltage is applied, the orientation of the dipoles and / or the electric field being able to determine the preferred direction of the changes in shape. The actuator material can preferably be a piezoelectric material, a polymer piezoelectrical material and / or electroactive polymers (EAP).

The piezoelectric material is particularly preferably selected from a group comprising lead zirconate titanate (PZT), aluminum nitride (AlN), aluminum scandium nitride (AIScN) and zinc oxide (ZnO).

The polymer piezoelectric materials preferably include polymers which have internal dipoles and thus piezoelectric properties. This means that when an external electrical voltage is applied, the piezoelectric polymer materials (analogous to the aforementioned classic piezoelectric materials) experience a change in shape (e.g. compression, stretching or shear). An example of a preferred piezoelectric polymer material is polyvinylidene fluoride.

This enables a macroscopic solution to be implemented in which a polymer piezoelectrical material layer is applied to a mechanical support layer and is wound over an upper and lower comb. A polymer piezoelectrical material layer (including electrode) is preferably first provided on a support layer (possibly including a counter electrode). Subsequently, an upper and a lower comb (preferably a MEMS structure) are moved against one another in such a way that a folded membrane with actuatable vertical sections is created.

In the context of the invention, the “layer comprising a mechanical support material” is preferably also referred to as a support layer or support layer. The mechanical support material or the support layer preferably serves as a passive layer which can withstand a change in shape of the actuator layer. In contrast to an actuator layer, the mechanical support material preferably does not change its shape when an electrical voltage is applied. The mechanical support material is preferably electrically conductive, so that it can also be used directly for contacting the actuator layer. However, in some embodiments it can also be non-conductive and, for example, be coated with an electrically conductive layer.

The mechanical support material is particularly preferably monocrystalline silicon, a polysilicon or a doped polysilicon.

While the actuator position undergoes a change in shape when an electrical voltage is applied, the position of the mechanical support material remains essentially unchanged. The resulting stress gradient between the two layers (mechanical bimorph) preferably causes a horizontal curvature. For this purpose, the thickness of the support layer in comparison to the thickness of the actuator layer should preferably be selected so that a sufficiently large stress gradient is generated for the curvature. For doped polysilicon as a mechanical support material and a piezoelectric material such as PZT or AlN, for example, a thickness of essentially the same size, preferably between 0.5 μm and 2 μm, has proven to be particularly suitable.

Terms such as essentially, approximately, about, etc. preferably describe a tolerance range of less than ± 20%, preferably less than ± 10%, even more preferably less than ± 5% and in particular less than ± 1%. Details of essentially, approximately, approximately, etc. disclose and always also include the exact stated value. With periodic control of the actuator position, for example by means of an alternating voltage, horizontal vibrations for sound emission can thus be generated quickly and precisely.

To ensure horizontal oscillation, the piezoelectric material can preferably have a C-axis orientation perpendicular to the surface of the vertical sections, so that a transverse piezoelectric effect is used. Other orientations and, for example, the use of a longitudinal piezoelectric effect to form the horizontal arches or vibrations (cf. FIG. 1) can also be preferred.

Contacting the actuator layer and / or the layer made of a mechanical support material and thus the application of an electrical voltage can take place directly via the end electrodes or be supported by a layer made of a conductive material,

In a preferred embodiment, the vibratory membrane therefore comprises at least one layer made of a conductive material.

In preferred embodiments, the conductive material is selected from a group comprising platinum, tungsten, (doped) tin oxide, monocrystalline silicon, polysilicon,

Molybdenum, titanium, tantalum, titanium-tungsten alloy, metal silicide, aluminum, graphite and copper.

The directional information vertical and horizontal (or lateral) preferably relate to a preferred direction in which the oscillatable membrane is oriented for generating or absorbing pressure waves of the fluid. The vibratory membrane is preferably suspended horizontally between at least two side regions of a carrier, while the vertical direction (direction of interaction with the fluid) for generating or absorbing pressure waves is orthogonal thereto. In the case of a MEMS loudspeaker, the vertical (interaction) direction corresponds to the vertical sound emission direction of the MEMS loudspeaker. In this case, vertical preferably means the direction of the sound emission, while horizontal means a direction orthogonal to it. In the case of a MEMS microphone, the vertical (interaction) direction corresponds to the vertical sound detection direction of the MEMS microphone. In this case, vertical preferably means the direction of sound detection or recording, while horizontal means a direction orthogonal thereto.

The vertical sections of the vibratable membrane thus preferably designate sections of the vibratable membrane which are essentially aligned in the emission direction of a MEMS loudspeaker or the detection direction of a MEMS microphone. The person skilled in the art understands that it does not have to be an exact vertical alignment, but rather the vertical sections of the oscillatable membrane are preferably aligned essentially in the emission direction of a MEMS loudspeaker or the detection direction of a MEMS microphone.

In a preferred embodiment, the vertical sections are aligned essentially parallel to the vertical direction, with essentially parallel a tolerance range of ±

30 °, preferably ± 20 °, particularly preferably ± 10 ° around the vertical direction means. The oscillatable membrane can therefore preferably not only have a rectangular meander shape in cross section, but also have a curved or wavy shape or a sawtooth shape (zigzag shape).

The vertical and / or horizontal sections are preferably rectilinear at least in sections or over their entire length, but the vertical and / or horizontal sections can also be designed to be curved at least in sections or over their entire length. In the case of a curved or wavy shape of the oscillatable membrane in cross section, the alignment preferably relates to a tangent to the curved vertical and / or horizontal sections at their respective midpoints.

While the vibratable membrane is preferably oriented horizontally to the direction of sound emission or sound detection, the sound waves are generated by actuation of the vertical sections or vice versa.

In a preferred embodiment of the invention, the carrier comprises two side areas between which the vibratable membrane is arranged in the horizontal direction.

The carrier is preferably a frame structure which is essentially formed by a continuous outer border in the form of side walls of a flat area that remains free. The frame structure is preferably stable and rigid.

In the case of an angular frame shape (triangular, square, hexagonal or generally polygonal outline), the individual side regions, which preferably essentially form the frame structure, are called in particular side walls.

The oscillatable membrane is preferably held by at least two side walls of the carrier. In the exemplary FIGS. 1-9, the two side walls can be seen in cross section. Preferably, however, the carrier preferably comprises four side areas, with additional end faces, as a rule parallel to the cross-section shown. These other two side walls span the frame structure.

The oscillatable membrane is preferably hung flat within the area that remains free. The two-dimensional expansion of the vibratable membrane characterizes a horizontal direction, while the vertical sections are essentially orthogonal to it. With regard to the end faces, the membrane can be adhered to these side walls or be slotted there for the purpose of greater mobility. The slot can advantageously represent a dynamic high-pass filter which, for example, couples a front volume and a rear volume to one another.

In a preferred embodiment of the invention, the carrier is formed from 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.

These materials are easy and inexpensive to process in semiconductor and / or microsystem production and are suitable for large-scale production. The carrier structure can be manufactured flexibly due to the materials and / or manufacturing methods become. In particular, it is preferably possible to manufacture the MEMS transducer comprising an oscillatable membrane together with a carrier in a (semiconductor) process, preferably on a wafer. This further simplifies and makes production cheaper, so that a compact and robust MEMS converter can be provided at low cost.

In a preferred embodiment, the vibratory membrane is formed by a lamellar structure or a meandering structure. The specification of a lamellar or meandering structure preferably relates to the shape of the oscillatable membrane in cross section.

A lamellar structure preferably denotes an arrangement of similar, parallel layers, which preferably form the vertical sections. The individual lamellae are preferably aligned with their area essentially parallel to the vertical direction, preferably an emission or detection direction. The lamellae are preferably constructed in several layers and form a mechanical biomorph. For example, the slats can each include an actuator layer and a passive layer made of a support material and / or two differently controllable actuator layers.

The person skilled in the art understands that the lamellae need not be aligned exactly parallel to the vertical direction, but rather the lamellae are preferably aligned essentially in the emission direction of a MEMS loudspeaker or the detection direction of a MEMS microphone.

In a preferred embodiment, the vertical sections or lamellae are aligned essentially parallel to the vertical direction, wherein essentially parallel means a tolerance range of ± 30 °, preferably ± 20 °, particularly preferably ± 10 ° around the vertical direction.

It can be preferred here that the lamellae are flat, which means in particular that their extension in each of the two dimensions (height, width) of their area is greater than in a dimension perpendicular thereto (the thickness). For example, size ratios of at least 2: 1, preferably at least 5: 1, 10: 1 or more can be preferred.

The oscillatable membrane preferably has a multiplicity of lamellae which form the vertical sections. For example, 2, 3, 4, 5, 10, 15, 20, 30, 40, 50 or more slats may be preferred. This results in a high degree of efficiency for a desired sound emission or sound detection in a very small space.

In the embodiment, the vibratory membrane is preferably formed by the lamellae as vertical sections, which are connected to one another via conductive bridges or horizontal sections. Metal bridges (cf. FIG. 10) or bridges made of other conductive materials are suitable as bridges. On the one hand, the conductive bridges ensure the mechanical integrity of the vibrating membrane. On the other hand, the conductive bridges advantageously allow contact to be made with all of the lamellae by means of electrodes at the end. Advantageously, the lamellas can be excited synchronously with horizontal vibrations or detect the same with little control and manufacturing effort. A meander structure preferably denotes a structure formed from a sequence of mutually essentially orthogonal sections in cross section. The mutually orthogonal sections are preferably vertical and horizontal sections of the oscillatable membrane. The meandering structure is particularly preferably rectangular in cross section. However, it can also be preferred that the meandering structure has a sawtooth shape (zigzag shape) in cross section or is designed in the shape of a curve or wave. This is particularly the case if the vertical sections are not aligned exactly parallel to the vertical emission or detection direction, but rather enclose an angle of, for example, ± 30 °, preferably ± 20 °, particularly preferably ± 10 °, with the vertical direction.

In preferred embodiments, the horizontal sections can likewise not be exactly at an orthogonal angle of 90 ° to the vertical emission or detection direction, but for example an angle between 60 ° and 120 °, preferably between 70 ° and 110 °, particularly preferably between 80 ° and Include 100 ° with the vertical direction.

In the case of a curved or wavy shape of the vertical and / or horizontal sections of the oscillatable membrane in cross section, the alignment preferably relates to a tangent to the vertical and / or horizontal sections at their respective midpoints.

The meander structure thus preferably corresponds to a membrane folded along its width. For the purposes of the invention, an oscillatable membrane can therefore preferably also be referred to as a bellows. The parallel folds of the bellows preferably form the vertical sections. The connecting sections between the folds preferably form the horizontal sections. The vertical sections are preferably longer than the horizontal sections, for example by a factor of 1, 5, 2, 3, 4 or more.

With regard to the function of a vibrating membrane in a meandering shape for generating or receiving sound waves, the vertical sections are decisive, analogously to the lamellae described above. The vertical sections are preferably constructed in several layers and form a mechanical biomorph. For example, the vertical sections can each include an actuator layer and a passive layer made of a support material and / or two differently controllable actuator layers. The horizontal sections of the folded membrane can preferably be constructed identically to the vertical sections (cf., inter alia, Figs. 3-7). However, it can also be preferred that the horizontal sections - in contrast to the vertical sections - do not have an actuator layer, but only a mechanical support layer and / or an electrically conductive layer.

In a preferred embodiment, the at least one layer made of an actuator material of the vibratable membrane is a continuous layer. Continuous preferably means that there are no interruptions in the cross-sectional profile. Accordingly, in the embodiment mentioned, it is preferred that there is a continuous layer of actuator material in both the vertical and horizontal sections. No structuring is therefore advantageously necessary. A continuous layer is particularly easy to manufacture and ensures synchronous actuation when operating a MEMS loudspeaker. The performance of the MEMS transducer, in particular of a MEMS loudspeaker or MEMS microphone, can essentially be determined by the number and / or dimensioning of the vertical sections.

In preferred embodiments, the vibratable membrane comprises more than 3, 4, 5, 10, 15, 20, 30, 40, 50, 100 or more vertical sections.

In preferred embodiments, the vibratable membrane comprises less than 10,000,

5000, 2000, or 1000 or less vertical sections.

The preferred number of vertical sections leads to high sound power on the smallest chip surfaces without the sound image or audio quality suffering.

The vertical sections are preferably flat, which means in particular that their extension in each of the two dimensions (height, width) of their area is greater than in a dimension perpendicular thereto (the thickness). For example, size ratios of at least 2: 1, preferably at least 5: 1, 10: 1 or more can be preferred.

According to the invention, the height of the vertical sections preferably corresponds to the dimension along the direction of the sound emission or sound detection, while the thickness of the vertical sections preferably corresponds to the sum of the layer thickness of the one or more layers that form the vertical sections. The length of the vertical sections preferably corresponds to a dimension orthogonal to the height or thickness. In the cross-sectional views of the figures below, the height and thickness are shown schematically (not necessarily true to scale), while the dimension of the length corresponds to a (not visible) drawing depth of the figures.

In a preferred embodiment, the height of the vertical sections is between 1 pm and 1000 pm, preferably between 10 pm and 500 pm. Intermediate ranges from the aforementioned ranges can also be preferred, such as, for example, 1 pm to 10 pm, 10 pm to 50 pm, 50 pm to 100 pm, 100 pm to 200 pm, 200 pm to 300 pm, 300 pm to 400 pm, 400 pm to 500 pm pm, 600 pm to 700 pm, 700 pm to 800 pm, 800 pm to 900 pm or 900 pm to 1000 pm. A person skilled in the art recognizes that the aforementioned range limits can also be combined in order to obtain further preferred ranges, such as, for example, 10 pm to 200 pm, 50 pm to 300 pm or also 100 pm to 600 pm.

In a preferred embodiment, the thickness of the vertical sections is between 100 nm and 10 μm, preferably between 500 nm and 5 μm. Intermediate ranges from the aforementioned ranges can also be preferred, such as, for example, 100 nm to 500 nm, 500 nm to 1 pm, 1 pm to 1.5 pm, 1.5 pm to 2 pm, 2 pm to 3 pm, 3 pm to 4 pm , 4 pm to 5 pm, 5 pm to 6 pm, 6 pm to 7 pm, 7 pm to 8 pm, 8 pm to 9 pm or 9 pm to 10 pm. A person skilled in the art recognizes that the aforementioned range limits can also be combined in order to obtain further preferred ranges, such as, for example, 500 nm to 3 pm, 1 pm to 5 pm or also 1500 nm to 6 pm.

In a preferred embodiment, the length of the vertical sections is between 10 μm and 10 mm, preferably between 100 μm and 1 mm. Also intermediate areas from the The above ranges can be preferred, such as 10 pm to 100 pm, 100 pm to 200 pm, 200 pm to 300 pm, 300 pm to 400 pm, 400 pm to 500 pm, 500 pm to 1000 pm, 1 mm to 2 mm, 3 mm to 4 mm, 4 mm to 5 mm, 5 mm to 8 mm or also 8 mm to 10 mm. A person skilled in the art recognizes that the aforementioned range limits can also be combined in order to obtain further preferred ranges, such as, for example, 10 μm to 500 μm, 500 μm to 5 μm or also 1 mm to 5 mm.

With the aforementioned preferred dimensions of the vibratory membrane or the vertical sections, a particularly compact MEMS transducer, in particular MEMS loudspeaker or MEMS microphone, can be provided which at the same time combines high performance with excellent sound image or audio quality.

In a preferred embodiment of the invention, the oscillatable membrane is formed by a meandering structure with alternating vertical and horizontal sections, with holding structures attached to at least two of the horizontal sections, which are connected directly or indirectly to the carrier. The holding structures can, for example, be provided by the substrate material of the carrier, i.e. the holding structures can be formed directly from the substrate of a bottom wafer. Alternatively, it is also possible for the holding structures to be connected to the horizontal sections as separate ridges or elevations of a top wafer.

The holding structures can preferably be present on one and / or two sides of the oscillatable membrane, i.e. preferably attached to upper and / or lower horizontal sections.

In particular when a larger-sized vibratory membrane is suspended between the side walls of a carrier, the use of holding structures advantageously allows stabilization without negatively affecting the generation or absorption of sound.

Since the horizontal sections in a meandering shape are at least essentially mechanically neutral, locking them by means of the holding structure advantageously does not lead to any undesirable stresses between the membrane and the holding structure or carrier.

Various layers can be provided for the construction of the vibratable membrane in order to ensure the described actuation and excitation of horizontal vibrations or their detection.

Contacting one or more actuator layer (s) and / or one or more layers made of a mechanical support material and thus the application or detection of an electrical voltage can take place directly via the end electrodes or be supported by a layer made of a conductive material,

Molybdenum, titanium, tantalum, titanium-tungsten alloy, metal silicide, aluminum, graphite and copper. In a preferred embodiment, the vibratory membrane comprises three layers, an upper layer being formed from a conductive material and connected to an upper electrode, a middle layer being formed from the actuator material and a lower layer being formed from a conductive material.

The conductive material of the upper and / or lower layer can preferably be a mechanical support material, so that this layer has a double function. On the one hand, the location ensures that the actuator layer is contacted with an electrical potential that can be applied to the electrodes at the end. On the other hand, it functions as a mechanical support layer in the manner described for generating horizontal curvatures or vibrations when the actuator position is actuated accordingly.

Such an embodiment can be obtained by means of simple manufacturing steps, as illustrated by way of example in FIG. In the preferred embodiment shown in FIGS. 2G, 3 and 4, the vibratable membrane has a meandering structure with a continuous upper layer made of a conductive material (metal), a continuous middle layer made of an actuator material and a lower layer made of a conductive mechanical material Support material. A reverse sequence of the layers or a further additional conductive layer in contact with the mechanical support layer and / or actuator layer for improved contacting can also be provided.

In a further preferred embodiment, the vibratory membrane comprises two layers made of an actuator material, which are separated by a middle layer made of a conductive material, preferably metal, the middle layer being connected to a first electrode and at least one of the two layers made of an actuator material another layer made of a conductive material, preferably a metal, is present in contact with a second electrode.

As explained above, in a preferred embodiment, two actuator positions can also be used in order, for example, to set the vertical sections into horizontal oscillations by means of different actuation. In order to transfer the electrical potential changes from the end electrodes to the respective actuator layer, two or more intermediate layers made of a conductive material can preferably be provided. The layers made of conductive material, for example made of a metal, in this case preferably serve exclusively for contacting and not as a mechanical support layer. The voltage required in the sense of a bimorph for a MEMS loudspeaker for bulging or oscillation is induced by a different control of the actuator positions themselves.

The layers made of a conductive material, such as metal, for example, can therefore be made particularly thin (less than 500 nm, preferably less than 200 nm).

Such a preferred embodiment is shown by way of example in FIG. This has an oscillatable membrane as a meandering structure with two layers made of an actuator material, which are separated by a middle layer made of a conductive material (metal). The middle layer is connected to a first electrode pad at the end, while the upper actuator layer is connected to a second end-side via a further layer made of a conductive material Electrode is in contact. A lower layer made of a conductive material is not in contact with any of the electrodes. A reverse order of the layers or a waiver of the lower layer made of conductive material, which is in no contact with the electrodes, can also be provided.

In the embodiment described above, it is preferred that the actuator layer (s) and possibly the mechanical support layers are continuous, ie in cross section from one end of the membrane (at which a first electrode is preferably present) over several alternating horizontal and vertical sections, up to a second end of the membrane (at which preferably a second electrode is present).

The inventors recognized that providing a mechanical biomorph in the vertical sections is sufficient for the operating principle of the MEMS transducer, preferably a MEMS loudspeaker.

In a preferred embodiment, the at least one actuator layer is not continuous, but is only present in the vertical sections, but not in the horizontal sections. In this case, it can be preferred that any mechanical support layer that may be present runs continuously or that it does not run continuously and is only provided in vertical sections, for example. In order to still be able to contact the vertical sections by means of electrodes at the end, it is preferred to apply one or more continuous layers of a conductive material (preferably metal).

A preferred production method for an embodiment with a discontinuous actuator position is illustrated in FIG. Here, a targeted spacer etching of the actuator layer can take place in horizontal sections, so that only the vertical sections of the membrane still have a layer made of an actuator material. A continuous layer made of a mechanical support material can at the same time be dielectric in order to avoid a short circuit between an upper and lower conductive layer (also referred to as top and bottom electrodes).

The embodiment is characterized by a particularly effective actuation and high power performance, in which only the vertical sections are specifically stimulated to alternate arching or swinging, while the horizontal sections remain mechanically neutral. The displaced volume can advantageously be increased again per phase of the actuation.

In the embodiments described above, an oscillatable membrane in a meandering shape is preferably obtained by applying or etching correspondingly functional layers.

Alternatively, a vibratory membrane can also be produced by providing vertical sections and connecting them by means of metal bridges.

In a preferred embodiment, the vertical sections of the vibratable membrane comprise two layers, a first layer consisting of an actuator material, a second layer consisting of a conductive support material and the vertical sections being connected via horizontal metal bridges. As illustrated in FIG. 10, a plurality of individual piezoceramic elements comprising a layer made of a mechanical support material and a layer made of a piezoelectric material, as well as a sacrificial layer, can preferably be provided for this purpose. A membrane with a high degree of efficiency can advantageously be obtained in a robust and process-efficient manner through several method steps comprising a through-hole plating and metal filling as well as stacking and dicing of the piezoceramic elements.

In the embodiment, a continuous, homogeneous conductive layer is not necessary. Instead, the contacting of the actuator layer in the vertical sections is ensured by the metal bridges and a conductive mechanical support material.

In a preferred embodiment, the vibratable membrane is coated with a layer made of a non-stick material. With non-stick materials are meant, in particular, materials with low surface energies, which are largely inert to the environment and thus avoid the deposition of dust or other undesirable particles. For example, the non-stick materials can be formed by carbon layers, e.g. diamond-like carbon (DLC) layers or also layers comprising perfluorocarbons (PFC), such as polytetrafluoroethylene (PTFT).

In a preferred embodiment of the invention, the MEMS converter, preferably a MEMS loudspeaker, comprises a control unit which is configured to control the at least one electrode so that the two or more vertical sections are excited to horizontal oscillations. The control unit is preferably configured to control the electrodes, which ensures a frequency of the horizontal oscillations between 10 Hz and 20 kHz.

In a preferred embodiment of the invention, the MEMS converter, preferably a MEMS microphone, comprises a control unit which is configured to detect an electrical signal provided by the at least one electrode which was generated by horizontal oscillations of the two or more vertical sections. The control unit of a MEMS microphone is preferably configured for picking up and processing an electrical signal which corresponds to a frequency of the horizontal oscillations between 10 Hz and 20 kHz and is therefore set up for sound detection in the audible range.

The control unit is therefore preferably configured and set up to control the vibratable membrane (or the actuator position (s) in the vertical sections) to horizontal vibrations and sound emission in the audible frequency range by means of electrical signals, or a corresponding electrical signal when the vibratable membrane is excited record and process.

For the MEMS loudspeaker, the vertical sections of the membrane are preferably controlled with audio signals. In contrast to the combined control of separate membrane units and a plurality of valves according to US 2019/011 64 17 A1, the control for generating sound is therefore significantly simplified.

For the purpose of generating or receiving electrical signals, the control unit can preferably comprise a data processing unit. In the context of the invention, a data processing unit preferably denotes a unit which is suitable and configured for receiving, sending, storing and / or processing data, preferably with a view to controlling the electrodes or receiving an electrical signal provided at the electrodes. The data processing unit preferably comprises an integrated circuit, for example also an application-specific integrated circuit, a processor, a processor chip, microprocessor or 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 the data.

In preferred embodiments, the control unit is integrated in addition to further components of the MEMS transducer (carrier, vibratable membrane) on a printed circuit board or printed circuit board. This means that the MEMS converter is preferably seamlessly integrated with the electronics required for control or detection. In addition to the control unit, other electronic components, such as a communication interface (preferably wireless, e.g. Bluetooth), an amplifier, a filter or a sensor system, can also be installed on one and the same printed circuit board.

A compact overall solution is advantageously obtained in which a MEMS converter, preferably a MEMS loudspeaker or MEMS microphone, can be provided together with the desired electronics in a very small space and preferably with inexpensive CMOS processing suitable for mass production.

In a further preferred embodiment, the vibratable membrane held by the carrier is arranged in a front side of a housing which encloses a rear-side resonance volume. The sound emission of such a MEMS loudspeaker is therefore preferably carried out towards the open front side (sound port), the sound image being improved in particular for lower frequencies by the resonance volume on the rear side.

In a further preferred embodiment, there is a ventilation opening in the housing to avoid acoustic short circuits and / or to support the sound image. The ventilation opening is preferably small compared to the sound port and can, for example, have a maximum dimension of less than 100 μm, preferably less than 50 μm.

In a further aspect, the invention relates to a manufacturing method for a MEMS converter, preferably a MEMS loudspeaker or MEMS microphone, as described above, comprising the following steps:

Etching of a substrate, preferably from a front side, to form a structure, preferably a meander structure

Optional application of an etch stop

Application of at least two layers, wherein at least first layer comprises an actuator material and a second layer comprises a mechanical support material or at least two layers comprise an actuator material

Contacting the first and / or second layer with an electrode Etching, preferably from the back, and optional removal of the etch stop, so that an oscillatable membrane, preferably in the form of a meander structure, is held by a carrier (4) formed by the substrate (8), the oscillatable membrane (1) for generating or Recording of pressure waves of the fluid in a vertical direction comprises at least two or more vertical sections (2) which are formed parallel to the vertical direction and so that the two or more vertical sections can be excited to horizontal oscillations by controlling the at least one electrode, or so that at Excitation of the two or more vertical sections to horizontal oscillations at the at least one electrode, an electrical signal can be generated.

The average person skilled in the art recognizes that technical features, definitions and advantages of preferred embodiments of the described MEMS transducer, preferably MEMS loudspeakers or MEMS microphones, also apply to the described manufacturing method and vice versa. The production method described is preferably used to provide a MEMS transducer with a folded, oscillatable membrane with a meander structure. Examples of preferred manufacturing steps are described in FIGS. 2A-G, 8A-J or FIG. 9.

As a substrate, for. B. one of the preferred materials mentioned above can be used. During the etching, a blank, for example a wafer, can be brought into the desired basic shape of the meander structure. In a next step, the layers for the vibratable membrane are preferably applied.

Applying the at least one layer of a conductive material preferably also includes applying a plurality of layers and in particular a layer system in addition to applying one layer. A layer system comprises at least two layers that are systematically applied to one another. The application of a layer or a layer system preferably serves to define the vibratable membrane comprising vertical sections which can be excited to horizontal vibrations.

The application can preferably 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). In particular, the application, for example, a deposition, z. B. in the case of a substrate made of polysilicon.

Etching and / or structuring can preferably be selected from the group comprising dry etching, wet chemical etching and / or plasma etching, in particular reactive ion etching, reactive ion deep etching (Bosch process).

In a preferred embodiment, the etching of a substrate, preferably from a front side, is characterized in that the substrate has a crystal structure and a plurality of pockets (trenches) are carried out by etching along a lattice vector of the crystal structure. Preferably the pockets are defined as parallel slots from the front of the substrate. After applying the functional Layers and corresponding rear side processing results in an oscillatable membrane as a bellows with a meandering structure in cross section (cf., inter alia, FIGS. 2 and 8).

The preferred etching along the orientation of a crystal substrate can advantageously result in smooth, quasi-crystalline pockets with a great depth of more than 200 μm, 300 μm, 400 μm, 500 μm or more with high-precision orientation and negligible roughness.

It is also advantageous that the surface normal of the side surfaces of the pockets is also aligned with the crystal structure, preferably an orthogonal lattice vector.

When applying a layer of an actuator material, preferably a piezoelectric material, to a substrate structured in this way, the orientation of the actuator material can also take place in a quasi-crystalline manner. In particular, piezoelectric materials such as AIN, AIScN or PZT advantageously have a columnar growth on the side walls of the pockets oriented in this way, which can ensure that the piezoelectric layer has a particularly precise c-axis orientation, which is perpendicular to the surface of the vertical sections of the resulting membrane is present.

The formation of horizontal vibrations through the transverse piezoelectric effect can therefore take place particularly effectively and precisely and ensure an improved sound image in the case of a MEMS loudspeaker or detection capability in the case of a MEMS microphone.

In a preferred embodiment, the etching of a substrate, preferably from a front side, to form a structuring, preferably a meander structure, is characterized in that the substrate has a crystal structure and a plurality of pockets (trenches) along a lattice vector are at least partially carried out by wet chemical etching , preferably anisotropic etching depending on the crystal orientation takes place.

For this purpose, an etchant is preferably used which has a clearly different etching rate in relation to the crystal orientation of the substrate in two orthogonal crystal orientations. For example, an etching rate can be higher by a factor of 50, 10, 150, 200 or more for the selected substrate in a first crystal orientation than in a second crystal orientation orthogonal thereto.

The substrate is preferably oriented in such a way that the first crystal orientation for which there is an increased etching rate is aligned with the surface normal of the substrate surface. Areas on the substrate surface which should not be etched can be defined by means of an etching mask. The etching mask can preferably define a frame in which slots or strips for forming the pockets remain free. Areas which remain between the parallel pockets to be formed can serve as a substrate for the horizontal sections of the membrane.

The anisotropic wet-chemical etching then preferably results in etching perpendicular to the substrate surface in order to form deep vertical pockets. The etching in an orthogonal (horizontal) direction is therefore reduced. The greater the anisotropy factor of the crystal orientation-dependent etching, the less underetching will be. Particularly good results can be achieved, for example, with potassium hydroxide (KOH) as an etchant for a silicon crystal substrate. For example, potassium hydroxide has a clear preferred direction for etching along a <110> orientation of a silicon crystal versus a <111> orientation. As in Sato et al. As shown in 1988, the etching rate for KOH on a silicon monocrystal in a <110> direction at 1.455 pm / min can be a factor of 291 higher than in an orthogonal <111> orientation (etching rate 0.005 pm / min).

9 illustrates how, by appropriate alignment of a silicon crystal, almost perfectly smooth and deep pockets can be produced in a robust manner, the side surfaces of which are crystal-oriented in order to ensure c-axis-oriented growth of piezoelectric materials.

The person skilled in the art understands that alternative etchants that depend on the crystal orientation, such as, for example, tetramethylammonium hydroxide (TMAH), can equally be used (cf. Seidel et al 1990, among others).

The process is advantageously not only suitable for upscaling for mass production. In addition, diaphragms capable of oscillating in this way in a meandering shape are also characterized by a particularly precise alignment of the vertical sections, which lead to improved oscillation behavior and therefore sound generation or detection.

Should a further structuring of the vibratable membrane be desired, this can be done, for. B. be made by further etching processes. Additional material can also be deposited or doping can be carried out using conventional methods.

For contacting the layers, suitable material, such as. B. copper, gold and / or platinum can be deposited by common processes. Physical vapor deposition (PVD), chemical vapor deposition (CVD) or electrochemical deposition can preferably be used for this purpose.

By means of the process steps, a finely structured vibratory membrane with a desired definition of vertical and horizontal sections can be provided, which are preferably suspended between two side areas of a stable carrier and have dimensions in the micrometer range. The manufacturing steps are part of standard method steps in semiconductor processing, so that they have proven themselves and are also suitable for mass production.

In a further aspect, the invention therefore also relates to a MEMS converter that can be produced by the production method as described above.

The person skilled in the art recognizes that special features of the manufacturing steps, such as, for example, a crystal orientation-dependent etching to form deep pockets with a quasi-crystalline smooth surface, are directly transferred to structural features of the MEMS change. In the case of a quasi-crystalline smooth surface of the side surfaces of the pockets, as explained above, an oscillatable membrane with a plurality of vertical sections in a meandering shape can be formed in a particularly precise manner. A c-axis orientation of an actuator material, preferably a piezoelectric material, can also follow directly from the application of the preferred manufacturing steps. In a further aspect, the invention relates to a manufacturing method for a MEMS transducer as described above, comprising the following steps:

Providing a plurality of individual piezoceramic elements, comprising a sacrificial layer, a layer made of a conductive material and a layer made of a piezoelectric material

Definition of holes for through-contacts in the piezoceramic elements and metal filling

Stacking the piezoceramic elements and optional cutting (dicing) so that a stack of piezoceramic elements is obtained, which is connected by metal bridges

Removal of the sacrificial layer and introduction of the stack of piezoceramic elements into a carrier, the piezoceramic elements being contacted with one electrode each, so that an oscillatable membrane, preferably in the form of a lamellar structure, is held by a carrier formed by the substrate, the oscillatable membrane for Generating or receiving pressure waves of the fluid in a vertical direction comprises at least two or more vertical sections, which are formed parallel to the vertical direction and so that by controlling the at least one electrode, the two or more vertical sections can be excited to horizontal oscillations or so when excited of the two or more vertical sections for horizontal oscillations at the at least one electrode, an electrical signal can be generated.

The average person skilled in the art recognizes that technical features, definitions and advantages of preferred embodiments of the described MEMS converter, preferably a MEMS loudspeaker or MEMS microphone, also apply to the described manufacturing method and vice versa. The production method described is preferably used to provide a MEMS transducer with an oscillatable membrane with a lamellar structure, the lamellae being mechanical bimorphs and being connected by metal bridges. Examples of preferred manufacturing steps are illustrated in FIGS. 10A-F and FIG.

In the alternative manufacturing process, several individual piezoceramic elements can advantageously be used in order to obtain an oscillatable membrane with lamellae as vertical sections, which are connected to one another by metal bridges, by means of a definition of holes, metal filling and stacking and dicing.

Piezoceramics are preferably ceramic materials which, when deformed by an external force, show charge separation or undergo a change in shape when an electrical voltage is applied. The piezoceramic elements preferably comprise a piezoelectric layer and a layer made of a mechanical support material, as described above, and also a sacrificial layer.

The sacrificial layer is used to process and provide the metal bridges and will not itself be part of the vibratory membrane. The sacrificial layer can preferably be a photoresist or photoresist, for example. These materials change their solubility when exposed to light, especially UV light. In particular, it can be a so-called positive varnish, the solubility of which increases as a result of UV irradiation. In this way, the sacrificial layer can be removed in a targeted manner after a metal filling to provide the metal bridges.

In a further aspect, the invention relates to a manufacturing method for a MEMS converter comprising the following steps:

Providing a plurality of individual piezoceramic elements, comprising a layer made of a mechanical support material, which is electrically conductive, and a layer made of a piezoelectric material

Provision of an upper and lower frame with recesses for the several individual piezoceramic elements

Fixation of the piezoceramic elements in the recesses of the upper and lower frame, preferably by means of an adhesive

Applying at least one continuous electrically conductive layer for contacting the piezoceramic elements by means of at least one electrode so that an oscillatable membrane, preferably in the form of a lamellar structure, is held by a carrier, which is formed by the upper and lower frame and where the oscillatable membrane is used to generate or receive of pressure waves of the fluid in a vertical direction comprises at least two or more vertical sections, which are formed parallel to the vertical direction, so that the two or more vertical sections can be excited to horizontal oscillations by controlling the at least one electrode or so that when the two or more vertical sections to horizontal oscillations at which at least one electrode an electrical signal can be generated.

The preferred embodiment is illustrated in FIG. Structured contacting is advantageously dispensed with in the embodiment. Instead, contact is made by means of a continuous conductive surface from a front side and / or a rear side of the MEMS converter.

In a preferred embodiment, the upper and lower frames are formed from an electrically non-conductive material, for example a polymer. A 3D printing process can preferably be used to shape the frames.

To make contact with the individual lamellae or piezoceramic elements, a continuous layer of a conductive material, preferably metal, is preferably applied from the front (front electrode) or from the rear (backside electrode). The application can take place, for example, using a sputtering process.

The average person skilled in the art recognizes that technical features, definitions and advantages of preferred embodiments of the described MEMS converter, preferably a MEMS loudspeaker or MEMS microphone, also apply to the described manufacturing method and vice versa. The production method described is preferably used to provide a MEMS transducer with an oscillatable membrane with a lamellar structure, the lamellae being mechanical bimorphs and being connected by a continuous layer made of a conductive material, preferably metal.

Detailed description

The invention is to be explained below with reference to further figures and examples. The examples and figures serve to illustrate preferred embodiment of the invention without restricting it.

Brief description of the images

Fig. 1 Schematic representation of a cross section of a preferred

Embodiment of a MEMS loudspeaker according to the invention A: at rest and B: during actuation.

Fig. 2 Schematic representation of a preferred manufacturing method for a

MEMS loudspeaker with an oscillating membrane, which has a meander shape in cross section.

3, 4 Schematic representation of preferred embodiments of a MEMS loudspeaker with an oscillatable membrane in a meandering shape, the horizontal sections of which are supported by holding structures.

Fig. 5 Schematic representation of a preferred embodiment of a MEMS

Loudspeaker with two actuator layers, which are separated by a middle layer made of a conductive material.

Fig. 6 Schematic representation of preferred controls for operating the MEMS

speaker

Fig. 7 Schematic representation of a preferred integration of a MEMS

Loudspeaker in the front of a housing with a resonance volume on the back.

Fig. 8 Schematic representation of a preferred manufacturing method for a

MEMS loudspeaker with an oscillatable membrane which has a meandering shape in cross section, only the vertical sections having a layer made of an actuator material.

9 Schematic representation of a preferred structuring of a substrate in

Crystal shape for the formation of deep trenches (pockets) by means of an etching process that depends on the crystal orientation

Fig. 10 Schematic representation of a preferred manufacturing method for a

MEMS loudspeaker with an oscillating membrane based on individual piezoceramic elements. 11 Schematic representation of a preferred electrical contacting of a

MEMS loudspeaker with an oscillating membrane based on individual piezoceramics.

Fig. 12 Schematic representation of a preferred manufacturing method for a

MEMS loudspeaker with an oscillating membrane based on individual piezoceramic elements.

Detailed description of the images

FIG. 1 illustrates a preferred embodiment of a MEMS loudspeaker according to the invention. FIG. 1 A shows an idle state, while FIG. 1 B illustrates two phases during the actuation of the MEMS loudspeaker.

The MEMS loudspeaker comprises an oscillatable membrane 1 for generating sound waves in a vertical emission direction, the oscillatable membrane 1 being held in a horizontal position by a support 4. The oscillatable membrane 1 has a meandering structure in cross section with horizontal 3 and vertical sections 2. The vertical sections are formed parallel to the emission direction and have at least one actuator layer, for example a layer made of a piezoelectric material. The vibratable membrane 1 and the actuator layer are preferably contacted by means of electrodes at the end. In addition, an electrode pad (not shown) can be located on the carrier 4, for example.

The vertical sections are preferably mechanical bimorphs, which can be excited to horizontal vibrations by suitable controls. For this purpose, the vertical sections 2 can comprise, for example, a first layer made of an actuator material and a second layer made of a mechanical support material. By controlling the actuator position, a stress gradient and thus a curvature or curvature can be created.

Vibration can be generated. Likewise, it can also be preferred that the vertical sections 2 comprise two actuator layers which are activated in opposite directions in order to bring about a curvature of the vertical sections 2 by means of a corresponding relative change in shape.

FIG. 1B illustrates, by way of example, two phases during an actuation. Advantageously, due to the plurality of vertical sections 2 of the vibratable diaphragm 1 with small horizontal movements (curvature) of a few micrometers, an increased total volume can be moved in the vertical emission direction and thus used to generate sound. The actuation allows a particularly efficient implementation, since during a phase almost the entire volume of air between the vertical sections can be moved up or down along the emission direction.

FIG. 2 schematically shows a preferred production method for providing a MEMS loudspeaker with an oscillatable membrane 1 which has a meander shape in cross section. A vibratory membrane with a meandering shape in cross section can also preferably be referred to as a folded membrane or bellows. 2A shows an etching of the substrate 8 from an upper or front side in order to form a structure. In the method step, parallel deep trenches (pockets) are etched into the substrate 8. The molded structure represents a bellows or a meander in cross section.

A layer of an etch stop 9 (FIG. 2B) is then applied, which may be TEOS or PECVD, for example. A layer of a mechanical support material 10 (FIG. 2C) and a layer of an actuator material 11 are applied to the etch stop 9. The mechanical support material 10 can be doped polysilicon, for example, while a piezoelectric material is used for the actuator material 10, for example can. For example, 1 μm may be preferred as layer thicknesses. The piezoelectric material can preferably have a C-axis orientation perpendicular to the surface, so that a transverse piezoelectric effect is used. Other orientations and, for example, the use of a longitudinal effect can also be preferred.

FIG. 2E shows the preferred application of a full-area top electrode as a layer made of a conductive material 12. End-side contacting can take place, for example, by means of an electrode pad 13 (FIG. 2F).

FIGS. 2F and 2G illustrate a further etching of the substrate 8 from the rear side or

Underside, and the removal of the etch stop.

The manufacturing steps 2A-G thus obtain an oscillatable membrane 1 which has a meandering structure in cross section. A continuous actuator layer 11 and the provision of end-side contacts 13 advantageously allow the vertical sections 2 to be efficiently actuated to produce horizontal oscillations (cf. FIG. 1). As can be seen in FIG. 2G, control is preferably carried out by means of two electrodes, so the actuator layer 12 is preferably contacted both from a front side (top electrode, conductive layer 12) and from a rear side (bottom electrode, via conductive mechanical support material 10) (see Fig. 6A).

To stabilize the membrane 1 suspended between the side walls of the carrier 4, retaining structures 14 can be provided. As shown in FIGS. 3 and 4, these preferably horizontal sections 3 of the vibratable membrane 1 can be supported. The horizontal sections 3 are advantageously mechanically neutral (cf. FIG. 1 B), so that no undesired stresses are induced between the membrane 1 and the holding structure 14 or carrier 4 during the actuation.

5 illustrates a preferred alternative embodiment of a MEMS loudspeaker, the vibratable membrane 1 comprising two actuator layers which are separated by a middle layer made of a conductive material 12, preferably metal. The middle layer is connected to a first electrode pad 13 at the end, while in the embodiment shown the upper actuator layer 11 is in contact with a second electrode pad 13 at the end via a further layer made of a conductive material 12.

6 illustrates preferred controls for operating the MEMS loudspeakers described. A preferred control for a MEMS loudspeaker with an actuator layer 11 and a passive mechanical support layer 10 is shown in FIG. 6A. The control is preferably carried out by means of two electrode pads 13 at one end, so that the horizontal vibrations can be generated by changing the shape of the actuator material in relation to the mechanical support material. The actuator layer 11 is preferably contacted both from a front side (top electrode 13, conductive layer 10) and from a rear side (bottom electrode 13, conductive mechanical support material 10). An alternating voltage as an audio input signal can, for example, be applied to the front-side electrode pad 13 (left), while the rear-side electrode pad 13 (right) is grounded.

6B shows a preferred control for a MEMS loudspeaker with two actuator layers 11, which are separated by a middle layer made of a conductive material 12, preferably metal.

An upper actuator layer 11 is preferably controlled from a front side (top electrode 13 and upper conductive layer 12) and the middle conductive layer 12. A lower actuator layer 11 is preferably controlled from a rear side (bottom electrode 13 and lower conductive layer 12) and the middle conductive layer 12. In the illustrated embodiment, an alternating voltage can be applied as an audio input signal, for example, to the one for the top and bottom electrode pads 13 (left), while the middle layer 12 is grounded via a further electrode pad 13 (right).

7 shows an example of a preferred integration of a MEMS loudspeaker according to the invention in a housing 15. The vibratable membrane 1 held by the carrier 4 is preferably arranged in a front side or front side of a housing (sound port). The housing also encloses a rear resonance volume (back volume 16). A ventilation opening 17 can be introduced to avoid acoustic short circuits or to support the sound image.

FIG. 8 illustrates an alternative production method for providing a MEMS loudspeaker with an oscillatable membrane 1 according to the invention. The method steps shown in FIGS. 8A-D are analogous to FIG. 2.

8A shows an etching of the substrate 8 from an upper or front side in order to form a structure, preferably a meander structure. In the method step, parallel deep trenches (pockets) are etched into the substrate 8. The molded structure represents a bellows or a meander in cross section.

A layer of an etch stop 9 (FIG. 2B) is then applied, which may be TEOS or PECVD, for example. A layer of a mechanical support material 10 (FIG. 2C) and an actuator material 11 is applied to the etch stop 9. The mechanical support material 10 can be, for example, doped polysilicon, while a piezoelectric material is preferably used for the actuator material 12.

In contrast to the embodiment shown in FIG. 2, the actuator layer 11 is not contacted as a continuous layer with an upper conductive layer. Instead, the actuator layer 11 is etched with a spacer (FIG. 8F) in the horizontal sections of the membrane, so that only the vertical sections of the membrane have a layer made of an actuator material 11.

A continuous dielectric layer 18 is then preferably applied to avoid a short circuit between the upper and lower electrodes to be applied later (FIG. 8G). A continuous conductive layer as top electrode 12 allows front-side contacting (FIG. 8H).

8 I and 8 J illustrate a further etching of the substrate 8 from the rear side or underside and optionally the application of a continuous conductive layer 12 as a rear side electrode.

FIG. 9 illustrates a preferred provision of a structured substrate 8. Analogously to the method step shown in FIG. 8 a, parallel deep trenches (pockets) are etched into the substrate 8. The molded structure represents a bellows or, in cross section, a meander, on which an oscillatable membrane can be applied in a meander shape.

The preferred provision of the structured substrate 8 in FIG. 9 is characterized by the utilization of a crystal structure of the substrate 8, the pockets being formed along a lattice vector of the crystal structure.

In this way, particularly smooth, quasi-crystalline pockets with a great depth of more than 200 μm, 400 μm or more with high-precision orientation can be obtained. It is also advantageous that the surface normal of the side surfaces of the pockets is aligned with a grid vector which is orthogonal to the grid vector in the direction of which the etching process took place.

If, for example, silicon is used as the substrate, the silicon substrate 8, as shown in FIG. 9, can preferably be aligned with a surface orientation of the Miller indices <110>. The lattice vector of the crystal structure <110> is therefore preferably perpendicular to the surface of the as yet unstructured substrate. By means of an etching mask 24, for example an S1O2 hard mask, horizontal areas or stripes can be defined on the substrate surface which should not be etched.

Anisotropic etching with a preferred direction along the <110> orientation of the silicon crystal versus a <111> orientation results in smooth and precisely oriented pockets. For this purpose, wet chemical processes can advantageously be used, which are suitable for mass production in a batch process. For example, potassium hydroxide has a clear preferred direction for etching along the <110> versus a <111> crystal orientation. We in Sato et al. As shown in 1988, the etching rate for KOH on a silicon monocrystal is <110> 1.455 pm / min, while the etching rate in the <111> orientation is only 0.005 pm / min. Due to the anisotropic etching rates, deep pockets with little undercutting can be achieved by means of the wet chemical process.

To form pockets 400 μm deep, for example, KOH can be applied to a <110> oriented silicon substrate for 275 min. Due to the etching rate reduced by a factor of 291 in the orthogonal <111> orientation, only an underetching of 1.37 pm will take place during this period. Even a variation in the local strength of the underetching process results in relation to the great depth of the pockets of 400 pm Orientation fluctuations of well below 1 °. Instead, the process can achieve almost perfectly vertical deep pockets with high accuracy, which are characterized by a smooth, quasi-crystalline orientation.

As a further advantage, the side walls of the pockets on which the vertical sections of the membrane are formed are in a crystal orientation (here: <111>). The fact favors a columnar growth of piezoelectric materials, such as AIN or PZT: This makes it possible to ensure in a particularly precise way that the piezoelectric material has a c-axis orientation perpendicular to the surface of the vertical sections, so that a transverse piezoelectric effect can be used.

10 illustrates a preferred manufacturing method for providing a MEMS loudspeaker with an oscillatable membrane based on individual piezoceramics.

First, several individual piezoceramic elements 19, comprising a layer made of a mechanical support material 10 (e.g. doped polysilicon) and a layer made of a piezoelectric material 11 and a sacrificial layer 20 are provided (see FIGS. 10A and 10B). The sacrificial layer 20 can be, for example, a photoresist (photoresist). Preferably, the layer made of a mechanical support material 10 can be designed to be electrically conductive in order to ensure contact. It is also possible to apply one or two layers of a conductive material 12 to the one layer of a piezoelectric material 11, which are used to make electrical contact with the piezoelectric material.

This is followed by a definition of holes for a through-hole plating and metal filling 21 (cf. FIG. 10 C). The piezoceramic elements 19 are stacked (FIG. 10 D) and cut (dicing 22, FIG. 10E), so that two or more stacks of Piezoceramic elements 19 are obtained, which are connected by metal bridges 21 (see. Fig. 10E).

After the sacrificial layer 20 (FIG. 10F) has been removed, the stacked piezoceramic elements 19 are introduced into a carrier 4, the first and last piezoceramic elements preferably being contacted with one electrode 13 each (FIG. 10E).

In this way, an oscillatable membrane 1 is also obtained between a carrier 4, which comprises at least two or more vertical sections 2 for generating sound waves in a vertical emission direction, which are formed parallel to the emission direction and can be excited to horizontal oscillations.

Here too, the actuator principle is preferably based on a relative change in shape of the actuator layer 11 compared to the mechanical support layer 10. A continuous actuator layer is not necessary for this. The metal bridges 23 in combination with conductive layers 12 ensure that all vertical sections 2 are contacted by end-side control.

11 illustrates a preferred electrical contacting of the MEMS loudspeaker with an oscillatable membrane based on individual piezoceramics. Figure 11A is a top view and Figure 11B is a side view of the MEMS speaker. Individual lamellae or vertical sections are controlled in parallel via the electrode pads 13, with U-shaped spacers being present on each side of the lamellae and creating a mechanical and electrical connection to the next lamella.

FIG. 12 illustrates an alternative production method for providing a MEMS loudspeaker with an oscillatable membrane based on individual piezoceramics.

In contrast to the embodiment according to FIGS. 10 or 11, structured contacting can advantageously be dispensed with in the embodiment shown. Instead, as explained below, contact can be made by means of a continuous conductive surface from the front (front electrode) or from the rear (backside electrode).

Analogously to the production method according to FIG. 10, several individual piezoceramic elements 19, comprising a layer made of a mechanical support material 10 (e.g. doped polysilicon) and a layer made of a piezoelectric material 11, are provided. The layer made of a mechanical support material 10 is preferably designed to be electrically conductive.

Furthermore, an upper frame 25 and a lower frame 26 are provided, which have recesses or grooves 27 for receiving the piezoceramic elements 19. The upper and lower frames are preferably made of an electrically non-conductive material, for example a polymer. A 3-D printing process can preferably be used to shape the frames.

To fix the piezoceramic elements 19, it can be preferred to use an adhesive. Which is preferably first introduced into depressions 27 (cf. FIG. 12A). After the piezoceramic elements 19 have been fastened in the respective depressions 27 of the lower frame 26, the adhesive can be applied to the piezoceramic elements 19 so that the upper frame fixes the piezoceramic elements 19 on the upper side (cf. FIG. 12B).

To make contact with the individual lamellae or piezoceramic elements 19, a continuous layer of a conductive material, preferably metal, (not visible) is preferably applied from the front (front electrode) or from the rear (backside electrode). For example, by means of a sputtering process.

In this way, an oscillatable membrane 1 can also be obtained, which comprises at least two or more vertical sections 2 for generating sound waves in a vertical emission direction, which are formed parallel to the emission direction and can be excited to produce horizontal oscillations. The composite frame 25, 26 can function as a support for the vertical sections 2. REFERENCE LIST

1 Vibrating membrane

2 vertical sections of the vibratable membrane

3 Horizontal sections of the vibratable membrane

4 carriers

5 volumes of air between the vertical sections

8 substrate

9 Etch stop

10 layer made of a mechanical support material, preferably doped polysilicon

11 layer made of an actuator material (actuator layer), preferably made of a piezoelectric material

12 layer made of a conductive material, preferably made of metal

13 Contacting the electrode, preferably electrode pad

14 holding structures

15 housing

16 rear resonance volume

17 ventilation opening

18 layer made of a dielectric material

19 piezoceramic element (s)

20 sacrifice

21 Defined holes for through-hole plating with metal filling

22 dicing

23 metal bridges

24 etching mask

25 Upper frame

26 Lower frame LITERATURE

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, TAIWAN, June 18-22, 2017.

Iman Shahosseini, Elie LEFEUVRE, Johan Moulin, Marion Woytasik, Emile Martincic, et al. Electromagnetic MEMS Microspeaker for Portable Electronic Devices. Microsystem Technologies, Springer Verlag (Germany), 2013, pp.10. <hal-01103612>.

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).

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 64 (1988) 87- 93).

Seidel, H., Csepregi, L., Heuberger, A. and Baumgartel, H. (1990). Anisotropy Etching of Crystalline Silicon in Alkaline Solutions. Journal of The Electrochemical Society 137.10.1149 / 1.2086277

Claims

PATENT CLAIMS

1. MEMS transducer for interaction with a volume flow of a fluid comprising

- a carrier (4),

- An oscillatable membrane (1) for generating or absorbing pressure waves of the fluid in a vertical direction, the oscillatable membrane (1) being held by the carrier (4), characterized in that the oscillatable membrane (1) has two or more vertical ones Sections (2) which are formed essentially parallel to the vertical direction and comprise at least one layer made of an actuator material (11), the vibratable membrane (1) being in contact with at least one electrode (13) at the end, so that by controlling the at least an electrode (13) the two or more vertical sections (2) can be excited to horizontal oscillations or so that when the two or more vertical sections (2) are excited to horizontal oscillations at the at least one electrode, an electrical signal can be generated.

2. MEMS converter according to the preceding claim, characterized in that the MEMS converter is a MEMS loudspeaker, preferably between the vertical sections (2) air volumes (5) are present, which are generated by the horizontal vibrations to generate sound waves along a vertical Emission direction are moved or the MEMS transducer is a MEMS microphone, air volumes (5) preferably being present between the vertical sections (2) which are moved along a vertical detection direction when sound waves are recorded.

3. MEMS transducer according to one of the preceding claims, characterized in that the two or more vertical sections (2) comprise at least two layers, wherein one layer (11) comprises an actuator material and a second layer (10) comprises a mechanical support material, wherein at least the layer (11) comprising the actuator material is in contact with an electrode (13) so that horizontal vibrations can be generated by a change in shape of the actuator material compared to the mechanical support material or so that horizontal vibrations lead to a change in shape of the actuator material compared to the mechanical support material and generate an electrical signal

4. MEMS transducer according to one of the preceding claims, characterized in that the two or more vertical sections (2) comprise at least two layers, both layers (11) comprising an actuator material and being in contact with one electrode (13) each the horizontal vibrations can be generated by changing the shape of one position in relation to the other position or horizontal vibrations lead to a change in shape of one position in relation to the other position and generate an electrical signal.

5. MEMS transducer according to one of the preceding claims, characterized in that the carrier (4) comprises two side areas between which the vibratable membrane

(1) is arranged in the horizontal direction and / or the carrier (4) is formed from 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.

6. MEMS transducer according to one of the preceding claims, characterized in that the oscillatable membrane (1) is formed by a lamellar structure or meander structure.

7. MEMS transducer according to one of the preceding claims, characterized in that the vibratory membrane (1) is formed by a meander structure with alternating vertical (2) and horizontal (3) sections, with holding structures (3) on at least two of the horizontal sections (3). 14) are attached, which are directly or indirectly connected to the carrier (4).

8. MEMS converter according to one of the preceding claims, characterized in that the actuator material comprises a piezoelectric material, a polymer piezoelectrical material and / or electroactive polymers (EAP), wherein the piezoelectric material is preferably selected from a group comprising lead-zirconate-titanate (PZT), aluminum nitride (AIN), aluminum scandium nitride (AIScN) and zinc oxide (ZnO).

9. MEMS transducer according to one of the preceding claims, characterized in that the vibratory membrane (1) comprises three layers, an upper layer (12) being formed by a conductive material, a middle layer (11) being formed by an actuator material and a lower layer (12) is formed from a conductive material, the conductive material of the upper and / or lower layer preferably being a mechanical support material.

10. MEMS converter according to one of the preceding claims, characterized in that the vibratable membrane (1) comprises two layers (11) made of an actuator material, which are separated by a middle layer (12) made of a conductive material, preferably metal the middle layer (12) being connected to a first electrode (13) and at least one of the two layers (11) made of an actuator material via a further layer (12) made of a conductive material, preferably a metal, with a second electrode (13) have been contacted.

11. MEMS transducer according to one of the preceding claims, characterized in that the vertical sections (2) of the vibratable membrane (1) comprise two layers, a first layer (11) consisting of an actuator material, a second layer (10) consisting of one There is conductive support material and wherein the vertical sections (2) are connected via horizontal metal bridges (23).

12. MEMS transducer according to one of the preceding claims, characterized in that the oscillatable membrane (1) is coated with a layer of a non-stick material.

13. MEMS transducer according to one of the preceding claims, characterized in that the vibratable membrane (1) held by the carrier (4) is arranged in a front side of a housing (15) which encloses a rear-side resonance volume (16), in which Housing (15) preferably has a ventilation opening (17) to avoid acoustic short circuits and / or to support the sound.

14. The manufacturing method for a MEMS transducer according to one of the preceding claims comprising the following steps:

Etching of a substrate (8), preferably from a front side, to form a structure, preferably a meander structure

Optional application of an etch stop

Application of at least two layers, at least one first layer (11) comprising an actuator material and a second layer (10) comprising a mechanical support material or at least two layers (11) comprising an actuator material

Contacting the first and / or second layer with an electrode (13)

Etching, preferably from the back, and optional removal of the etch stop, so that an oscillatable membrane (1), preferably in the form of a meander structure, is held by a carrier (4) formed by the substrate (8), the oscillatable membrane (1) comprises at least two or more vertical sections (2) for generating or receiving pressure waves of the fluid in a vertical direction, which are formed parallel to the vertical direction and so that the two or more vertical sections (2) are activated by controlling the at least one electrode (13) can be excited to horizontal vibrations or so that when the two or more vertical sections (2) are excited, they become horizontal Vibrations at the at least one electrode, an electrical signal can be generated.

15. Manufacturing method for a MEMS transducer according to one of the preceding claims 1 - 13 comprising the following steps:

Providing a plurality of individual piezoceramic elements (19), comprising a sacrificial layer (20), a layer (12) made of a conductive material and a layer (11) made of a piezoelectric material

Definition of holes (21) for through-contacts in the piezoceramic elements and metal filling

Stacking the piezoceramic elements (19) and optional cutting (dicing) (22) so that a stack of piezoceramic elements (19) is obtained which is connected by metal bridges (23)

Removal of the sacrificial layer (29) and introduction of the stack of piezoceramic elements (19) in a carrier (4), the first and last piezoceramic elements (19) being contacted with one electrode (13) each, so that an oscillatable membrane (1), preferably in the form of a lamellar structure, is held by a carrier (4) formed by the substrate (8), the vibratable membrane (1) for generating or absorbing pressure waves of the fluid in a vertical direction at least two or more vertical sections (2) which are formed parallel to the vertical direction and so that the two or more vertical sections (2) can be excited to horizontal vibrations by controlling the at least one electrode (13) or so that when the two or more vertical sections (2) are excited to horizontal ones Vibrations at the at least one electrode, an electrical signal can be generated.