Method for manufacturing irregular electrode, film bulk acoustic resonator with irregular electrode

By generating irregular electrodes in a thin-film bulk acoustic resonator and utilizing sinusoidal wave superposition and neural network optimization, stray modes were suppressed, the resonator performance was improved, and a high-frequency and wide-bandwidth thin-film bulk acoustic resonator was realized.

CN118868835BActive Publication Date: 2026-05-29SHANGHAI UNIV

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SHANGHAI UNIV
Filing Date
2024-07-03
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Existing thin-film bulk acoustic resonators suffer from stray mode problems, which lead to a decrease in the resonator's Q value. Existing improvement methods are complex and result in a loss of device performance.

Method used

An irregular electrode generation method is adopted, which forms the electrode shape by superimposing sine waves with random amplitude and phase in polar coordinate system, and uses a feedforward neural network to optimize the electrode shape to suppress stray modes, thereby generating an irregular polygonal electrode.

Benefits of technology

It effectively suppresses the generation of transverse standing waves, reduces stray modes in the spectrum, improves the resonant frequency and effective electromechanical coupling coefficient, and maintains the high performance of the resonator.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118868835B_ABST
    Figure CN118868835B_ABST
Patent Text Reader

Abstract

This invention relates to a method for generating irregular electrodes and a thin-film bulk acoustic resonator with irregular electrodes, belonging to the field of microelectromechanical systems (MEMS) technology. The generation method includes: S1, superimposing multiple sinusoidal waves with random amplitudes and phases in a polar coordinate system; S2, calculating the number of stray modes near the resonant frequency point in the impedance frequency curve, setting a fluctuation threshold Δs, and if the amplitude S of the impedance frequency curve > Δs, it is recorded as a stray mode. The total number of stray modes near the resonant point in a device is denoted as M; S3, establishing a feedforward neural network to encode the coordinates of a single irregular polygon obtained in S1 into a one-dimensional... Vector V serves as the input layer, and the number M of its corresponding spurious modes serves as the output layer. Before training, both the output and input layers are normalized to match the neural network structure. An Adam optimizer is used, with the optimization objective set to minimize the number of spurious modes M. Training is then performed to obtain a relationship model between irregular polygons and the number of spurious modes. This model is validated using a validation set. In step S4, a new set of irregular polygons is generated using the validated relationship model from step S3. These generated polygons are used as the electrode shapes for a thin-film bulk acoustic resonator, resulting in irregular electrodes for the thin-film bulk acoustic resonator. Compared to existing technologies, the irregular electrode shape structure used in this invention has been verified to effectively suppress the generation of transverse standing waves, thereby reducing spurious modes in the spectrum.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of microelectromechanical systems (MEMS) technology, specifically to a method for generating irregular electrodes and a thin-film bulk acoustic resonator with irregular electrodes. Background Technology

[0002] Film Bulk Acoustic Wave Resonators (FBARs) are now widely used in communication equipment as an important component of radio frequency circuits such as filters and transceivers. They offer advantages such as a high quality factor, high resonant frequency, and compatibility with CMOS semiconductor processes, and have experienced rapid development in recent years. They achieve resonance by converting electrical energy into acoustic energy through the inverse piezoelectric effect of the piezoelectric thin film.

[0003] Existing thin-film bulk acoustic resonators mainly employ a pentagonal electrode structure, where electrodes are deposited on a piezoelectric thin film using a metal sputtering process. During device operation, the longitudinal vibration of the piezoelectric layer is excited by the electrodes to achieve the conversion of electro-mechanical energy. However, the excited piezoelectric layer exhibits multiple vibration modes. In this structure, the longitudinal mode is the dominant mode, while the transverse mode is a secondary mode. The longitudinal mode is characterized by vibration waves propagating along the thickness direction of the piezoelectric layer, while the transverse mode is characterized by vibration waves propagating along the plane of the piezoelectric layer. This transversely propagating wave is called a Lamb wave. Lamb wave leakage causes energy loss, and its reflection at the resonator edges can also generate spurious modes, leading to a decrease in the Q value of the resonator. Therefore, suppression is necessary.

[0004] Stray modes are influenced by various factors, such as the thickness of the electrode layer and piezoelectric layer films, the materials of the electrodes and piezoelectric layers, and the electrode distribution. To obtain better resonator performance and suppress stray modes, these parameters can be modified. Compared to changing the film thickness and material, the electrode distribution can be more easily modified using lithographic patterning.

[0005] Traditionally, thin-film bulk acoustic resonators (TIRs) use pentagonal electrodes with unequal side lengths, where no two sides are parallel, to eliminate standing waves and reduce stray modes to some extent. However, the effect is limited and cannot effectively suppress lateral vibrations. Several improvements have been made, such as the use of interdigitated electrodes in Chinese patent CN116131803B, which incorporates holes of varying shapes to alter the excitation mode and suppress stray modes. However, this method involves extremely complex electrode structures, requiring drilling and other fabrication work, making it very difficult to manufacture. While it can suppress stray modes to some extent, it also results in some performance degradation. Therefore, a method is urgently needed to suppress stray modes without excessively compromising resonator performance. Summary of the Invention

[0006] The purpose of this invention is to overcome at least one of the defects of the prior art by providing a method for generating irregular electrodes and a thin-film bulk acoustic resonator with irregular electrodes, which can suppress stray modes commonly found in existing resonators and improve the quality factor of the resonator.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] One objective of this invention is to provide a method for generating irregular electrodes, comprising the following steps:

[0009] S1. In the polar coordinate system, multiple sine waves with random amplitudes and random phases are superimposed to form a curve after the superposition of sine waves:

[0010] Where H is the number of harmonics, randomly generated by a computer, ranging from 1 to n, where n is a positive integer; ρ h It is the amplitude of the h-th harmonic, which is a random number between 0 and 1; It is the phase of the h-th harmonic, which is obtained by random number generation;

[0011] The curve resulting from the superposition of sine waves is sampled at equal intervals within the range of 0-2π. The sampling points are then transformed to a rectangular coordinate system, and their x and y coordinates are generated using the following transformation formula, thus forming the electrode shape:

[0012]

[0013] S2. Calculate the number of stray modes near the resonant frequency point in the impedance frequency curve, and set the fluctuation threshold Δ. s If the amplitude S of the impedance frequency curve is greater than Δ s If a stray mode is denoted as a stray mode, then the total number of stray modes near the resonant point in a device is denoted as M.

[0014] S3. Establish a feedforward neural network. Encode the coordinates of a single irregular polygon obtained in S1 into a one-dimensional vector V, which serves as the input layer. The number of corresponding spurious modes M serves as the output layer. Before training, the output and input layers are normalized to match the structure of the neural network. Use the Adam optimizer, setting the optimization objective to minimize the number of spurious modes M, and train to obtain a connection model reflecting the relationship between the irregular polygon and the number of spurious modes. Validate the model using a validation set to obtain a model that accurately reflects the relationship between the irregular polygon and the number of spurious modes.

[0015] Specifically, during data preparation, a training dataset and a validation dataset were prepared, accounting for 70% and 30% of the total dataset, respectively. After model training, the device model for validation was input into the neural network. The neural network calculated the device's performance, and the results were compared with test data of the same device to obtain the error between the neural network model's calculation results and the actual test results. Statistical tools, namely the variance (MSE) and the correlation coefficient (R²), can be used to characterize the range of this error. Generally, a smaller variance between the neural network's calculation results and experimental results is better; in this case, the variance is 0.03%. A correlation coefficient closer to 1 is also better; generally, a value above 0.9 indicates that the neural network model accurately reflects the relationship between the input and output.

[0016] S4. Using the verified relational model in S3, generate a new set of irregular polygons. The rules for generating the polygons are the same as those used for neural network training. Use the generated polygons as the electrode shapes of the thin-film bulk acoustic resonator to obtain the irregular electrodes for the thin-film bulk acoustic resonator.

[0017] Furthermore, the shape of the irregular electrode must satisfy at least one of the following conditions:

[0018] ①The irregular electrode is an irregular polygon with at least 3 sides;

[0019] ②The interior angles of the irregular polygon are greater than 90° and less than 180°;

[0020] ③The ratio of the longest side to the shortest side of the irregular polygon does not exceed 2:1.

[0021] A second objective of this invention is to provide a thin-film bulk acoustic resonator with irregular electrodes, comprising:

[0022] piezoelectric layer;

[0023] A first support portion and a second support portion are symmetrically arranged on both sides of the piezoelectric layer;

[0024] A top electrode structure includes a top electrode and a first lead-out portion disposed on the side of the top electrode. The top electrode is disposed on the top surface of the piezoelectric layer, and the portion of the first lead-out portion extending out of the piezoelectric layer is disposed on a first support portion.

[0025] The bottom electrode structure includes a bottom electrode and a second lead-out portion disposed on the side of the bottom electrode. The bottom electrode is disposed on the bottom surface of the piezoelectric layer, and the portion of the second lead-out portion extending out of the piezoelectric layer is disposed on the second support portion.

[0026] The top electrode and / or bottom electrode are irregular electrodes obtained by the generation method described above.

[0027] Specifically, the piezoelectric layer of the present invention converts electrical energy into mechanical energy through the inverse piezoelectric effect, causing the resonator to vibrate and resonate when it reaches near the device's natural frequency; the top electrode and bottom electrode serve as the excitation source for the vibration of the piezoelectric material, and alternating voltages with different frequencies and amplitudes are passed through corresponding leads; the support part fixes the entire structure on the cavity below the resonator, so that the upper and lower sides of the resonator are air mediums, and it is given a certain range of free vibration; the irregular electrode suppresses the propagation of transverse waves in the resonator and thus suppresses the generation of stray modes through the irregular shape of the electrode edge.

[0028] Furthermore, the portion of the first lead extending out of the piezoelectric layer completely covers the first support portion; the portion of the second lead extending out of the piezoelectric layer completely covers the second support portion.

[0029] Furthermore, the first support portion and / or the second support portion and / or the first lead-out portion and / or the second lead-out portion are rectangular.

[0030] Furthermore, both the top electrode and the bottom electrode are irregularly shaped electrodes of the same type, and they are aligned vertically; the piezoelectric layer is rectangular and completely covers the top electrode.

[0031] Furthermore, the top electrode is an irregular electrode; both the piezoelectric layer and the bottom electrode are rectangular, with equal length and width and aligned vertically; the piezoelectric layer completely covers the top electrode.

[0032] Furthermore, both the top electrode and the bottom electrode are irregularly shaped electrodes of the same type, and they are aligned vertically; the piezoelectric layer has the same shape as the top electrode, and they are aligned vertically.

[0033] Furthermore, the piezoelectric layer is made of a material with the inverse piezoelectric effect, including one or more of AlN, AlScN, or ZnO.

[0034] Furthermore, the top electrode and the bottom electrode are made of conductive metals, including alloys of one or more of Cu, Al, Mo, Au, Ag or Pt.

[0035] Compared with the prior art, the present invention has the following advantages:

[0036] (1) Spurious Mode Suppression: In traditional resonators, transverse waves generated by the inverse piezoelectric effect are reflected when they encounter specific boundaries during propagation. When the incident and reflected waves have the same frequency, amplitude, and opposite propagation directions, they superimpose to form a stable and high-energy standing wave. The irregular electrode shape structure used in this invention has been verified to effectively suppress the generation of transverse standing waves, thereby reducing spurious modes in the spectrum.

[0037] (2) High operating frequency: Since the fabrication process of this thin film bulk acoustic wave is compatible with CMOS semiconductor process, the piezoelectric layer can be as thin as hundreds of nanometers while maintaining the stability of the structure, thus it can obtain a high resonant frequency, usually above 2GHz.

[0038] (3) High frequency bandwidth: The bandwidth of thin film bulk acoustic resonator is an important indicator of its performance. It is usually represented by the effective electromechanical coupling coefficient. When multiple resonators with irregular electrode structures are tested, their effective electromechanical coupling coefficient can reach more than 6%. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the thin-film bulk acoustic resonator with irregular electrodes in Example 1;

[0040] Figure 2 This is a top view of the thin-film bulk acoustic resonator with irregular electrodes in Example 1;

[0041] Figure 3 This is a side view of the thin-film bulk acoustic resonator with irregular electrodes in Example 1;

[0042] Figure 4 This is a schematic diagram of the thin-film bulk acoustic resonator with irregular electrodes in Example 2;

[0043] Figure 5 This is a top view of the thin-film bulk acoustic resonator with irregular electrodes in Example 2;

[0044] Figure 6 This is a side view of the thin-film bulk acoustic resonator with irregular electrodes in Example 2;

[0045] Figure 7 This is a schematic diagram of the thin-film bulk acoustic resonator with irregular electrodes in Example 3;

[0046] Figure 8 This is a top view of the thin-film bulk acoustic resonator with irregular electrodes in Example 3;

[0047] Figure 9 This is a side view of the thin-film bulk acoustic resonator with irregular electrodes in Example 3;

[0048] Figure 10 These are comparison diagrams of stray modes of thin-film bulk acoustic resonators using irregular electrodes and conventional pentagonal electrodes, respectively, in Example 1;

[0049] The numbers in the figure indicate: 1-Piezoelectric layer; 2-Top electrode; 3-Bottom electrode; 4-First support part; 5-First lead-out part; 6-Second support part; 7-Second lead-out part. Detailed Implementation

[0050] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are implemented based on the technical solution of the present invention, providing detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0051] Example 1

[0052] A method for generating irregular electrodes.

[0053] S1. In the polar coordinate system, multiple sine waves with random amplitudes and random phases are superimposed to form a curve after the superposition of sine waves:

[0054]

[0055] Where H is the number of harmonics, randomly generated by a computer, ranging from 1 to n, where n is a positive integer; ρ h It is the amplitude of the h-th harmonic, which is a random number between 0 and 1; It is the phase of the h-th harmonic;

[0056] The curve resulting from the superposition of sine waves is sampled at equal intervals within the range of 0-2π. The sampling points are then transformed to a rectangular coordinate system, and their x and y coordinates are generated using the following transformation formula, thus forming the electrode shape:

[0057]

[0058] S2. Calculate the number of stray modes near the resonant frequency point in the impedance frequency curve, set the fluctuation threshold Δs. If the amplitude S of the impedance frequency curve is greater than Δs, it is recorded as one stray mode. The total number of stray modes near the resonant point in a device is recorded as M.

[0059] S3. Establish a feedforward neural network. Encode the coordinates of a single irregular polygon into a one-dimensional vector V, which serves as the input layer. The number of corresponding spurious modes M serves as the output layer. Before training, the output and input layers are normalized to match the structure of the neural network. Use the Adam optimizer, setting the optimization objective to minimize the number of spurious modes M, and train the network to obtain a connection model reflecting the relationship between the irregular polygon and the number of spurious modes. Then, validate the model using a validation set to obtain a model that accurately reflects the relationship between the irregular polygon and the number of spurious modes.

[0060] Specifically, during data preparation, a training dataset and a validation dataset were prepared, accounting for 70% and 30% of the total dataset, respectively. After model training was completed, the device model for validation was input into the neural network. The neural network calculated the device's performance, and the results were compared with test data of the same device to obtain the error between the neural network model's calculation results and the actual test results. Statistical tools, namely the variance MSE and the correlation coefficient R, were used to further analyze the results. 2 This can characterize the range of the error. Generally speaking, the smaller the variance between the neural network calculation results and the experimental results, the better; in this case, the variance is 0.03%. The closer the correlation coefficient is to 1, the better; it is generally accepted that a correlation coefficient above 0.9 indicates that the neural network model correctly reflects the relationship between the input and output. In this embodiment, the correlation coefficient is at least 0.94.

[0061] S4. Using the relational model obtained in S3, generate a new set of irregular polygons. The rules for generating the polygons are the same as those used for neural network training. Use the generated polygons as the electrode shapes of the thin-film bulk acoustic resonator to obtain the irregular electrodes for the thin-film bulk acoustic resonator.

[0062] In this embodiment, the shape of the irregular electrode satisfies at least one of the following conditions:

[0063] ①The irregular electrode is an irregular polygon with at least 3 sides;

[0064] ②The interior angles of the irregular polygon are greater than 90° and less than 180°;

[0065] ③The ratio of the longest side to the shortest side of the irregular polygon does not exceed 2:1.

[0066] This embodiment provides a thin-film bulk acoustic resonator with irregular electrodes, comprising:

[0067] Piezoelectric layer 1;

[0068] The first support portion 4 and the second support portion 6 are symmetrically arranged on both sides of the piezoelectric layer 1;

[0069] The top electrode structure includes a top electrode 2 and a first lead-out portion 5 disposed on the side of the top electrode 2. The top electrode 2 is disposed on the top surface of the piezoelectric layer 1, and the portion of the first lead-out portion 5 extending out of the piezoelectric layer 1 is disposed on the first support portion 4.

[0070] The bottom electrode structure includes a bottom electrode 3 and a second lead-out portion 7 disposed on the side of the bottom electrode 3. The bottom electrode 3 is disposed on the bottom surface of the piezoelectric layer 1, and the portion of the second lead-out portion 7 extending out of the piezoelectric layer 1 is disposed on the second support portion 6.

[0071] The top electrode 2 and the bottom electrode 3 are irregular electrodes obtained by the above-described generation method.

[0072] In this embodiment, the portion of the first lead-out portion 5 extending out of the piezoelectric layer 1 completely covers the first support portion 4; the portion of the second lead-out portion 7 extending out of the piezoelectric layer 1 completely covers the second support portion 6.

[0073] In this embodiment, the first support portion 4, the second support portion 6, the first lead-out portion 5, and the second lead-out portion 7 are rectangular.

[0074] In this embodiment, the top electrode 2 and the bottom electrode 3 are both irregularly shaped electrodes and are aligned vertically; the piezoelectric layer 1 is rectangular and completely covers the top electrode 2.

[0075] In this embodiment, the piezoelectric layer 1 is made of a material with inverse piezoelectric effect, including one or more of AlN, AlScN, or ZnO.

[0076] In this embodiment, the top electrode 2 and the bottom electrode 3 are made of conductive metals, including alloys of one or more of Cu, Al, Mo, Au, Ag or Pt.

[0077] Working principle:

[0078] The top electrode 2 of the resonator is set as the positive electrode, and the bottom electrode 3 as the negative electrode. An alternating current is applied. When the frequency is near the device's resonant frequency, the piezoelectric layer 1 is excited by the electrodes and vibrates, thus converting electrical energy into acoustic energy and realizing the function of the resonator. During resonance, the piezoelectric layer 1 exhibits multiple vibration modes after being excited. In this structure, the longitudinal mode is the primary mode, and the transverse mode is the secondary mode. The longitudinal mode is characterized by vibration waves propagating along the thickness direction of the piezoelectric layer 1, while the transverse mode is characterized by vibration waves propagating along the plane of the piezoelectric layer 1. This transversely propagating wave is called a Lamb wave. Lamb wave leakage leads to a decrease in the Q value of the resonator, and its reflection at the resonator's edge also causes the generation of stray modes, which need to be suppressed. This invention employs an irregular electrode structure to suppress the propagation and reflection of transverse waves, thereby suppressing the generation of stray modes and achieving a higher resonator quality factor. Figure 10 This diagram compares the stray modes of the thin-film bulk acoustic resonators using irregular electrodes and traditional pentagonal electrodes in this embodiment. The use of irregular electrode structures significantly reduces the number of stray modes. This embodiment uses a validated relational model to generate multiple models showing that irregular electrodes significantly reduce stray modes. Figure 10 The irregular electrode used is representative of this.

[0079] Example 2

[0080] This embodiment is basically the same as embodiment 1, except that: in this embodiment, the top electrode 2 is an irregular electrode; the piezoelectric layer 1 and the bottom electrode 3 are both rectangular, with equal length and width and aligned vertically; the piezoelectric layer 1 completely covers the top electrode 2.

[0081] Example 3

[0082] This embodiment is basically the same as embodiment 1, except that in this embodiment: the top electrode 2 and the bottom electrode 3 are both irregular electrodes with the same shape and are aligned vertically; the piezoelectric layer 1 has the same shape as the top electrode 2 and is aligned vertically.

[0083] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for generating an irregular electrode, characterized in that, Includes the following steps: S1. In the polar coordinate system, multiple sine waves with random amplitudes and random phases are superimposed to form a curve after the superposition of sine waves: Where H is the number of harmonics, randomly generated by a computer, ranging from 1 to n, where n is a positive integer; ρ h It is the amplitude of the h-th harmonic, which is a random number between 0 and 1; It is the phase of the h-th harmonic, which is obtained by random number generation; The curve resulting from the superposition of sine waves is sampled at equal intervals within the range of 0-2π. The sampled points are then transformed to a rectangular coordinate system, and their x and y coordinates are generated using the following transformation formula, forming the coordinate code for the electrode shape: S2. Calculate the number of stray modes near the resonant frequency point in the impedance frequency curve, and set the fluctuation threshold Δ. s If the amplitude S of the impedance frequency curve is greater than Δ s If a stray mode is denoted as a stray mode, then the total number of stray modes near the resonant point in a device is denoted as M. S3. Establish a feedforward neural network. Encode the coordinates of a single irregular polygon obtained in S1 into a one-dimensional vector V, which serves as the input layer. Use the number M of its corresponding spurious modes as the output layer. Before training, normalize the output and input layers to match the structure of the neural network. Use the Adam optimizer with the optimization objective set to minimize the number M of spurious modes. Train the network to obtain a model relating the irregular polygon to the number of spurious modes. Validate the model using a validation set to obtain a model that correctly reflects the relationship between the irregular polygon and the number of spurious modes. S4. Using the verified relational model in S3, generate a new set of irregular polygons. The rules for generating the polygons are the same as those used for neural network training. Use the generated polygons as the electrode shapes of the thin-film bulk acoustic resonator to obtain the irregular electrodes for the thin-film bulk acoustic resonator.

2. The method for generating an irregular electrode according to claim 1, characterized in that, The shape of the irregular electrode must satisfy at least one of the following conditions: ①The irregular electrode is an irregular polygon with at least 3 sides; ②The interior angles of the irregular polygon are greater than 90° and less than 180°; ③The ratio of the longest side to the shortest side of the irregular polygon does not exceed 2:

1.

3. A thin-film bulk acoustic resonator with irregular electrodes, characterized in that, It includes: piezoelectric layer (1); The first support portion (4) and the second support portion (6) are symmetrically arranged on both sides of the piezoelectric layer (1); The top electrode structure includes a top electrode (2) and a first lead-out portion (5) disposed on the side of the top electrode (2). The top electrode (2) is disposed on the top surface of the piezoelectric layer (1), and the portion of the first lead-out portion (5) extending out of the piezoelectric layer (1) is disposed on the first support portion (4). The bottom electrode structure includes a bottom electrode (3) and a second lead-out portion (7) disposed on the side of the bottom electrode (3). The bottom electrode (3) is disposed on the bottom surface of the piezoelectric layer (1), and the portion of the second lead-out portion (7) extending out of the piezoelectric layer (1) is disposed on the second support portion (6). The top electrode (2) and / or bottom electrode (3) are irregular electrodes obtained by any one of the generation methods described in claims 1-2.

4. A thin-film bulk acoustic resonator with irregular electrodes according to claim 3, characterized in that, The portion of the first lead-out portion (5) extending out of the piezoelectric layer (1) completely covers the first support portion (4); The portion of the second lead-out portion (7) extending out of the piezoelectric layer (1) completely covers the second support portion (6).

5. A thin-film bulk acoustic resonator with irregular electrodes according to claim 3, characterized in that, The first support portion (4) and / or the second support portion (6) and / or the first lead-out portion (5) and / or the second lead-out portion (7) are rectangular.

6. A thin-film bulk acoustic resonator with irregular electrodes according to claim 3, characterized in that, The top electrode (2) and the bottom electrode (3) are both irregular electrodes of the same shape, and they are aligned vertically. The piezoelectric layer (1) is rectangular and completely covers the top electrode (2).

7. A thin-film bulk acoustic resonator with irregular electrodes according to claim 3, characterized in that, The top electrode (2) is an irregular electrode; The piezoelectric layer (1) and the bottom electrode (3) are both rectangular, with equal length and width and aligned vertically. The piezoelectric layer (1) completely covers the top electrode (2).

8. A thin-film bulk acoustic resonator with irregular electrodes according to claim 3, characterized in that, The top electrode (2) and the bottom electrode (3) are both irregular electrodes of the same shape, and they are aligned vertically. The piezoelectric layer (1) has the same shape as the top electrode (2), and the two are aligned vertically.

9. A thin-film bulk acoustic resonator with irregular electrodes according to claim 3, characterized in that, The piezoelectric layer (1) is made of a material with inverse piezoelectric effect, including one or more of AlN, AlScN or ZnO.

10. A thin-film bulk acoustic resonator with irregular electrodes according to claim 3, characterized in that, The top electrode (2) and bottom electrode (3) are made of conductive metals, including alloys of one or more of Cu, Al, Mo, Au, Ag or Pt.