WO2022180161A1

WO2022180161A1 - Mems acoustic transducer array

Info

Publication number: WO2022180161A1
Application number: PCT/EP2022/054647
Authority: WO
Inventors: Fabian LOFINK; Fabian STOPPEL; Malte Florian Niekiel; Bernhard Wagner
Original assignee: Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V.
Priority date: 2021-02-25
Filing date: 2022-02-24
Publication date: 2022-09-01
Also published as: DE102021201784A1; EP4298798A1

Abstract

A MEMS acoustic transducer array (10*) comprises at least two acoustic transducers (12, 13, 14), a first acoustic transducer (14) of the at least two acoustic transducers being designed to reproduce an amplitude characteristic of a first resonant frequency, and a second acoustic transducer (12) of the at least two acoustic transducers being designed to reproduce an amplitude characteristic of a second resonant frequency. The second resonant frequency is higher than the first resonant frequency. The second acoustic transducer has a quality factor Q ≥ 1 or ≥ 3.

Description

MEMS transducer array

description

Embodiments of the present invention relate to a MEMS sound transducer array and to a system with a MEMS sound transducer array. Before ferred embodiments relate to a micro-loudspeaker with high sound pressure by utilizing resonance peaks. In general, the invention is in the field of micro-loudspeakers in MEMS technology.

Loudspeakers are used to generate airborne sound in the audible range for interaction with the human sense of hearing. Micro loudspeakers are characterized by the smallest possible dimensions and are used in particular in portable devices in the entertainment and telecommunications industry, e.g. e.g. smartphones, tablets and wearables. Micro-loudspeakers are also used in medical technology, e.g. B. in hearing aids to support the hearing impaired.

The technical challenge with micro-loudspeakers lies in achieving high sound pressure levels (sound pressure level, SPL). For a piston vibrator, the sound pressure level achieved in the free field at a distance r is at frequency f to

with A active area, Ax deflection of the active area, p air density and p _ref reference pressure (20 pPa).

In a closed volume V ₀ the so-called pressure chamber effect occurs, the sound pressure level reached can be calculated

with po pressure in the closed volume.

The sound pressure level achieved is therefore directly proportional to the displaced volume A·Dc (before conversion to the logarithmic scale) both in the free field and in the closed volume (e.g. for in-ear applications). Thus, with increasing miniaturization (A— >0), both free-field and in-ear speakers have to be larger and larger Deflections Dc of the loudspeaker membrane are realized in order to achieve sufficiently high sound pressure levels for audio applications (Dr > 120 dB). In the case of micro loudspeakers for free-field applications, the frequency dependency of the sound pressure level achieved also has a significant impact. At low frequencies, the sound pressure level drops quickly (12 dB per frequency halving). In conventional loudspeakers, this effect is balanced across the surface, but this is not an option with micro-loudspeakers. Therefore, micro-loudspeakers in the free field usually show a sharp drop in the SPL at low frequencies.

Due to the demands on loudspeakers for mobile and usually also body-hugging applications (hearables, wearables, ...) for moderate driver voltages of < 30 V, a very low energy consumption (> 120 dB/mW) and a compact design (< 20 mm ² ) Classic solutions for generating high sound pressure levels (> 120 dB) are now reaching their limits.

As a further development of conventional loudspeakers, micro-loudspeakers have emerged from a miniaturization of the established electrodynamic drive. In the most common voice coil arrangement, a coil is attached to the back of the membrane, which moves in the magnetic field of a fixed permanent magnet when a current signal is applied, thus deflecting the membrane.

A development from hearing aid applications are the so-called balanced armature converters (BA converters). A coil-wound rod sits in the decoupling slot of a ring-shaped permanent magnet and is connected to a diaphragm. A current signal to the coil magnetizes the bar, which is then subjected to a torque by the magnetic field of the permanent magnet. The rotation is transmitted to the membrane via a rigid connection. In the ground state, the rod is in an unstable balance of magnetic attractive forces. Due to this unstable state, higher deflections can be achieved with little effort (drive forces, energy). BA converters are therefore characterized by the higher achievable sound pressure levels and are preferably used for in-ear applications due to their size. Driven by the need for miniaturization and inspired by the successes in the field of microphones, microsystems technology has taken on the subject of micro-loudspeakers. A development by Fraunhofer ISIT together with the company USound resulted in a MEMS loudspeaker based on piezoelectric bending actuators, which deflect a hybrid applied membrane [1].

Further developments by Fraunhofer ISIT are based on piezoelectric bending actuators that do not require an additional membrane [2,3]. The actuators are mechanically decoupled via thin slots and function as an acoustically radiating membrane. Narrow decoupling slot widths of a few microns and optional flow baffles prevent an acoustic short-circuit and enable high sound pressure levels despite the mechanically open design.

Various concepts of electrodynamically actuated MEMS loudspeakers are also known. The work at the Universite Paris-Sud and the Universite du Maine [4,5] is particularly worth mentioning. A stiffened Si membrane suspended from Si springs forms a piston oscillator. The coil is applied directly to the Si membrane as a planar coil and moves the membrane in the magnetic field of a hybrid applied permanent magnet.

A related approach followed by several groups [6,7,8,9,10,11] is that the planar coil is applied to a soft polymer membrane instead of the stiffened Si membrane.

In contrast to piezoelectrically actuated MEMS loudspeakers with electrodynamic drives are still a long way from commercial use. Due to the hybrid assembly of the required magnets, there are no cost advantages compared to the prior art. The small winding cross-section of integrated planar coils and the poor heat dissipation through the thin membrane limit the coil current so that the sound pressure level of conventional micro-loudspeakers is not reached. The problem of current limitation can be reduced if the planar coil is placed on the substrate and the magnet is placed on the moving membrane instead. Thanks to the high thermal conductivity of silicon, current densities that are orders of magnitude higher are then possible in the coil. In the component described in Ref. [12, 13], the micromagnets were integrated at the substrate level. For this, NdFeB powder was etched into micro molds introduced and then solidified with wax. Due to the insufficient resistance of the wax-bound structures, however, this development did not go beyond a demonstrator.

The concept of a magnetostrictively driven micro-loudspeaker is proposed by Albach et al. [14] pursued. The sound transducer consists here of a two-part structure. The first part is a micro-loudspeaker chip that carries the loudspeaker's magnetostrictive membrane. By applying a magnetic field, the membrane is deflected out of the plane of the chip and sound is generated. The second part of the micro-loudspeaker is a current-carrying coil that generates the magnetic field required for operation. The concept proposed here provides for a second chip that carries the corresponding flat micro coils.

Another micro-loudspeaker concept is based on the nanoscopic electrostatic drive (NED) [15]. The device comprises clamped electrostatic bending actuators arranged in pairs in rows and decoupling slots within the device layer of a SOI wafer (Silicon on Insulator) and covered with another wafer bonded to the SOI wafer with a small gap. Between each adjacent row of actuators, acoustically effective openings are alternately integrated into the top and bottom of the wafer to allow sound to be emitted from the component without acoustic short-circuiting.

In summary, it should be noted that the existing MEMS loudspeaker concepts require predominantly complex, expensive and sometimes hybrid manufacturing processes and mostly show inadequate performance characteristics. From these points of view, the approach using piezoelectric bending transducers and fixed flow screens [2] is currently particularly promising, as it combines good performance characteristics with comparatively good manufacturability in MEMS technology.

The object of the present invention is to improve the playback quality, in particular the sound pressure capability of micro-loudspeakers.

The object is solved by the subject matter of the independent patent claims. Embodiments of the present invention create a MEMS acoustic transducer array with at least two acoustic transducers. The first sound transducer is designed to have an amplitude response with a first resonant frequency (fres), e.g. B. 3 kHz to reproduce. The second, third, etc. sound transducers are designed to have an amplitude response with a second, third, etc. resonant frequency, e.g. 7 kHz, 12 kHz, etc. In general, the second resonant frequency is higher than the first resonant frequency.

Furthermore, the second, third, etc. sound transducer is designed in such a way that it has a quality factor Q of >1 or >3. According to exemplary embodiments, the first sound transducer can also have a Q factor of >1 or >3. According to further exemplary embodiments, it would also be conceivable for the Q factor for the first and second sound transducer to be >5.

Exemplary embodiments of the present invention are based on the knowledge that the use of (one or) several sound transducers with a quality >1 or quality >3 and different resonance frequencies in a component makes it possible to efficiently increase the sound pressure level in a wide frequency range. In this respect, the increased deflection of a sound transducer is utilized in its mechanical resonance to increase the sound pressure level (sound pressure level, SPL). The quality factors used of at least three enable the resonances to be exploited. Because the sound transducers are designed for different amplitude responses, they also have different resonant frequencies, so that the overall sound pressure level can be increased over the broad frequency range.

Due to the dependence of the sound pressure on the deflected membrane area of a MEMS sound transducer, at least two sound transducers of the at least 3 sound transducers can form a first group in the array, e.g. to increase the membrane area per frequency range. The sound transducers of the first group are defined, for example, via a "common" bandwidth. For example, one of the sound transducers can define a first bandwidth fres/Q around its resonance frequency, in which case the other sound transducer(s) in the same group lie within this first bandwidth with their resonance frequency/resonance peak. This means that the resonance frequencies of the sound transducers in this group differ by a maximum of fres/Q.

According to exemplary embodiments, with more than three transducers in the MEMS transducer array, several transducers can form one or more groups insofar as than that the sound transducers of a group deviate from each other in their resonant frequencies by a maximum of fres/Q. In this case, the number of sound transducers belonging to the second group is greater than or equal to the number of sound transducers belonging to the first group. This means that the second group z. B. may have one or two or more sound transducers.

According to a further exemplary embodiment, the array can comprise at least one or two third sound transducers of a third group, the at least one or two third sound transducers being designed to reproduce an amplitude response with a third resonance frequency. This third resonant frequency is higher than the second. All amplitude responses belonging to the first/second/third group form the amplitude response of the transmission behavior of the MEMS array. The number of sound transducers belonging to another group can be greater than or equal to the number of sound transducers belonging to the first group.

According to further embodiments, the number of groups of sound transducers is selected as an odd number of sound transducers. In this case, the groups which are adjacent according to the value of their resonant frequency can then be driven in different phases or in anti-phase. The present invention is based on the finding that the sound pressure level is proportional to the membrane area of the sound transducer and that, for higher frequencies, several bending transducers with almost identical properties should therefore be operated in parallel in order to optimize the amplitude response of such a group of sound transducers for the frequency range addressed increase and the sound amplitude of the lower frequency transducers to approximate, so by superimposing the amplitude responses of several groups, the amplitude response of the entire array can be adjusted and / or optimally amplified. This is particularly relevant for the in-ear case of a closed or almost closed ear volume, since there is no compensating, frequency-dependent contribution to the sound amplitude as in the free-field case.

In other words, it means that according to exemplary embodiments in the in-ear case, the number of sound transducers per group can increase with increasing resonance frequency in order to achieve a frequency response which is as uniform as possible. If a drop in the SPL z. B. is tolerable in a very high frequency range (> 15 kHz), a correspondingly smaller number of sound transducers can be selected for this group with the high resonant frequency. In the free field case it would be conceivable that also the number of sound transducers for a group with a high resonant frequency is no longer increased, since the compensating frequency-dependent contribution to the sound amplitude can be used here.

According to a further exemplary embodiment, the three resonance frequencies or at least one of the three resonance frequencies can lie in a frequency range of 1 to 20 kHz, that is to say in the audible range. Of course, the application can also be extended to ultrasonic transducers or the like. According to exemplary embodiments, the frequency spacing between the first and the second resonant frequency (first resonant frequency, for example, lower than the second (or highest) resonant frequency) or generally between the resonant frequencies is selected in such a way that from a quality of Q > 3, a resonance peak is generated in order to to increase the sound pressure at least in a frequency band above the first resonance frequency and/or below the highest resonance frequency/at least partially or to increase it in the entire range below the (higher) highest resonance frequency. This advantageously enables a targeted setting of the sound pressure level and target frequency response by superimposing the amplitudes of a number of sound transducers. This leads to an increase in the sound pressure level in the target frequency response due to the targeted superimposition of the amplitudes of several sound transducers, e.g. B. with a mechanical quality of Q> 3 or Q> 5. As an alternative to increasing the sound pressure, this also creates the possibility of using this procedure to reduce the energy consumption of the component and/or to reduce the size.

In the following, the activation of the sound transducers will be discussed. In order to optimally exploit the amplification effect by superimposing several sound transducers while utilizing their resonance peak, the different groups can be operated either in different phases or anti-phase. This results in a predominantly constructive interference of the various amplitude responses of the different sound transducers in the entire frequency range of interest and thus an optimal increase in the sound pressure level of the MEMS sound transducer array. In a simple manner, the anti-phase control in a piezoelectric bending transducer can be realized by alternately interchanged contacting of the top and bottom electrodes of the piezo capacitor of the respectively adjacent groups. As an alternative to different-phase control, a crossover can also be used. According to one embodiment, an in-ear earphone or a hearing aid based on the MEMS transducer array is created. This application benefits from the targeted improvement in the performance of the MEMS sound transducer for application-specific, individual frequency ranges. Example hearing aid: Here, a high sound level is required, particularly in the low-frequency range, and this is achieved by the MEMS sound transducer array. The performance in the high-frequency range is of secondary importance here. Furthermore, for hearing aids as well as for mobile applications, a reduction in energy consumption is particularly important, which is also achieved in accordance with the above exemplary embodiments, as already explained.

According to exemplary embodiments, for example, bending transducers are used as sound transducers belonging to the first and second groups. According to further exemplary embodiments, the bending transducer can be implemented in such a way that a decoupling slot is provided between it and the surrounding structure or between it and the next bending transducer. Compared to previous MEMS systems, which were mostly based on closed membranes, the present concept enables a significant increase in performance, since due to the decoupling of the bending transducer, no energy has to be used to deform additional mechanical membrane elements, which means that higher deflections and forces are possible. In addition, non-linearities only occur with significantly larger movement amplitudes. Due to the concept and material-related low oscillating mass, systems with an extraordinarily wide frequency range and at the same time high movement amplitudes can be implemented. According to exemplary embodiments, the decoupling slot is sealed by using screens. This leads to high cost savings, since the hybrid assembly and process integration of a membrane can be dispensed with. Here, one or more screens extend along the one or more shaped decoupling slots.

A further exemplary embodiment creates a system comprising a controller and a MEMS acoustic transducer array, as has already been explained. The controller is designed to control the MEMS sound transducer array accordingly. According to exemplary embodiments, at least one or two sound transducers are controlled directly. According to further exemplary embodiments, each individual sound transducer can also be controlled directly. This has the advantage that, as already explained above, an anti-phase control of the frequency of adjacent sound transducers or groups of sound transducers is possible in order to further increase the sound pressure level. According to one embodiment, the controller may include a crossover, e.g. B. to separate the frequency band belonging to the first/second/third resonant frequency from one another or to control the corresponding first/second/third sound transducer. At this point, however, it should also be noted that the sound converters themselves can act as filters due to their quality and therefore no further electrical crossovers are necessary for effective operation. In this respect, multi-path operation is given without the need for a crossover, which means that the effort on the system side is reduced, which offers potential for miniaturization and savings.

In general, by using several sound transducers as a one-chip solution, with the sound transducers being used as a type of resonator with high mechanical quality, it is possible to address applications that could otherwise only be achieved with great effort or with conventional technologies. In contrast to previous systems, which usually have no sensors or only record the deflection of the drive (not the membrane), this principle allows the actual position of the sound-generating element to be determined with the help of easily integrated sensors. This is of great advantage and enables a much more precise and reliable detection. This forms z. B. the basis for a controlled excitation (closed loop), with which external influences, aging effects and non-linearities can be electronically compensated.

While conventional systems sometimes require membranes or magnets with complex shapes, which up to now could not be realized in MEMS technology but could only be integrated hybridly with great effort, the present concept can be realized with common methods of silicon technology. This offers significant advantages in terms of location and costs.

Further formations are defined in the dependent claims. Embodiments of the invention vorlie are explained below with reference to the accompanying drawings. Show it:

1a shows a schematic representation of a MEMS sound transducer according to a basic exemplary embodiment;

1b shows a schematic representation of a MEMS sound transducer using groups of sound transducers according to exemplary embodiments; Fig. 2a and 2b the illustration of dependencies for the phase (2a) and amplitude (2b) of a resonator for different quality factors to elucidate tion of embodiments;

3a and 3b schematic diagrams of a bending line of a bending transducer (3a) for different forms of excitation and the change in the statically generable sound pressure (3b) of a bending transducer when increasing the resonance frequency for a piezoelectrically excited sound transducer (case A) to explain exemplary embodiments;

3c three schematic representations of an idealized bending transducer for the explanation of the diagrams from FIGS. 3a and 3b;

Fig. 4a is a schematic diagram showing the sound pressure as

Function of the frequency for different arrays from FIGS. 4b to 4d to illustrate the sound pressure level increase according to exemplary embodiments.

4b to 4d schematic representations of acoustic transducer arrays in plan view according to embodiments;

Fig. 5 is a schematic diagram illustrating sound pressure as

Function of the frequency for the resonator array from FIG. 4d for different resonator qualities Q;

6a and 6b show schematic diagrams to explain the number (a) of the required sound transducers for the same maximum amplitude in the resonance and exemplary frequency response (b) for the anti-phase excited set resulting from the points in (a) according to exemplary embodiments;

Fig. 7 is a schematic representation of an embodiment of a

Microsound transducer with multiple bending transducers according to extended th embodiments with aperture and decoupling slots; and 8a to 8c schematic cross sections of possible bending transducers according to exemplary embodiments. unimorph (a); Bimorph with passive interlayer (b); Bimorph (c). The piezoelectric layers can be divided into any number of layers (dashed line) and provided with electrodes and separating layers

Before exemplary embodiments of the present invention are explained below with reference to the accompanying drawings, it should be noted that elements and structures that have the same effect are provided with the same reference symbols, so that the description of them can be applied to one another or are interchangeable.

Fig. 1a shows a MEMS transducer array 10 with three transducers 12, 13 and 14 un ferent resonant frequency. The acoustic transducer 13 can be considered optional, but advantageously allows the maximum utilization of the chip area. All sound transducers 12, 13 and 14 can be realized, for example, as bending sound transducers that oscillate out of a substrate plane (cf. surrounding structure 11). The sound transducers 12, 13 and 14 have a very high mechanical quality, e.g. >1.

The quality is determined with a quality factor Q. The quality factor Q is in the range of >1, >3 or even >5. The quality factor, also known as resonance sharpness or Q factor, is a technical measure of the damping or energy loss of an oscillating system. A high quality of a system means that the system is weakly damped. The figure of merit is a measure of the ratio of energy lost to stored energy per cycle. If more energy is stored in the system per oscillation period than is consumed by the damping, then the value for the quality factor is > 1 . This means that with high quality factors, such as factor 3, the damping is so weak that significantly more energy can be stored in the system, which leads to an increased deflection of the sound transducer in its mechanical resonance and thus to an increase in sound pressure. The quality indicates by how much the amplitude in the resonance is increased compared to the amplitude in the static case. In other words, by utilizing the increased deflection from a sound transducer 14 in its mechanical resonance, the sound pressure increases ge (comp. Fig. 2b). In particular, the use of several sound transducers with a quality > 1 and different resonance frequencies (cf. different sound transducer designs of the sound transducer 12 compared to the sound transducer 14 of the Efficiently increased sound pressure levels over a wide frequency range. The increase in sound pressure in the low-frequency range < 1 kHz is of central interest. In this way, on the one hand, the drop in the sound pressure achieved towards low frequencies when radiating into the free field can be compensated. On the other hand, sound transducers for closed volumes (in-ear) with a significantly increased sound pressure level in the bass range can be implemented without a dip in the treble range.

The underlying principle is explained in detail below with reference to FIGS. 2a, 2b, 3a, 3b and 3c.

1b shows a MEMS transducer array 10 ^* with a first group of transducers 12 and a second group of transducers 14. The first group of transducers 12 includes at least one transducer 12a and optionally a second transducer 12b. The second group of sound transducers comprises at least one sound transducer 14a or, in accordance with further exemplary embodiments, at least two second, here three sound transducers 14a to 14c. All sound converters 12a to 12b and 14a to 14c can be implemented as bending sound converters, for example, which oscillate out of a substrate plane (cf. surrounding structure 11).

For this purpose, the bending sound transducers 12a to 12b or 14a to 14c are at least partially connected to the substrate 12. A clamped end of each transducer 12a to 12b or 14a to 14c is marked here as an example, with between the individual transducers 12a and 12b or 14a and 14b or 14b and 14c and between the transducers 12a to 12b and 14a to 14c and the surrounding structure 11 a decoupling slot can be provided.

In general, it can be stated that, according to exemplary embodiments, the number of bending sound transducers 14a to 14c in the second group 14 is greater than or equal to the number of bending sound transducers 12a and 12c. It should also be noted at this point that preferably the bending sound converters 14a to 14c, but alternatively also all sound converters 12a to 12b and 14a to 14c, have a very high mechanical quality, as explained above. The basis for this is formed on the one hand by the relationship between excitation amplitude A ₀ and deflection A of a resonator as a function of its oscillation frequency w, given by

where Q is the mechanical quality of the resonator and w ₀ is its resonant frequency, as well as the course of its oscillation phase, given by f(w) = arctan

2a and 2b show the dependencies for phase and deflection for different quality factors.

2a shows the frequency dependency for phase (a) and amplitude (b) of a resonator plotted for different quality factors. The sound transducer described here is also designed as a bending transducer (cf. FIG. 3c), which can be excited to vibrate piezoelectrically, thermally, magnetically and electrostatically.

In Fig. 3c, three load cases are shown on a rigidly suspended bending transducer 12, as can occur for different excitation mechanisms (see above). The bending transducer 12 has a length I, in case A a continuous load on the surface, in case B a point load at the force application point I and in case C a torque load is shown. The corresponding resulting bending curves are assigned to case A/B/C in FIG. 3a.

The area A _Biege under the bending lines of a bending transducer (Fig. 3a) can be used as a measure of the sound pressure level generated, neglecting the resonance peak (Fig. 2). Because with an identical layer structure, the frequency w of a bending transducer according to / =

1 a ² d !Y 1

2p \[Ϊ ^2 _[ 1 ₂ ^Z -p directly proportional to / oc — is can then

(Free field. See equation (1)) (closed volume or in-ear case. Compare

Eq. (2)) as a measure of the frequency-dependent change in the generated sound pressure level (Fig. 3b).

As can be seen in FIG. 3b, the statically producible sound pressure decreases sharply for higher resonators with higher resonant frequencies, so that it initially seems advantageous to use resonators with the lowest possible resonant frequency in order to generate the highest possible sound pressure. This applies to both the free field and the in-ear case. However, if you look at the graphic in Fig. 2b, it becomes clear that, taking into account the phase behavior of a resonator, the sound pressure above the resonance frequency decreases significantly, so that it already drops by a factor of 3 above the resonance frequency by about 10 log = 10 dB has fallen off. Therefore it seems to be true

to use a sound transducer with the lowest possible resonance frequency as cheaply as possible in order to generate the highest possible sound pressure in the low-frequency range, but then to place further resonators for sound generation at a suitable frequency spacing above and across the entire frequency spectrum. Due to the 180° phase jump shown in FIG. 2a when passing through the resonance, adjacent resonators can be excited in antiphase (FIG. 4), since otherwise the sound level between the resonances collapses due to destructive interference (compare FIG. 4 dotted lines). When compared with a resonator placed at a higher frequency (black curve in Fig. 4), this approach (V2 and V1 curve in Fig. 4) can significantly increase the sound level in the entire range below the highest resonance frequency. Without anti-phase operation, an even greater increase in amplitude is achieved in the low-frequency range below the lowest resonance frequency. However, this is at the expense of the previously mentioned dips in the upper frequency range (dashed line). Both operating modes are justified depending on the application. Alternatively, a crossover can also be used, which meaningfully assigns certain sections of the entire frequency spectrum to the individual sound transducers so that there is no destructive interference. This offers the possibility of an even more individual influencing of the amplitude response of the MEMS sound transducer array. However, this is associated with greater technical complexity and the associated higher system costs, especially since the resonators act as frequency filters anyway due to their high quality. 4 shows the sound pressure level as a function of frequency for different arrays. The three exemplary arrays are illustrated in Figures 4b-d. Fig. 4b shows a bending transducer array 10 'with 20 bending transducers 14. These bending transducers 14 are arranged opposite an adjacent structure or a substrate 11, namely in 4 rows of 5 bending transducers 14. Between the bending transducers 14 decoupling slots 15 are provided, these Decoupling slots are sealed by aperture 17.

In the configuration shown, this array 10' therefore has 20 x 12 kFIz sound transducers, with 12 kFIz precisely representing the resonant frequency. These 20 flexural transducers form a group, since the individual flexural transducers deviate from one another only slightly in terms of their resonant frequency, but at most by the value fres/Q. With parallel activation of all elements 14, the curve vOa sets in.

Another configuration of an array 10" is shown in FIG. 4c, the total area of which, however, is identical to 10'. In the configuration v1, this array comprises 16×12 kFIz converters (cf. reference number 14) and 2×3 kFIz converters (cf. reference number 12). As can be seen, the number of 3 kFIz converters 12 is reduced compared to the number of 12 kFIz converters 14 . In addition, each (single) transducer 12 is provided with a greater length and/or area (eg, 1.x times the length/area or twice the length/area) compared to the (single) transducer 14 .

Between the converters 14 and 14 or 14 and 12 or 12 and 12 the decoupling slots 15 or diaphragms 17 explained in connection with FIG.

In Fig. 4a the resulting sound pressure level is shown with v1g for in-phase and vi a for anti-phase.

4b shows an array 10"" in a configuration v2 with 8×12 kFIz (cf. reference number 14), 2×3 kFIz (cf. reference number 12) and 6×7 kFIz (cf. reference number 16), their total area However, 10' and 10" are identical. The number of sound transducers decreases from 12 kFIz (reference number 14) to 3 kFIz (see reference number 12). The number of sound converters and the size of the sound converters 16 is between 12 and 14. The total area of the arrays 10' to 10'' (cf. v1 to v3) is always comparable. Next will be the Quality of all resonators set to 5, for example. With more complex arrays, an increase in the sound level can be brought about to an ever more complete extent in the entire frequency range below the highest resonance frequency, as can be seen from FIG. 4a. Here, for example, reference is made to the comparison of v0 to v2. For the reference curve vOa, the maximum value in the resonance and the SPL value at 20 kHz correspond exactly to the measured values of an actually measured MEMS bending transducer.

The dimensioning of such arrays 10' to 10'' is explained below according to an exemplary embodiment. The frequency spacing between the different resonators can be selected in such a way that from a certain resonator quality Q, the resonant boost (Fig. 2b or Fig. 5) leads to a significant increase in the generated sound pressure in the frequency range below the highest-frequency resonator (Fig .5). In the example shown, this is the case from a quality of Q=5.

The bending line (FIG. 6a) can be used to calculate the number of sound transducers that is necessary in order to generate the same maximum amplitude in resonance for all resonators with the same quality. For example, the set (1 x 3 kHz + 6 x 6 kHz + 32 x 12 kHz, in Fig. 6a) shows such behavior.

6a shows the relationship between the frequency ratio of different bending transducers (w/wi) and the required number of transducers for the same maximum amplitude in resonance. In Fig. 6b is an example of the frequency response for 3 arrays corresponding to the 3 points x1 to x3 shown in Fig. 6a. Here, a set of 1 x 3 kHz + 6 x 6 kHz + 32 x 12 kHz (x3) and a reduced set of 1 x 3 kHz + 3 x 6 kHz + 6 x 12 kHz (cf. X2) are assumed and this compared to a 1 x 3 kHz converter.

As can be seen from the comparison, the number of resonators can be reduced, particularly in the high-frequency range, since the SPL is not limited by the maximum values but by the minimum values in the frequency response. These may fall down to the pO line (chosen here as 100 dB in Fig. 5). In the present case, the set can therefore be reduced to (1 x 3 kHz + 3 x 6 kHz + 6 x 12 kHz).

FIG. 7 shows a possible exemplary embodiment in relation to FIG. 6b (X2). In the illustrated example, each sound transducer is designed as a bending transducer that can be deflected out of the plane. On the surrounding substrate there are acoustic baffle elements that protrude from the plane and prevent an acoustic short circuit when the bending transducers move. Alternatively, these screen elements could also be performed on the bending transducers. The individual sound transducers are further characterized in that they are spaced apart from other bending transducers as well as from the substrate by a narrow decoupling slot, which on the one hand mechanically decouples the bending transducer from the substrate, but is also sufficiently narrow to prevent acoustic short circuit. Since bending transducers with the same resonance frequency always move in the same way in this micro-loudspeaker, they can in principle also be designed as one element or connected at the tips or, as shown, mechanically decoupled through a narrow slot to avoid any deviations in amplitude and frequency behavior from a narrow resonator to avoid. Nevertheless, the slots must be made sufficiently narrow to prevent an acoustic short circuit. An aperture structure is only necessary between bending transducers with different resonant frequencies or towards the substrate. The screens can be implemented on the substrate as well as on the bending transducers.

Referring to FIG. 7, a micro-loudspeaker 10″″ consisting of several bending transducers 14, 16 and 12 will now be explained. The bending transducer X2 from FIG. 6b is shown here in detail. This has a 3 kFIz converter (cf. reference number 12), 3 6 kFIz converter (cf. reference number 16) and 6 12 kFIz converter (cf. reference number 14). The transducers 14 are arranged in a field in two rows of 3 transducers each. A further field with the 3 converters 16 is arranged adjacent to the field. Decoupling slots (cf. reference numeral 15) are provided between the transducers. Screen elements 19 which protrude vertically from the substrate 11 are provided both laterally to each row and between the individual rows. In this exemplary embodiment, the diaphragm elements 19 are arranged on the adjacent structure, so that the diaphragm elements 19 act from the substrate plane and either the diaphragm element 19 continues into the substrate plane or the substrate itself is guided as a type of diaphragm. In addition, the slot between the bending transducer 14 and the screen 19 or the substrate 11 is correspondingly small, so that good acoustic decoupling of the front volume and rear volume is possible due to the high flow losses. According to other embodiments, the panels can of course also be arranged differently, z. B. on the flexural transducer structure 14 itself. This then also makes it possible for the decoupling slot between the individual flexural transducers 14 to be sealed by the screen elements 19 . In addition to the field consisting of the converters 14 and 16, a converter 12 is provided.

Regardless of the embodiment, the bending transducers are preferably driven or read out piezoelectrically. Alternatively, electrostatic, thermal or magnetic conversion principles are also possible. With the piezoelectric conversion principle, the converters consist of at least two layers, with at least one layer being piezoelectric. The piezoelectric layers can be designed as multilayer systems with additional separating layers and are contacted via flat or interdigital electrodes. According to FIGS. 8a, 8b and 8c, the transducer elements can have one or more passive functional layers in addition to piezoelectrically active layers. In the case of the thermal conversion principle, errors! Reference source could not be found, the thermally active layers the piezoelectric layers.

In the implementations from FIG. 8a, b, c, each of which represents a cross section of possible bending transducers, the piezoelectric layer is provided with 14p and the passive intermediate layer with 14z. 8a shows a unimorph structure, FIG. 8b a bimorph structure with a passive intermediate layer and FIG. 8c a bimorph structure without an intermediate layer. The piezoelectric layers can be divided into any number of layers (see dashed line) and with electrodes and separating layers be provided (not shown).

Further exemplary embodiments create an array of miniaturized sound transducers for sound generation (audio or audible sound and ultrasound), a. each sound transducer has a mechanical quality Q > 3 (alternatively Q > 5) b. where the number of sound transducers in the array is n > 3 (alternatively: with at least 2 sound transducers in the range between 1 kFIz and 20 kFIz) with each sound transducer being a bending transducer that can be deflected out of the plane and having at least one acoustic baffle element protruding vertically out of the plane that is located either on the bending transducer or on the adjacent substrate area c. the individual sound transducers being characterized in that the flow baffle is spaced apart from a surrounding structure by a narrow decoupling slot. According to exemplary embodiments, the height and geometry of the screen element is designed in such a way that an acoustic short circuit through the decoupling slot in the audio and ultrasonic frequency range (20 Hz to 300 kHz) is largely or completely prevented.

According to exemplary embodiments, the decoupling slot width between the panel element and the surrounding structure is made sufficiently small that an acoustic short circuit through the decoupling slot in the audio and ultrasonic frequency range (20 Hz to 300 kHz) is largely or completely prevented. According to exemplary embodiments, one or more crossovers can be used here to further improve the performance (not necessarily).

According to exemplary embodiments, the drive of the sound transducer is realized piezoelectrically, magnetically, electrostatically or thermally.

A further exemplary embodiment creates a sound converter, in which case the sound pressure level can be improved over the entire target frequency response by driving adjacent sound converters in different phases (special case: anti-phase).

Here, the following cases can be provided: a. in the anti-phase case, an odd number of sound transducers can be used (advantageous) b. In the case of a piezoelectric drive, a 180° phase shift (anti-phase control) can be achieved by simply swapping over the electrical connections of the top and bottom electrodes. Different-phase control can also be achieved here by using additional electronic components.

According to exemplary embodiments, the bending transducer is intended for generating sound in air. Corresponding exemplary embodiments can also be provided with a sensor element for position or phase determination.

The bending transducer structures described are suitable for areas of application in which sound in a frequency range between 10 Hz and 500 kHz is to be generated with the smallest possible component volumes (<10 cm ³ ). This applies primarily to miniaturized speakers for wearables, smartphones, tablets, laptops, headphones, but hearing aids also ultrasonic transducers too. Overall, other applications in which fluids are displaced (e.g. fluid mechanical and aerodynamic drive and guide structures, inkjets) can also be considered.

bibliography

[1] Patent specification US 10349182B2, "Micromechanical piezoelectric actuators for realizing high forces and deflections"

[2] Patent application EP 3632135A2, "Micromechanical sound transducer"

[3] F. Stoppel, A. Männchen, F. Niekiel, D. Beer, T. Giese, B. Wagner, "New integrated full-range MEMS Speaker for in-ear applications", IEEE Micro Electro Mechanical Systems (MEMS), 2018

[4] US Pat. No. 9,237,961 B2

[5] I Shahosseini, E Lefeuvre, J Moulin, E Martincic, M Woytasik, G Lemarquand, IEEE Sens J 13 (2013), pp 273-284

[6] F L Ayatollahi, B Y Majlis, "Materials Design and Analysis of Low-Power MEMS Microspeaker Using Magnetic Actuation Technology", Adv. Mater. Res. 74 (2009), pp. 243-246

[7] Chen YC, Cheng YT, “A low-power milliwatt electromagnetic microspeaker us ing a PDMS membrane for hearing aids application”, IEEE Int. conf Micro Electro Mech. Syst., 24th (2011), pp. 1213-1216

[8] M.-C. Cheng, W.-S. Huang, S.R.-S. Huang, "A Silicon microspeaker for hearing instruments", J. Micromech. Microeng. 14 (2004), pp. 859-866

[9] S.-S. Je, F Rivas, RE Diaz, J Kwon, J Kim, B Bakkaloglu, S Kiaei, J Chae, "A Compact and Low-Cost MEMS Loudspeaker for Digital Hearing Aids", IEEE Trans. Biomed. about system 3 (2009), pp. 348-358

[10] B Y Majlis, G Sugandi, M M Noor, “Compact electrodynamics MEMS-speaker”, China Semiconductor Technology International Conference (CSTIC), 2017

[11] P. R. Jadhav, Y. T. Cheng, SK Fan, C. Y. Liang, "A sub-mW Electromagnetic Microspeaker with Bass Enhancement using Parylene/Graphene/Parylene Composite Membrane", IEEE Micro Electro Mechanical Systems (MEMS), 2018

[12] C. Shearwood, MA Harradine, TS Birch, JC Stevens, "Applications of Polyimide Membranes to MEMS Technology", Microelectron. Closely. 30 (1996), pp. 547-550

[13] Q. Zhang, E.S. Kim, “Fully-microfabricated electromagnetically-actuated membrane for microspeaker”, Transducers Ί5, Int. conf Solid. State Sens., Actuators Microsyst., 2015, pp. 2125-2128

[14] Albach, TS, Horn, P., Sutor, A. & Lerch, R. Sound Generation Using a Magnetostrictive, Micro Actuator. J.Appl. physics 109(7), (2011) [15] B. Kaiser, S. Langa, L. Ehrig, M. Stolz, H. Schenk, H. Conrad, H. Schenk, K

Schimmanz, D. Schuffenhauer, Concept and proof for an all-silicon MEMS micro Speaker utilizing air Chambers, Microsystems & Nanoengineering (2019)

Claims

patent claims

1. MEMS transducer array (10, 10', 10", 10'", 10"") comprising at least two transducers (12, 14), wherein a first transducer (12) of the at least two transducers is formed with an amplitude response to reproduce with a first resonant frequency, and wherein a second sound transducer (14) of the at least two sound transducers (12, 14) is designed to reproduce an amplitude response with a second resonant frequency; wherein the second resonant frequency is higher than the first resonant frequency; wherein the second sound transducer (14) has a quality factor Q of >1 or >3.

2. MEMS acoustic transducer array (10, 10', 10", 10"', 10"") according to claim 1, wherein the second acoustic transducer (14) has a quality Q of >5; and/or wherein the first acoustic transducer (12) has a Q-factor of > 3; and/or wherein the second sound transducer (14) has a quality Q different from the Q factor of the first sound transducer (12).

3. MEMS transducer array (10, 10', 10", 10"', 10"") according to claim 1, wherein at least three transducers in the MEMS transducer array (10, 10', 10", 10 ”', 10””) are provided.

4. MEMS transducer array (10, 10', 10", 10"', 10"") according to claim 3, wherein at least two transducers of the at least 3 transducers form a first group; or with at least two sound transducers of the at least three sound transducers forming a first group and with one of the sound transducers in the first group defining a first bandwidth fres/Q around its resonant frequency and the resonant frequency of the other sound transducer(s) in the group lying within the first bandwidth.

5. MEMS transducer array (10, 10′, 10”, 10′”, 10″”) according to claim 4, wherein at least two distinguishable groups of transducers are formed, the resonance frequencies of the transducers of the second group of the two groups are above the resonant frequencies of the first group of the two groups.

6. MEMS transducer array (10, 10', 10", 10'", 10"") according to claim 5, wherein the number of transducers (14) belonging to the second group is greater than the number of transducers (12 ) belonging to the first group.

7. MEMS acoustic transducer array (10, 10′, 10″, 10″″, 10″″) according to any one of the preceding claims, wherein the frequency spacing between the first and the second resonant frequency and/or wherein the first resonant frequency is lower than the second resonant frequency is selected such that from a resonator quality factor of Q>3, an amplitude increase is generated in order to at least partially increase the sound pressure in a frequency range below the second resonant frequency.

8. MEMS acoustic transducer array (10, 10′, 10”, 10″′, 10″″) according to one of the preceding claims, wherein the acoustic transducers (12, 14) or groups of acoustic transducers ( 12, 14) are operated in different phases or are operated anti-phase.

9. MEMS transducer array (10, 10', 10", 10"', 10"") according to claim 8, wherein the array has an odd number of transducers or an odd number of groups of transducers.

10. MEMS transducer array (10, 10', 10", 10"', 10"") according to one of the preceding claims, wherein at least one of the transducers (12, 14) is formed by a bending transducer.

11. MEMS acoustic transducer array (10, 10', 10", 10"', 10"") according to claim 10, wherein a decoupling slot (15) is provided between the bending acoustic transducers or between the bending acoustic transducers and a surrounding structure.

12. Device according to claim 11, wherein the gap width of the decoupling slot (15) between the panel element and the surrounding structure is sufficiently small that an acoustic short circuit through the gap in the audio and ultrasonic frequency range (20 Hz to 300 kHz) is largely or completely prevented.

13. MEMS sound transducer array (10, 10', 10", 10'", 10"") according to claim 11 or 12, wherein along the decoupling slot (15) at least one acoustic diaphragm element (17, 19 ) is formed that is located either on the bending transducer or on the adjacent substrate area.

14. The device according to claim 13, wherein the height and geometry of the acoustic diaphragm element (17, 19) is designed such that an acoustic short circuit through the gap in the audio and ultrasonic frequency range (20 Hz to 300 kHz) is at least largely or completely is prevented.

15. MEMS transducer array (10, 10', 10", 10"', 10"") according to one of the preceding claims, wherein at least two transducers (12, 14) have a resonant frequency in the range from 1 to 20 kHz.

16. System comprising a controller and a MEMS transducer array (10, 10', 10", 10"', 10"") according to one of the preceding claims, wherein the controller is designed to control the MEMS transducer array (10 , 10', 10”, 10”', 10””).

17. System according to claim 16, wherein the controller comprises a crossover network, which is designed to the at least one first sound transducer (12a, 12b) and the at least one second sound transducer (14a, 14b, 14c) with a frequency band associated with the first and to control the second resonant frequency.

18. Device according to one of the preceding claims, wherein the drive of at least two sound transducers (12, 14) is realized piezoelectrically, magnetically, electrostatically or thermally.

19. Device according to claim 8 or one of the preceding claims with reference back to claim 8, wherein the anti-phase sound transducer (12, 14) of different resonance frequency is realized by swapping the electrical connections of the upper and lower electrodes.

20. Device according to one of the preceding claims, wherein the bending transducer is intended for sound generation in air.

21. Device according to one of the preceding claims with sensor elements for position or phase detection.