CN113163311B

CN113163311B - MEMS speaker and electronic device

Info

Publication number: CN113163311B
Application number: CN202110388993.4A
Authority: CN
Inventors: 张孟伦; 孙晨
Original assignee: ROFS Microsystem Tianjin Co Ltd
Current assignee: ROFS Microsystem Tianjin Co Ltd
Priority date: 2021-04-12
Filing date: 2021-04-12
Publication date: 2023-02-17
Anticipated expiration: 2041-04-12
Also published as: CN113163311A

Abstract

The invention provides an MEMS (micro-electromechanical systems) loudspeaker and electronic equipment, wherein the MEMS loudspeaker has better performance. The MEMS loudspeaker is provided with an upper bottom surface and a lower bottom surface which are parallel to each other, a side wall is arranged between the upper bottom surface and the lower bottom surface, and one or more of the upper bottom surface, the lower bottom surface and the side wall is provided with a through hole; the actuator in the loudspeaker is connected to the side wall, and the vibration direction is generally parallel to the upper bottom surface and the lower bottom surface; the actuator has a plurality of parallel plate-like branches.

Description

MEMS speaker and electronic device

Technical Field

The present invention relates to a MEMS speaker and an electronic apparatus.

Background

The micro-speaker is widely applied to various miniaturized acoustic devices and electronic equipment at present, and is used in multimedia and electronic entertainment equipment. The MEMS speaker based on MEMS (micro electro mechanical system) actuator is a new important means for realizing micro speaker, and its core working principle is to use piezoelectric material to realize the coupling and mutual conversion of sound energy (mechanical energy) to electric energy.

At present, miniaturization of speakers is one of the concerns in the industry. Due to the small size, the performance (e.g., output sound pressure) of the speaker is limited. How to optimally design the internal structure of the speaker in a small size/space is a key element affecting the performance of the micro-speaker.

Disclosure of Invention

In view of this, the present invention provides a MEMS speaker and an electronic device, and the MEMS speaker has better performance. The invention provides the following technical scheme:

a MEMS loudspeaker, its body has upper bottom and lower bottom parallel to each other, have sidewalls between the two, one or more in upper bottom, lower bottom, sidewall have via holes; the actuator in the loudspeaker is connected to the side wall, and the vibration direction is generally parallel to the upper bottom surface and the lower bottom surface; the actuator has a plurality of parallel plate-like branches.

Optionally, there are only 1 through hole between adjacent branches according to the vibration direction; the air flow rate of the through hole is not less than that of the gap between the shell and the actuator branch.

Optionally, piezoelectric layers are directly attached to two side surfaces of the silicon-based skeleton of each branch of the actuator, and electrode layers are attached to the piezoelectric layers; or, directly attaching bottom electrodes to two side surfaces of the silicon-based framework of each branch of the actuator, and sequentially attaching piezoelectric layers and top electrodes to the bottom electrodes; or, seed crystal layers are directly attached to two side faces of the silicon-based framework of each branch of the actuator, and a bottom electrode, a piezoelectric layer and a top electrode are sequentially attached to the seed crystal layers.

Optionally, the silicon-based framework is made of N-type or P-type doped silicon.

Optionally, the doped silicon has an electrical conductivity of less than 2000 ohm-cm.

Optionally, the side surface of each branch of the actuator is perpendicular to the upper bottom surface and/or the lower bottom surface; and the side surface between the branches of two adjacent actuators is provided with an extension part extending towards the inside of the loudspeaker at the position close to the connection part of the actuator branch and the side surface, and the projection of the extension part in the horizontal direction is pentagonal.

Optionally, the roughness of the silicon-based backbone side of the actuator is no greater than 50nm.

Optionally, the speaker is fabricated from <110> type silicon.

Optionally, the edge of the top of the lower bottom surface of the speaker has a ring-shaped protrusion, and/or the edge of the bottom of the upper bottom surface of the speaker has a ring-shaped protrusion.

Optionally, the annular protrusion is formed by etching the material of the upper bottom surface or the lower bottom surface, or is a deposition layer deposited on the upper bottom surface or the lower bottom surface, or is a silicon dioxide layer of the SOI type wafer.

Optionally, the actuator is multi-layered; a circle of annular bulge is arranged between the side walls of all layers of the loudspeaker, and the annular bulge is formed by a deposited material layer or a silicon dioxide layer of an SOI (silicon on insulator) type wafer or by etching a material of the side walls.

Optionally, the actuator is 1 layer; the distance between the upper top surface of each branch of the actuator and the upper bottom surface is not more than 20um, or not more than 5um, and/or the distance between the lower top surface of each branch of the actuator and the lower bottom surface is not more than 20um, or not more than 5um.

Optionally, the actuator is 2 or more layers distributed up and down; in two adjacent layers of actuators, the distance between the lower top surface of the upper layer actuator and the upper top surface of the lower layer actuator is not more than 20um, or not more than 5um.

Optionally, one end of each branch of the actuator is connected to a first side wall of the speaker, the other end is a free end, and a distance from a side wall opposite to the first side wall to the free end is not more than 100um, or not more than 30um; or the branches of the actuator are divided into a first group and a second group which are respectively connected to two opposite side walls of the loudspeaker, each branch of the first group and each branch of the second group are aligned in pairs, and the distance between the top ends of the two aligned branches is not more than 100um or not more than 30um; or the branches of the actuator are divided into a first group and a second group which are respectively connected to two opposite side walls of the loudspeaker, each branch of the first group and each branch of the second group are staggered in pairs, and the distance between the side walls of the two staggered branches is not more than 100um or not more than 30um.

Optionally, both ends of each branch of the actuator are connected to a set of opposite side walls of the speaker, respectively.

An electronic device comprising a MEMS speaker according to the present invention.

According to the technical scheme of the invention, the through holes and the actuators of the MEMS loudspeaker are specifically arranged, and the size requirement of each internal gap is met, so that the output sound pressure of the loudspeaker is improved. When the MEMS loudspeaker is applied to electronic equipment, the occupied space is small, and the volume is large. The gap between the loudspeaker shell and the actuator branch can improve the displacement sensitivity of the actuator, but the output sound pressure of the loudspeaker is influenced by the overlarge gap. The <110> type silicon wafer is etched by adopting a wet method, and the actuator side wall with good verticality and smoothness is formed by utilizing the anisotropy of the crystal, so that the output sound pressure of the loudspeaker is improved. On the basis, the air tightness of the loudspeaker can be improved by combining the Bosch process.

Drawings

For purposes of illustration and not limitation, the present invention will now be described in accordance with its preferred embodiments, particularly with reference to the accompanying drawings, in which:

fig. 1A is a schematic diagram of an external three-dimensional structure of a MEMS speaker according to an embodiment of the present invention;

FIG. 1B is a schematic diagram of an exploded structure of a MEMS speaker according to an embodiment of the present invention;

fig. 1C is a schematic view of the structure of an actuator of a MEMS speaker according to an embodiment of the present invention;

FIG. 1D is a schematic diagram of the overall structure of a speaker actuator according to an embodiment of the present invention;

FIGS. 1E and 1F are schematic diagrams of lateral and longitudinal expansion, respectively, of a functional substrate according to an embodiment of the present invention;

FIG. 2A is a schematic diagram of a cross-section of a MEMS speaker according to an embodiment of the present invention;

FIG. 2B is a schematic diagram of one configuration of a material layer of an actuator of a MEMS speaker according to an embodiment of the present invention;

FIG. 2C is a schematic illustration of another configuration of a material layer of an actuator of a MEMS speaker according to an embodiment of the present invention;

FIG. 2D is a schematic illustration of a vibration mode of an actuator of a MEMS speaker according to an embodiment of the present invention;

FIGS. 2E and 2F are schematic views showing the direction of air flow when the actuator vibrates in the configuration shown in FIG. 2A;

FIG. 2G is a detail taken from FIG. 2E;

FIG. 2H is a schematic diagram of the air flow between the end of the actuator branch and the inner wall of the speaker of a MEMS speaker according to an embodiment of the present invention;

fig. 2I is a schematic view of a U-shaped through hole of a housing of a MEMS speaker according to an embodiment of the present invention;

FIG. 2J is a schematic top view of portion D1 of FIG. 2I;

fig. 3A to 3E are schematic views of an internal gap of a MEMS speaker according to an embodiment of the present invention;

FIG. 3F is a schematic diagram of the relationship between the longitudinal gap between the actuator branch and the speaker housing and the output sound pressure of the speaker according to an embodiment of the present invention;

FIG. 3G is a schematic diagram of the relationship between the lateral gap between the actuator branch and the speaker housing and the output sound pressure of the speaker according to an embodiment of the present invention;

fig. 4A to 4C are schematic views illustrating a formation manner of a longitudinal gap of a MEMS speaker according to an embodiment of the present invention;

fig. 4D is a schematic diagram of the structure of an SOI-type wafer according to an embodiment of the present invention;

FIGS. 5A and 5B are schematic illustrations of intermediate states of etching on an SOI-type wafer to form actuator branches, in accordance with an embodiment of the present invention;

fig. 6A to 6E are schematic views of crystal orientations according to embodiments of the present invention;

FIG. 6F is a schematic illustration of the intersection of the <110> crystal plane and the <111> crystal plane on the wafer;

FIG. 7A is a schematic view of a mask according to an embodiment of the present invention;

FIG. 7B is a schematic illustration of the effect of wet etching using the mask of FIG. 7A;

FIG. 8 is a schematic diagram of an etching process according to an embodiment of the invention;

FIG. 9A is a scanning electron microscope image of a silicon surface obtained by etching according to the prior art;

FIG. 9B is a scanning electron microscope image of a silicon surface obtained by etching in accordance with an embodiment of the present invention;

FIGS. 10A and 10B are schematic diagrams of actuator branches formed during wet etching according to embodiments of the present invention, varying distances from the inner wall of the speaker;

fig. 11A and 11B are schematic views of forming an actuator bifurcated end gap according to embodiments of the present invention.

Detailed Description

In the embodiment of the present invention, the structure and the manufacturing method of the MEMS speaker of the present invention are illustrated. Fig. 1A is a schematic diagram of an external three-dimensional structure of a MEMS speaker according to an embodiment of the present invention. As shown in fig. 1A, the MEMS speaker has a rectangular box shape, and has upper and lower bottom surfaces parallel to each other and four side walls therebetween. Of course, a polygonal bottom surface and thus a plurality of side walls may be provided. In terms of basic structure, the speaker includes a first package substrate C1 forming a lower bottom surface, and a functional substrate D1 forming a sidewall and an actuator, and a second package substrate C2 forming an upper bottom surface. Where M3 is a hypothetical cross-section. For ease of understanding, the coordinate systems identified in the various figures herein are coordinated coordinate systems.

If the structure in fig. 1A is disassembled in the Z direction, the exploded view shown in fig. 1B can be obtained. FIG. 1B is a schematic diagram of an exploded structure of a MEMS speaker according to an embodiment of the present invention, wherein a row of through holes is distributed on C1, and one of the through holes is Vj; similarly, one of the through holes in the row distributed on the C2 is Uk; in addition, in D1, in addition to the peripheral frame structure, a plurality of parallel plate-like branches of the actuator are visible inside, and the whole of each branch constitutes the actuator of the speaker, and vibrates under the piezoelectric effect to output sound pressure.

The vibration direction of the actuator of the loudspeaker is in an XY plane and can be visually understood to be similar to the left-right swing of a fin, so the loudspeaker can also be called a fin type loudspeaker, one branch of the actuator is equivalent to one fin, and the actuator is in a fin array structure as a whole. Fig. 1C is a schematic view of the actuator of the MEMS speaker according to the embodiment of the present invention, in which a part of the peripheral frame of D1 is cut away, and the structure of each fin, for example, one of the fins Fn, is seen, and it is in the form of a long plate, one end of which is connected to the inner wall of D1 and the other end of which is a free end. Other possible configurations of the fins can also be seen hereinafter.

D1 is shown in fig. 1D in a top view, and fig. 1D is a schematic diagram of the overall structure of the loudspeaker actuator according to the embodiment of the present invention, wherein the region F is the region where the fins are located. It can also be noted from fig. 1D that one end (upper end in the view of the figure) of each fin is connected to the perimeter frame and the other end (lower end in the view of the figure) is suspended and maintains a slight gap with the inner wall of the frame.

Fig. 1E and 1F are schematic views of lateral and longitudinal expansion, respectively, of a functional substrate according to an embodiment of the present invention. The functional substrate D1 can be replicated to D2 and then spliced horizontally as shown in fig. 1E, or vertically as shown in fig. 1F.

Further details of the structure of the MEMS speaker according to the embodiment of the present invention will be described below, in conjunction with the operation of the actuator.

Fig. 2A is a schematic diagram of a cross section of a MEMS speaker according to an embodiment of the present invention. As shown in fig. 2A, if fig. 1A is cut along a section M3, a structural section shown in fig. 2A can be obtained: here, for clarity of illustration, the number of fins and through holes is reduced, and only the fins F1 to F7, and the upper through holes V1 to V4 and the lower through holes U1 to U4 are shown. As can be seen from fig. 2A, the upper and lower end surfaces of each fin in the cross-sectional view maintain a slight gap from the inner wall of C1 or C2.

Fig. 2B is a schematic diagram of one configuration of material layers of an actuator of a MEMS speaker according to an embodiment of the present invention. As shown in fig. 2B, one branch of the actuator, for example Fn, is as follows from inside to outside:

f00: the silicon-based framework is made of monocrystalline silicon or polycrystalline silicon;

f01: a seed layer comprising AlN or other thin film material that assists in the growth of the piezoelectric layer;

f02: the bottom electrode is made of Mo, pt and other common electrode materials;

f03: piezoelectric layer, alN, doped AlN, PZT and other common piezoelectric film material;

f04: the top electrode is made of Mo, pt, au, al and other common electrode materials.

The material layers F02, F03 and F04 on each side of the framework F00 form a piezoelectric sandwich structure, when alternating voltage is applied between the electrodes F02 and F04, the piezoelectric layer F03 can vibrate under the action of piezoelectric effect, the voltage phase difference of the sandwich structures on the two sides of each framework F00 is properly adjusted, the sandwich structures on the two sides can drive the framework F00 to move towards the same direction, and therefore vibration of the fins is achieved. Similarly, by properly adjusting the voltage phase relationship between adjacent fins, it is possible to realize the opposite movement of the adjacent fins, for example, as shown in fig. 2D, fig. 2D is a schematic diagram of the vibration state of the actuator of the MEMS speaker according to the embodiment of the present invention, in which the movement direction of a certain fin Fn is opposite to that of the two adjacent fins Fn-1 and Fn + 1.

Furthermore, the electrodes and piezoelectric layers on the side of the silicon-based skeleton F00 can also be distributed in the manner of fig. 2C, fig. 2C being a schematic view of another structure of the material layers of the actuator of the MEMS speaker according to the embodiment of the present invention, which differs from the structure shown in fig. 2B in that: in fig. 2C, the piezoelectric layer F03 is grown directly on the F00 side, but not on the seed layer F01 and the electrode layer F02. F00 is realized by patterning low-resistance silicon, and the F00 is used as an electrode by utilizing the conductive property of the low-resistance silicon. In the structure shown in 2C, a reference potential is applied to the silicon-based skeleton F00, and a potential different from the reference potential is applied to the electrode layers F04 on both sides; when the potentials of the two sides F04 are opposite in phase, the same vibration mode as that of the structure of fig. 2B can be realized, and the vibration mode is shown in fig. 2D, where fig. 2D is a schematic diagram of the vibration mode of the actuator of the MEMS speaker according to the embodiment of the present invention. The structural design can greatly reduce the cost by directly using low-resistance silicon, save a part of metal electrodes, and reduce the electrical loss formed on the silicon substrate F00, thereby improving the conversion efficiency of electrical energy and mechanical energy.

Based on the motion rules, the situation shown in fig. 2E and 2F occurs when the static structure shown in fig. 2A starts to operate. Fig. 2E and 2F are schematic views of the direction of air flow when the actuator vibrates in the configuration shown in fig. 2A. In fig. 2A, 2E, and 2F, according to the distribution direction of the actuator branches, i.e. the x direction in the drawing, the distribution rule of the through holes and the actuator branches is: U1-F1-V1-F2, U2-F3-V2-F4, \8230;. That is, there is only one through hole between adjacent branches, and the through hole is an upper through hole or a lower through hole. According to the characteristics of the distribution, for the moving state of the fins shown in fig. 2E, taking the fins F1 to F3 as an example, where F1 and F2 move toward each other and F2 and F3 move away from each other, so that the space between F1 and F2 is compressed and the space between F2 and F3 is stretched, according to the gas equilibrium method, a part of air originally existing between F1 and F2 will be forced to the atmosphere through the through hole above C1 and between F1 and F2 (as shown by the upward arrow in C1 in the figure), and a part of air in the atmosphere will also be sucked into the space between F2 and F3 through the through hole above C2 and between F2 and F3 (as shown by the upward arrow in C2 in the figure).

Similarly, for the moving state of the fins shown in fig. 2F, still taking the fins F1-F3 as an example, where F2 and F3 move toward each other and F1 and F2 move away from each other, so that the space between F2 and F3 is compressed and the space between F1 and F2 is stretched, according to the gas equilibrium method, a part of air originally between F2 and F3 tends to be expelled into the atmosphere through the through hole above C2 and between F2 and F3 (as shown by the downward arrow in C1 in the figure), and a part of air in the atmosphere also tends to be sucked into the space between F1 and F2 through the through hole above C2 and between F1 and F2 (as shown by the downward arrow in C2 in the figure).

Thus, the combination of the motions of fig. 2E and 2F results in one complete cycle of air vibration at each through-hole exit, and the cycle of air vibration results in sound waves propagating to the external environment. The larger the gas flux in the through-hole per unit time, the higher the sound pressure output from the speaker (the louder the sound). In the above process, assuming that the movement frequency of the fins is constant and the sealing state among C1, C2 and D1 is good, the following three factors directly affecting the amount of gas exchanged with the external atmosphere in unit time of the through hole are mainly included:

(1) Amplitude of the fin

The larger the fin amplitude, the larger the pressing/stretching amplitude of the space between the adjacent fins, and the more the amount of air discharged/sucked per unit time.

(2) The connectivity of the space between every two fins and the space between the adjacent fins

The poorer the connectivity of the space between the fins, the better the air tightness of the space, and when the fins on the two sides extrude or expand the air in the space, the air leakage can be prevented.

(3) The sound resistance of the through hole or the connectivity between the outside and the space between the fins

The smaller the sound resistance of the through hole is, or the better the connectivity between the outside and the space between the fins is, when the fins on the two sides extrude or expand the air in the space, the air vibration can be ensured to be transmitted from the space between the fins to the outside through the sound hole.

The influence of "connectivity" in the factor (2) on the sound pressure will be described below with reference to fig. 2G and 2H. Fig. 2G is a partial view taken from fig. 2E, taking F1-F3 as an example, when the fins F1 and F2 move toward each other, if the distance between the upper and lower end surfaces of the fins and the inner surface of the package substrate C1C2 is too large, a considerable amount of air molecules enter the "partition wall" space from the space between F1F2 along the direction of the arrow in the figure, so that the amount of air discharged from the V1 hole is reduced, and the amount of air molecules entering the space between F2 and F3 along the direction of the arrow is reduced, so that the output sound pressure of the speaker is reduced, and this kind of gap is called a longitudinal gap or a Z-direction gap. Similarly, for the longitudinal expansion shown in fig. 1F, the above-mentioned longitudinal gap also exists between fins of adjacent layers. In addition, as shown in fig. 2H, fig. 2H is a schematic diagram of the air flow between the end of the actuator branch and the inner wall of the speaker of the MEMS speaker according to the embodiment of the present invention, in which air also passes through the gap between the free end of the fin and the side wall of the frame, the air flow direction is similar to that shown by the arrow in the figure, and therefore the size of the gap also has the same influence on the sound pressure output, and the gap is called a lateral gap or an XY-direction gap.

When the amplitude of the fin is fixed, the larger the air flow of the through hole is, the larger the air flow of a gap between the shell and the actuator branch is, the higher the sound pressure output by the loudspeaker is; in other words, when the amplitude of the fins is constant, the acoustic resistance of the through hole is smaller than that of the gap between the housing and the actuator branch, and the higher the sound pressure output by the speaker. The air flow rate of the through-opening should therefore be greater than the air flow rate of the interspace between the housing and the actuator branch, and the acoustic resistance of the through-opening should be less than the acoustic resistance of the interspace between the housing and the actuator branch.

Fig. 2I is a schematic view of a U-shaped through-hole of a housing of a MEMS speaker according to an embodiment of the present invention; fig. 2J is a schematic top view of the portion D1 in fig. 2I. There may also be variations in the positions of the through holes Vj and Uk, for example, as shown in fig. 2I and 2J, fig. 2I being a schematic view of a U-shaped through hole of a housing of a MEMS speaker according to an embodiment of the present invention, in which the through holes Vj and Uk are respectively provided on two frame side walls opposite to D1, and no through hole is provided on C1 and C2.

Fig. 2J is a schematic top view of the portion D1 in fig. 2I (for clarity of illustration, the number of fins and through holes is intentionally reduced in the top view), and from the top view, the through holes V1-V4 and the through holes U1-U5 are distributed in a staggered manner, which means that when V through holes are provided on the frame D1 between two adjacent fins (e.g., F1 and F2), U through holes cannot be provided between the two adjacent fins, and vice versa. This ensures that when the V-via is admitting air, the U-via is venting and vice versa.

It should be noted that there may be other variations in the arrangement of the via locations, such as the combination of Vj on C1 and Uk on D1 or Vj on D1 and Uk on C2.

In order to ensure that the air flow rate of the through hole is greater than the air flow rate of the gap between the housing and the actuator branch and that the acoustic resistance of the through hole is smaller than the acoustic resistance of the gap between the housing and the actuator branch, embodiments of the present invention provide alternative and preferred values for some dimensions of such a MEMS speaker, see fig. 3A to 3E, which are schematic illustrations of the internal gap of a MEMS speaker according to embodiments of the present invention.

For the longitudinal gap, the gap widths d1 and d2 (fig. 3A) of the upper and lower top surfaces of each fin and the inner surfaces of the upper and lower package substrates, respectively, do not exceed 20um, preferably 5um. For the longitudinal stacking expansion structure, it is required that the gap width d3 (as shown in fig. 3B) of the upper/lower top surface of a certain fin in a certain layer and the lower/upper top surface of a fin located directly above/below the fin in an adjacent layer is not more than 20um, preferably not more than 5um.

As for the lateral gap, there may be several variations shown in fig. 3C to 3E depending on the process used, in which the lateral gap D4 is defined as the distance from the end face of each fin to the inner wall of the frame in fig. 3C, the lateral gap D5 is defined as the distance from the end face of each fin to the end face of the opposite short fin in fig. 3D, and the lateral gap D6 is defined as the distance from the side wall of each fin to the side wall of the opposite short fin in fig. 3E. The values of d4, d5, d6 do not exceed 50um, preferably 20um.

Fig. 3F is a schematic diagram of the relationship between the longitudinal gap between the actuator branch and the speaker housing and the output sound pressure of the speaker according to the embodiment of the present invention. In fig. 3F, the 6 fold lines from top to bottom correspond to longitudinal gaps of 1, 3, 5, 10, 20, 50 microns, respectively. As can be seen, the larger d1 and d2, the lower the sound pressure level, and especially the more pronounced the low frequency step difference. At 500Hz, a 5 micron spacing resulted in a sound pressure loss of 3dB and a 20 micron spacing resulted in a sound pressure loss of 20 dB.

Fig. 3G is a schematic diagram of the relationship between the lateral gap between the actuator branch and the speaker housing and the output sound pressure of the speaker according to the embodiment of the present invention. In fig. 3G, the 7 fold lines from top to bottom correspond to lateral gaps of 1, 3, 5, 10, 20, 50, 100 microns, respectively. As can be seen from the figure, the larger d4, the lower the sound pressure level, especially the more pronounced the low frequency step. At 500Hz, a 20 micron spacing resulted in a sound pressure loss of 10dB and a 50 micron spacing resulted in a sound pressure loss of 20 dB.

The vertical gap may be formed in the manner described with reference to fig. 4A-4C, in which SOI-type wafers may be used. Fig. 4A to 4C are schematic views of a manner of forming a longitudinal gap of a MEMS speaker according to an embodiment of the present invention. Fig. 4D is a schematic diagram of a structure of an SOI wafer according to an embodiment of the present invention, where TS is top silicon, OX is a silicon dioxide layer, and BS is bottom silicon.

For the formation of d2, referring to fig. 4A, a circle of protrusions CR2 made of silicon dioxide is formed on the edge of the upper surface of C2, and the height of CR2 is d2. May be implemented using the silicon dioxide layer of an SOI type wafer. I.e., the top silicon of the SOI type wafer is processed to form D1, the middle silicon dioxide layer forms CR2, and the bottom silicon forms C2. In this case, there is no separation between D1 and CR2, which is only illustrated schematically. Alternatively, instead of a SOI type wafer, C2 may be formed on a silicon substrate and D1 etched on the silicon substrate, or a material may be deposited on the silicon substrate to form D1.

For the formation of d1, it may be the height of the protrusion of the lower surface edge of C1, and the formation manner thereof may refer to the formation of d2. If C1 is formed from a single piece of silicon substrate, turning can be performed after processing is complete to cover D1, as shown in FIG. 4B. In the case where one of C1 and C2 is a SOI type wafer, the other may be fabricated solely from a single silicon substrate.

For the formation of D3, referring to fig. 4C, if the actuator is formed of two layers D1 and D2, a circle of protrusions DR1 may be formed on the edge of one of the layers, for example, D1, and then turned over to cover D2. Similar to D1 and D2, D1 may be fabricated by using an SOI type wafer, where DR1 is formed as a silicon dioxide layer; it is also possible to fabricate D1 using a silicon substrate, where DR1 may be formed by etching on the silicon substrate, or DR1 may be formed by depositing material on the silicon substrate.

For the lateral gap, a Bosch process in combination with a patterned mask can be used, the gap size being determined by the mask window size. In addition, the transverse gap can be realized by a machining mode, a disc silicon wafer scribing knife is commonly adopted, the tail end of the fin structure is cut by high-speed rotation, and the gap size is determined by parameters such as blade thickness, blade rotating speed and feeding speed. For fig. 3C and 3D, the actuator branch not including D4 and D5 can be made by using a graphic mask, i.e. both ends of the actuator branch are connected to the side wall of the speaker, and then cutting is performed by feeding at the positions marked with D4 and D5 in the figure.

For fig. 3E, a patterned mask may be used to form a state without the gap d6, i.e. a zigzag silicon-based skeleton is formed, and then a cutting step is performed at the connection point of the upper and lower actuator branches in fig. 3E, for example, the marked d 6.

The actuator processing method of the MEMS speaker according to the embodiment of the present invention will be further described below.

In an embodiment of the present invention, as mentioned above, the SOI type wafer shown in fig. 4D is used to process an actuator. Referring now to fig. 5A and 5B, fig. 5A and 5B are schematic illustrations of intermediate states of etching on SOI-type wafers to form actuator fingers, in accordance with an embodiment of the present invention. Fig. 5A is a perspective view, and fig. 5B is a line view corresponding to fig. 5A. During processing, generally speaking, a plurality of parallel grooves are formed on the top silicon of the SOI type wafer, and the walls of the grooves are used as actuator branches; the bottom of the groove is a SiO2 layer, which needs to be removed as a sacrificial layer completely or only one circle around, and can be referred to as CR2 in FIG. 4A. The actuator branches are suspended vertically so as to swing.

When etching, the embodiment of the invention provides that a <110> type silicon wafer is selected to be etched so as to form the actuator side face with good verticality. Fig. 6A to 6E can be referred to for the crystal orientation, and fig. 6A to 6E are schematic views of the crystal orientation according to the embodiment of the present invention. In fig. 6A to 6E, each cube is a lattice unit cell, the shaded plane in fig. 6A is the <110> crystal plane, and the shaded plane in fig. 6B to 6E is the <111> crystal plane, where the shaded planes in fig. 6C and 6D are parallel, and the intersection line with the <110> crystal plane is Q1; the shaded planes in fig. 6B and 6E are parallel and the intersection with the <110> crystal plane is Q2, where Q1 and Q2 are shown in fig. 6F, which is a schematic illustration of the intersection of the <110> crystal plane and the <111> crystal plane on the wafer. As shown in fig. 6F, the wafer W1 is parallel to the <110> crystal plane, and the plane perpendicular to the wafer W1 (shown as a straight line in the figure) can be divided into two groups, one group being parallel to Q1 and the other group being parallel to Q2. The included angle between the two sets is about 70.53. Also shown in fig. 6F is a chamfer B1 of the wafer, which is substantially parallel to Q1.

Therefore, in the embodiment of the present invention, when etching is performed, referring to fig. 4D, in the viewing angle of fig. 4D, the upper surface is the <110> crystal plane, and etching is performed in the direction perpendicular to the <110> crystal plane, that is, in the vertically downward direction, a side surface extending strictly downward, that is, in the direction perpendicular to the etching direction, that is, a side surface having no inclination (if operation is not performed in the above direction, a side surface in a non-vertical direction may be generated due to the anisotropy of the crystal), and the side surface is parallel to Q1 or Q2 in the horizontal direction.

In operation, the <110> type silicon wafer can be wet etched in alkaline solution (potassium hydroxide (KOH) or sodium hydroxide (NaOH) solution with the mass concentration of 20% -60% or other alkaline solution) at high temperature (80-120 ℃). Wherein the masks used are as shown in fig. 7A and 7B, fig. 7A is a schematic view of a mask according to an embodiment of the present invention, and fig. 7B is a schematic view of the effect of wet etching using the mask of fig. 7A. The mask in the embodiment of the present invention has a rectangular outer contour, and other shapes can be adopted in the implementation, but the mask window needs to correspond to the shape of the actuator branch. Mask M2 in fig. 7A is laid on the TS layer upper surface of the SOI type wafer of fig. 4D, and the L direction is parallel to the Q1 or Q2 direction described above. After etching, the state of fig. 7B is obtained, and the perspective views of the state are fig. 5A and 5B (fig. 5A and 5B show a part of a wafer). Inclined planes P4, and <111> crystal planes P5 are formed at the left and right ends of the stripe-shaped mask window due to the effect of crystal anisotropy. The P1 plane and the P0 plane are <110> crystal planes, the P2 plane and the P3 plane are parallel, and <111> crystal planes, that is, if the direction in which the P2 plane and the P3 plane extend along the Y axis is the Q1 direction, then P5 is the Q2 direction.

The P2 and P3 planes form an ideal right angle with the P1 plane, but not during etching, and referring to fig. 8, fig. 8 is a schematic diagram of an etching process according to an embodiment of the present invention, which corresponds to the XZ plane of fig. 5A or fig. 5B. During the etching process, the P1 plane is continuously lowered in the top silicon to form a trench, the side wall of the trench is a crystal plane with good verticality <111>, and a tiny inclined plane with a crystal direction <112> is generated at the intersection of the bottom of the trench and the side wall. Since this <112> plane adversely affects subsequent processes, it needs to be removed. Experimental results show that the <112> crystal plane is extremely unstable in a wet environment, and is quickly decomposed in an etching solution after being formed and replaced by a new <112> crystal plane located therebelow, so that if a single crystal silicon material below a plane P1 where the bottom of the trench is located in fig. 8 is replaced by an etching stop material (the etching stop material refers to a material which is highly resistant to etching compared with single crystal silicon, for example, for a KOH etching environment, the etching stop material may be silicon dioxide, silicon nitride, or the like), and the etching time is appropriately prolonged, so that the <112> crystal plane can be removed, and the side wall <111> plane remains stable, thereby forming an ideal right angle between the P1 at the bottom of the trench and the <111> crystal plane. Therefore, the above processing can be performed using an SOI type wafer, and the SiQ2 layer in the middle thereof is used as an etch stop layer. Other means, such as a silicon material with a silicon dioxide layer, etc., may of course be used.

By adopting the mode, the verticality of the side wall of the groove is good, and a smooth surface can be obtained. And the trench sidewalls are the sides of the actuator branches. Referring to fig. 9A and 9B, fig. 9A is a scanning electron microscope image of a silicon surface obtained by etching according to the prior art, and fig. 9B is a scanning electron microscope image of a silicon surface obtained by etching according to an embodiment of the present invention. The silicon surface in fig. 9A is typically achieved by deep silicon etch (DRIE) of the Bosch system of the prior art, which actually consists of several isotropic etch cycles, and therefore typically leaves many scales or undulations on the sidewalls, which are not only dense but also typically tens or even hundreds of nanometers in distance from the peak to the valley of the undulations, as shown in fig. 9A. The surface state of the silicon-based framework has a very adverse effect on the lattice state of a film layer subsequently grown on the silicon-based framework, and particularly, the crystal orientation of the aluminum nitride piezoelectric layer is disordered and the flatness of the film layer is deteriorated, so that the electromechanical coupling coefficient of the piezoelectric layer and the utilization efficiency of input electric energy are greatly reduced, and finally, the amplitude of a fin is low and sufficient sound pressure cannot be output.

In the MEMS industry, hydrogen atmosphere high temperature annealing (e.g., heat treatment at 1100 ℃ for more than 20 minutes in 100% hydrogen atmosphere) is usually used to greatly reduce the roughness of the silicon-based surface after the Bosch process, but the hydrogen annealing furnace is often expensive, and hydrogen belongs to high-risk gas, which has severe requirements for peripheral safety protection measures, so the solution of the Bosch process combined with hydrogen annealing is very low in cost performance for laboratory-level processing and small-scale mass production. In addition, bosch needs to find a group of processing parameters for ensuring the verticality of the side wall of the silicon-based framework and the horizontal plane through multiple process tests, the process parameters are sensitive to the regular maintenance process of the deep silicon etching equipment, and considerable time and materials are consumed for parameter debugging after each maintenance.

The scheme for etching the 110 type silicon wafer by the wet method in the embodiment of the invention has the advantages that the anisotropic characteristic of the monocrystalline silicon in the alkali solution can be effectively utilized, compared with the Bosch process, the method can generate the fin side surface with better steepness, the smoothness of the side surface is also obviously superior to the processing result of the Bosch process (as shown in figure 9B, after the etching by the high-temperature KOH solution, the roughness of most areas is less than 10 nm), the requirement of the subsequent film growth can be met without a hydrogen annealing process, and the market price of strong alkali such as potassium hydroxide with electronic grade purity is low, so the cost can be greatly reduced.

A disadvantage of wet etching <110> type silicon wafers is poor handling of certain profiles, inevitably producing additional profiles in addition to the steep <111> crystal planes available. Referring to fig. 10A and 10B, fig. 10A and 10B are schematic diagrams illustrating a distance change between an actuator branch and an inner wall of a speaker formed during wet etching according to an embodiment of the present invention. The fin mask of the patterned mask M3 shown in fig. 10A has a length L1 and a free end E1 with a gap between E1 and the other edge E2 of the mask. Experiments show that the free end of the fin structure generated by wet etching under the coverage of the mask M3 can not form a stable <111> crystal plane in etching liquid, so that the free end E1 can be continuously decomposed in a wet environment, and the actual length L2 of the fin is shortened in the direction of 4 left arrows and is shorter than the designed length L1; at the same time, the edge E2 will also move to the right (as shown by the arrow on the right in the figure) under the action of the etching liquid, so that the gap between the end of the fin and the boundary is larger than the designed size, as shown in fig. 10B, resulting in poor air tightness of the speaker.

To solve the above problem, referring to fig. 11A and 11B, there is a schematic view of forming the actuator branch end gap according to an embodiment of the present invention. The structure of fig. 7B can be fabricated using the mask of fig. 7A and the functional layers deposited, and then the gaps X1 shown in fig. 11A and 11B can be fabricated on each fin using a Bosch process with the addition of a mask. In this process, the Bosch process can achieve good control accuracy for the slit width d5 and the distance t1 between the slit and the apex of the <111> crystal plane P4 at the bottom of the trench. Since the two sidewalls of X1 do not need to deposit any film layer, the local roughness from the Bosch process does not have any impact on the device performance. By combining the wet method and the Bosch etching process, on one hand, a main molded surface with ideal smoothness and steepness can be obtained, and on the other hand, the precise control of a local process surface can be realized; meanwhile, the <110> type silicon wafer is selected to be compatible with the two processes.

According to the technical scheme of the embodiment of the invention, the through holes and the actuators of the MEMS loudspeaker are specially arranged, and each internal gap has a size requirement, which is beneficial to improving the output sound pressure of the loudspeaker. When the MEMS loudspeaker is applied to electronic equipment, the occupied space is small, and the volume is large.

The above-described embodiments should not be construed as limiting the scope of the invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions can occur, depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A MEMS speaker having a housing with upper and lower parallel bottom surfaces with a sidewall therebetween,

one or more of the upper bottom surface, the lower bottom surface and the side wall is provided with a through hole;

the actuator in the loudspeaker is connected to the side wall, and the vibration direction is generally parallel to the upper bottom surface and the lower bottom surface;

the actuator has a plurality of parallel plate-like branches;

the actuator is 1 layer, the distance between the upper top surface and the upper bottom surface of each branch of the actuator is not more than 20um, and/or the distance between the lower top surface and the lower bottom surface of each branch of the actuator is not more than 20um;

or the actuators are 2 or more layers distributed up and down, and in the two adjacent layers of actuators, the distance between the lower top surface of the upper layer of actuator and the upper top surface of the lower layer of actuator is not more than 20 mu m.

2. The MEMS speaker of claim 1,

according to the vibration direction, only 1 through hole is arranged between every two adjacent branches;

the air flow rate of the through hole is not less than the air flow rate of the gap between the shell and the actuator branch.

3. The MEMS speaker of claim 1,

piezoelectric layers are directly attached to two side faces of the silicon-based framework of each branch of the actuator, and electrode layers are attached to the piezoelectric layers;

or, directly attaching bottom electrodes to two side surfaces of the silicon-based framework of each branch of the actuator, and sequentially attaching piezoelectric layers and top electrodes to the bottom electrodes;

or, seed crystal layers are directly attached to two side faces of the silicon-based framework of each branch of the actuator, and a bottom electrode, a piezoelectric layer and a top electrode are sequentially attached to the seed crystal layers.

4. The MEMS loudspeaker of claim 3, wherein the material of the silicon-based framework is N-type or P-type doped silicon.

5. The MEMS speaker of claim 4, wherein the doped silicon has a conductivity of less than 2000 ohm cm.

6. The MEMS speaker of claim 1,

the side surface of each branch of the actuator is vertical to the upper bottom surface and/or the lower bottom surface;

for the side wall connected with the actuator branches, the side between the branches of two adjacent actuators is provided with an extension part extending towards the inside of the loudspeaker, and the projection of the extension part in the horizontal direction is pentagonal.

7. The MEMS speaker of claim 1,

the roughness of the side surface of the silicon-based framework of the actuator is not more than 50nm.

8. The MEMS speaker of claim 6,

the speaker is made of <110> type silicon.

9. The MEMS speaker of claim 1,

the edge of the top of the lower bottom surface of the loudspeaker has a ring of annular protrusions and/or the edge of the bottom of the upper bottom surface of the loudspeaker has a ring of annular protrusions.

10. The MEMS speaker of claim 9,

the annular bulge is formed by etching materials of the upper bottom surface or the lower bottom surface, or is a deposition layer deposited on the upper bottom surface or the lower bottom surface, or is a silicon dioxide layer of the SOI type wafer.

11. MEMS loudspeaker according to any of claims 1 to 10,

the actuator is a plurality of layers distributed up and down;

a circle of annular bulge is arranged between the side walls of all layers of the loudspeaker, and the annular bulge is formed by a deposited material layer or a silicon dioxide layer of an SOI (silicon on insulator) type wafer or by etching a material of the side walls.

12. MEMS loudspeaker according to any of claims 1 to 10,

in the case of 1 layer of actuator, the distance between the upper top surface and the upper bottom surface of each branch of the actuator is not more than 5um, and/or the distance between the lower top surface and the lower bottom surface of each branch of the actuator is not more than 5um.

13. MEMS loudspeaker according to any one of claims 1 to 10,

in the case of 2 or more actuators distributed above each other,

in two adjacent layers of actuators, the distance between the lower top surface of the upper layer actuator and the upper top surface of the lower layer actuator is not more than 5um.

14. MEMS loudspeaker according to any of claims 1 to 10,

one end of each branch of the actuator is connected to a first side wall of the loudspeaker, the other end of each branch is a free end, and the distance from the side wall opposite to the first side wall to the free end is not more than 100um;

or the branches of the actuator are divided into a first group and a second group which are respectively connected to two opposite side walls of the loudspeaker, each branch of the first group and each branch of the second group are aligned in pairs, and the distance between the top ends of the two aligned branches is not more than 100um;

or the branches of the actuator are divided into a first group and a second group which are respectively connected to two opposite side walls of the loudspeaker, each branch of the first group and each branch of the second group are staggered pairwise, and the distance between the side walls of the two staggered branches is not more than 100um.

15. MEMS loudspeaker according to any of claims 1 to 10,

one end of each branch of the actuator is connected to a first side wall of the loudspeaker, the other end of each branch of the actuator is a free end, and the distance from the side wall opposite to the first side wall to the free end is not more than 30um;

or the branches of the actuator are divided into a first group and a second group which are respectively connected to two opposite side walls of the loudspeaker, each branch of the first group and each branch of the second group are aligned pairwise, and the distance between the top ends of the two aligned branches is not more than 30um;

or the branches of the actuator are divided into a first group and a second group which are respectively connected to two opposite side walls of the loudspeaker, each branch of the first group and each branch of the second group are staggered in pairs, and the distance between the side walls of the two staggered branches is not more than 30um.

16. MEMS loudspeaker according to any of claims 1 to 10, wherein each branch of the actuator is connected at both ends to a respective set of opposite side walls of the loudspeaker.

17. An electronic device comprising the MEMS speaker of any one of claims 1 to 16.