Piezoelectric-electrostatic composite MEMS resonator and MEMS resonator array

Abstract

The application provides a piezoelectric-electrostatic composite MEMS resonator and a MEMS resonator array. The MEMS resonator comprises a silicon substrate, a main resonator body made of a low-loss material and comprising a main mechanical energy storage area and a peripheral edge function area, a plurality of anchor points fixed on the silicon substrate, a plurality of support arms arranged in the same layer as the main resonator body and spaced apart from the silicon substrate to form a release cavity, each support arm being made of a low-loss material and having a front end connected to the main resonator body and a tail end connected to a corresponding anchor point, a plurality of piezoelectric transduction layers, each piezoelectric transduction layer being arranged on a corresponding support arm or extending only to the edge function area of the main resonator body to form a piezoelectric transduction structure, and a plurality of fixed electrodes adjacent to the main resonator body and forming an air gap, the fixed electrodes being used to form an electrostatic transduction structure. Based on the piezoelectric-electrostatic composite structure, the piezoelectric and electrostatic transduction mechanisms are combined to achieve high Q value, low impedance and high frequency performance.

Description

Technical Field

[0001] This application relates to the field of MEMS (Micro electro Mechanical Systems) technology, specifically to a piezoelectric-electrostatic composite MEMS resonator and a MEMS resonator array. Background Technology

[0002] Currently, MEMS resonators are gradually replacing traditional quartz crystal oscillators as core fundamental components in fields such as wireless communication, inertial navigation, time and frequency references, and biochemical sensing. With their advantages of small size, light weight, low power consumption, mass production capability, and high compatibility with CMOS (Complementary Metal-Oxide-Semiconductor) processes, MEMS resonators have become an important candidate technology path for next-generation on-chip time and frequency systems.

[0003] Key performance indicators of MEMS resonators include resonant frequency ( Mechanical quality factor (Q), kinematic resistance (R) m The factors considered include static capacitance (C0), electromechanical coupling coefficient (η), phase noise, frequency stability, and manufacturability. However, in engineering practice, there is an inherent and irreconcilable contradiction between high frequency, high Q value, and low motion impedance. That is, there is a natural performance trade-off among the three, making it difficult to achieve optimal performance simultaneously. This constitutes the core obstacle for the current development of MEMS resonator technology towards high performance. Summary of the Invention

[0004] This application provides a piezoelectric-electrostatic composite MEMS resonator and MEMS resonator array, which achieves high Q value, low impedance and high frequency performance through the synergy of piezoelectric and electrostatic transduction mechanisms.

[0005] This application provides a piezoelectric-electrostatic composite MEMS resonator, comprising: silicon substrate; A main resonator, made of low-loss material and used to generate resonance, the main resonator includes a main mechanical energy storage region and an edge functional region disposed around the main mechanical energy storage region, the edge functional region being integrally formed with the main mechanical energy storage region; Multiple anchor points are used to fix the silicon substrate; Multiple support arms are disposed in the same layer as the main resonator, and the multiple support arms and the main resonator are spaced apart from the silicon substrate to form a release cavity. Each support arm is made of a low-loss material, and its front end is connected to the edge functional area of ​​the main resonator and its end is connected to a corresponding anchor point. Multiple piezoelectric transducer layers are disposed on a corresponding support arm, or disposed on a corresponding support arm and extending only to the edge functional area of ​​the main resonator; each piezoelectric transducer layer includes a lower electrode, a piezoelectric film and an upper electrode sequentially disposed on the support arm, and the three form a piezoelectric transducer structure. Multiple fixed electrodes are disposed adjacent to the main resonator and an air gap is provided between them. The fixed electrodes are used to induce the main resonator to resonate when energized, so as to form an electrostatic transduction structure.

[0006] Optionally, a single-crystal silicon layer is disposed in a corresponding region on the support arm, and the single-crystal silicon layer is subjected to heavy doping treatment to form a heavily doped single-crystal silicon layer, which serves as the lower electrode.

[0007] Optionally, the support arm is a single-crystal silicon structure layer, and the corresponding region on the support arm is heavily doped to form a heavily doped single-crystal silicon layer, which serves as the lower electrode.

[0008] Optionally, the lower electrode, the main resonator, and the support arm all belong to the single-crystal silicon material system.

[0009] Optionally, the piezoelectric film is deposited directly on the lower electrode.

[0010] Optionally, the lower electrode is electrically connected through an intermediate layer, which is a conductive adhesion layer, a diffusion barrier layer, or a seed layer, and the thickness of the intermediate layer is less than 50 nm.

[0011] Optionally, the length of the support arm is Ls along the direction of the main resonator toward the corresponding anchor point; the wavelength of the sound wave in the main resonator under the target vibration mode is λ, and satisfies: Ls = λ / 4 or λ / 2.

[0012] Optionally, the target vibration mode includes at least one of the following: Radial breathing mode with n=0, elliptical mode with n=2, Lamé mode, contour mode, width extension mode, higher-order volumetric acoustic mode, and in-plane vibration mode with definite vibration nodes.

[0013] Optionally, the piezoelectric transducer layer extends above the edge functional region of the main resonator. The piezoelectric transducer layer is divided into a readout electrode region and a tuning electrode region. The readout electrode region is used to connect an oscillator or detection circuit to perform the function of the piezoelectric transducer structure. The tuning electrode region is used to connect an adjustable termination impedance network. By changing the impedance value of the adjustable termination impedance network, the dynamic effective modulus of the tuning electrode region is adjusted to change the equivalent dynamic stiffness of the support arm located in the tuning electrode region.

[0014] Optionally, the piezoelectric-electrostatic composite MEMS resonator includes four support arms and four fixed electrodes. The four support arms are arranged symmetrically in pairs, and the four fixed electrodes are arranged symmetrically in pairs around the main resonator. Each fixed electrode is disposed between two adjacent support arms.

[0015] Optionally, the upper and / or lower electrodes located on the two support arms symmetrically arranged therein serve as differential detection terminals, which are used to output detection signals with opposite polarities. The upper and / or lower electrodes located on the other two symmetrically arranged support arms serve as differential drive terminals, which are used to output drive signals with opposite polarities.

[0016] Optionally, the piezoelectric-electrostatic composite MEMS resonator further includes multiple acoustic reflectors and multiple coupling beams. The multiple support arms and the multiple coupling beams are spaced apart from the silicon substrate to form a release cavity. The two ends of the coupling beams are connected to the ends of the support arms and the acoustic reflectors. The acoustic reflectors are disposed between the coupling beams and the front end of the anchor point and are used to reflect the acoustic energy leaking along the support arms.

[0017] Optionally, the reflection coefficient of the acoustic reflector satisfies the following relationship:

[0018] in, The acoustic impedance of the acoustic reflector is... The acoustic impedance of the support arm is given.

[0019] Optionally, the fixed electrode is also used to apply a DC bias voltage as a tuning electrode, and the resonant frequency of the main resonator is adjusted using the following spring softening effect relationship:

[0020] in, The absolute change in the resonant frequency of the main resonator when the DC bias voltage is applied. This is the inherent mechanical resonant frequency of the main resonator when the DC bias voltage is not applied. The vacuum permittivity, The area of ​​the fixed electrode. The DC bias voltage is mentioned above. The width of the air gap. The equivalent mechanical stiffness of the main resonator, and the tuning range achieved by the fixed electrode are within... The range is from 0.01% to 1%.

[0021] This application provides a MEMS resonator array, comprising a plurality of piezoelectric-electrostatic composite MEMS resonators as described in any of the preceding claims, wherein the piezoelectric-electrostatic composite MEMS resonators are electrically connected in parallel, mechanically coupled, piezoelectrically coupled, or capacitively coupled to each other.

[0022] As described above, the main resonator of this application is made of low-loss material, which acts as the high-performance mechanical vibration body, while the piezoelectric transducer layer serves only as a high-efficiency signal transducer. Since the vibration of the main resonator no longer depends on the piezoelectric transducer layer, its thickness, electrode design, etc., can be simplified or even removed to reduce losses, thereby effectively suppressing the degradation of the quality factor (Q value) by the piezoelectric transducer layer. Furthermore, this application forms a piezoelectric transducer structure and an electrostatic transducer structure. The high-efficiency electromechanical conversion of the piezoelectric transducer structure ensures low motion impedance, while the non-contact characteristics of the electrodes of the electrostatic transducer structure with the movable structure, including the main resonator, provide a low-loss, low-feedthrough detection or driving path. Through the synergy of the piezoelectric and electrostatic transducer mechanisms, the composite MEMS resonator of this application can maintain a high frequency while also possessing excellent quality factor and low impedance level, that is, taking into account high Q value, low impedance, and high-frequency performance. Attached Figure Description

[0023] Figure 1 This is a top view of the structure of a piezoelectric-electrostatic composite MEMS resonator according to an embodiment of this application; Figure 2 yes Figure 1 The piezoelectric-electrostatic composite MEMS resonator shown is a cross-sectional view along A-A'. Figure 3 This is a schematic diagram of the connection between the four-port differential drive and the detection signal according to an embodiment of this application; Figure 4 This is an equivalent schematic diagram of the piezoelectric-electrostatic composite MEMS resonator of this application in PP mode; Figure 5 This is an equivalent schematic diagram of the piezoelectric-electrostatic composite MEMS resonator of this application in PC mode; Figure 6 This is an equivalent schematic diagram of the piezoelectric-electrostatic composite MEMS resonator of this application in CP mode; Figure 7 This is an equivalent schematic diagram of the piezoelectric-electrostatic composite MEMS resonator of this application in CC mode; Figure 8 This is a top view of the structure of a piezoelectric-electrostatic composite MEMS resonator according to another embodiment of this application; Figure 9 yes Figure 8 A schematic diagram of one of the sound energy reflection paths is shown; Figure 10This is a top view of the structure of a MEMS resonator array according to an embodiment of this application.

[0024] Figure reference numerals: Composite resonator 1, silicon substrate 10, release cavity Z0, main resonator 11, main mechanical energy storage region Z1, edge functional region Z2, support arm 12, piezoelectric transducer layer 13, lower electrode 131, piezoelectric thin film 132, upper electrode 133, fixed electrode 14, anchor point 15, acoustic reflector 16, coupling beam 17, buried oxide layer 18, mechanical coupling beam 19, MEMS resonator array 100. Detailed Implementation

[0025] To address the problems of the prior art, this application provides a piezoelectric-electrostatic composite MEMS resonator and a MEMS resonator array. These two protection subjects are based on the same concept, and the principles for solving the problems are basically the same or similar. The implementation methods of each protection subject can be referred to mutually, and repeated details will not be elaborated.

[0026] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly described below in conjunction with specific embodiments and corresponding drawings. Obviously, the embodiments described below are only a part of the embodiments of this application, and not all of them. Unless otherwise specified, the following embodiments and their technical features can be combined with each other, and also belong to the technical solutions of this application.

[0027] In the description of the embodiments of this application, the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the technical solutions of the corresponding embodiments, and are not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. They should not be construed as limitations on this application.

[0028] Figure 1 This is a top view of the structure of a piezoelectric-electrostatic composite MEMS resonator according to an embodiment of this application. Figure 2 yes Figure 1 The diagram shows a cross-sectional view of the piezoelectric-electrostatic composite MEMS resonator along line A-A'. For better description, the piezoelectric-electrostatic composite MEMS resonator will be referred to as a "composite resonator" in the following description, although it may also be called a "composite MEMS resonator" in certain contexts. Figure 1 In the orientation shown, the so-called top-down direction can be called the second direction y, which is the direction perpendicular to the plane where the composite resonator is located. Figure 1The first direction x and the third direction z shown can be regarded as the other two coordinate axes of the three-dimensional Cartesian coordinate system in which the composite resonator is located.

[0029] See also Figure 1 and Figure 2 As shown, the composite resonator 1 includes a silicon substrate 10, a main resonator 11, multiple support arms 12, multiple piezoelectric transducer layers 13, multiple fixed electrodes 14, and multiple anchor points 15. Figure 1 The image shows four support arms 12, four piezoelectric transducer layers 13, four fixed electrodes 14, and four anchor points 15. This is merely an illustrative example and does not constitute a limitation on the scope of protection of this application.

[0030] The silicon substrate 10 is a planar structure as a whole.

[0031] The main resonator 11, made of a low-loss material, can also be called an oscillator, and is used to generate resonance at the desired frequency, i.e., resonance at a preset resonant frequency. The main resonator 11 includes a main mechanical energy storage region Z1 and an edge functional region Z2 disposed on the periphery of the main mechanical energy storage region Z1, and is integrally formed with the main mechanical energy storage region Z1, including lateral edge surfaces disposed opposite to the fixed electrode 14. The main mechanical energy storage region Z1 is the region where mechanical vibration energy, elastic strain energy, or kinetic energy is mainly concentrated under the target resonance mode, and is used to determine the intrinsic frequency, equivalent stiffness, and equivalent mass of the composite resonator 1. The edge functional region Z2 is located on the periphery of the main mechanical energy storage region Z1 and is integrally formed with the main mechanical energy storage region Z1 by the same low-loss structural layer. The edge functional region Z2 is used to connect the support arm 12, form an electrostatic transduction boundary opposite to the fixed electrode 14, bear at least part of the piezoelectric detection or tuning structure, form a stress relief transition region, and / or constitute an acoustic impedance transition region, thereby reducing the mechanical loss contribution of the piezoelectric thin film 132 and electrode (including the lower electrode 131 and upper electrode 133 described below) materials of the piezoelectric transduction layer 13 to the main mechanical energy storage region Z1. The edge functional region Z2 can also be used to apply electrostatic driving force, electrostatic tuning force or electrostatic detection coupling to the main resonator 11 to adjust the modal shape of the target resonance mode.

[0032] Multiple anchor points 15 are fixed on the silicon substrate 10.

[0033] Multiple support arms 12 are arranged in the same layer as the main resonator 11, and the multiple support arms 12 and the main resonator 11 are spaced apart from the silicon substrate 10 to form a release cavity Z0. Each support arm 12 is made of low-loss material, and its front end is connected to the main resonator 11 and its end is connected to a corresponding anchor point 15.

[0034] exist Figure 1 and Figure 2In the example shown, each piezoelectric transducer layer 13 is disposed on a corresponding support arm 12 and extends only to the edge functional region Z2 of the main resonator 11, but does not extend to the main mechanical energy storage region Z1. In other examples, each piezoelectric transducer layer 13 may be disposed on a corresponding support arm 12, that is, it does not extend to any region of the main resonator 11.

[0035] The piezoelectric transducer layer 13 includes a lower electrode 131, a piezoelectric film 132, and an upper electrode 133 sequentially disposed on the support arm 12, and these three form a piezoelectric transducer structure.

[0036] Multiple fixed electrodes 14 are disposed adjacent to the main resonator 11, and each fixed electrode 14 is spaced apart from the main resonator 11 to form an air gap. The fixed electrodes 14 are used to induce the main resonator 11 to resonate when energized, so as to form an electrostatic transduction structure.

[0037] Based on this, the main resonator 11 of this application is made of a low-loss material, which acts as the high-performance mechanical vibration body, while the piezoelectric transducer layer 13 serves only as a high-efficiency signal transducer. Since the vibration of the main resonator 11 no longer depends on the piezoelectric transducer layer 13, its thickness, electrode design, etc., can be simplified or even removed to reduce losses, thereby effectively suppressing the degradation of the quality factor (Q value) by the piezoelectric transducer layer 13. Furthermore, this application forms a piezoelectric transducer structure and an electrostatic transducer structure. The high-efficiency electromechanical conversion of the piezoelectric transducer structure ensures low motion impedance, while the non-contact characteristics of the electrodes (i.e., fixed electrodes 14) of the electrostatic transducer structure with the movable structure including the main resonator 11 (and of course, the support arm 12 and the piezoelectric transducer layer 13) provide a low-loss, low-feedthrough detection or driving path. Through the synergy of the piezoelectric and electrostatic transducer mechanisms, the composite resonator 1 of this application can maintain a high frequency while also possessing excellent quality factor and low impedance level, that is, taking into account high Q value, low impedance, and high-frequency performance.

[0038] The following section will introduce each of the above-mentioned structural components of the composite resonator 1.

[0039] Main resonator 11 The main resonator 11 can vibrate as the core of mechanical energy storage. Its main structure (i.e., the main mechanical energy storage region Z1) does not have a piezoelectric transducer layer 13 to avoid internal friction and Q-value degradation caused by the introduction of heterogeneous materials. The main resonator 11 can be a disk, a polygonal disk, a ring structure, a rectangular plate, a square plate, or other shapes. It can be made of monocrystalline silicon (including SOI device layer monocrystalline silicon, bulk silicon, low-doped high-resistivity silicon, and heavily doped monocrystalline silicon), or polycrystalline silicon, diamond, silicon carbide, silicon nitride, or other low-loss materials.

[0040] The main resonator 11 can operate in one of the following vibration modes: n=0 radial breathing mode (Button-like mode), n=2 elliptical mode (wine-glass mode), Lamé mode, contour mode, width-extensional mode, high-order BAW mode, or other in-plane vibration modes with a defined vibration node. This application preferably uses the n=0 radial breathing mode, which has a high frequency, high Q value, a defined node position (i.e., a specific radial position near the functional area Z2 at the edge of the disk), and high symmetry, making it suitable for node anchoring and differential readout.

[0041] For example, for Figure 1 The disk-shaped main resonator 11 shown has a resonant frequency of (i.e., the inherent mechanical resonant frequency) satisfies the following relationship:

[0042] in, The equivalent mechanical stiffness of the main resonator 11 is given by [reference to a specific component]. The effective mass of the main resonator 11. For the radial breathing mode with n=0, the resonant frequency is... Approximately:

[0043] in, The diameter of the disk of the main resonator 11. For Young's modulus, The density of the main resonator 11, Poisson's ratio, These are dimensionless coefficients related to the mode, such as the n=0 radial breathing mode for monocrystalline silicon. Up to 0.373, The specific value depends on the crystal orientation. The resonant frequency can be adjusted over a wide range by adjusting the disk diameter of the main resonator 11. .

[0044] piezoelectric transducer layer 13 In the composite resonator 1 of this application, the support arm 12 and the piezoelectric transducer layer 13 can be collectively referred to as the support transducer arm. Figure 1 and Figure 2 In the example, the support arm 12 connects the main resonator 11 and the anchor point 15, and the piezoelectric transducer layer 13 does not cover the main mechanical energy storage area Z1 of the main resonator 11.

[0045] The piezoelectric thin film 132 of the piezoelectric transducer layer 13 is preferably made of scandium-doped aluminum nitride (ScAlN), wherein the molar fraction of scandium (Sc) can be selected in the range of 10% to 43% to maintain low mechanical loss while improving the electromechanical coupling coefficient. The material of the piezoelectric thin film 132 can also be aluminum nitride (AlN), lead zirconate titanate (PZT), zinc oxide (ZnO), lithium niobate (LiNbO3), lithium tantalate (LiTaO3), or other piezoelectric materials. The materials of the lower electrode 131 and the upper electrode 133 can be the same, preferably molybdenum (Mo), but platinum (Pt), ruthenium (Ru), aluminum (Al), titanium (Ti), titanium nitride (TiN), or other conductive materials can also be used. The thickness of the lower electrode 131 and the upper electrode 133 can be the same, preferably in the range of 50 nm to 300 nm, to minimize the influence of these two metal layers on the Q value while ensuring conductivity.

[0046] In other examples, the lower electrode 131 can be formed from a heavily doped monocrystalline silicon layer, and is not necessarily limited to a metal. Specifically, a monocrystalline silicon layer is disposed in a corresponding region on the support arm 12, and this monocrystalline silicon layer is heavily doped to form a heavily doped monocrystalline silicon layer to serve as the lower electrode 131; or, the support arm 12 is a monocrystalline silicon structure layer, and a corresponding region on the support arm 12 is heavily doped to form a heavily doped monocrystalline silicon layer to serve as the lower electrode 131. In this example, a low-resistivity heavily doped monocrystalline silicon layer is formed by ion implantation, in-situ doped epitaxy, diffusion doping, etc. The heavily doped monocrystalline silicon layer can be an n-type heavily doped silicon layer or a p-type heavily doped silicon layer. The n-type dopant can be selected from phosphorus, arsenic, antimony, or combinations thereof, and the p-type dopant can be selected from boron, aluminum, gallium, or combinations thereof. The doping concentration of the heavily doped monocrystalline silicon layer can be 1 × 10⁻⁶. 18 cm -3 Up to 1×10 21 cm -3 Preferably 1×10 20 cm -3 Up to 3×10 20 cm -3 This is to maintain a high mechanical quality factor of the single-crystal silicon structure layer while ensuring low sheet resistance and good electrode conductivity.

[0047] Compared to the metal-based lower electrode 131, the heavily doped monocrystalline silicon lower electrode 131, along with the main resonator 11 and the support arm 12, belongs to the same monocrystalline silicon material system, exhibiting better lattice continuity and acoustic continuity. This reduces internal friction, interface losses, mass loading, and abrupt changes in acoustic impedance caused by the metal layer. Especially in local regions of the piezoelectric transducer layer 13, using a heavily doped monocrystalline silicon layer as the lower electrode 131 further reduces the proportion of elastic strain energy in the metal, which is beneficial for improving the mechanical quality factor.

[0048] The piezoelectric thin film 132 can be directly deposited on the lower electrode 131, or it can be electrically connected to the lower electrode 131 through an intermediate layer, which can be a conductive adhesion layer, a diffusion barrier layer, or a seed layer. The conductive adhesion layer or diffusion barrier layer can be selected from Ti, TiN, Mo, W, Pt, Ru, or a combination thereof, with a thickness of less than 50 nm, preferably less than 20 nm, to reduce additional mechanical losses.

[0049] The geometry of the support arm 12 can be a rectangular beam, a T-beam, an L-beam, a stepped beam, or other shapes. Figure 1 The example is a T-beam structure, whose laterally widened portion can increase the transducer area and improve the electromechanical coupling coefficient. Simultaneously, its longitudinally thinner beam portion can reduce the constraint on the vibration of the main resonator 11 and reduce anchor point losses. Further optionally, the length of the support arm 12 is Ls, which refers to the length along the direction from the main resonator 11 towards the corresponding anchor point 15. The wavelength of the sound wave in the target vibration mode of the main resonator 11 is λ, and it satisfies: Ls = λ / 4 or λ / 2. The target vibration mode includes at least one of the following: n=0 radial breathing mode, n=2 elliptical mode, Lamé mode, contour mode, width extension mode, higher-order volume acoustic wave mode, and in-plane vibration mode with a defined vibration node. Based on this structural design, the support arm 12 acts as an acoustic waveguide connecting the main resonator 11 and the anchor point 15. The equivalent acoustic impedance at its input end (the side connected to the main resonator 11) changes periodically with its length. The anchor point 15 can be considered a rigid fixed end, and its acoustic impedance tends to infinity. When Ls = λ / 4, the high impedance of anchor point 15, after being transformed by a quarter wavelength, presents an acoustic short circuit at the input end of support arm 12 (equivalent impedance approaches zero), causing the vibration energy to be completely reflected back to the main resonator 11 at the entrance, preventing it from entering support arm 12. When Ls = λ / 2, the impedance of anchor point 15, after being transformed by a half wavelength, presents an acoustic open circuit at the input end of support arm 12 (equivalent impedance approaches infinity), similarly blocking the propagation of acoustic energy to anchor point 15. Both length designs can achieve extremely high acoustic energy reflection efficiency and significantly reduce the loss of anchor point 15. The difference between the two lies in the different boundary conditions presented at the input end of support arm 12: the λ / 4 length design can provide an approximately fixed end constraint, where the displacement is minimal and the stress is relatively large, while the λ / 2 length design can provide an approximately free end constraint, where the stress is minimal and the displacement is relatively large. In practical scenarios, the optimal design can be selected based on the strain distribution at the connection of the target vibration mode at the main resonator 11.

[0050] Fixed electrode 14 A fixed electrode 14 is disposed near the main resonator 11 and the support arm 12. An air gap, also known as an air gap capacitor, is formed between the fixed electrode 14 and the movable structure (e.g., the main resonator 11). The width of the air gap can be selected in the range of 50 nm to 5 µm. A release cavity Z0 is located below both the main resonator 11 and the support arm 12, forming a movable structure.

[0051] The composite resonator 1 can achieve the following multiple functions through the fixed electrode 14: 1. Electrostatic drive: An alternating voltage is applied between the fixed electrode 14 and the main resonator 11, i.e., a superimposed DC bias is applied, and the main resonator 11 is driven to vibrate by electrostatic force.

[0052] II. Electrostatic detection: Detects the dynamic change in capacitance between the fixed electrode 14 and the main resonator 11, and senses the vibration displacement of the main resonator 11.

[0053] III. Frequency Tuning: Adjust the DC bias voltage on the fixed electrode 14 and use the electrostatic spring softening effect to change the effective stiffness of the main resonator 11 to achieve resonant frequency adjustment.

[0054] IV. Non-contact piezoelectric-electrostatic composite transducer: This type utilizes an air gap capacitor to achieve coordinated operation of electrostatic drive and piezoelectric detection, or coordinated operation of piezoelectric drive and electrostatic detection, thus combining the low-loss characteristics of capacitive transducer with the high-efficiency characteristics of piezoelectric transducer. In the non-contact piezoelectric-electrostatic transducer mode, the voltage distribution relationship between the air gap capacitor and the piezoelectric transducer layer 13 is as follows:

[0055] in, The voltage across the piezoelectric transducer 13 is... Input voltage, The thickness of the piezoelectric transducer layer 13 is... Here is the dielectric constant of piezoelectric transducer 13. The width of the air gap. is the vacuum permittivity. Through optimization... and The ratio of the two values ​​can be used to adjust the effective driving voltage on the piezoelectric transducer layer 13, thereby optimizing the transducer efficiency.

[0056] V. Feedthrough Compensation: By using differential electrode configuration or dedicated compensation electrode, the electrical feedthrough caused by the air gap capacitance acting as static capacitance C0 is offset.

[0057] VI. Array Coupling Adjustment: By adjusting the air gap capacitance, the modal coupling strength between each unit in the MEMS resonator array based on the composite resonator 1 is controlled.

[0058] VII. Differential Readout: Differential electrostatic detection is achieved by using a diagonal differential electrode configuration for multiple fixed electrodes 14, effectively suppressing common-mode noise and feedthrough.

[0059] Therefore, air gap capacitors can achieve a variety of optional functions, such as non-contact piezoelectric-electrostatic composite transduction, frequency tuning, differential readout, and feedthrough compensation. Air gap capacitors can operate independently (i.e., purely electrostatic drive or detection) or in conjunction with piezoelectric transduction (i.e., piezoelectric drive + electrostatic detection, electrostatic drive + piezoelectric detection, etc.), combining the low-loss characteristics of capacitive transduction with the high-efficiency characteristics of piezoelectric transduction.

[0060] Anchor point 15 Anchor point 15 is connected to the end of support arm 12 and serves as the fixed end of the entire movable structure chain, fixedly connected to silicon substrate 10. In acoustic design, anchor point 15 can be positioned at the displacement node of the target vibration mode or in the region of minimum acoustic flux density (e.g., adapted to the length of support arm 12) to reduce acoustic energy leakage loss at anchor point 15. The criterion for anchoring this displacement node is: ,in The amplitude of the vibration displacement at anchor point 15. This is the maximum vibration displacement amplitude of the main resonator 11. Preferably, this ratio should be less than 0.05 to further reduce the loss of the anchor point 15.

[0061] The geometry of anchor point 15 can be square or circular, and its area can be in the range of 1µm×1µm to 5µm×5µm, in order to minimize the cross-sectional area of ​​the acoustic energy leakage channel at anchor point 15.

[0062] Figure 3 This is a schematic diagram showing the connection between the four-port differential drive and detection signals according to an embodiment of this application. Please refer to it as well. Figures 1 to 3 In the example where the composite resonator 1 includes four support arms 12 and four fixed electrodes 14, the four support arms 12 are arranged symmetrically in pairs, that is, the four support arms 12 are evenly distributed along the symmetry axis direction of the main resonator 11 (i.e., the 0°, 90°, 180°, and 270° directions), as shown below. Figure 1 As shown, the four fixed electrodes 14 are arranged symmetrically in pairs around the main resonator 11, with each fixed electrode 14 positioned between two adjacent support arms 12.

[0063] Electrodes (e.g., upper electrode 133 and / or lower electrode 131) located on two symmetrically arranged support arms 12 can serve as differential detection terminals, which are used to output detection signals with opposite polarities; electrodes (e.g., upper electrode 133 and / or lower electrode 131) located on two other symmetrically arranged support arms 12 can serve as differential driving terminals, which are used to output driving signals with opposite polarities. Specifically, in a four-port fully differential structure, preferably two diagonally opposite (e.g., 0° and 180° directions) support arms 12 and their electrodes serve as differential driving terminals (i.e.,... Figure 3 Ports 1 and 3 in the middle), and the other two diagonally opposite (such as 90° and 270° directions) support arms 12 and their electrodes serve as differential detection terminals (i.e. Figure 3 (Ports 2 and 4 in the diagram). A drive signal of opposite polarity is applied to the differential drive terminal, also known as a differential drive signal. The differential detection terminal outputs a detection signal of opposite polarity, also known as a sensing voltage or differential voltage. The output differential voltage is... .

[0064] The main advantages of this four-port fully differential structure include: First, differential driving enables the main resonator 11 to be symmetrically excited, effectively enhancing the target vibration mode while suppressing asymmetric spurious modes; Second, differential detection cancels common-mode feedthrough, significantly reducing the interference of electrical feedthrough on the resonant signal; Third, the amplitude of the output differential signal is twice that of the single-ended output, effectively reducing the motion impedance; Fourth, this differential structure has good suppression capability against external common-mode interference (such as power supply noise and substrate noise).

[0065] In this four-port fully differential structure, the piezoelectric transducer layers 13 extending from each support arm 12 to the edge functional region Z2 can be polarity-grouped according to the strain polarity of the target vibration mode of the main resonator 11. For example, for the n=2 elliptical mode, two piezoelectric transducer layers 13 extending to the edge functional region Z2 opposite to each other can be connected as in-phase ports, and two piezoelectric transducer layers 13 extending to the edge functional region Z2 in orthogonal directions can be connected as out-of-phase ports, thereby matching the piezoelectric driving force or detection charge distribution with the spatial function of the target vibration mode. For the n=0 radial breathing mode, multiple piezoelectric transducer layers 13 extending to the edge functional region Z2 can be connected in parallel in-phase to enhance radially symmetrical driving and detection response.

[0066] In other examples, this application may provide other numbers of support arms 12, fixed electrodes 14 and piezoelectric transducer layers 13 so that the composite resonator 1 supports multiple operating modes and can be flexibly selected according to specific application scenarios. According to the differences in transducer mechanisms, it can be divided into the following seven operating modes.

[0067] Mode 1: PP mode (piezoelectric drive + piezoelectric detection), combined with Figure 4 As shown, the composite resonator 1 has two symmetrically arranged support arms 12 and two piezoelectric transducer layers 13 located on them. The fixed electrode 14 can be omitted. In this case, one of the support arms 12 and the piezoelectric transducer layer 13 serves as a differential drive terminal for inputting the drive voltage V. drive Another support arm 12 and the piezoelectric transducer layer 13 located thereon serve as a differential detection terminal for outputting a differential voltage V. out Both the differential drive end and the differential detection end adopt piezoelectric transduction, which has the highest electromechanical coupling efficiency and is suitable for low motion impedance application scenarios.

[0068] Mode 2: PC mode (piezoelectric drive + electrostatic detection), combined with Figure 5 As shown, the composite resonator 1 is provided with a support arm 12 and a piezoelectric transducer layer 13 located thereon. A fixed electrode 14 is arranged opposite to the support arm 12. An air gap (i.e., forming an air gap capacitor) is formed between the fixed electrode 14 and the main resonator 11. Figure 5 (The capacitor shown in the figure), at this time, the support arm 12 and the piezoelectric transducer layer 13 located thereon serve as differential drive terminals for inputting drive voltage V. drive The fixed electrode 14, also known as the fixed detection electrode, serves as the differential detection terminal and is used to output the differential voltage V. out The differential driving end uses piezoelectric transducer (non-contact type), and the differential detection end uses capacitive transducer. At this time, the differential detection end does not have piezoelectric film 132 and metal layer (i.e., upper electrode 133 and lower electrode 131 made of metal material), which can reduce its influence on Q value and reduce static capacitance C0 to suppress feedthrough.

[0069] Mode 3: CP mode (electrostatic drive + piezoelectric detection), combined with Figure 6 As shown, the composite resonator 1 is provided with a support arm 12 and a piezoelectric transducer layer 13 located thereon. A fixed electrode 14 is arranged opposite to the support arm 12. An air gap (i.e., forming an air gap capacitor) is formed between the fixed electrode 14 and the main resonator 11. Figure 6 (The capacitor shown in the figure), at this time, the fixed electrode 14 is also called the fixed drive electrode as a differential drive terminal, used for input drive voltage V. drive +V DC V DC The DC bias voltage is used, and the support arm 12 and the piezoelectric transducer layer 13 located on it serve as differential detection terminals for outputting differential voltage V. out The differential driving end uses capacitive transduction (non-contact), and the differential detection end uses piezoelectric transduction. The capacitive transduction used in the differential driving end is an electrostatic drive, which does not introduce additional losses and is suitable for scenarios sensitive to Q-value.

[0070] Mode 4: CC mode (electrostatic drive + electrostatic detection), combined with Figure 7 As shown, the composite resonator 1 has two symmetrically arranged fixed electrodes 14, one of which is called the fixed drive electrode and serves as the differential drive terminal for inputting the drive voltage V. drive +V DC Another electrode serves as a fixed detection electrode, acting as a differential detection terminal for outputting a differential voltage V. out Both the differential driving end and the differential detection end employ capacitive transduction. In this case, the composite resonator 1 does not require the piezoelectric thin film 132 and the metal layer, achieving the highest Q value, but with relatively high motion impedance, making it suitable for applications with extremely high Q value requirements.

[0071] Mode 5: Combined driving mode, that is, the differential driving end is simultaneously provided with a fixed electrode 14 and a support arm 12 and a piezoelectric transducer layer 13 located thereon, so that piezoelectric transducer and capacitor transducer are used at the same time. The two mechanisms work together to improve driving efficiency and reduce driving voltage requirements.

[0072] Mode 6: Joint detection mode, that is, the differential detection end is simultaneously provided with a fixed electrode 14, a support arm 12 and a piezoelectric transducer layer 13 located thereon, so as to simultaneously use piezoelectric transducer and capacitive transducer, and weighted summation of the detection signals generated by the two transducers to optimize the signal-to-noise ratio.

[0073] Mode 7: Piezoelectric transduction + electrostatic frequency tuning: A fixed electrode 14 and a support arm 12 and a piezoelectric transducer layer 13 are simultaneously set on either side of the main resonator 11. While using piezoelectric transduction, the fixed electrode 14 is used as a dedicated electrostatic tuning electrode. Frequency tuning is achieved by adjusting the DC bias voltage applied to the fixed electrode 14, thereby compensating for process deviations, temperature drift and aging drift.

[0074] Mode 8: Piezoelectric stress tuning + capacitor-assisted detection: A fixed electrode 14, a support arm 12, and a piezoelectric transducer layer 13 on the differential detection end are set simultaneously. Stress is generated by applying a DC electric field to the piezoelectric transducer layer 13 to adjust the effective stiffness of the main resonator 11 to achieve frequency tuning. At the same time, capacitor-assisted detection is achieved by the air gap capacitance formed between the fixed electrode 14 and the main resonator 11. The detection sensitivity is improved by using capacitor-assisted detection.

[0075] Among them, Mode 7 and Mode 8 can be regarded as auxiliary function type transducer modes.

[0076] The electrostatic frequency tuning and the piezoelectric stress tuning are described below.

[0077] Electrostatic frequency tuning The fixed electrode 14 near the main resonator 11 and the support arm 12 can serve as a tuning electrode, which is used to apply a DC bias voltage. After the DC bias voltage is applied, the effective stiffness (also known as "equivalent mechanical stiffness") of the main resonator 11 is reduced by the electrostatic spring softening effect, thereby achieving frequency downregulation. For example, the fixed electrode 14, as the tuning electrode, is disposed near the circular main resonator 11, and is spaced apart from the main resonator 11 to form an air gap capacitance. The width of the air gap generated by the spaced-apart arrangement is in the range of 200nm to 2µm. By adjusting the DC bias voltage on the tuning electrode, the resonant frequency of the main resonator 11 is adjusted using the following spring softening effect relationship:

[0078] in, This refers to the absolute change in the resonant frequency of the main resonator 11 when the DC bias voltage is applied. The inherent mechanical resonant frequency of the main resonator 11 when the DC bias voltage is not applied. The vacuum permittivity, The area of ​​the fixed electrode 14 is... The DC bias voltage is mentioned above. The width of the air gap. The equivalent mechanical stiffness of the main resonator 11, and the tuning range achieved by the fixed electrode 14 are within... The range is from 0.01% to 1%.

[0079] Piezoelectric stress tuning For the structural design where the piezoelectric transducer layer 13 extends only to the edge functional region Z2 of the main resonator 11, for ease of description, the portion of the piezoelectric transducer layer 13 extending to this edge functional region Z2 can be referred to as the extension. By applying a DC bias voltage to this extension, in-plane tensile strain, compressive strain, or bending strain is generated in the extension, thereby applying a static tuning force to the main resonator 11, changing the equivalent stiffness, boundary constraints, and strain energy distribution of the target mode of the main resonator 11, and achieving frequency adjustment, i.e., realizing so-called piezoelectric stress tuning, which belongs to active tuning. The local piezoelectric transducer generated by this extension in the edge functional region Z2 of the main resonator 11 can enhance the stress application efficiency and improve the frequency tuning sensitivity under unit tuning voltage. Piezoelectric stress tuning can achieve bidirectional tuning according to the electric field direction, piezoelectric polarization direction, and support arm 12 structure, i.e., frequency up or down, with the tuning range preferably within the resonant frequency. The range is between 0.01% and 0.5%.

[0080] The composite resonator 1 of this application can also achieve piezoelectric stiffness enhancement tuning. Unlike the aforementioned active tuning method that generates static piezoelectric stress through DC bias voltage, piezoelectric stiffness enhancement tuning is a passive tuning mechanism. Its core principle is: by changing the equivalent termination impedance between the lower electrode 131 and the upper electrode 133, the dynamic effective modulus of the piezoelectric film 132 under the target vibration mode is adjusted, thereby changing the equivalent dynamic stiffness of the support arm 12, and thus affecting the equivalent stiffness and output frequency of the entire composite resonator 1 mechanical system. This method does not rely on static stress, but modulates the dynamic contribution of the piezoelectric material (i.e., the piezoelectric film 132 of the piezoelectric transducer layer 13) to the vibration by changing the electrical boundary conditions.

[0081] These two types of tuning are electrically independent but mechanically work together on the main resonator 11. Piezoelectric stress tuning handles low-voltage, wide-range coarse tuning, while electrostatic frequency tuning handles high-precision fine tuning. The combination of the two achieves both a wide tuning range and high tuning resolution, avoiding the limitations of a single tuning method. In addition, electrostatic frequency tuning can also be used to compensate for nonlinear effects introduced by piezoelectric transduction (such as Duffing nonlinearity). By adjusting the DC bias voltage to change the nonlinear stiffness coefficient, it cancels out the nonlinearity caused by the deformation of the piezoelectric film 132, thereby extending the linear operating range and improving power capacity and dynamic range.

[0082] In one example, for a structural design where the piezoelectric transducer layer 13 extends over the edge functional region Z2 of the main resonator 11, the piezoelectric transducer layer 13 can be divided into a readout electrode region and a tuning electrode region. The readout electrode region is used to connect an oscillator or detection circuit to perform the function of the piezoelectric transducer structure. The tuning electrode region is used to connect an adjustable termination impedance network (such as a varactor diode, capacitor array, or variable resistor network). By changing the impedance value of the adjustable termination impedance network, the dynamic effective modulus of the tuning electrode region is adjusted to change the equivalent dynamic stiffness of the support arm 12 located in the tuning electrode region. Since the tuning electrode region is only located on the support arm 12 and does not cover the main mechanical energy storage region Z1 of the main resonator 11, this tuning method achieves fine-tuning of the target modal frequency without introducing additional mechanical losses, and has minimal impact on the high Q value of the main resonator 11.

[0083] The core advantages of the aforementioned passive tuning mechanism include: First, zero DC power consumption, meaning that the tuning process does not consume static power, making it suitable for low-power applications; Second, multi-level collaborative capability, meaning that the tuning range and resolution depend on the adjustment accuracy of the impedance network, and it can be used in conjunction with electrostatic frequency tuning or piezoelectric stress tuning to form a multi-level tuning strategy of "coarse tuning + fine tuning"; Third, simple structure and easy integration, meaning that no additional large-area electrodes or complex process steps are required, and it is highly compatible with existing MEMS process flows.

[0084] The principle of coarse tuning via piezoelectric stress tuning is as follows: by applying a DC bias electric field to the piezoelectric transducer layer 13, in-plane strain or bending strain is generated, changing the stress state at the edge of the support arm 12 or the main resonator 11, thus achieving a large range of frequency pre-adjustment. This is used to compensate for initial frequency deviations caused by process deviations, packaging stress, etc., and to bring the frequency into the target window. The principle of fine tuning via electrostatic frequency tuning is as follows: by generating electrostatic force through the air gap capacitor, the spring softening effect is used to achieve fine frequency adjustment with small steps, high resolution, continuous operation, and low hysteresis. This is used for real-time temperature compensation, aging compensation, or closed-loop frequency locking.

[0085] Based on the above, the composite resonator 1 can be used to compensate for the following four types of frequency deviations: (1) Frequency deviation caused by process deviation: compensates for frequency deviation caused by dimensional errors and material parameter fluctuations introduced by processing steps such as photolithography, etching, and deposition. The compensation range is ±0.1% to ±1% of the resonant frequency. This compensation range covers the typical dispersion of conventional MEMS manufacturing processes and can effectively improve wafer-level frequency consistency.

[0086] (2) Frequency deviation caused by temperature drift: compensates for frequency drift caused by changes in the elastic modulus of monocrystalline silicon with temperature, with a compensation range of ±50 ppm / ℃ to ±200 ppm / ℃. This capability enables the composite resonator 1 to maintain a stable frequency output within an operating temperature range of -40℃ to +85℃.

[0087] (3) Frequency deviation caused by packaging stress: compensate for the influence of residual stress generated during chip bonding, surface mounting, injection molding and other packaging processes on the resonant frequency, and eliminate frequency uncertainty introduced by the packaging process.

[0088] (4) Frequency deviation caused by array unit frequency mismatch: When designing the array of composite resonators 1, the resonant frequency of each unit is independently adjusted to reduce the frequency mismatch between units to below ±100ppm, usually below ±10ppm, to avoid the decrease in equivalent Q value and increase in motion impedance caused by frequency dispersion.

[0089] Figure 8 This is a top view of a piezoelectric-electrostatic composite MEMS resonator according to another embodiment of this application. For structural elements with the same names, this embodiment uses the same reference numerals as the aforementioned embodiments. For example... Figure 8 As shown, the composite resonator 1 in this embodiment further includes a plurality of acoustic reflectors 16 and a plurality of coupling beams 17. The plurality of support arms 12 and the plurality of coupling beams 17 are spaced apart from the silicon substrate 10 to form a release cavity Z0. Figure 8 The four acoustic reflectors 16 and four coupling beams 17 shown are merely illustrative examples and do not constitute a limitation on the scope of protection of this application.

[0090] Each of the coupling beams 17 has its two ends connected to the end of a corresponding support arm 12 and a corresponding acoustic reflector 16. The acoustic reflector 16 is disposed between the coupling beam 17 and the front end of the anchor point 15 and is used to reflect the acoustic energy leaking along the support arm 12, thereby further reducing the loss of the anchor point 15.

[0091] The acoustic reflector 16 may be configured in one or a combination of the following forms: (1) Pure silicon reflector: The acoustic reflector 16 is a single-crystal silicon disk or polygonal disk in the same layer as the main resonator 11. Through acoustic impedance matching design, an acoustic reflection interface is formed at the end of the support arm 12.

[0092] (2) Bragg reflector: The acoustic reflector 16 is a periodic acoustic reflector composed of alternating high acoustic impedance layers and low acoustic impedance layers, each layer having a thickness of about one-quarter of the sound wave wavelength.

[0093] (3) Phononic crystal: The acoustic reflector 16 has an acoustic bandgap structure with a periodic geometric structure (such as a periodic aperture array or a periodic mass block array) that prohibits the propagation of sound waves within a specific frequency range.

[0094] (4) Local acoustic bandgap structure: The acoustic reflector 16 forms a local acoustic bandgap in the area where the support arm 12 is located through local geometric changes (such as local necking and local expansion).

[0095] (5) Stepped impedance reflection structure: The acoustic reflector 16 forms an acoustic impedance change through the stepped change of the cross section of the support arm 12, thereby realizing acoustic energy reflection.

[0096] In one example, the reflection coefficient of the acoustic reflector 16 satisfies the following relationship:

[0097] in, The acoustic impedance of the acoustic reflector 16, The acoustic impedance of support arm 12.

[0098] For example, see also Figure 9As shown, the acoustic reflector 16 is a pure silicon reflector disk with a diameter similar to that of the main resonator 11. For example, the diameter of the acoustic reflector 16 is 80% to 120% of the diameter of the main resonator 11. The acoustic reflector 16 is connected to the support arm 12 through a coupling beam 17 (whose width is approximately 20% to 50% of the width of the support arm 12), forming an acoustic impedance abrupt change interface (shown by the dotted line in the figure). When the acoustic energy propagating along the direction of sound energy propagation reaches the acoustic impedance abrupt change interface, the acoustic impedance abrupt change interface reflects the sound energy once, and the reflected sound energy is transmitted towards the main resonator 11. The sound energy passing through the acoustic impedance abrupt change interface (called transmitted sound energy) is relatively small, and when it reaches the acoustic reflector 16, the acoustic reflector 16 performs a secondary reflection, thereby achieving efficient sound energy reflection.

[0099] Furthermore, the connection point between the support arm 12 and the main resonator 11 can be located at the displacement node or near-node region of the target vibration mode, so that the acoustic energy entering the support arm 12 is already at a low level. The support arm 12 extends outward from this displacement node position, passing successively through the coupling beam 17 and the acoustic reflector 16, and finally connects to the anchor point 15. Among them, the acoustic reflector 16 further reflects the residual acoustic energy propagating along the support arm 12 back to the direction of the main resonator 11, significantly attenuating the acoustic energy reaching the anchor point 15 and reducing the loss of the anchor point 15.

[0100] The manufacturing principle and process of the composite resonator 1 of this application are described below.

[0101] S1. Providing a substrate: An SOI wafer is selected, which includes a silicon support substrate, a buried oxide layer and a device layer stacked sequentially. The device layer is single-crystal silicon, and its thickness is preferably in the range of 2µm to 20µm, for example, it can be 10µm in actual manufacturing. The buried oxide layer is SiO2, and its thickness is preferably in the range of 0.5µm to 5µm, for example, it can be 2µm in actual manufacturing. The thickness of the silicon support substrate is preferably in the range of 200µm to 700µm, for example, it can be 500µm in actual manufacturing.

[0102] S2. Forming the lower electrode 131 of the piezoelectric transducer layer 13: On the upper surface of the substrate, an electrode metal layer is deposited by a process such as physical vapor deposition. The material of the electrode metal layer is preferably Mo, and its thickness is preferably in the range of 50 nm to 200 nm. Then, the electrode metal layer is patterned by photolithography and dry etching processes, so that the patterned electrode metal layer only covers the area of ​​the support arm 12, or extends to the edge functional area Z2 of the main resonator 11, thereby obtaining the lower electrode 131.

[0103] S3. Forming a piezoelectric thin film 132 of the piezoelectric transducer layer 13: A piezoelectric material layer is deposited on the lower electrode 131 by reactive magnetron sputtering. The material is preferably ScAlN and the thickness is in the range of 0.2µm to 2µm, for example, 0.5µm in actual manufacturing. The scandium content of the piezoelectric material layer is controlled by adjusting the sputtering power ratio of the Sc target and the Al target, preferably with a Sc molar fraction in the range of 10% to 43%. Then, the piezoelectric material layer is patterned by photolithography and wet etching (preferably phosphate-based etchant) or dry etching, so that the patterned piezoelectric material layer only covers the lower electrode 131.

[0104] S4. Forming the upper electrode 133 of the piezoelectric transducer layer 13: Deposit an electrode metal layer by physical vapor deposition, preferably of Mo, with a thickness in the range of 50 nm to 200 nm; then, pattern the electrode metal layer by photolithography and dry etching, so that the patterned electrode metal layer only covers the piezoelectric thin film 132 and extends to the pad area to form electrode traces for connection.

[0105] The piezoelectric transducer layer 13 of this application can be manufactured in one or a combination of the following structural forms: 1. Edge-extending piezoelectric transducer layer 13: piezoelectric film 132, upper electrode 133 and lower electrode 131 extend from support arm 12 to edge functional region Z2 of main resonator 11.

[0106] II. Electrode-extended piezoelectric transducer layer 13: The piezoelectric thin film 132 is located on the support arm 12, while the upper electrode 133 or the lower electrode 131 extends to the edge functional region Z2 of the main resonator 11. This structural design can adjust the electric field distribution or stress transmission path.

[0107] III. Piezoelectric transducer layer 13 with piezoelectric extension and electrode cut-off: The piezoelectric thin film 132 extends to the edge functional region Z2 of the main resonator 11, but the upper electrode 133 and the lower electrode 131 do not extend to the edge functional region Z2 of the main resonator 11. This structural design can reduce static capacitance C0 and feedthrough.

[0108] The areas of the piezoelectric film 132, the upper electrode 133, and the lower electrode 131 should be minimized while meeting the transduction efficiency requirements to reduce static capacitance C0 and minimize electrical feedthrough. The layout of the upper electrode 133 and the lower electrode 131 should maintain strict symmetry to ensure the balance of the differential signal and maximize the common-mode feedthrough rejection ratio.

[0109] S5. Forming the planar contours of the main resonator 11, support arm 12, and anchor point 15: The device layer from step S1 is patterned using deep reactive ion etching (preferably Bosch process), with the etching depth reaching the surface of the buried oxide layer from step S1, to form the planar contours of the main resonator 11, support arm 12, and anchor point 15. The deep reactive ion etching process should ensure the perpendicularity of the sidewalls, for example, the sidewall tilt angle should be less than 2°, to ensure the geometric accuracy of the main resonator 11. It should be noted that the acoustic reflector 16 and coupling beam 17 can also be formed simultaneously in this step.

[0110] S6. Structure Release (i.e., hydrofluoric acid etching): The buried oxide layer beneath the movable structure (e.g., main resonator 11, support arm 12, acoustic reflector 16, and coupling beam 17) is removed by gas-phase or liquid-phase hydrofluoric acid etching, causing the main resonator 11, support arm 12, acoustic reflector 16, and coupling beam 17 to suspend above the silicon substrate 10. The etching time should be precisely controlled to ensure that the buried oxide layer in the anchor point 15 area is preserved, such as... Figure 2 As shown, the buried oxide layer 18 below the anchor point 15 fixes the anchor point 15 to the silicon substrate 10. This application preferably uses a vapor-phase hydrofluoric acid etching process to avoid structural adhesion problems caused by capillary forces in the liquid-phase hydrofluoric acid etching process.

[0111] S7. Vacuum Packaging: This step is optional. To eliminate gas damping losses and improve the Q value, the device prepared through steps S1 to S6 should be packaged under vacuum conditions. Vacuum packaging can employ wafer-level bonding technology, preferably anodic bonding or metal eutectic bonding, and the gas pressure inside the packaging cavity is preferably below 1 Pa.

[0112] The above manufacturing method is based on a mature SOI MEMS process platform. The main process steps are all standard MEMS industry processes with good process compatibility. For example, the ScAlN thin film deposition process has been widely verified in FBAR / SMR products and has good batch manufacturing consistency.

[0113] The manufacturing method described above in this application may further include any combination of the following steps (1) to (6): (1) Temperature compensation layer deposition step: deposit a temperature compensation layer on the main resonator 11 or the support arm 12, for example, the material is SiO2, and use the positive temperature coefficient elastic modulus of SiO2 to compensate the negative temperature coefficient elastic modulus of single crystal silicon, thereby achieving temperature compensation.

[0114] Alternatively, the main resonator 11 and support arm 12 of the monocrystalline silicon can be heavily doped. The temperature-dependent resonant frequency of monocrystalline silicon mainly originates from the temperature dependence of silicon's elastic constant, thermal expansion, thermal stress, changes in elastic constant caused by doping, and changes in packaging stress. For heavily doped monocrystalline silicon, carrier concentration and changes in band structure caused by doping alter the elastic coefficient of silicon and its temperature dependence. Therefore, by selecting n-type or p-type doping, controlling the doping concentration, selecting the crystal orientation, and optimizing the vibration mode, the first-order TCF (First-order Temperature Coefficient of Frequency) can be reduced, even achieving near-zero first-order TCF near a specific temperature point. Specifically, the monocrystalline silicon containing the main resonator 11 or support arm 12 can be heavily doped using ion implantation, diffusion doping, plasma doping, in-situ doped epitaxy, or by using a pre-heavily doped SOI wafer, with a doping concentration of 1×10⁻⁶. 19 cm -3 Up to 1×10 21 cm -3 By adjusting the doping concentration, crystal orientation, and device geometry, the first-order TCF of the composite resonator 1 can be made to be less than ±10 ppm / ℃ within the target operating temperature range, and in the best case, less than ±1 ppm / ℃, or even close to zero near a specific temperature point.

[0115] (2) Preparation of electrostatic tuning electrode (i.e. fixed electrode 14): A fixed electrostatic tuning electrode is prepared at a position adjacent to the main resonator 11 by a separate photolithography and metal deposition step. The air gap between the electrostatic tuning electrode and the movable structure can be controlled by the thickness of the sacrificial layer.

[0116] (3) Multilayer metal interconnect: The ports (e.g., Figure 3 The electrode traces of ports 1 to 4 shown are led out to the pads. In this application, aluminum (Al) or copper (Cu) can be used to realize the interconnect metal, and the interlayer planarization can be achieved by chemical mechanical planarization (CMP) process.

[0117] (4) CMOS back-end integration: The fabrication process of this application is compatible with the standard CMOS (Complementary Metal Oxide Semiconductor) process. It can be integrated with the CMOS oscillator circuit, amplifier circuit or signal processing circuit on the same chip through back-end integration to achieve monolithic integration. Back-end integration is a prior art in this field and will not be described in detail here.

[0118] (5) Fabrication steps of the lower electrode 131 of heavily doped single-crystal silicon: Before forming the piezoelectric thin film 132 in step S3, the step of forming the lower electrode 131 of heavily doped single-crystal silicon may also be included. Specifically, this includes: defining the region of the lower electrode 131 on the device layer by photolithography; forming a heavily doped single-crystal silicon region in the region of the lower electrode 131 by ion implantation, diffusion doping, plasma doping, in-situ doped epitaxy, or by using a pre-heavily doped SOI wafer, followed by rapid thermal annealing or laser annealing to activate the dopant and reduce the wafer resistance. For example, the doping type is n-type or p-type, and the doping concentration is 1×10⁻⁶. 18 cm -3 Up to 1×10 21 cm -3 After forming the lower electrode 131 of heavily doped single-crystal silicon, a piezoelectric thin film 132 can be deposited directly, or a very thin conductive seed layer, diffusion barrier layer or adhesion layer can be deposited first, and then the piezoelectric thin film 132 can be deposited.

[0119] Furthermore, the lower electrode 131 of the heavily doped single-crystal silicon can simultaneously serve as an electrode, a local temperature-regulating resistor, and a TCF control structure. By applying a controlled current to the heavily doped single-crystal silicon region, local self-heating can be achieved, thereby enabling the composite resonator 1 to operate at a temperature point with a lower TCF or a smaller slope in the frequency-temperature curve, thus forming a temperature-controlled frequency reference source. This design allows this application to achieve not only a high Q value and low motion impedance, but also improved temperature stability through material doping.

[0120] (6) Post-frequency tuning step: The post-frequency tuning step can be performed after wafer-level release, before vacuum packaging in step S6, after vacuum packaging, or during the system-level testing phase. First, the actual resonant frequency of each composite resonator 1 is measured and compared with the target frequency; then, one or more of the following are selected based on the frequency deviation: Joule thermal tuning structure, heating resistor tuning structure, laser tuning structure, mass loading tuning structure, and plasma tuning structure, so that the composite resonator 1 can still be tuned after fabrication. This design can compensate for frequency deviations caused by photolithography size error, thin film thickness error, residual stress error, doping concentration error, and packaging stress, and improve wafer-level frequency consistency and synchronization of the MEMS resonator array composed of composite resonators 1. For arrayed low-impedance composite resonators 1, post-frequency tuning can significantly reduce the frequency mismatch between composite resonators 1 (at this time, each composite resonator 1 is called a unit), and avoid the decrease in equivalent Q value and increase in motion impedance caused by frequency dispersion in parallel arrays.

[0121] The frequency post-tuning structure may include at least one of the following (a1) to (a5): (a1) Joule heating tuning structure. A heavily doped single-crystal silicon conductive path is formed in the main resonator 11, the support arm 12, the lower electrode 131 of the heavily doped single-crystal silicon, and the acoustic reflector 16. By applying a controlled current to generate local Joule heating, local areas undergo thermal annealing, stress release, defect passivation, adjustment of the doped electroactivation state, or improvement of the interface state, thereby changing the effective stiffness, effective mass, residual stress, or boundary conditions of the composite resonator 1 and achieving resonant frequency tuning.

[0122] It should be noted that when using the Joule thermal tuning structure for tuning, the Joule heat treatment does not necessarily change the total chemical concentration of doped atoms in silicon. More precisely, it can change the equivalent elastic parameters and boundary conditions of the composite resonator 1 by altering factors such as the electroactive ratio of doped atoms, local defect states, lattice damage, residual stress, effective contribution of carrier concentration, thin film stress, and interface state, thereby achieving frequency tuning. Under higher temperature or longer heat treatment conditions, limited dopant redistribution or diffusion effects may also occur, which can serve as an auxiliary mechanism for frequency tuning.

[0123] (a2) Heating resistor adjustment structure. Local metal heating resistors or polycrystalline silicon heating resistors are set near the main resonator 11 or on the support arm 12. The specific area is subjected to short-term high-temperature treatment by electrothermal method, so that the piezoelectric film 132, metal electrodes (including upper electrode 133 and lower electrode 131), corresponding silicon structure layer or their interface undergo controlled stress redistribution, thereby achieving frequency adjustment.

[0124] (a3) Laser-tuned structure. The mass, stiffness, or residual stress of the main resonator 11 or its auxiliary structures can be altered by local laser ablation, local oxidation, local annealing, or local material modification to achieve frequency up-tuning or down-tuning. Laser tuning can be performed at the wafer level or under package window conditions.

[0125] (a4) Plasma tuning structure. By means of plasma etching, reactive ion etching, focused ion beam etching or ion implantation, the material of the main resonator 11 or the support arm 12 is locally removed, the surface stress is changed or the elastic parameters of the material are changed, so as to achieve frequency tuning.

[0126] (a5) Mass loading tuning structure. Controlled mass loading is introduced into the main resonator 11 or its associated structure by locally depositing metal, dielectric, polymer or other functional thin films to reduce the resonant frequency; or the resonant frequency is increased by subsequently removing the mass loading layer by local etching.

[0127] In one example, the edge functional region Z2 of the main resonator 11 and the supporting transducer arm composed of the supporting arm 12 and the piezoelectric transducer layer 13 can serve as a frequency post-tuning region. By altering the residual stress, equivalent mass, piezoelectric film stress state, or interface state of this region, fine-tuning of the frequency before or after packaging can be achieved. Since this tuning region is close to the supporting transducer arm or the edge of the main resonator 11, it has high adjustment efficiency for the target modal frequency while having minimal impact on the loss of the main mechanical energy storage region Z1 of the main resonator 11. The target range for frequency post-tuning can be ±10ppm to ±10000ppm, preferably ±50ppm to ±1000ppm. For a MEMS resonator array composed of a composite resonator 1, each unit can be tuned individually to reduce the frequency mismatch between units to below ±100ppm, or even below ±10ppm, thereby improving array synchronization and reducing equivalent motion impedance.

[0128] Figure 10 This is a top view of the structure of a MEMS resonator array according to an embodiment of this application. Figure 10 As shown, the MEMS resonator array 100 includes multiple composite resonators 1. Only three composite resonators 1 are illustrated in the figure, and this does not constitute a limitation on the scope of protection of this application. The composite resonators 1 can adopt the structural design of any of the foregoing embodiments, thereby enabling the MEMS resonator array 100 to produce the beneficial effects of the composite resonators 1 in the corresponding embodiments. The composite resonators 1 are electrically connected in parallel, mechanically coupled, piezoelectrically coupled, or capacitively coupled.

[0129] For a MEMS resonator array 100 consisting of multiple composite resonators 1 connected in parallel, the equivalent moving impedance satisfies the following relationship:

[0130] in, The moving impedance of the MEMS resonator array 100, i.e., the equivalent moving impedance, Let N be the moving impedance of a single composite resonator 1, and N be the number of composite resonators 1. Through array design, the equivalent moving impedance can be reduced to a practical range close to 50Ω by selecting an appropriate N (preferably 4 to 100) without changing the resonant frequency and Q value of the composite resonator 1. This application flexibly constructs a MEMS resonator array 100 using composite resonators 1, and the array size can be flexibly adjusted according to the target moving impedance requirements, exhibiting good scalability.

[0131] Figure 10The diagram illustrates the mechanical coupling between the various composite resonators 1, where adjacent composite resonators 1 are mechanically connected via a mechanical coupling beam 19. This allows for synchronous vibration of all composite resonators 1, preventing performance degradation caused by frequency mismatch between units. The length of the mechanical coupling beam 19 is preferably designed to be one-quarter of the wavelength of the sound wave λ to achieve optimal mechanical coupling efficiency.

[0132] It should be understood that the composite resonator 1 and MEMS resonator array 100 provided in the embodiments of this application are corresponding complete devices and also have the structures of known devices. Here, only the components of the device that are related to the solution of this application will be described, and other components will not be described in detail.

[0133] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. For those skilled in the art, any equivalent structural transformations made using the content of this specification and drawings are similarly included within the patent protection scope of this application.

[0134] Although this document uses terms such as "first," "second," etc., to describe various types of information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. Furthermore, the singular forms "a," "an," and "the" are intended to also include the plural forms. The terms "or" and "and / or" are interpreted as inclusive, or meaning either one or any combination thereof. Exceptions to this definition only arise when combinations of elements, functions, steps, or operations are inherently mutually exclusive in some way.

Claims

1. A piezoelectric-electrostatic composite MEMS resonator, characterized in that, include: silicon substrate; A main resonator, made of low-loss material and used to generate resonance, the main resonator includes a main mechanical energy storage area and an edge functional area disposed on the periphery of the main mechanical energy storage area, the edge functional area being integrally formed with the main mechanical energy storage area. Multiple anchor points are used to fix the silicon substrate; Multiple support arms are disposed in the same layer as the main resonator, and the multiple support arms and the main resonator are spaced apart from the silicon substrate to form a release cavity. Each support arm is made of a low-loss material, and its front end is connected to the edge functional area of ​​the main resonator and its end is connected to a corresponding anchor point. Multiple piezoelectric transducer layers are disposed on a corresponding support arm, or disposed on a corresponding support arm and extending only to the edge functional area of ​​the main resonator; each piezoelectric transducer layer includes a lower electrode, a piezoelectric film and an upper electrode sequentially disposed on the support arm, and the three form a piezoelectric transducer structure. Multiple fixed electrodes are disposed adjacent to the main resonator and an air gap is provided between them. The fixed electrodes are used to induce the main resonator to resonate when energized, so as to form an electrostatic transduction structure.

2. The piezoelectric-electrostatic composite MEMS resonator according to claim 1, characterized in that, A single-crystal silicon layer is disposed in a corresponding area on the support arm, and the single-crystal silicon layer is subjected to heavy doping treatment to form a heavily doped single-crystal silicon layer, which serves as the lower electrode. Alternatively, the support arm may be a single-crystal silicon structure layer, and the corresponding region on the support arm may be heavily doped to form a heavily doped single-crystal silicon layer, which may serve as the lower electrode.

3. The piezoelectric-electrostatic composite MEMS resonator according to claim 2, characterized in that, The lower electrode, the main resonator, and the support arm all belong to the single-crystal silicon material system.

4. The piezoelectric-electrostatic composite MEMS resonator according to claim 2, characterized in that, The piezoelectric thin film is deposited directly on the lower electrode; Alternatively, the lower electrode can be electrically connected via an intermediate layer, which is a conductive adhesion layer, a diffusion barrier layer, or a seed layer, and the thickness of the intermediate layer is less than 50 nm.

5. The piezoelectric-electrostatic composite MEMS resonator according to claim 1, characterized in that, Along the direction of the main resonator toward the corresponding anchor point, the length of the support arm is Ls; the wavelength of the sound wave in the main resonator under the target vibration mode is λ, and satisfies: Ls = λ / 4 or λ / 2.

6. The piezoelectric-electrostatic composite MEMS resonator according to claim 5, characterized in that, The target vibration mode includes at least one of the following: Radial breathing mode with n=0, elliptical mode with n=2, Lamé mode, contour mode, width extension mode, higher-order volumetric acoustic mode, and in-plane vibration mode with definite vibration nodes.

7. The piezoelectric-electrostatic composite MEMS resonator according to claim 1, characterized in that, The piezoelectric transducer layer extends above the edge functional region of the main resonator. The piezoelectric transducer layer is divided into a readout electrode region and a tuning electrode region. The readout electrode region is used to connect an oscillator or detection circuit to perform the function of the piezoelectric transducer structure. The tuning electrode region is used to connect an adjustable termination impedance network. By changing the impedance value of the adjustable termination impedance network, the dynamic effective modulus of the tuning electrode region is adjusted to change the equivalent dynamic stiffness of the support arm located in the tuning electrode region.

8. The piezoelectric-electrostatic composite MEMS resonator according to claim 1, characterized in that, The piezoelectric-electrostatic composite MEMS resonator includes four support arms and four fixed electrodes. The four support arms are arranged symmetrically in pairs, and the four fixed electrodes are arranged symmetrically in pairs around the main resonator. Each fixed electrode is disposed between two adjacent support arms.

9. The piezoelectric-electrostatic composite MEMS resonator according to claim 8, characterized in that, The upper and / or lower electrodes located on the two symmetrically arranged support arms serve as differential detection terminals, which are used to output detection signals with opposite polarities. The upper and / or lower electrodes located on the other two symmetrically arranged support arms serve as differential drive terminals, which are used to output drive signals with opposite polarities.

10. The piezoelectric-electrostatic composite MEMS resonator according to claim 1, characterized in that, The piezoelectric-electrostatic composite MEMS resonator also includes multiple acoustic reflectors and multiple coupling beams. The multiple support arms and the multiple coupling beams are spaced apart from the silicon substrate to form a release cavity. The two ends of the coupling beams are connected to the ends of the support arms and the acoustic reflectors. The acoustic reflectors are disposed between the coupling beams and the front end of the anchor point and are used to reflect the acoustic energy leaking along the support arms.

11. The piezoelectric-electrostatic composite MEMS resonator according to claim 10, characterized in that, The reflection coefficient of the acoustic reflector satisfies the following relationship: in, The acoustic impedance of the acoustic reflector is... The acoustic impedance of the support arm is given.

12. The piezoelectric-electrostatic composite MEMS resonator according to claim 1, characterized in that, The fixed electrode is also used to apply a DC bias voltage as a tuning electrode, and the resonant frequency of the main resonator is adjusted using the following spring softening effect relationship: in, The absolute change in the resonant frequency of the main resonator when the DC bias voltage is applied. This is the inherent mechanical resonant frequency of the main resonator when the DC bias voltage is not applied. The vacuum permittivity, The area of ​​the fixed electrode. The DC bias voltage is mentioned above. The width of the air gap. The equivalent mechanical stiffness of the main resonator, and the tuning range achieved by the fixed electrode are within... The range is from 0.01% to 1%.

13. A MEMS resonator array, characterized in that, It includes multiple piezoelectric-electrostatic composite MEMS resonators as described in any one of claims 1 to 12, wherein each of the piezoelectric-electrostatic composite MEMS resonators is electrically connected in parallel, mechanically coupled, piezoelectrically coupled, or capacitively coupled.

Images