Multi-diaphragm piezoelectric MEMS transducer and manufacturing method thereof

Abstract

The invention discloses a multi-diaphragm type piezoelectric MEMS (Micro Electro Mechanical System) transducer and a manufacturing method thereof. The multi-diaphragm type piezoelectric MEMS transducer can carry out frequency correction on the multi-diaphragm type piezoelectric MEMS transducer and realize a multi-diaphragm type piezoelectric MEMS transducer structure with high frequency consistency. The method comprises the following steps: determining the resonant frequency of each vibrating diaphragm in the transducer, and executing the following frequency correction step on each vibrating diaphragm to reduce the difference between the resonant frequency of the vibrating diaphragm and a specified frequency: if the resonant frequency of the vibrating diaphragm is lower than the specified frequency, reducing the mass of the vibrating diaphragm; and if the resonant frequency of the vibrating diaphragm is higher than the specified frequency, reducing the equivalent stiffness of the vibrating diaphragm.

Description

Technical Field

[0001] This invention relates to the field of microelectromechanical systems (MEMS) technology, and particularly to multi-diaphragm piezoelectric MEMS transducers and their manufacturing methods. Background Technology

[0002] Piezoelectric microelectromechanical systems (MEMS) ultrasonic transducers (hereinafter referred to as MEMS ultrasonic transducers or transducers) are a type of MEMS device that uses the direct and inverse piezoelectric effects of piezoelectric materials to vibrate a piezoelectric thin film, thereby emitting or receiving ultrasonic signals. When used to emit ultrasonic waves, a MEMS ultrasonic transducer is an actuator; when used to receive ultrasonic waves, it is a sensor. These characteristics allow the transducer to function as both an actuator (emitting sound waves) and a sensor (receiving sound waves). This means that in commercial products, identical transducers can perform two opposing functions, such as ultrasonic reversing radars for automobiles. Whether it's a single transducer that emits and receives ultrasonic waves, or two identical transducers, one emitting and one receiving, the same device design is used. This significantly reduces the cost of transducer devices from design to production, packaging, testing, and system integration. Furthermore, the standardized mass production processes of MEMS ultrasonic transducers make them highly suitable for commercial applications. In applications, MEMS ultrasonic transducers are typically in array form, containing multiple array elements, each capable of emitting / receiving ultrasonic waves. This structure presents a frequency consistency problem: because each element moves relatively independently, and manufacturing deviations occur during actual processing, resonant frequency differences arise between the elements. Furthermore, since transducers typically operate in the kHz to MHz range of resonant frequencies, even minute dimensional deviations (micrometer-level) can lead to considerable frequency differences, severely impacting the frequency consistency of MEMS ultrasonic transducer devices and causing a decline in device performance. Summary of the Invention

[0003] In view of this, the present invention proposes a multi-diaphragm piezoelectric MEMS transducer and its manufacturing method, which can perform frequency correction and achieve a high frequency consistency multi-diaphragm piezoelectric MEMS transducer structure. This solution allows for independent frequency correction of each diaphragm within the transducer device array, ensuring high frequency consistency and thus enabling the transducer to exhibit excellent performance. The present invention provides the following technical solution:

[0004] A method for manufacturing a multi-diaphragm piezoelectric MEMS transducer, wherein the transducer is an array of multiple array elements, each array element having multiple diaphragms, the diaphragms being cantilever beam structures, characterized in that the method includes: determining the resonant frequency of each diaphragm in the array elements, and performing the following frequency correction steps for each diaphragm to reduce the difference between the resonant frequency of the diaphragm and a specified frequency: if the resonant frequency of the diaphragm is lower than the specified frequency, then reducing the mass of the diaphragm; if the resonant frequency of the diaphragm is higher than the specified frequency, then reducing the equivalent stiffness of the diaphragm.

[0005] Optionally, before determining the resonant frequency of each diaphragm in the array element, the method further includes: fabricating a frequency-correcting layer above the piezoelectric structure of each diaphragm of the array element; and the step of reducing the mass of the diaphragm includes: partially removing the material of the frequency-correcting layer.

[0006] Optionally, the step of partially removing the material from the frequency correction layer includes removing the material from the frequency correction layer by means of focused particle beam bombardment or laser beam irradiation.

[0007] Optionally, the frequency correction layer includes a protective film layer above the piezoelectric structure of each diaphragm and a metal film on the surface of a predetermined frequency correction region near the free end of the protective film layer; the step of partially removing the material of the frequency correction layer includes: partially removing the metal film.

[0008] Optionally, the step of partially removing the metal film includes: oxidizing a portion of the metal film; irradiating the oxidized metal film with a laser beam, thereby partially or completely removing the oxidized metal film.

[0009] Optionally, the step of partially removing the metal film includes: oxidizing a portion of the metal film or oxidizing all of the metal film; and partially or completely removing the oxidized metal film.

[0010] Optionally, the step of partially or completely removing the oxidized metal film includes: irradiating the oxidized metal film with a laser beam or bombarding the metal film with a particle beam, thereby partially or completely removing the oxidized metal film.

[0011] Optionally, the step of oxidizing a portion of the metal thin film includes one of the following: oxidizing a portion of the metal thin film in the thickness direction, wherein the oxidized region is a continuous region or multiple separate regions in the planar extension direction of the metal thin film; forming multiple separate oxidized regions on the metal thin film, wherein the thickness of all oxidized regions is equal to or less than the thickness of the metal thin film, or only a portion of the oxidized regions has a thickness equal to the thickness of the metal thin film.

[0012] Optionally, the step of reducing the equivalent stiffness of the diaphragm includes: forming a qualitative change region in a preset frequency correction region near the fixed end of the diaphragm, thereby reducing the equivalent stiffness of the diaphragm.

[0013] Optionally, the qualitative change region is a series of regions distributed along the edge of the fixed end.

[0014] Optionally, the step of forming the qualitative change region includes: irradiating the frequency-correcting region with a laser beam.

[0015] Optionally, before the step of reducing the mass of the diaphragm, the method further includes: making the resonant frequency of all diaphragms of the array element lower than the specified frequency; before the step of reducing the equivalent stiffness of the diaphragm, the method further includes: making the resonant frequency of all diaphragms of the array element higher than the specified frequency.

[0016] Optionally, the plurality of diaphragms are centrally symmetrically distributed, with the cavity of the vibrating element disposed below; before performing the frequency correction step, the method further includes: determining the offset vector between the geometric center of the cavity boundary and the symmetry center of the plurality of diaphragms in the horizontal and vertical directions respectively using photolithographic alignment marking method, the offset vector being used to determine the magnitude relationship between the resonant frequency of the diaphragm and a specified frequency.

[0017] Optionally, before determining the resonant frequency of each diaphragm in the array element, the method further includes: setting four congruent right-angled isosceles triangular diaphragms in the array element, the four diaphragms being centrally symmetrically distributed to form a square, and a cavity of the array element being set below, wherein two branches are formed at the ends of the gaps between the four diaphragms to form a Y-shape, the two branches being perpendicular to each other, and the edges of the two branches away from the right-angled isosceles triangles being straight lines, and parallel to the edges of the cavity respectively.

[0018] Optionally, before performing the frequency correction step, the method further includes: determining an offset vector between the geometric center of the cavity boundary and the center of symmetry of the plurality of diaphragms based on the distance between the straight line and the boundary of the cavity parallel to and closest to the straight line, the offset vector being used to determine the magnitude relationship between the resonant frequency of the diaphragm and a specified frequency.

[0019] Optionally, before determining the resonant frequency of each diaphragm in the array element, the method further includes: setting four congruent right-angled isosceles triangular diaphragms in the array element, the four diaphragms being centrally symmetrically distributed to form a square, and a cavity of the array element being set below it. The gap between adjacent diaphragms has a first segment and a second segment. The beginning of the first segment is located at the right-angle vertex of the right-angled isosceles triangle, and the end is connected to the beginning of the second segment. The end of the second segment is the end of the gap between adjacent diaphragms, and the end of the first segment has a horizontal line segment and a vertical line segment, which are parallel to a set of adjacent sides of the square, respectively.

[0020] Optionally, before performing the frequency correction step, the method further includes: determining an offset vector between the geometric center of the cavity boundary and the symmetry center of the plurality of diaphragms based on the lateral offset and the longitudinal offset, wherein the offset vector is used to determine the magnitude relationship between the resonant frequency of the diaphragm and a specified frequency; wherein the lateral offset is the distance between the lateral line segment and the boundary of the cavity parallel to the lateral line segment and closest to the lateral line segment, and the longitudinal offset is the distance between the longitudinal line segment and the boundary of the cavity parallel to the longitudinal line segment and closest to the longitudinal line segment.

[0021] A multi-diaphragm piezoelectric MEMS transducer, wherein the transducer is an array of multiple array elements, each array element having multiple diaphragms, the diaphragms being cantilever beam structures, characterized in that the array elements are manufactured using the method described in this invention.

[0022] A multi-diaphragm piezoelectric MEMS transducer, wherein the transducer is an array of multiple array elements, each array element having multiple diaphragms, the diaphragms being cantilever beam structures, characterized in that, for one or more array elements of the transducer, each diaphragm of the array element has a frequency correction layer above its piezoelectric structure, and the quality of the frequency correction layer on each diaphragm in the array element is inconsistent.

[0023] Optionally, the frequency correction layer includes a protective film layer above the piezoelectric structure of each diaphragm and a metal film on the surface of a predetermined frequency correction region near the free end of the protective film layer; the quality of the metal film on each diaphragm in the same element is inconsistent.

[0024] A multi-diaphragm piezoelectric MEMS transducer is disclosed. The transducer is an array of multiple array elements, each array element having multiple diaphragms. The diaphragms are cantilever beam structures. For one or more array elements of the transducer, there is a qualitative change region in a preset frequency correction region near the fixed end of one or more diaphragms of the array element, which reduces the equivalent stiffness of the diaphragm.

[0025] Optionally, the qualitative change region is a series of regions distributed along the edge of the fixed end.

[0026] Optionally, the array element comprises four congruent right-angled isosceles triangular diaphragms, which are arranged to form a square, with the cavity of the array element below. The ends of the gaps between the four diaphragms have two branches to form a Y-shape, and the two branches are perpendicular to each other. The edges of the two branches away from the right-angled isosceles triangles are straight lines and are parallel to the edges of the cavity.

[0027] Optionally, the array element comprises four congruent right-angled isosceles triangular diaphragms, which are arranged to form a square, with the cavity of the array element below. The gap between adjacent diaphragms has a first segment and a second segment. The beginning of the first segment is located at the right-angle vertex of the right-angled isosceles triangle, and the end is connected to the beginning of the second segment. The end of the second segment is the end of the gap between adjacent diaphragms. The end of the first segment has a horizontal line segment and a vertical line segment, which are parallel to a set of adjacent sides of the square, respectively.

[0028] According to the technical solution of the present invention, for a multi-diaphragm piezoelectric MEMS transducer, each diaphragm in the array element is independently frequency-corrected. During frequency correction, the frequency of the diaphragm is increased by reducing the mass of the diaphragm or decreased by reducing the equivalent stiffness of the diaphragm, so as to ensure that the multiple diaphragms in the array element and the multiple array elements in the array maintain a high frequency consistency. Attached Figure Description

[0029] For illustrative and not limiting purposes, the invention will now be described with reference to preferred embodiments thereof, particularly the accompanying drawings, in which:

[0030] Figure 1A A top view of a typical structure of a single array element in a piezoelectric MEMS ultrasonic transducer;

[0031] Figure 1B for Figure 1A A cross-sectional structure;

[0032] Figure 1C for Figure 1A Another cross-sectional structure;

[0033] Figure 1D for Figure 1A Another type of cross-sectional structure;

[0034] Figure 2A yes Figure 1A A schematic diagram with some parts of the structure omitted;

[0035] Figure 2B It shows Figure 2A The geometry of the diaphragm R100 in the middle;

[0036] Figure 2C for Figure 2AA magnified schematic diagram of the central region;

[0037] Figure 2D for Figure 2A Another enlarged diagram of the central region;

[0038] Figure 2E This is a schematic diagram of the orthogonal decomposition of vector OP;

[0039] Figure 2F This is a schematic diagram illustrating the effect of vector offset on the area of ​​an equilateral triangular diaphragm.

[0040] Figure 2G The measured impedance curve of one element of a multi-diaphragm piezoelectric MEMS transducer is shown.

[0041] Figure 3 This is a schematic diagram of the frequency correction region near the free end of a triangular diaphragm.

[0042] Figure 4A This is a top view that includes the frequency correction region near the free end of the diaphragm;

[0043] Figure 4B This is a schematic diagram of a protective film deposited on the upper surface of the diaphragm.

[0044] Figure 4C A schematic diagram of a metal thin film deposited on the upper surface of a protective coating layer;

[0045] Figure 4D A schematic diagram showing the complete conversion of a metal thin film into an oxide layer of the metal;

[0046] Figure 4E This is a schematic diagram of partial oxidation along the thickness direction of a metal thin film.

[0047] Figure 4F This is a schematic diagram of patterned oxidation of a metal thin film in the horizontal direction;

[0048] Figure 4G This is a schematic diagram illustrating the oxidation process that combines the thickness of a metal thin film with its horizontal orientation.

[0049] Figure 4H A schematic diagram of one method of patterned oxidation of a metal thin film;

[0050] Figure 4I This is a schematic diagram of an oxide region of various sizes in a metal thin film.

[0051] Figure 4J for Figure 4I Top view;

[0052] Figure 5A and Figure 5B This is a schematic diagram illustrating frequency correction using a laser beam or particle beam.

[0053] Figure 5C This is a schematic diagram of the free end of the diaphragm after frequency correction.

[0054] Figure 5D For Figure 4F A schematic diagram of a thin metal film being irradiated with a laser beam;

[0055] Figure 5E To adopt Figure 5D A schematic diagram of the free end of the diaphragm after processing as shown;

[0056] Figure 6A This is a schematic diagram showing the area near the fixed end of the diaphragm illuminated by a laser beam;

[0057] Figure 6B A schematic diagram of the frequency correction region near the fixed end of the diaphragm;

[0058] Figure 7A This is a schematic diagram of the diaphragm frequency correction area and a magnified portion thereof;

[0059] Figure 7B This is a schematic diagram showing frequency correction applied only to the free end region of the diaphragm.

[0060] Figure 7C This is a schematic diagram showing frequency correction applied only to the fixed-end region of the diaphragm.

[0061] Figure 8A This is a schematic diagram of the first type of photolithographic alignment mark;

[0062] Figure 8B This is a schematic diagram of the second type of photolithographic alignment mark;

[0063] Figure 8C This is a schematic diagram illustrating a method of stacking the first type of photolithographic alignment mark and the second type of photolithographic alignment mark.

[0064] Figure 8D This is a schematic diagram illustrating another method of superimposing the first type of photolithographic alignment mark and the second type of photolithographic alignment mark;

[0065] Figure 8E A schematic diagram and a magnified view of the alignment marks distributed on the substrate wafer;

[0066] Figure 9A This is a magnified view of the cross slit between the diaphragms and the ends of the slit branches.

[0067] Figure 9B A schematic diagram showing the offset of the cross slit relative to the back cavity;

[0068] Figure 9C A schematic diagram showing the curve appearing at the end of the slit;

[0069] Figure 9D A schematic diagram showing a Y-shaped structure at the end of the slit;

[0070] Figure 10 This is a top view of an array that includes multiple array elements.

[0071] In the XYZ coordinate system shown in the attached diagram, the points inside the circles represent the positive direction of the axis as they are closer, and the crosses represent the positive direction of the axis as they are farther away. Detailed Implementation

[0072] The structure of the piezoelectric MEMS ultrasonic transducer according to the embodiments of the present invention will be described first, and then the frequency correction method in the embodiments of the present invention will be described.

[0073] Figure 1A This is a top view of a typical structure of a single element of a piezoelectric MEMS ultrasonic transducer. Its core working part contains four triangular diaphragms R100-R400, and each transducer comprises multiple... Figure 1A The array shown consists of four triangular diaphragms. In each element, the adjacent sides of four triangular diaphragms form a cross-shaped slit F1. Each triangular diaphragm also includes at least one piezoelectric sandwich structure and one supporting membrane structure. A back cavity structure is located at the bottom of the diaphragm structure (i.e., in the negative Z-axis direction shown in the figure), and the boundary of this cavity is B1. Beside the core diaphragm (taking R400 as an example), there are also electrical connection structures M101, M111, M121, M131, and M100, M110, M120, which serve to achieve electrical connections between the core diaphragm and external circuitry. Clearly, this demonstrates… Figure 1A The structure described is for illustrative purposes only and not for limiting the structure and method of the embodiments of the present invention. For example, the transducer array elements involved in the embodiments of the present invention are not limited to triangles, their number is not limited to four, the slits are not limited to cross shapes, and the supporting membrane structure is not mandatory. The arrangement of the array elements is not limited to a centrally symmetrical arrangement, but can also be other symmetrical forms; it can also be an asymmetrical arrangement.

[0074] Figure 1B for Figure 1A A cross-sectional structure (along) Figure 1A(Cut open along straight line AA'), where D100 is the support layer structure; the bottom electrode B100, piezoelectric film P100, and top electrode T100 constitute the first sandwich structure and are located below the support layer; the bottom electrode B200, piezoelectric film P200, and top electrode T200 constitute the second sandwich structure and are located above the support layer. A slit F1 is formed between the diaphragm on the left and the diaphragm on the right (a cross-section of the central part of the cross-shaped slit in the top view). The lower part of the diaphragm structure is a cavity structure, the boundary B1 of which is defined by the intersection line formed by the contact between the dielectric material layer Y100 and the surface of the top electrode T100 of the first sandwich structure.

[0075] Figure 1C for Figure 1A Another cross-sectional structure (along) Figure 1A (Cut open along straight line AA'), where F200 is the support layer structure, the bottom electrode B100, the piezoelectric film P100 and the top electrode T100 form the first sandwich structure and are located at the bottom of the support layer; the bottom electrode B200, the piezoelectric film P200 and the top electrode T200 form the second sandwich structure and are located at the top of the support layer. A slit F1 is formed between the diaphragm on the left and the diaphragm on the right (a cross-section of the central part of the cross-shaped slit in the top view). The lower part of the diaphragm structure is a cavity structure, the boundary B1 of which is defined by the intersection line formed by the contact between the substrate material layer S100 and the surface of the bottom electrode B100 of the first sandwich structure.

[0076] Figure 1D for Figure 1A Another cross-sectional structure is shown, where D100 is the support layer structure, which has only a single-layer piezoelectric sandwich film structure. The bottom electrode B100, the piezoelectric film P100, and the top electrode T100 constitute the first sandwich structure and are located on top of the support layer D100. A slit F1 is formed between the diaphragm on the left and the diaphragm on the right (the cross-section of the central part of the cross-shaped slit in the top view). The lower part of the diaphragm structure is a cavity structure, the boundary B1 of which is defined by the intersection line formed by the contact between the dielectric material layer Y100 and the surface of the support layer D100.

[0077] Figure 2A By omitting Figure 1A The electrodes and other electrical connection structures within the cavity are extracted to visualize the edge F1 of the cross slit and the edge B1 of the back cavity, allowing for a description and analysis of their relative positional relationship. Theoretically, the cross slit F1 is centrally symmetrical, and the cavity edge B1 is square.

[0078] Figure 2B It shows Figure 2AThe diaphragm R100 has a geometry that is an isosceles right triangle ABC, where AB and AC are the legs, and the hypotenuse BC is part of the cavity edge B1. Triangle ABC has a certain height h, and its hypotenuse BC has a length of 2h. In one embodiment, diaphragms R100-R400 are congruent triangles, and the four branches of the cross slit have equal widths. The right-angle vertex A is the free end or suspended end of diaphragm R100, and the hypotenuse BC is its fixed end or root end.

[0079] If Figure 2A If the central region E1 is magnified, then the following can be obtained: Figure 2C Details of the positional relationships shown. Figure 2C In the diagram, D1 and D2 are the diagonals of square B1, intersecting at point O. Point P is the geometric center of the cross slit F1. Figure 2C It is known that, in order to ensure that the four triangular diaphragms have the same dimensions (i.e., the resonant frequencies of the four triangular diaphragms are the same), it is required that, on the one hand, the geometric center P of the cross slit F1 should coincide with the geometric center O of the cavity boundary B1, and on the other hand, the diagonal of the cavity boundary B1 should be the centerline of each branch of the cross slit F1. However, in actual manufacturing, due to limitations in the precision of the process equipment (mainly the photolithography patterning stage) and other factors, there may be a relative positional offset between the slit F1 and the cavity edge B1.

[0080] like Figure 2D As shown, the center P of the cross slit F1 will usually be slightly offset by a vector relative to the center O of the cavity boundary B1. At this time, the points O and P will no longer coincide, and the points O and P will form a vector OP of length ε.

[0081] like Figure 2E As shown, if the offset ε is orthogonally decomposed along the x and y directions, the components ε along the x and y directions can be obtained. x and ε y .

[0082] The aforementioned minute vector offset will have the following effect on the area of ​​each equilateral triangular diaphragm: (e.g.) Figure 2F As shown, with Figure 1A and Figure 2A Taking the diaphragm R100 as an example, when the center P of the cross slit shifts relative to O, OP = (ε x ,ε y When h is changed to h+ε, the height of the original triangle ABC will change from h to h+ε. y Its hypotenuse length will change from 2h to 2(h+ε) y For a diaphragm R100, the area S' of the new triangle A'B'C' is... R100 Compared to the area S of the original triangle ABC R100 The resulting change ΔSR100 It can be approximated as:

[0083] ΔS R100 =(h+ε y ) 2 -h 2 ≈2hε y (1)

[0084] Similarly, under the same center offset, the area change ΔS of the diaphragm R300 located on the opposite side of R100 is obtained. R300 It can be approximated as:

[0085] ΔS R300 =(h-ε y ) 2 -h 2 ≈-2hε y (2)

[0086] Similarly, for the area change ΔS of R200 and R400... R200 and ΔS R400 We can also obtain approximate expressions:

[0087] ΔS R200 =(h-ε x ) 2 -h 2 ≈-2hε x (3)

[0088] ΔS R400 =(h+ε x ) 2 -h 2 ≈2hε x (4)

[0089] As can be seen from expressions (1)-(4), the area change of each triangular diaphragm is linearly related to its corresponding offset component, and the area change of a certain diaphragm triangle is always equal in magnitude but opposite in sign to the area change of the diaphragm triangle on the opposite side.

[0090] For example, when the offset component ε y >0, ε x When ΔS > 0, R100 >0, ΔS R300 <0, ΔS R200 <0, ΔS R400 >0, that is, S R100 The area increases, S R300 The area becomes smaller, S R200 The area becomes smaller, S R400 The area has increased.

[0091] When the areal density of the diaphragm is considered constant, the mass change of each diaphragm due to displacement is proportional to the change in its area, that is, proportional to the component of the displacement ε in the direction perpendicular to the hypotenuse of the diaphragm. For example, if the hypotenuse of R100 is in the x-direction, then the mass change of R100 due to displacement is proportional to ε. y .

[0092] Furthermore, changes in the height or area of ​​the diaphragm triangle caused by the offset will also cause changes in the stiffness of the diaphragm. The rule is that the greater the height or area of ​​the diaphragm, the smaller the stiffness of the diaphragm, and the easier the diaphragm is to bend.

[0093] If the vibration of the diaphragm in a small amplitude range is approximated as simple harmonic motion, then the resonant frequency can be expressed as:

[0094]

[0095] Where s is the diaphragm area, k is the stiffness of the equivalent harmonic oscillator, m is the mass of the diaphragm, and ψ is the mass of the equivalent harmonic oscillator, with ψ increasing as m increases. From expression (5), it can be seen that the change in diaphragm area s caused by the offset will cause a change in the resonant frequency of the diaphragm. Taking R100 as an example, when ε... y When the value is greater than 0, the area of ​​R100 increases, its mass m increases, and the mass ψ of its equivalent harmonic oscillator also increases, while the stiffness k of R100 decreases as the area increases. Thus, the net effect of the above changes is that the fractional value on the right side of expression (5) decreases, and therefore the resonant frequency ν of R100 decreases.

[0096] Ideally, the four diaphragms of each piezoelectric micromechanical ultrasonic transducer should have the same resonant frequency. However, as the above analysis shows, the offset will cause the four diaphragms of the same piezoelectric micromechanical ultrasonic transducer to have different resonant frequencies. Since the operating frequency of piezoelectric micromechanical ultrasonic transducers is usually in the kHz to MHz range (40kHz to 40MHz), this means that even a small offset (on the order of micrometers) can cause the actual resonant frequency of the diaphragm to deviate significantly from the theoretical value, resulting in significant frequency differences between the diaphragms and thus a severe decline in the performance of the piezoelectric micromechanical ultrasonic transducer.

[0097] Figure 2G The figure shows the measured impedance curve of one element in a multi-diaphragm piezoelectric MEMS transducer. The dashed line represents the ideal case with only one resonance peak; the solid line represents the measured resonance peak, showing four peaks with different resonant frequencies. Therefore, to ensure the performance of the piezoelectric micromechanical ultrasonic transducer, a frequency correction process is needed during actual manufacturing to minimize the difference between the actual operating frequency and the theoretical frequency of the diaphragm and to improve the frequency consistency of the multiple diaphragms.

[0098] Specifically, in this embodiment, based on expression (5) and combined with the actual processing structure, the following solution is proposed:

[0099] 1. If the resonant frequency ν is lower than the theoretical value, the frequency can be corrected by reducing the equivalent mass ψ, that is, by reducing the diaphragm mass m.

[0100] 2. If the resonant frequency ν is higher than the theoretical value, the frequency can be corrected by reducing the equivalent stiffness k of the diaphragm.

[0101] Furthermore, for Scheme 1, a focused particle beam bombardment or laser beam irradiation method is used to remove part of the material on or inside the diaphragm surface. Simultaneously, since the equivalent mass ψ is more sensitive to mass changes near the free end of the triangular diaphragm, removing a small amount of material near the free end can significantly change the equivalent mass ψ. This reduces processing costs and improves processing efficiency, while also minimizing the adverse effects of laser or particle beams on the stability of the membrane structure during frequency correction. Here, "near the free end" refers to... Figure 3 The isosceles right-angled triangular diaphragm ABC is shown with a triangular region AM1M2 bounded by the right-angle vertex A and the midline M1M2 parallel to the hypotenuse BC. For scheme 2, laser irradiation can be used to alter the microstructure of the diaphragm near the fixed end, thereby reducing the equivalent resonator stiffness k of the diaphragm. Here, "near the fixed end" refers to... Figure 3 The isosceles right-angled triangular diaphragm ABC shown has a trapezoidal strip region N1N2CB formed by parallel line segments N1N2 to the hypotenuse BC, and the height of the trapezoid is 5%-10% of the distance h from the right-angle vertex A of triangle ABC to the hypotenuse BC. It should be noted that because the piezoelectric micromechanical ultrasonic transducer elements are small, with each diaphragm within the element being less than 1 mm or even less than 100 micrometers, focused ion beams or laser beams can precisely locate the position of the diaphragm "near the free end" or "near the fixed end" and remove material in a high-throughput manner, which is impossible with other methods.

[0102] Obviously Figure 3 The area where the scheme shown is implemented is also applicable to other diaphragms in piezoelectric micromechanical ultrasonic transducer devices. Figure 3 The frequency correction structure near the free end can include several specific forms, which are explained below.

[0103] Figure 4A This is a top view including the frequency-correcting region AM1M2 near the free end of the diaphragm. A series of cross-sectional flowcharts and structural diagrams can be obtained along line AD.

[0104] like Figure 4BAs shown, a protective film layer PT100 is first deposited on the upper surface of the diaphragm, where PT100 mainly covers the entire upper surface and sides of the top electrode T200. The material of PT100 can be silicon dioxide, aluminum nitride, silicon carbide, silicon nitride, etc., and its function is to reduce the damage to the underlying structures such as electrodes and piezoelectric films by subsequent frequency-correcting ion beams or lasers.

[0105] like Figure 4C As shown, a frequency correction layer MD100 is deposited on the upper surface of PT100, and its coverage area is limited to the region near the free end of the diaphragm by graphical representation (within the triangle AM1M2 in the top view, and to the right of the boundary M1M2 in the cross-sectional view). The material of the frequency correction layer MD100 can be a metal such as molybdenum, aluminum, or copper.

[0106] As mentioned earlier, when the resonant frequency ν is lower than the theoretical value, it is necessary to reduce the diaphragm mass. During manufacturing, for each diaphragm, regardless of whether its mass needs to be reduced in the future, the aforementioned protective film layer PT100 and metal thin film MD100 are pre-formed on it, so that the mass of each diaphragm can be further reduced to adjust the frequency.

[0107] like Figure 4D As shown. The metal frequency correction layer is processed through an oxidation process. Figure 4C The MD100 in the metal is completely converted into the oxide layer AD100 of the metal.

[0108] like Figure 4E As shown, partial oxidation in the thickness direction (z direction) can also be achieved by controlling the oxidation process parameters, so that the upper surface of MD100 is partially oxidized and transformed into AD100.

[0109] like Figure 4G As shown, it can also be combined with the control of oxidation process in the thickness and horizontal direction to achieve incomplete patterned oxidation in the thickness.

[0110] Figure 4H A possible embodiment of the patterned oxide is shown in a top view, where the shaded dots within the triangle AM1M2 represent the oxide layer AD100, and the blank area within the triangle represents the metal thin film MD100.

[0111] exist Figure 4G Based on the structure, further changes can be made, such as... Figure 4I As shown, the size of the oxide region AD100-1 is larger than the size of AD100-2. Optionally, the thickness of AD100-1 can also be greater than the thickness of AD100-2.

[0112] Figure 4J Given Figure 4IA possible embodiment of the patterned oxide structure in a top view, wherein the size of the triangular AM1M2 range oxide layer AD100-1 is larger than that of AD100-2, and AD100-1 and AD100-2 are arranged alternately in the y direction.

[0113] The following explains the removal operation of the frequency trimming layer. For example... Figure 5A , Figure 5B As shown, for Figure 4D The frequency-correcting layer structure shown can have part of its mass removed by irradiating or bombarding the metal oxide layer AD100 with a laser beam or particle beam LB. Figure 5A In the process, within the energy focusing range, the metal oxides are completely removed in terms of thickness, while... Figure 5B In this process, partial removal of metal oxides on the thickness is achieved by controlling the beam power and the interaction time.

[0114] Figure 5C A possible top view pattern is given after partial removal of the metal oxide AD100 by a beam, where the shaded area within triangle AM1M2 represents the metal oxide region untreated by the beam; the striped blank areas represent regions where the metal oxide has been completely or partially removed by the beam in thickness, exposing the underlying protective layer PT100 or the deeper metal oxide AD100. The oxide-removed region PT100 may have symmetry with respect to the center line AD of the triangle in the top view.

[0115] like Figure 5D As shown, for Figure 4F ( Figure 4G and Figure 4I The frequency correction structure in this method can also be used, where the metal oxide AD100 is partially sublimated by irradiation with a laser beam LB, while the metal MD100 is retained because it has a relatively low absorption rate of laser energy.

[0116] Figure 5E A possible top view pattern is given after partial beam removal of the patterned metal oxide AD100, where the shaded dots within triangle AM1M2 represent the metal oxide region untreated by the beam, and the blank dotted areas represent the regions where the metal oxide has been removed by the beam. The oxide removal areas (blank dots) may have symmetry with respect to the center line AD of the triangle in the top view.

[0117] The frequency modulation layer structure preferably uses a laser beam for frequency correction because if the particle beam bombards the metal retained in the frequency modulation layer, it may cause metal particles or debris to splash onto the surrounding electrical structure, causing short circuits or other adverse effects.

[0118] The above describes the technical solution for frequency correction by reducing the diaphragm mass m. The following describes the solution for frequency correction in this invention by reducing the equivalent stiffness k of the diaphragm.

[0119] Figure 6A This is a schematic diagram showing the application of a laser beam to the area near the fixed edge of the diaphragm. The area being irradiated is the frequency correction region. Figure 6A As shown, irradiation with a laser beam LB causes microstructural changes in the internal film structure near the fixed end of the diaphragm, such as the support layer D100, resulting in the localized formation of a polycrystalline or porous region PL100. Here, the laser beam LB enters from the substrate S100 towards the support layer D100. It should be noted that because the piezoelectric micromechanical ultrasonic transducer elements are small, with each diaphragm within the element being less than 1 mm or even less than 100 micrometers in size, using a laser beam allows for precise positioning of the diaphragm "near the fixed end" and high-throughput alteration of the internal film structure, which is impossible with other methods.

[0120] Unlike frequency correction methods that partially remove additional structures near the free end, frequency correction processes near the fixed end alter or degrade the internal structure of the diaphragm itself. Essentially, this constitutes localized damage to the diaphragm. When the degraded area is too large or too concentrated, it will adversely affect the reliability of the diaphragm. Therefore, the qualitative change region PL100 generated by frequency correction should preferably be distributed in segments. Figure 6B A top-view distribution of a possible qualitative change region formed by frequency correction near a fixed end is given. The region near the fixed end is a volume N1N2CB, and the qualitative change region PL100 within this region is a series of shaded dots distributed parallel to the base BC of the trapezoid.

[0121] The following explains this frequency correction scheme from a quantitative perspective. For example... Figure 7A As shown ( Figure 7A The left half of the diagram exaggerates the size of the cross slit F1 to more clearly show its offset relative to the diagonals of its lower cavity square boundary B1, when the diaphragm of the piezoelectric micromechanical ultrasonic transducer device (represented by the center P of its cross slit F1) undergoes a small vector offset relative to the center O (two diagonals D1 and D2) of its lower cavity square boundary B1 (see...). Figure 7A (The right half of the image), based on the analysis above, the area of ​​the four diaphragms will change slightly, and this change is approximately proportional to the projection of the offset vector OP onto the direction perpendicular to the hypotenuse of one of the diaphragms. Therefore, according to... Figure 7AFrom the offset shown in the example, it can be seen that the areas of R100 and R400 increase, while the areas of R200 and R300 decrease. Furthermore, it can be seen that the actual resonant frequencies of R100 and R400 will decrease compared to the theoretical values, while the frequencies of R200 and R300 will increase. And according to... Figure 7A From the direction of vector OP in the right half, we can see that the area increment of R400 is greater than that of R100, while the area reduction of R300 is less than that of R200. Therefore, according to formula (7), the resonant frequency is:

[0122] ν R200 >ν R300 >ν0>ν R100 >ν R400 Where ν0 is the theoretical resonant frequency. Therefore, frequency trimming is needed to make the actual resonant frequencies of each diaphragm as close as possible to a single frequency, such as ν0. Since frequency trimming near the free end or the fixed end can only reduce mass or stiffness, for R100 and R400 with frequencies lower than the theoretical value, frequency trimming near the free end can reduce the diaphragm mass, while for R200 and R300 with frequencies higher than the theoretical value, frequency trimming near the fixed end can reduce the diaphragm stiffness. For example... Figure 7A As shown, graphic modifier structures AD100-1 and AD100-4 (represented by shaded dots) are distributed near the free ends of R100 and R400. Through modifier treatment, a certain mass of modifier structures (represented by hollow dots) is removed from R100 and R400, respectively. Graphic modifier structures PL100-2 and PL100-3 (represented by hollow dots) are distributed near the fixed ends of R200 and R300, respectively. These structures are used to reduce the equivalent stiffness of the diaphragms R200 and R300.

[0123] Furthermore, since ν0 > ν R100 >ν R400 Therefore, as Figure 7A As shown in the left half of the diagram, the mass removed from the region near the free end of R400 (3 hollow dots) is greater than the mass removed from the region near the free end of R100 (1 hollow dot). However, it should be noted that the number of hollow dots on R400 and R100 in the diagram is only used to qualitatively represent the magnitude relationship and does not represent specific numerical values. Figure 7A The left half can also be seen, due to ν R200 >ν R300 >ν0, the density of the ground frequency correction quality change point PL100-2 arranged near the fixed end of R200 is higher than the density of the ground frequency correction quality change point PL100-3 arranged near the fixed end of R300.

[0124] Furthermore, the diaphragm frequency in the device initially obtained through actual processing can be controlled through design, so that... Figure 7A In the example, the uncorrected frequencies of each diaphragm are shifted relative to ν0, resulting in the relationship: ν0 > ν0. R200 >ν R300 >ν R100 >ν R400 In this case, it is possible to... Figure 7B As shown, frequency correction is performed only on the free end region of each diaphragm without operating near the fixed end of the diaphragm. For example, based on the frequency relationship of each diaphragm, less mass can be removed from the diaphragm with higher frequency, such as forming 3 hollow circles on R200, and more mass can be removed from the diaphragm with lower frequency, such as forming 6 hollow circles on R400.

[0125] Furthermore, the diaphragm frequency in the device initially obtained through actual processing can be controlled through design, so that... Figure 7A In the example, the uncorrected frequencies of each diaphragm are shifted relative to ν0, forming the following relationship: ν R200 >ν R300 >ν R100 >ν R400 >ν0. In this case, it can be as follows: Figure 7C As shown, frequency correction is performed only in the area near the fixed end of each diaphragm without operating on the free end area of ​​each diaphragm. For example, based on the frequency relationship of each diaphragm, more or denser frequency correction points can be formed at the root of the diaphragm with higher frequency, such as forming a frequency correction point array PL100-2 on R200, and fewer or sparser frequency correction points can be formed at the root of the diaphragm with lower frequency, such as forming a frequency correction point array PL100-4 on R400. The number and distribution density of frequency correction points in array PL100-4 are both less than those in PL100-2.

[0126] As analyzed above, the projected lengths of the offset vector OP in the x and y directions determine the magnitude of the offset in geometric or physical quantities such as the area, mass, stiffness, and frequency of each diaphragm. This, in turn, determines the degree of frequency correction applied to each diaphragm, such as the amount of mass removed and the density of the quality change points. However, since features such as the crosshair center P and the cavity boundary center O do not actually exist on the silicon wafer, the offset OP needs to be reflected by the deviation between measurable features that actually exist on the wafer surface during the manufacturing process. This provides a quantitative basis for frequency correction.

[0127] This invention proposes two methods for determining the offset: the photolithographic alignment marking method and the on-chip feature comparison method. The photolithographic alignment marking method will be described first. This method is applicable when the cavity machining photolithographic patterning and the cross-slit photolithographic patterning are located on the same side of the substrate wafer. Figure 8AThe scale-type photolithographic alignment marks along the x-direction of the corresponding cavity photolithographic patterning layer G1 are given, where P0 is the main scale, and several sub-scales L1-L5 and R1-R5 are symmetrically distributed on the left and right sides of P0.

[0128] Figure 8B The scale-type photolithographic alignment marks corresponding to the cross slit photolithographic patterning layer G2 are given, where P0' is the main scale, and several sub-scales L1'-L5' and R1'-R5' are symmetrically distributed on the left and right sides of P0'.

[0129] like Figure 8C As shown, the alignment marks of layer G2 have unique dimensional design features relative to layer G1: In the ideal state where scales P0 and P0' are perfectly aligned in the x-direction (corresponding to the crosshair center and cavity edge center being perfectly aligned in the x-direction), the corresponding scale on the right, R1', is offset to the right by δ relative to R1; R2' is offset to the right by 2 times δ relative to R2; R3' is offset to the right by 3 times δ relative to R3, and so on. The relative offset of the corresponding scale on the left is symmetrical to that of the scale on the right about the main scale, i.e., L1' is offset to the left by δ relative to L1, and so on. The value of δ can be designed according to the required accuracy.

[0130] In actual processing, P0' and P0 usually deviate in the x-direction. The rule for reading this deviation is as follows: find the set with the smallest deviation among the scale divisions, and take the negative of the theoretical design deviation of this set of scale divisions under the ideal alignment state of P0' and P0 as the current deviation value of P0' relative to P0. For example, in Figure 8D In the right-hand scale group, R3' has the smallest deviation relative to R3. Therefore, in this state, P0' is -3δ relative to P0, meaning that P0' is offset by 3δ in the negative x-direction relative to P0. The offset in the y-direction can be obtained in the same way, which will not be repeated here.

[0131] like Figure 8E As shown in (a), several alignment marks MK are distributed on a substrate wafer. A rectangular or square jurisdiction AR corresponding to each alignment mark is delineated around it. The offsets in the x and y directions measured at the alignment mark are used as the offsets of the cross slit center of each piezoelectric micromechanical ultrasonic transducer device within that jurisdiction relative to its cavity center. Taking mark MK23 and its jurisdiction AR23 as an example, from... Figure 8E-(b) shows the magnified microscopic image of mark MK23. Following the aforementioned rules, the x- and y-direction offsets at this mark are read. Based on this data, the frequency deviations of the diaphragm devices of all piezoelectric micromechanical ultrasonic transducers within the jurisdiction AR23 are estimated, and frequency correction is performed based on the estimation results. The dimensions of the aforementioned rectangular or square jurisdiction can be adjusted according to accuracy requirements to ensure that the offset at the alignment mark represents the offset of the piezoelectric micromechanical ultrasonic transducer structure within the jurisdiction within the required accuracy range.

[0132] The on-chip feature comparison method is explained below. This method can be used when the patterning process for creating cavities and the patterning process for creating slits are located on opposite sides of the wafer.

[0133] like Figure 9A As shown, in Figure 9A -(a) The end of each branch of the cross slit F1 (taking the end in the upper right region E23 as an example) is designed as shown in the enlarged view of region E23. Figure 9A The structure shown in (b) is used for measuring offset. Its structural features are: a horizontal line segment S3x and a vertical line segment S3y on the contour of the end edge F1-3. Figure 9A -(b) indicates the bold black line segment), and in theoretical design, S3x and S3y coincide with the upper horizontal side B1-3 and the right vertical side B1-2 of cavity B1, respectively.

[0134] like Figure 9B As shown, when the cross slits shift relative to the back cavity, S3x and B1-3, as well as S3y and B1-2, no longer coincide. The shifts of the slits relative to the cavity in the x and y directions can be measured, which are respectively the shifts of S3y relative to B1-2 (σ3x) and S3x relative to B1-3 (σ3y). Before observation, the back cavity was patterned and the front slits were fabricated; therefore, observing the exposed portions of B1-3 and B1-2 within the slits and measuring the distances between them and S3x and S3y is practically feasible.

[0135] like Figure 9C As shown, in actual processing, it is often not easy to obtain such results. Figure 9B The slit shown is entirely enclosed by straight line segments at its end, but a chamfer or rounded corner appears at the inflection point of its end profile. The presence of these features may cause the straight line segments S3x and S3y to shorten or even disappear, thus affecting measurement accuracy. Since the theoretical design lengths of S3x and S3y in this structure are constrained by device performance indicators (increasing the slit linewidth within the cavity boundary affects device performance) and fabrication capabilities (narrowing the slit linewidth outside the cavity boundary is not feasible in terms of fabrication capabilities), increasing the theoretical design lengths of S3x and S3y in this structure is difficult.

[0136] To overcome Figure 9C The problems that exist can be addressed by modifying the structure at the end of the slit. Figure 9D The improvements shown Figure 9D The slit end has a "Y"-shaped structure, with the horizontal side S3x and vertical side S3y used for measurement located at the two bifurcations of the "Y"-shaped structure. Theoretically, the intersection of the horizontal side S3x and the vertical side S3y coincides with the intersection of the upper horizontal side B1-3 and the right vertical side B1-2 of the back cavity B1, and S3x coincides with the extension of B1-3, while S3y coincides with the extension of B1-2. Since the lengths of S3x and S3y in this structure are not limited by device performance or fabrication capabilities, the length of the bifurcated structure can be adjusted to ensure sufficient length of S3x and S3y for measurement even with rounded corners.

[0137] Similarly, the offsets at the other three ends of the cross slit can be obtained, and the average of the algebraic sum of the offsets at the four ends can be taken as the offset σ at the center point. x and σ y Estimate:

[0138]

[0139] The sign of the two offset components at each end is determined by the relative position of the measuring edge on the crosshair to the horizontal or vertical edge of the back cavity. (Still using...) Figure 9B and Figure 9D Taking the end F1-3 as an example, when S3x is above B1-3, σ 3y The value is positive, and negative otherwise; when S3y is to the right of B1-2, σ 3x The value is positive, and vice versa. This rule applies to all four ends of the cross slit.

[0140] It should be noted that the above technical solution is applicable not only to adjusting the diaphragm in a single array element, but also to adjusting the entire array. For example... Figure 10 The array shown includes a total of 9 array elements in a 3×3 array. Each array element has multiple diaphragms, and the technical solution of this application can be used to perform frequency correction on the different diaphragms in these 9 array elements.

[0141] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for manufacturing a multi-diaphragm piezoelectric MEMS transducer, wherein the transducer is an array of multiple elements, each element having multiple diaphragms, the diaphragms being cantilever beam structures, characterized in that... The method includes: The resonant frequency of each diaphragm in the array element is determined, and for each diaphragm, the following frequency correction steps are performed to reduce the difference between the resonant frequency of that diaphragm and the specified frequency: If the resonant frequency of the diaphragm is lower than the specified frequency, the mass of the diaphragm is reduced; if the resonant frequency of the diaphragm is higher than the specified frequency, the equivalent stiffness of the diaphragm is reduced.

2. The method according to claim 1, characterized in that, Before the step of determining the resonant frequency of each diaphragm in the array element, the method further includes: fabricating a frequency correction layer above the piezoelectric structure of each diaphragm of the array element; Furthermore, the step of reducing the mass of the diaphragm includes partially removing the material of the frequency correction layer.

3. The method according to claim 2, characterized in that, The steps for partially removing the material from the frequency correction layer include: removing the material from the frequency correction layer by bombardment with a focused particle beam or irradiation with a laser beam.

4. The method according to claim 2, characterized in that, The frequency correction layer includes a protective film layer above the piezoelectric structure of each diaphragm and a thin metal film on the surface of a predetermined frequency correction region near the free end of the protective film layer. The step of partially removing the material from the frequency correction layer includes: partially removing the metal thin film.

5. The method according to claim 4, characterized in that, The step of partially removing the metal film includes: Oxidize a portion of the metal film; The oxidized metal film is irradiated with a laser beam, thereby partially or completely removing the oxidized metal film.

6. The method according to claim 4, characterized in that, The step of partially removing the metal film includes: Oxidize a portion of the metal film, or oxidize the entire metal film; The oxidized metal film is partially or completely removed.

7. The method according to claim 6, characterized in that, The step of partially or completely removing the oxidized metal film includes: irradiating the oxidized metal film with a laser beam or bombarding the metal film with a particle beam, thereby partially or completely removing the oxidized metal film.

8. The method according to claim 5 or 6, characterized in that, The step of oxidizing a portion of the metal thin film includes one of the following: A portion of the metal thin film in the thickness direction is oxidized, and the oxidized region is either a continuous region or multiple separate regions in the planar extension direction of the metal thin film; Multiple separate oxidation regions are formed on the metal thin film, and the thickness of all oxidation regions is equal to or less than the thickness of the metal thin film, or only some oxidation regions have a thickness equal to the thickness of the metal thin film.

9. The method according to claim 1, characterized in that, The step of reducing the equivalent stiffness of the diaphragm includes: A qualitative change region is formed in a preset frequency correction region near the fixed end of the diaphragm, thereby reducing the equivalent stiffness of the diaphragm.

10. The method according to claim 9, characterized in that, The qualitative change region is a series of regions distributed along the edge of the fixed end.

11. The method according to claim 9 or 10, characterized in that, The step of forming the qualitative change region includes: irradiating the frequency correction region with a laser beam.

12. The method according to claim 1, characterized in that, Before the step of reducing the mass of the diaphragm, the method further includes: making the resonant frequency of all diaphragms of the array element lower than the specified frequency; Before the step of reducing the equivalent stiffness of the diaphragm, the method further includes: making the resonant frequency of all diaphragms of the array element higher than the specified frequency.

13. The method according to any one of claims 1 to 7, 9, 10, and 12, characterized in that, The multiple diaphragms are centrally symmetrically distributed, and the cavity of the vibrating element is set below them; Before performing the frequency correction step, the method further includes: determining the offset vector between the geometric center of the cavity boundary and the symmetry center of the plurality of diaphragms in the horizontal and vertical directions using photolithographic alignment marking methods, respectively. The offset vector is used to determine the magnitude relationship between the resonant frequency of the diaphragm and a specified frequency.

14. The method according to any one of claims 1 to 7, 9, 10, and 12, characterized in that, Before the step of determining the resonant frequency of each diaphragm in the array element, the method further includes: Four congruent right-angled isosceles triangular diaphragms are arranged in the array element. The four diaphragms are centrally symmetrically distributed to form a square. A cavity of the array element is arranged below it. At the end of the gap between the four diaphragms, two branches are formed to form a Y-shape. The two branches are perpendicular to each other, and the edges of the two branches away from the right-angled isosceles triangles are straight lines, which are parallel to the edge of the cavity.

15. The method according to claim 14, characterized in that, Before performing the frequency correction step, the method further includes: determining an offset vector between the geometric center of the cavity boundary and the center of symmetry of the plurality of diaphragms based on the distance between the straight line and the boundary of the cavity that is parallel to and closest to the straight line, the offset vector being used to determine the magnitude relationship between the resonant frequency of the diaphragm and a specified frequency.

16. The method according to any one of claims 1 to 7, 9, 10, and 12, characterized in that, Before the step of determining the resonant frequency of each diaphragm in the array element, the method further includes: Four congruent right-angled isosceles triangular diaphragms are arranged in the array element. The four diaphragms are centrally symmetrically distributed to form a square. A cavity of the array element is arranged below it. The gap between adjacent diaphragms has a first segment and a second segment. The beginning of the first segment is located at the right-angle vertex of the right-angled isosceles triangle, and the end is connected to the beginning of the second segment. The end of the second segment is the end of the gap between adjacent diaphragms. The end of the first segment has a horizontal line segment and a vertical line segment, which are parallel to a set of adjacent sides of the square.

17. The method according to claim 16, characterized in that, Before performing the frequency correction step, the method further includes: determining an offset vector between the geometric center of the cavity boundary and the symmetry center of the plurality of diaphragms based on the lateral and longitudinal offsets. This offset vector is used to determine the magnitude relationship between the resonant frequency of the diaphragm and a specified frequency; wherein, The lateral offset is the distance between the lateral line segment and the boundary of the cavity that is parallel to and closest to the lateral line segment. The longitudinal offset is the distance between the longitudinal line segment and the boundary of the cavity that is parallel to and closest to the longitudinal line segment.

18. A multi-diaphragm piezoelectric MEMS transducer, wherein the transducer is an array of multiple elements, each element having multiple diaphragms, the diaphragms being cantilever beam structures, characterized in that... The array element is manufactured using the method described in any one of claims 1 to 17.

19. A multi-diaphragm piezoelectric MEMS transducer, wherein the transducer is an array of multiple elements, each element having multiple diaphragms, the diaphragms being cantilever beam structures, characterized in that... For one or more elements of the transducer, each diaphragm of the element has a frequency correction layer above its piezoelectric structure, and the quality of the frequency correction layer on each diaphragm of the element is inconsistent.

20. The transducer according to claim 19, characterized in that, The frequency correction layer includes a protective film layer above the piezoelectric structure of each diaphragm and a metal film on the surface of a preset frequency correction region near the free end of the protective film layer; the quality of the metal film on each diaphragm in the same element is inconsistent.

21. A multi-diaphragm piezoelectric MEMS transducer, wherein the transducer is an array of multiple array elements, each array element having multiple diaphragms, the diaphragms being cantilever beam structures, characterized in that... For one or more array elements of the transducer, there is a qualitative change region in a preset frequency correction region near the fixed end of one or more diaphragms of the array element, which reduces the equivalent stiffness of the diaphragm.

22. The transducer according to claim 21, characterized in that, The qualitative change region is a series of regions distributed along the edge of the fixed end.

23. The transducer according to any one of claims 19 to 22, characterized in that, The array element comprises four congruent right-angled isosceles triangular diaphragms, which are arranged into a square, with the cavity of the array element below. The gap between the four diaphragms has two branches at its end, forming a Y-shape. The two branches are perpendicular to each other. The edges of the two branches away from the right-angled isosceles triangle are straight lines and are parallel to the edge of the cavity, respectively.

24. The transducer according to any one of claims 19 to 22, characterized in that, The array element comprises four congruent right-angled isosceles triangular diaphragms, which are arranged into a square, with the cavity of the array element below. The gap between adjacent diaphragms has a first segment and a second segment; The first segment begins at the right-angle vertex of the right-angled isosceles triangle and ends at the beginning of the second segment; the end of the second segment is the end of the gap between adjacent diaphragms. The end of the first segment has a horizontal line segment and a vertical line segment, which are parallel to a set of adjacent sides of the square.

Images