A MEMS wafer integrated with a restoring force test unit and a test method

Abstract

The application provides a MEMS wafer integrated with a restoring force test unit and a test method, and relates to the technical field of MEMS wafer testing. The test unit comprises, from bottom to top, a substrate layer, a movable structure layer and a cover layer. The movable structure layer comprises a single-end fixed cantilever beam, the anchor point of the cantilever beam is fixed to the upper surface of the substrate layer, the cover layer is arranged above the cantilever beam, a bump is further arranged between the cover layer and the cantilever beam, the bump is located between the anchor point and the free end of the cantilever beam, when the cantilever beam bends downward away from the cover layer, the bending length of the cantilever beam is the length of the cantilever beam, when the cantilever beam bends upward toward the cover layer, the bump serves as a temporary mechanical support point, so that the bending length of the cantilever beam becomes the distance from the bump to the free end, and the interface adhesion force borne by the cantilever beam when the cantilever beam is in contact with the cover layer is equal to the interface adhesion force borne by the cantilever beam when the cantilever beam is contacted with the substrate layer. The application can realize restoring force tests of multiple bending lengths on the premise of occupying a small chip area.

Description

Technical Field

[0001] This invention relates to the field of MEMS wafer testing technology, and in particular to a MEMS wafer with an integrated restoring force testing unit and a testing method thereof. Background Technology

[0002] MEMS devices often contain movable structures such as cantilever beams and sensing mass blocks. During operation, these movable structures are highly susceptible to adhesion failure due to capillary forces, van der Waals forces, electrostatic forces, or hydrogen bonding. When the adhesion force exceeds the structure's own restoring force, the movable structure will be unable to return to its initial position, leading to permanent device failure. Therefore, it is necessary to effectively and accurately test and verify the restoring force and adhesion force of MEMS wafers during mass production to monitor process fluctuations, screen defective products, and guide design optimization.

[0003] Currently, the common testing method involves setting up multiple sets of single-end fixed cantilever beams and detection masses with different stiffnesses within the test unit of the wafer. During testing, a sufficiently large voltage is applied between the detection mass of the cantilever beam and the polysilicon detection electrode located on the substrate, causing them to short-circuit due to electrostatic attraction. The voltage is then removed, and the cantilever beam's ability to spring back to its initial position is determined by checking whether the insulation resistance between the detection mass and the polysilicon detection electrode recovers.

[0004] However, the existing testing schemes described above have the following significant drawbacks: First, they occupy a large chip area. To cover the different adhesion force ranges that may occur due to the complex internal structure of the device, multiple cantilever beams of different lengths or stiffnesses need to be designed within the test unit. This occupies valuable chip area and is detrimental to the miniaturization of MEMS chips and the control of mass production costs. Second, process fluctuations interfere with test accuracy. During the manufacturing process, multiple cantilever beams inevitably introduce process fluctuations such as linewidth and thickness, resulting in variable interference between test results of different beams. This makes it impossible to guarantee that the test data has a strictly single variable and comparability, thus affecting the accuracy of judging the magnitude of adhesion force and making it difficult to accurately reflect process adjustments. Summary of the Invention

[0005] This invention provides a MEMS wafer and testing method with an integrated recovery force testing unit, which can obtain multiple sets of effective recovery force and adhesion force test data with fewer test structures while reducing the chip test area, and ensure that the test results have higher accuracy and comparability.

[0006] In a first aspect, embodiments of the present invention provide a MEMS wafer with an integrated recovery force testing unit, wherein the testing unit comprises, from bottom to top, a substrate layer, a movable structure layer and a capping layer; The movable structural layer includes a cantilever beam fixed at one end, and the anchor point of the cantilever beam is fixed to the upper surface of the substrate layer. The cover layer is disposed above the cantilever beam; a protrusion is also disposed between the cover layer and the cantilever beam; the protrusion is located between the anchor point and the free end of the cantilever beam; When the cantilever beam bends downward away from the capping layer, its bending length is the length of the cantilever beam; when the cantilever beam bends upward toward the capping layer, the protrusion acts as a temporary mechanical support point, making the bending length of the cantilever beam the distance from the protrusion to the free end; the interfacial adhesion force experienced by the cantilever beam is equal when it is in contact with the capping layer and when it is in contact with the substrate layer.

[0007] In one possible implementation, the bump is fixedly disposed on the lower surface of the cover layer.

[0008] In one possible implementation, the protrusion is fixedly disposed on the upper surface of the cantilever beam.

[0009] In one possible implementation, the height of the bump is equal to the distance from the lower surface of the capping layer to the upper surface of the cantilever beam.

[0010] In one possible implementation, an upper detection electrode is further provided on the lower surface of the cover layer; the upper detection electrode is positioned above the free end of the cantilever beam. A lower detection electrode is also disposed on the upper surface of the substrate layer; the lower detection electrode is disposed below the free end of the cantilever beam; The upper detection electrode is used to apply electrostatic force to drive the cantilever beam to bend upward; The lower detection electrode is used to apply electrostatic force to drive the cantilever beam to bend downwards.

[0011] In one possible implementation, the upper detection electrode is also used to apply an electrostatic force to disengage the cantilever beam adhered to the lower detection electrode; The lower detection electrode is also used to apply electrostatic force to disengage the cantilever beam adhered to the upper detection electrode.

[0012] In one possible implementation, there are multiple cantilever beams, and the lengths of each cantilever beam are different; the distance from the protrusion of any cantilever beam to its free end is not equal to the length of other cantilever beams, nor is it equal to the distance from the protrusion of other cantilever beams to their free ends.

[0013] Secondly, embodiments of the present invention provide a testing method for MEMS wafers, applied to testing MEMS wafers with integrated restoring force testing units as described in any one of the first aspects; the method includes: The cantilever beam is controlled to bend upwards until it contacts the capping layer, and it is tested whether the cantilever beam can de-adhere to the capping layer under the restoring force corresponding to the first bending length; the first bending length is the distance from the protrusion to the free end; If the cantilever beam can detach from the capping layer, the cantilever beam is controlled to bend downwards until it contacts the substrate layer, and it is tested whether the cantilever beam can detach from the substrate layer under the restoring force corresponding to the second bending length; the second bending length is the length of the cantilever beam.

[0014] In one possible implementation, an upper detection electrode is further disposed on the lower surface of the capping layer; the upper detection electrode is disposed above the free end of the cantilever beam; a lower detection electrode is further disposed on the upper surface of the substrate layer; the lower detection electrode is disposed below the free end of the cantilever beam; the step of controlling the cantilever beam to bend upward to contact the capping layer, and testing whether the cantilever beam can de-adhere to the capping layer under the restoring force corresponding to the first bending length includes: The upper detection electrode is controlled to apply electrostatic force to drive the cantilever beam to bend upwards until it contacts the cover layer; Remove the electrostatic force applied by the upper detection electrode and test whether the cantilever beam can de-adhere to the capping layer under the restoring force corresponding to the first bending length; Accordingly, after testing whether the cantilever beam can de-adhere to the cap layer under the restoring force corresponding to the first bending length, the method further includes: If the cantilever beam cannot de-adhere to the capping layer, control the lower detection electrode to apply electrostatic force to drive the cantilever beam to de-adhere to the capping layer, and obtain the electrostatic force value at the time of de-adhesion; The magnitude of the adhesion force is determined based on the electrostatic force value during the disadhesion.

[0015] In one possible implementation, there are multiple cantilever beams, and the lengths of each cantilever beam are different; the distance from the protrusion of any cantilever beam to its free end is not equal to the length of other cantilever beams, nor is it equal to the distance from the protrusion of other cantilever beams to its free end. Before controlling the cantilever beam to bend upwards to contact the cap layer, the method further includes: The first bending length and the second bending length of each cantilever beam are arranged in ascending order to obtain the test sequence. Accordingly, after controlling the cantilever beam to bend upwards to contact the cap layer and testing whether the cantilever beam can de-adhere to the cap layer under the restoring force corresponding to the first bending length, the method further includes: The bending lengths of other cantilever beams in the screening test sequence between the first and second bending lengths of the cantilever beam; In the selected test sequence, the other cantilever beams are tested one by one in ascending order to see if they can adhere to the delaminations they are in contact with under the restoring force at the corresponding bending length.

[0016] This invention provides a MEMS wafer and testing method with an integrated restoring force testing unit. By setting up a stacked structure of a substrate layer, a movable structure layer, and a capping layer, and placing bumps between the capping layer and the cantilever beam, the cantilever beam bends downwards at its full length, while upwards the bumps act as temporary mechanical support points, reducing the bending length to the distance from the bump to the free end. Thus, the same cantilever beam can provide two different bending lengths, eliminating the need for multiple cantilever beams of different lengths to obtain two sets of corresponding restoring force test results. This reduces the number of cantilever beams while maintaining the amount of test data, helping to reduce the area of ​​the testing unit and lower chip manufacturing costs.

[0017] Meanwhile, the interfacial adhesion force experienced by the cantilever beam is equal when it contacts the capping layer and when it contacts the substrate layer. That is, the adhesion force at the two interfaces in contact with the same cantilever beam is consistent when it bends upwards and downwards. Thus, when testing the recovery force at two different bending lengths, the influencing factors of adhesion force are only related to the bending length, eliminating the interference of variables introduced by different interface states or process fluctuations between different structures (such as differences in cantilever beam linewidth and thickness). This improves the singularity of the test variables and the comparability of the test results, enabling a more accurate assessment of the impact of bending length variations on the relationship between recovery force and adhesion force.

[0018] Therefore, the MEMS wafer of this invention can achieve multiple sets of bending length recovery force tests while occupying a small chip area, and ensure the accuracy and comparability of the test results. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the structure of a restoring force testing unit provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure when bending downwards, provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure when bent upwards, provided in an embodiment of the present invention; Figure 4 This is a flowchart illustrating the implementation of a MEMS wafer testing method provided in an embodiment of the present invention. Detailed Implementation

[0020] To enable those skilled in the art to better understand this solution, the technical solutions in the embodiments of this solution will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this solution, not all of them. Based on the embodiments of this solution, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this solution.

[0021] The term "comprising" and any other variations thereof in the specification, claims, and accompanying drawings of this invention mean "including but not limited to," and are intended to cover a non-exclusive inclusion, not limited to the examples listed herein. Furthermore, the terms "first" and "second," etc., are used to distinguish different objects, not to describe a specific order.

[0022] The implementation of the present invention will be described in detail below with reference to the accompanying drawings: This invention provides two different bending lengths using the same cantilever beam, thus eliminating the need for multiple cantilever beams of different lengths to obtain two sets of corresponding recovery force test results. This allows for the acquisition of multiple sets of recovery force and adhesion force test data with fewer test structures while reducing the chip test area, and ensuring higher accuracy and comparability of the test results.

[0023] Figure 1 This is a schematic diagram of the structure of a restoring force testing unit provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure when bending downwards, provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure when bent upwards, provided in an embodiment of the present invention. (Refer to...) Figure 1 , Figure 2 , Figure 3 This invention provides a MEMS wafer with an integrated recovery force testing unit. The testing unit, from bottom to top, includes: a substrate layer 1, a movable structural layer 2, and a capping layer 3. The movable structural layer 2 includes a cantilever beam 21 fixed at one end, with its anchor point fixed to the upper surface of the substrate layer 1. The capping layer 3 is disposed above the cantilever beam 21. A protrusion 4 is also provided between the capping layer 3 and the cantilever beam 21. The protrusion 4 is located between the anchor point and the free end of the cantilever beam 21. When the cantilever beam 21 bends downwards away from the capping layer 3, its bending length is equal to the length of the cantilever beam 21. When the cantilever beam 21 bends upwards towards the capping layer 3, the protrusion 4 acts as a temporary mechanical support point, making the bending length of the cantilever beam 21 equal to the distance from the protrusion to the free end. The interfacial adhesion force experienced by the cantilever beam when in contact with the capping layer and when in contact with the substrate layer is equal.

[0024] In some embodiments, a MEMS wafer with an integrated restoring force testing unit includes, from bottom to top, a substrate layer 1, a movable structure layer 2, and a capping layer 3.

[0025] For example, the surface of a MEMS wafer can be functionally divided into a normal chip area and a test cell area. The test cell is a wafer-level process monitoring structure (PCM) and can be located in the dicing lanes or idle areas of the wafer. The test cell is used to characterize the relationship between the restoring force and interface adhesion of the movable structure of the normal chip in the online stage of wafer manufacturing.

[0026] For example, the test unit adopts a layered architecture, which includes a substrate layer 1, a movable structure layer 2 and a cover layer 3 from bottom to top. Each layer works together to realize the function of outputting multiple levels of restoring force test for a single cantilever beam 21.

[0027] The idea behind this embodiment is to achieve restoring force testing at multiple bending lengths on a single cantilever beam 21 by using switchable mechanical support points, and to control the interference of adhesion force variables on the test results by unifying the surface state of the contact interface, thereby reducing the area occupied by the test structure and improving the comparability of the test data.

[0028] The substrate layer 1 is located at the bottom of the structure, providing mechanical support for the entire test unit and serving as the contact surface for the cantilever beam 21 when it bends downwards. The anchor points of the cantilever beam 21 are fixed to the upper surface of the substrate layer 1, providing a stable support foundation for the bending deformation of the beam.

[0029] The movable structural layer 2 is arranged between the substrate layer 1 and the capping layer 3, and the core component is a cantilever beam 21 fixed at one end. From a mechanical perspective, the restoring force of the cantilever beam 21 is determined by its material properties, cross-sectional dimensions, and bending length. It should be noted that, in this application, the bending length refers to the length of the beam segment that actually participates in bending and generates elastic restoring force during bending deformation, i.e., the distance from the support point to the free end. On the same beam, the material and cross-sectional parameters are formed by the same process steps, resulting in high parameter consistency. Therefore, the difference in restoring force is mainly determined by the bending length—the shorter the bending length, the higher the corresponding structural stiffness, and the stronger the restoring force under the same deformation. In the design of the test structure, the single-end fixed cantilever beam 21 used in this embodiment of the invention has lower structural stiffness than complex support structures such as double-end fixed beams and folded beams commonly used in MEMS devices, under the same dimensions (such as the same length and cross-section). After being attracted by electrostatic force, the single-end fixed beam with lower stiffness is more prone to adhesion, thus amplifying the differences in adhesion / recovery states under different process conditions. This amplification of differences effectively improves the sensitivity of the testing system to minute changes in adhesion force, thereby increasing the test resolution and enabling more accurate determination of the anti-adhesion ability of different processes and surface treatments. For example, in order to further reduce the structural stiffness of the single-end fixed beam, in addition to increasing the length of the cantilever beam 21, the width of the main body of the cantilever beam 21 can be reduced while the width of the free end of the cantilever beam 21 is increased, i.e., a cantilever beam 21 with a flap-like structure.

[0030] The capping layer 3 is disposed above the cantilever beam 21 and is arranged opposite to the substrate layer 1, together enclosing a cavity that allows the cantilever beam 21 to move up and down. The lower surface of the capping layer 3 corresponds to the area of ​​the free end of the cantilever beam 21, serving as the contact surface when the cantilever beam 21 bends upward.

[0031] Between the capping layer 3 and the cantilever beam 21, a protrusion 4 is also provided, the position of which corresponds to the section between the anchor point and the free end of the cantilever beam 21. The core function of this protrusion 4 is to provide the cantilever beam 21 with a temporary mechanical support point that can be switched according to the direction of movement, thereby changing the bending length of the beam.

[0032] For example, when the cantilever beam 21 bends downwards and moves away from the cover layer 3, the beam body and the protrusion 4 do not contact each other, and the protrusion 4 does not exert mechanical constraints on the beam body. At this time, the cantilever beam 21 only uses the anchor point at its root as the only support point, and the entire beam body participates in bending deformation. The bending length is equal to the total length of the cantilever beam 21, corresponding to a relatively low restoring force.

[0033] For example, when the cantilever beam 21 bends upward and moves towards the capping layer 3, it will come into contact with the protrusion 4 after rising to a certain extent. At this time, the protrusion 4 will mechanically limit the corresponding position of the beam, preventing the beam segment near the anchor point from continuing to bend upward; the subsequent bending deformation will be mainly concentrated on the beam segment from the protrusion 4 to the free end of the cantilever beam 21, which is equivalent to shortening the bending length of the cantilever beam 21, corresponding to a relatively higher restoring force.

[0034] It should be noted that the protrusion 4 here is not a conventional displacement limit stop. Its core function is to act as a temporary mechanical anchor point: after the cantilever beam 21 bends upward and contacts the protrusion 4, the protrusion 4 forms a new mechanical support boundary. Only the beam segment from the protrusion 4 to the free end participates in bending deformation. The bending length is shortened and the structural stiffness is improved, so that a single cantilever beam 21 can output two levels of restoring force, reducing the number of test structures.

[0035] In terms of interface design, the interface between the cap layer and the cantilever beam, and the interface between the substrate layer and the cantilever beam, are configured such that the interfacial adhesion force experienced by the cantilever beam when in contact with the cap layer and when in contact with the substrate layer are equal. For example, the interface between the cap layer 3 and the cantilever beam 21, and the interface between the substrate layer 1 and the cantilever beam 21, employ the same surface condition. The core logic of this design is to control the consistency of the adhesion force variable.

[0036] It should be noted that the same surface state mentioned here refers to the interface between the capping layer 3 and the cantilever beam 21, and the interface between the substrate layer 1 and the cantilever beam 21, where the adhesion force between the two interfaces is consistent during the upward and downward contact processes of the cantilever beam 21. That is, the adhesion force between the capping layer 3 and the substrate layer 1 to the free end of the cantilever beam 21 is the same.

[0037] For example, the interface between the capping layer and the cantilever beam and the interface between the substrate layer and the cantilever beam have the same surface roughness and contact area.

[0038] For example, within the vertical projection range of the free end of the cantilever beam 21, the surface of the cover layer 3 and the surface of the substrate layer 1 have the same roughness, and the roughness of the upper and lower surfaces of the cantilever beam 21 is the same.

[0039] For example, the roughness and contact area of ​​the two interface contact regions are the same. The two contact surfaces employ the same preparation process and surface treatment technology, ensuring that the surface roughness parameters and micro-uneven morphology are at similar levels. The nominal contact area of ​​the contact region is consistent. In terms of structural design, the dimensions of the contact area at the free end of the cantilever beam 21 are kept uniform, and the corresponding contact platform dimensions on the capping layer 3 and the substrate layer 1 are also matched, making the nominal contact area of ​​the cantilever beam 21 comparable when it contacts upwards and downwards.

[0040] The essence of subsequent restoring force testing is to verify whether the structure's own restoring force can overcome the adhesive forces between interfaces and achieve de-adhesion. If the surface states of the upper and lower contact interfaces are different, then the adhesive forces between the interfaces will differ when the cantilever beam 21 contacts upwards and downwards. The difference in test results will simultaneously include the effects of both restoring force changes and adhesive force changes, making it difficult to accurately distinguish the role of bending length in de-adhesion capability. Setting the two contact interfaces to have the same surface state can make the adhesive force conditions of the upward and downward tests more uniform. The difference in test results will correspond more to the restoring force changes caused by bending length, which helps to improve the accuracy and comparability of the test results.

[0041] Meanwhile, since the two bending length tests are based on the same cantilever beam 21, the beam's material, thickness, line width and other parameters all come from the same process. This can also reduce the parameter differences between multiple cantilever beams 21 caused by process fluctuations, and further enhance the reference value of the test data.

[0042] It should be noted that the bending of the cantilever beam 21 can be achieved through two methods: internal electrostatic force and external inertial excitation. Internal electrostatic force driving relies on the detection electrode on the lower substrate layer 1 and the detection electrode on the upper capping layer 3: when driving upwards, a voltage is applied between the upper detection electrode and the cantilever beam 21, using electrostatic attraction to pull the beam upwards, reaching the bump 4 to form a short bending length; when driving downwards, a voltage is applied between the lower detection electrode and the cantilever beam 21, pulling the beam downwards and maintaining a full-length bending state. This method requires no external mechanical equipment and is suitable for wafer-level batch online testing. External inertial excitation applies acceleration through a centrifuge stage and impact stage, relying on the inertial force of the beam's own mass to generate deformation, which can closely approximate the stress scenario of actual device operation and verify the restoring force and adhesion characteristics under real working conditions.

[0043] In embodiments of the present invention, the movable structure may be adhered to the lower layer or the upper layer. A representative example is the sandwich structure Z-axis capacitive MEMS accelerometer, which is a typical MEMS chip structure that requires variable capacitors to be built on both the upper and lower layers.

[0044] These chips employ a vertical stacked architecture: a movable detection mass block with elastic support is located in the middle, a lower detection electrode is arranged on the substrate layer 1, and an upper detection electrode is arranged on the side of the capping layer 3. The upper and lower electrodes and the movable mass block form two sets of variable capacitors. The bidirectional variable capacitors serve two core functions: first, they form a differential detection capacitor, utilizing the reverse changes in the upper and lower capacitors as the mass block moves to output a signal, effectively improving detection sensitivity while suppressing common-mode interference such as temperature drift and process fluctuations; second, they support a closed-loop force balance operating mode, where electrostatic forces applied to the upper and lower electrodes can counteract inertial forces, significantly improving the device's measurement range and linearity.

[0045] In real-world operating conditions, when a device encounters impact overload, electrostatic discharge, or drive failure, the movable mass block may adhere to the substrate electrode downwards or to the capping electrode upwards. Both types of adhesion can lead to permanent device failure. The test unit of this invention is perfectly suited to the bidirectional adhesion failure scenario of such devices, enabling simultaneous characterization of the adhesion characteristics and structural resilience of the upper and lower interfaces at the wafer level, thus more closely reflecting the failure modes of real chips.

[0046] Based on the integrated recovery force testing unit of this embodiment, two core testing functions can be realized: qualitative determination of adhesion failure and quantitative characterization of adhesion force. The two functions are realized based on the mechanical properties and interface conditions of the same cantilever beam 21 structure, and can be adapted to different testing needs such as wafer-level process monitoring and failure analysis.

[0047] 1. Qualitative test for adhesion failure to determine whether adhesion occurs. The essence of this test is to verify whether the elastic restoring force of the cantilever beam 21 can overcome the adhesion between interfaces and achieve self-reset. During the test, the cantilever beam 21 is first bent by external drive to form contact with the corresponding interface. Then, all external driving force is removed, and it is observed whether the cantilever beam 21 can detach from the contact interface by its own elastic recovery. This determines whether adhesion failure will occur at this level of restoring force.

[0048] By incorporating the dual-bending-length design of this structure, the same cantilever beam 21 can cover two levels of adhesion determination based on the recovery force: when the cantilever beam 21 bends upwards to contact the capping layer 3, the protrusion 4 acts as a temporary support point, shortening the bending length, corresponding to a higher structural recovery force, allowing for testing the de-adhesion capability at the high recovery force level; when the cantilever beam 21 bends downwards to contact the substrate layer 1, the bending length is the full length of the beam, corresponding to a lower structural recovery force, allowing for testing the de-adhesion capability at the low recovery force level. By combining the results of the two tests, the approximate range of the interfacial adhesion force can be preliminarily determined, enabling rapid screening of adhesion risks caused by process fluctuations.

[0049] 2. Quantitative testing of adhesion force to measure the magnitude of adhesion force. When the cantilever beam 21 cannot detach itself after the drive is removed, it indicates that the restoring force corresponding to the current bending length is insufficient to overcome the interfacial adhesion force. In this case, the adhesion force can be quantitatively measured by reverse drive.

[0050] The core logic is as follows: an external driving force opposite to the adhesion direction is applied to the cantilever beam 21, and the amplitude of the driving force is gradually increased; when the resultant force of the reverse driving force and the structural restoring force can counteract the interfacial adhesion, the cantilever beam 21 will dissociate from the contact interface. By recording the driving force value at the moment of dissociation and combining it with the restoring force characteristics of the cantilever beam 21 at that bending length, the magnitude of the adhesion force at the corresponding interface can be calculated.

[0051] Since the upper and lower contact interfaces of this structure maintain a consistent surface state, the interference caused by interface properties is relatively limited when the adhesion force test is performed on the 3rd side of the capping layer and the 1st side of the substrate layer. The test results are easier to compare laterally, which can not only improve the reference value of a single set of test data, but also be used to cross-verify the uniformity of the interface adhesion characteristics.

[0052] This invention employs a stacked structure of a substrate layer 1, a movable structural layer 2, and a capping layer 3, with a bump 4 positioned between the capping layer 3 and the cantilever beam 21. This allows the cantilever beam 21 to bend downwards at its full length, while bending upwards, the bump 4 acts as a temporary mechanical support point, reducing the bending length to the distance from the bump 4 to the free end. Consequently, the same cantilever beam 21 can provide two different bending lengths, eliminating the need for multiple cantilever beams 21 of varying lengths to obtain two sets of corresponding restoring force test results. This reduces the number of cantilever beams 21 while maintaining sufficient test data, thus helping to shrink the area of ​​the test unit and lower chip manufacturing costs.

[0053] Meanwhile, the interfaces between the capping layer 3 and the cantilever beam 21 and the substrate layer 1 and the cantilever beam 21 have the same surface roughness and the same contact area, ensuring that the adhesion force between the two interfaces is consistent when the same cantilever beam 21 is bent upwards and downwards. Thus, when testing the recovery force at two different bending lengths, the influencing factors of adhesion force are only related to the bending length, eliminating the interference of variables introduced by different interface states or process variations between different structures (such as differences in the linewidth and thickness of the cantilever beam 21). This improves the singularity of the test variables and the comparability of the test results, enabling a more accurate assessment of the impact of bending length variations on the relationship between recovery force and adhesion force.

[0054] Therefore, the MEMS wafer of this invention can achieve multiple sets of bending length recovery force tests while occupying a small chip area, and ensure the accuracy and comparability of the test results.

[0055] The bump 4 is the core component for switching the bending length of the cantilever beam 21. This embodiment provides two structural configuration schemes. The mechanical principles of the two schemes are the same, but the process implementation paths and applicable scenarios have different focuses. They can be selected in combination with the overall process architecture.

[0056] In one possible implementation, the protrusion 4 is fixedly disposed on the upper surface of the cantilever beam 21.

[0057] The protrusion 4 is fixedly installed on the upper surface of the cantilever beam 21. The protrusion 4 and the cantilever beam 21 are an integral structure and move synchronously with the beam. When the cantilever beam 21 bends upward, the protrusion 4 rises with the beam and abuts against the lower surface of the cover layer 3, thereby forming a new mechanical support boundary.

[0058] There are two main approaches to achieving this process. One is the additive deposition method: after the main morphology of the movable structural layer 2 is prepared, the placement area of ​​the bumps 4 is defined using a photolithographic mask, and then structural material is deposited in this area to form bumps 4 protruding from the upper surface of the beam. The other is the area thinning method: first, a movable structural layer 2 with a thickness higher than the design value is deposited, and then, through photolithography and etching processes, the material thickness of the area corresponding to the bumps 4 is retained, while the remaining areas of the beam are etched to the target thickness, thereby forming a protruding structure on the beam.

[0059] In one possible implementation, the protrusion 4 is fixedly disposed on the lower surface of the cover layer 3.

[0060] The protrusion 4 is fixedly installed on the lower surface of the cover layer 3. The protrusion 4 is a fixed structure of the cover layer 3 and does not move with the cantilever beam 21. When the cantilever beam 21 bends upward, the upper surface of the beam contacts the protrusion 4 hanging down from the cover layer 3, thereby forming a temporary mechanical support point and shortening the bending length of the beam.

[0061] In terms of process implementation, the bump 4 can be processed simultaneously with the cavity structure of the capping layer 3: during the wafer fabrication stage of the capping layer 3, the bump structure is directly fabricated on the side of the capping layer 3 facing the cantilever beam 21 through photolithography and etching processes. This process can be integrated with the fabrication of the detection electrode and cavity forming on the capping layer 3, without needing to be inserted into the processing flow of the movable structure layer 2.

[0062] This approach generally offers advantages in terms of process compatibility and structural stability. Firstly, the protrusion 4 is machined independently of the movable structural layer 2, thus maintaining the uniformity of the cantilever beam 21's thickness and minimizing the introduction of additional residual stress, resulting in less interference with the beam's original mechanical properties. Secondly, the height of the protrusion 4 can be precisely controlled by the etching depth of the capping layer 3, making it easier to match the design requirements of the cavity height. From a process timing perspective, the machining of the capping layer 3 and the preparation of the sacrificial layer below the movable structural layer 2 belong to different process stages, and there is no process conflict between them.

[0063] In one possible implementation, the height of the bump 4 is equal to the distance from the lower surface of the cover layer 3 to the upper surface of the cantilever beam 21.

[0064] In this embodiment, the height of the protrusion 4 is designed to be equal to the initial distance from the lower surface of the cover layer 3 to the upper surface of the cantilever beam 21. This design revolves around the precise switching of the bending length, and the core is to ensure that the role of the temporary mechanical support point can be intervened in a timely manner.

[0065] From a structural perspective, taking the scheme where protrusion 4 is located on the lower surface of the cover layer 3 as an example, protrusion 4 protrudes from the lower surface of the cover layer 3 toward the cantilever beam 21. The vertical height of its protrusion corresponds exactly to the vertical gap between the upper surface of the cantilever beam 21 and the lower surface of the cover layer 3 when the cantilever beam 21 is in its initial horizontal state. At this time, in its initial position, the upper surface of the cantilever beam 21 is in contact with or very close to the end face of protrusion 4.

[0066] From a mechanical perspective, when the cantilever beam 21 begins to bend under an upward driving force, the protrusion 4 immediately applies a mechanical constraint to the corresponding position of the beam, restricting the beam from the anchor point to the protrusion 4 section from which it continues to bend upward. This concentrates the deformation directly on the beam section from the protrusion 4 to the free end. This design avoids the idle stroke phase where the cantilever beam 21 bends upward a certain distance before contacting the protrusion 4, allowing the bending length to switch to a short length position at the initial stage of upward deformation. This ensures that the structural stiffness and restoring force characteristics corresponding to this position are closer to the design values.

[0067] If bump 4 is placed on the upper surface of cantilever beam 21, the underlying logic of this height design remains the same: the height of bump 4 protruding upwards from the beam body exactly fills the initial gap, thus achieving immediate support. For wafer-level testing, this height-matching design can reduce test deviations introduced by transition deformation, make the distinction between the two restoring forces clearer, and help improve the consistency and reference value of test results.

[0068] In one possible implementation, an upper detection electrode 31 is further disposed on the lower surface of the cover layer 3; the upper detection electrode 31 is disposed above the free end of the cantilever beam 21; a lower detection electrode 11 is further disposed on the upper surface of the substrate layer 1; the lower detection electrode 11 is disposed below the free end of the cantilever beam 21; the upper detection electrode 31 is used to apply electrostatic force to drive the cantilever beam 21 to bend upward; the lower detection electrode 11 is used to apply electrostatic force to drive the cantilever beam 21 to bend downward.

[0069] In this embodiment, the test unit is configured with a pair of detection electrodes, and the cantilever beam 21 is driven to bend bidirectionally by electrostatic force. This is compatible with the test architecture with double bending length and double contact surface, and supports wafer-level automated testing process.

[0070] The upper detection electrode 31 is arranged on the lower surface of the cover layer 3, corresponding to the area above the free end of the cantilever beam 21; the lower detection electrode 11 is arranged on the upper surface of the substrate layer 1, corresponding to the area below the free end of the cantilever beam 21. The electrodes are placed at the corresponding positions at the free end mainly because the displacement of the free end is the largest when the cantilever beam 21 bends. The electrostatic interaction between the electrode and the beam can obtain a longer lever arm, resulting in relatively higher driving efficiency and making it easier to accurately control the contact state between the beam and the contact surface.

[0071] The core function of the upper detection electrode 31 is to provide an upward electrostatic driving force. When a driving voltage is applied between the electrode and the cantilever beam 21, an electrostatic attraction is generated between the variable capacitors formed by the two, pulling the free end of the cantilever beam 21 to bend towards the capping layer 3. As the beam deforms upward, the protrusion 4 gradually intervenes to form a temporary support point, allowing the cantilever beam 21 to enter a short bending length working state. Finally, the free end of the beam contacts the contact interface on the side of the capping layer 3, which can be used for deadhesion testing at the corresponding restoring force level.

[0072] The core function of the lower detection electrode 11 is to provide a downward electrostatic driving force. When a driving voltage is applied between the electrode and the cantilever beam 21, the electrostatic attraction pulls the free end of the cantilever beam 21 towards the substrate layer 1. During this process, the cantilever beam 21 disengages from the bump 4, and the beam is supported only by the root anchor point. The entire beam participates in bending deformation, maintaining the working state of the full bending length. Finally, the free end of the beam contacts the contact interface of the substrate layer 1, which can be used for deadhesion testing at the corresponding restoring force level.

[0073] This symmetrical electrode configuration allows the same test unit to independently perform driving and testing in both upward and downward directions, eliminating the need for external inertial loading equipment and making the testing process easier to automate at the wafer level. The symmetrical electrode stack-up layout also saves chip area.

[0074] It should be noted that the essence of electrostatic force driving is the electrostatic attraction effect of parallel-plate capacitors. The detection electrode and the cantilever beam 21 form a pair of capacitor plates. When a voltage is applied between them, a potential difference is created, generating a mutual attractive electrostatic force that pulls the cantilever beam 21 to bend and deform towards the electrode. When the cantilever beam 21 comes into contact with the corresponding electrode, they become conductive and reach the same potential, the potential difference between the plates disappears, and the electrostatic force dissipates. At this point, the bonding effect between the cantilever beam 21 and the contact surface is provided only by the interfacial adhesion force. Therefore, under conditions without electrostatic interference, the ability of the structural restoring force to overcome the adhesion effect can be tested.

[0075] For example, the cantilever beam 21 is covered by polycrystalline silicon layers both above and below it. The polycrystalline silicon layers can be modified in situ by local ion implantation to form conductive detection electrodes in a designated area, which also serve as the contact interface with the cantilever beam 21.

[0076] It should be noted that the core design logic for ensuring consistent contact area is large-area contact with small-area: the contact area of ​​the capping layer 3 and the substrate layer 1 is designed to be larger than the contact area of ​​the free end of the cantilever beam 21, or vice versa. In this way, even if there is a lateral alignment deviation in the process, the actual contact area is always determined by the side with the smaller area, and will not change significantly due to alignment offset.

[0077] In this structure, the upper and lower contact surfaces follow the same design rules, and the contact area of ​​the same cantilever beam 21 when it contacts upward and downward can be kept consistent; multiple cantilever beams 21 can also be connected by unifying the contact size of the free ends to further ensure the uniformity of the contact area and reduce the interference of process deviations on the adhesion test.

[0078] For example, the cantilever beam 21 adopts a narrow beam body with a wide free end in a paddle-like structure. The free end detection mass is rectangular, and its width is greater than the width of the main body of the cantilever beam 21. This design can increase the facing area between the free end and the upper and lower detection electrodes 11, improve the efficiency of electrostatic drive, and expand the dimensional redundancy of the contact area, reducing the impact of process alignment deviation on the contact area. On the other hand, the narrower beam body makes it easier to adjust the structural stiffness within a limited length, making the design range of the restoring force more flexible and adaptable to the needs of multi-level adhesion testing.

[0079] In one possible implementation, the upper detection electrode 31 is further configured to apply electrostatic force to disengage the cantilever beam 21 adhered to by the lower detection electrode 11; the lower detection electrode 11 is further configured to apply electrostatic force to disengage the cantilever beam 21 adhered to by the upper detection electrode 31.

[0080] In this embodiment, the upper detection electrode 31 and the lower detection electrode 11 adopt a functional reuse design. In addition to driving the cantilever beam 21 to bend to the corresponding side, they can also apply a reverse dissociation force to the adhesion state of the other side interface.

[0081] When the cantilever beam 21 adheres to the interface with the substrate layer 1 and its own restoring force is insufficient to achieve autonomous reset, a driving voltage can be applied to the upper detection electrode 31. The electrostatic attraction formed between the upper detection electrode 31 and the cantilever beam 21 will provide an additional upward force to the beam, which, together with the restoring force of the cantilever beam 21 itself, helps to overcome the interface adhesion force on the substrate side, allowing the beam to detach from the adhered state.

[0082] Correspondingly, when the cantilever beam 21 adheres to the interface of the cover layer 3 and cannot reset itself, a driving voltage can be applied to the downward detection electrode 11 to provide a downward auxiliary force to the beam by means of electrostatic attraction, which, together with the structural restoring force, realizes the separation of the interface of the cover layer 3.

[0083] The purpose of this design is to reuse existing electrode structures without adding additional process layers and functional units. On the one hand, it can reduce the probability of permanent adhesion after a single test, making the test structure reusable; on the other hand, by gradually adjusting the amplitude of the reverse driving voltage and recording the voltage parameters at the moment of interface dissociation, the quantitative value of the interface adhesion force can be further calculated, expanding the qualitative determination of adhesion failure to a quantitative characterization of adhesion force.

[0084] In one possible implementation, there are multiple cantilever beams 21, and the lengths of each cantilever beam 21 are different; the distance from the protrusion 4 of any cantilever beam 21 to its free end is not equal to the length of other cantilever beams 21, nor is it equal to the distance from the protrusion 4 of other cantilever beams 21 to its free end.

[0085] In this embodiment, multiple cantilever beams 21 can be set up to form a test array. By arranging the cantilever beams 21 with different lengths and protrusions 4, the test range of the restoring force can be expanded within a limited layout area, covering a wider range of adhesion forces.

[0086] The total length of each cantilever beam 21 is different, and the position of the protrusion 4 follows a unified design rule: the distance from the protrusion 4 of any cantilever beam 21 to the free end is neither equal to the total length of the other cantilever beams 21 nor equal to the distance from the protrusion 4 of the other cantilever beams 21 to the free end.

[0087] The core logic of this design is to increase the density of bending length ranges. A single cantilever beam 21 can output two bending length ranges: the total length and the length from the bump 4 to the free end, corresponding to two levels of restoring force. Through the above arrangement rules, the bending lengths output by all cantilever beams 21 are different. N cantilever beams 21 can form 2N differentiated restoring force test ranges, reducing the area loss caused by range redundancy. At the same time, the bending lengths of each range can be sorted by size to form a continuous test sequence, which facilitates the gradual positioning of the numerical range of interface adhesion force, improving the range coverage and resolution accuracy of wafer-level testing.

[0088] Figure 4 This is a flowchart illustrating the implementation of a MEMS wafer testing method provided in an embodiment of the present invention. (Refer to...) Figure 4 This invention provides a testing method for MEMS wafers, applied to testing MEMS wafers with integrated restoring force testing units as described in any of the possible implementations above; the method includes: Step 201: Control the cantilever beam 21 to bend upwards until it contacts the cover layer 3, and test whether the cantilever beam 21 can detach from the cover layer 3 under the restoring force corresponding to the first bending length; the first bending length is the distance from the protrusion 4 to the free end; It should be noted that de-adhesion means separating the originally stuck-together contact surfaces and removing the adhesive state.

[0089] When controlling the cantilever beam 21 to bend upwards, an upward external force is first applied to the free end of the cantilever beam 21, for example, by applying electrostatic force through the upper detection electrode 31, which pulls the beam body to deform in the direction of the cover layer 3. During the bending process, the corresponding position of the beam body contacts the protrusion 4, and the protrusion 4 forms a temporary mechanical support point. After that, the bending deformation is mainly concentrated in the beam segment from the protrusion 4 to the free end, and the bending length at this time is the distance from the protrusion 4 to the free end.

[0090] The force is applied continuously until the free end of the cantilever beam 21 contacts the capping layer 3. After the external force is removed, the cantilever beam 21 generates a restoring force based on its own elasticity: if the restoring force can overcome the interfacial adhesion force, the beam separates from the capping layer 3; if the restoring force is insufficient, the beam remains in an adhered state, thereby verifying the de-adhesion capability at this setting.

[0091] Step 202: If the cantilever beam 21 can detach from the capping layer 3, control the cantilever beam 21 to bend downwards until it contacts the substrate layer 1, and test whether the cantilever beam 21 can detach from the substrate layer 1 under the restoring force corresponding to the second bending length; the second bending length is the length of the cantilever beam 21.

[0092] In this embodiment, the testing process is executed in the order of first bending upwards to contact, and then bending downwards to contact. That is, the disadhesion test corresponding to the first bending length is completed first, and then the test for the second bending length is determined based on the test results. This sequential design is based on the difference in the strength of the two levels of restoring force. The core is to improve testing efficiency through progressive judgment, while realizing the gradual positioning of the adhesion force range.

[0093] The first bending length is the distance from the protrusion 4 to the free end, corresponding to a shorter beam segment, higher structural stiffness, and relatively stronger restoring force; the second bending length is the total length of the cantilever beam 21, corresponding to a longer beam segment and relatively weaker restoring force.

[0094] The test begins at the setting with the stronger restoring force. If the cantilever beam 21 cannot detach from the cap layer 3 at this setting, it indicates that the interfacial adhesion force is higher than the restoring force at that setting. For the second setting with the weaker restoring force, self-detachment is usually difficult to achieve either, so the test of the current cantilever beam 21 can be terminated directly to avoid redundant testing. If the cantilever beam 21 can successfully detach at the strong restoring force setting, it indicates that the interfacial adhesion force is lower than the restoring force at that setting. In this case, continuing the test at the weaker restoring force setting can further narrow the range of adhesion force judgment, determining whether it falls between the two restoring force levels.

[0095] Since the surface conditions of the upper and lower contact interfaces are kept consistent, the adhesion benchmarks of the two tests tend to be unified, and the progressive test results are comparable. More specific adhesion range information can be obtained within a limited number of test steps, which is suitable for the efficiency requirements of wafer-level batch testing.

[0096] Regarding how to utilize the test results in subsequent steps: The main test object of this test unit is not the absolute value of the restoring force of cantilever beam 21, but rather the relative magnitude relationship between the restoring force and the interfacial adhesion force. The restoring force of cantilever beam 21 can be calculated through material properties, cross-sectional dimensions, and bending length, and is a structural parameter that can be pre-designed and quantified; however, the interfacial adhesion force is affected by the surface condition of the process and is difficult to calculate directly and accurately, so it needs to be characterized through testing.

[0097] By conducting bending tests in both upward and downward directions, the same cantilever beam 21 can output two levels of restoring force corresponding to different bending lengths. When bending upward, the protrusion 4 forms a temporary support point, resulting in a shorter bending length and a higher restoring force level; when bending downward, the beam participates in deformation along its entire length, corresponding to a lower restoring force level. Since the surface conditions of the upper and lower contact interfaces remain consistent, the adhesion force benchmarks for the two tests tend to be unified, providing a basis for lateral comparison of the test results.

[0098] Based on the deadhesion results of the two tests, the range of interfacial adhesion force can be determined. If the cantilever beam 21 still cannot deadhere at the high recovery force setting, it indicates that the interfacial adhesion force is higher than the recovery force of that setting; if deadhesion is possible at the high recovery force setting but not at the low recovery force setting, it indicates that the interfacial adhesion force is between the two recovery force settings; if deadhesion is possible at the low recovery force setting, it indicates that the interfacial adhesion force is lower than the recovery force of that setting. Therefore, the numerical range of adhesion force can be quickly determined through a limited number of testing steps.

[0099] The core value of this testing method lies in utilizing the structural characteristics of bidirectional bending to obtain multiple sets of relative size relationships with fewer test structures. It can be used for wafer-level process qualification screening to determine whether the adhesion force is within the design allowable range. It can also be combined with multiple cantilever beams 21 of different lengths to form a multi-level test sequence, further narrowing the determination range of adhesion force and adapting to different precision testing requirements.

[0100] In one possible implementation, an upper detection electrode 31 is further disposed on the lower surface of the capping layer 3; the upper detection electrode 31 is disposed above the free end of the cantilever beam 21; a lower detection electrode 11 is further disposed on the upper surface of the substrate layer 1; the lower detection electrode 11 is disposed below the free end of the cantilever beam 21; the step of controlling the cantilever beam 21 to bend upward to contact the capping layer 3 and testing whether the cantilever beam 21 can detach from the capping layer 3 under the restoring force corresponding to the first bending length includes: controlling the upper detection electrode 31 to apply electrostatic force to drive the cantilever beam 21 to bend upward to contact the capping layer 3; removing the electrostatic force applied by the upper detection electrode 31 and testing whether the cantilever beam 21 can detach from the capping layer 3 under the restoring force corresponding to the first bending length; Accordingly, after testing whether the cantilever beam 21 can detach from the cover layer 3 under the restoring force corresponding to the first bending length, the method further includes: if the cantilever beam 21 cannot detach from the cover layer 3, controlling the lower detection electrode 11 to apply electrostatic force to drive the cantilever beam 21 to detach from the cover layer 3, and obtaining the electrostatic force value at the time of detachment; and determining the magnitude of the adhesion force based on the electrostatic force value at the time of detachment.

[0101] This embodiment relies on paired detection electrodes to realize a test process for qualitative deadhesion determination and quantitative adhesion force characterization within the same test unit. The test process revolves around the recovery force level corresponding to the first bending length and can be logically divided into two stages.

[0102] The first stage is a qualitative test of deadhesion in the upward direction. During the test, a driving voltage is first applied to the upward detection electrode 31, forming an electrostatic attraction between the electrode and the cantilever beam 21, pulling the free end of the cantilever beam 21 to bend and deform towards the capping layer 3. During the bending process, the corresponding position of the beam contacts the protrusion 4 and forms a temporary mechanical support point, switching the bending length to the short length range from the protrusion 4 to the free end. The driving voltage is continuously applied until the free end of the cantilever beam 21 is in full contact with the interface of the capping layer 3; at this time, the beam and the electrode are connected to form an equipotential, and the electrostatic driving force dissipates, leaving only adhesion between the interfaces. Then the driving voltage of the upper detection electrode 31 is removed to determine whether the cantilever beam 21 can detach from the contact interface by its own restoring force, thus determining whether adhesion failure will occur at this restoring force level.

[0103] The second stage is a quantitative test of adhesion force under the adhesive state. If the cantilever beam 21 cannot detach itself after the drive is removed, it indicates that the restoring force of the current gear is insufficient to overcome the interfacial adhesion force. In this case, reverse dissociation can be implemented through the lower detection electrode 11. The specific process is as follows: a driving voltage is gradually applied to the lower detection electrode 11, causing a downward electrostatic attraction between the electrode and the cantilever beam 21. This force is superimposed in the same direction as the elastic restoring force of the cantilever beam 21, jointly resisting the interfacial adhesion effect. As the driving voltage gradually increases, the net external force gradually increases. When the net external force is sufficient to overcome the interfacial adhesion force, the cantilever beam 21 and the capping layer 3 dissociate. The electrostatic force value corresponding to the moment of dissociation is recorded. Combined with the restoring force characteristics of the cantilever beam 21 at this bending length, the corresponding interfacial adhesion force value can be calculated.

[0104] The embodiments of the present invention reuse the existing bidirectional electrode structure without adding additional functional units. It can quickly screen for adhesion failure risks through qualitative testing and quantitatively characterize adhesion force through reverse dissociation, adapting to the testing requirements of different precision in wafer-level batch testing.

[0105] Furthermore, when the cantilever beam 21 cannot detach itself from the contact interface, a reverse electrostatic force is applied to dissociate it, and the value of the applied electrostatic force at the moment of dissociation is recorded. Since the structural parameters of the cantilever beam 21 (such as length, width, thickness, and material properties) are known, its own structural restoring force can be pre-calculated using mechanical formulas. According to the mechanical equilibrium relationship, the interface adhesion force is equal to the sum of the reverse electrostatic force and the self-restoring force of the cantilever beam 21 at that bending length, that is: adhesion force = reverse electrostatic force + structural restoring force.

[0106] After obtaining the adhesion force, the adhesion force per unit area can be calculated by combining the actual contact area between the free end of the cantilever beam 21 and the corresponding layer (capping layer 3 or substrate layer 1) (adhesion force per unit area = adhesion force ÷ contact area). This adhesion force per unit area eliminates the interference of differences in structural dimensions and contact area, and can objectively characterize the surface adhesion characteristics of silicon materials under current processing technology. Through this parameter, a quantitative relationship between process conditions and adhesion characteristics can be established. On the one hand, it can be used to provide feedback to guide the structural design of MEMS devices, such as designing an elastic beam with sufficient restoring force to resist the adhesion force; on the other hand, it can be used to verify the actual adjustment effect of anti-adhesion processes (such as surface hydrophobic treatment and roughening treatment), and realize closed-loop optimization of process and design.

[0107] In one possible implementation, after testing whether the cantilever beam 21 can detach from the capping layer 3 under the restoring force corresponding to the first bending length, the method further includes: if the cantilever beam 21 can detach from the capping layer 3, controlling the lower detection electrode 11 to apply an electrostatic force to drive the cantilever beam 21 to bend downwards until it contacts the substrate layer 1; removing the electrostatic force applied by the lower detection electrode 11, and testing whether the cantilever beam 21 can detach from the substrate layer 1 under the restoring force corresponding to the second bending length.

[0108] Accordingly, after testing whether the cantilever beam 21 can detach from the substrate layer 1 under the restoring force corresponding to the second bending length, the method further includes: if the cantilever beam 21 cannot detach from the substrate layer 1, controlling the upper detection electrode 31 to apply electrostatic force to drive the cantilever beam 21 to detach from the substrate layer 1, and obtaining the electrostatic force value at the time of detachment; and determining the magnitude of the adhesion force based on the electrostatic force value at the time of detachment.

[0109] The testing method for multiple cantilever beams 21 can be flexibly selected according to testing requirements. It is not necessary to complete all gear tests one by one before switching to the next beam. The core objective is to efficiently locate the critical range of adhesion force.

[0110] One execution method is to test each cantilever beam sequentially: Test each cantilever beam 21 in both upward and downward directions, obtaining the deadhesion results for the corresponding two bending lengths, before moving on to the next cantilever beam 21. The advantage of this method is that the two test results for a single beam are closely correlated, facilitating rapid confirmation of the adhesion force range covered by the single beam, and the process flow is relatively intuitive.

[0111] Another approach is to test sequentially by overall length: All bending lengths of all cantilever beams 21 are arranged in ascending order, starting with the shortest length corresponding to the highest restoring force, and then proceeding to the longer lengths and lower restoring force levels. This eliminates the need to bind the two tests for a single beam together. This method is better suited for quickly locating critical levels.

[0112] From a testing logic perspective, shorter bending lengths result in stronger structural resilience and easier deadhesion; conversely, longer bending lengths lead to weaker resilience and easier adhesion. Therefore, the testing process from short to long lengths corresponds to a progression from strong to weak resilience. When the critical threshold where deadhesion becomes non-deadhesion first appears during the test, it can be determined that the interfacial adhesion force is between the resilience force corresponding to that critical threshold and the previous threshold. The testing objective has been largely achieved, and further testing at longer thresholds can be discontinued, helping to reduce the overall time for wafer-level batch testing.

[0113] In one possible implementation, there are multiple cantilever beams 21, and the lengths of each cantilever beam 21 are different; the distance from the protrusion 4 of any cantilever beam 21 to its free end is not equal to the length of other cantilever beams 21, nor is it equal to the distance from the protrusion 4 of other cantilever beams 21 to its free end; before controlling the cantilever beam 21 to bend upward to contact the cover layer 3, the method further includes: arranging the first bending length and the second bending length of each cantilever beam 21 in ascending order to obtain a test sequence; Accordingly, after controlling the cantilever beam 21 to bend upwards to contact the capping layer 3 and testing whether the cantilever beam 21 can de-adhere to the capping layer 3 under the restoring force corresponding to the first bending length, the method further includes: screening the bending lengths of other cantilever beams 21 between the first and second bending lengths of the cantilever beam 21 in the test sequence; and in the screened test sequence, testing in ascending order whether the other cantilever beams 21 can de-adhere to the contacted layer under the restoring force corresponding to the bending length.

[0114] This embodiment uses multiple cantilever beams 21 of varying lengths to construct a multi-level test array. Through a progressive test logic of unified sorting and interval filtering, it achieves gradual and accurate positioning of adhesion force within a wide range, balancing test efficiency and resolution accuracy.

[0115] During the test preparation phase, the first and second bending lengths of all cantilever beams 21 are summarized and arranged in ascending order of length value to form a continuous test sequence. Based on the design rules, all bending lengths are unique, and each range corresponds to a unique structural stiffness and restoring force level; the shorter the length, the higher the structural stiffness and the stronger the corresponding restoring force, thus forming a gradient distribution of the restoring force test range.

[0116] During formal testing, the de-adhesion test was first performed on a single cantilever beam 21 at two bending lengths to obtain the adhesion determination results for the short and long bending lengths. Subsequently, in the complete test sequence, the remaining bending lengths between the first and second bending lengths of the beam were selected, and the de-adhesion test was carried out one by one in ascending order for the corresponding bending lengths. By gradually approximating the bending lengths, the numerical range of the adhesion force was narrowed, and higher positioning accuracy could be obtained without traversing all bending lengths.

[0117] If the initial cantilever beam 21 selected has a first bending length that is the shortest value in the entire test sequence, corresponding to the highest restoring force setting across the entire range, then the test can start from the highest restoring force setting. If the cantilever beam 21 cannot de-adhere at this shortest setting, it can be directly determined that the interface adhesion force is higher than the maximum restoring force across the entire range; if the beam can still de-adhere at the second bending length setting, then continue testing along the ascending direction of the test sequence, testing subsequent settings with larger lengths in sequence, until the critical setting between adhesion and de-adhesion is located, thus completing the determination of the adhesion force range across the entire range.

[0118] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A MEMS wafer with an integrated restoring force testing unit, characterized in that, The test unit, from bottom to top, includes: a substrate layer, a movable structure layer, and a capping layer; The movable structural layer includes a cantilever beam fixed at one end, and the anchor point of the cantilever beam is fixed to the upper surface of the substrate layer. The cover layer is disposed above the cantilever beam; a protrusion is also disposed between the cover layer and the cantilever beam; the protrusion is located between the anchor point and the free end of the cantilever beam; When the cantilever beam bends downward away from the capping layer, its bending length is the length of the cantilever beam; when the cantilever beam bends upward toward the capping layer, the protrusion acts as a temporary mechanical support point, making the bending length of the cantilever beam the distance from the protrusion to the free end; the interfacial adhesion force experienced by the cantilever beam is equal when it is in contact with the capping layer and when it is in contact with the substrate layer.

2. The MEMS wafer with an integrated restoring force testing unit as described in claim 1, characterized in that, The protrusion is fixedly disposed on the lower surface of the cover layer.

3. The MEMS wafer with an integrated restoring force testing unit as described in claim 1, characterized in that, The protrusion is fixedly mounted on the upper surface of the cantilever beam.

4. The MEMS wafer with an integrated restoring force testing unit as described in claim 1, characterized in that, The height of the protrusion is equal to the distance from the lower surface of the capping layer to the upper surface of the cantilever beam.

5. The MEMS wafer with an integrated restoring force testing unit as described in claim 1, characterized in that, The lower surface of the cover layer is also provided with an upper detection electrode; the upper detection electrode is located above the free end of the cantilever beam; A lower detection electrode is also disposed on the upper surface of the substrate layer; the lower detection electrode is disposed below the free end of the cantilever beam; The upper detection electrode is used to apply electrostatic force to drive the cantilever beam to bend upward; The lower detection electrode is used to apply electrostatic force to drive the cantilever beam to bend downwards.

6. The MEMS wafer with an integrated restoring force testing unit as described in claim 5, characterized in that, The upper detection electrode is also used to apply electrostatic force to dissociate the cantilever beam adhered to the lower detection electrode; The lower detection electrode is also used to apply electrostatic force to disengage the cantilever beam adhered to the upper detection electrode.

7. The MEMS wafer with an integrated restoring force testing unit as described in claim 1, characterized in that, There are multiple cantilever beams, and the lengths of each cantilever beam are different; The distance from the protrusion of any cantilever beam to its free end is not equal to the length of any other cantilever beam, nor is it equal to the distance from the protrusion of any other cantilever beam to its free end.

8. A testing method for MEMS wafers, characterized in that, The method is applied to testing a MEMS wafer with an integrated restoring force testing unit as described in any one of claims 1 to 7; the method includes: The cantilever beam is controlled to bend upwards until it contacts the capping layer, and it is tested whether the cantilever beam can de-adhere to the capping layer under the restoring force corresponding to the first bending length; the first bending length is the distance from the protrusion to the free end; If the cantilever beam can detach from the capping layer, the cantilever beam is controlled to bend downwards until it contacts the substrate layer, and it is tested whether the cantilever beam can detach from the substrate layer under the restoring force corresponding to the second bending length; the second bending length is the length of the cantilever beam.

9. The testing method for MEMS wafers as described in claim 8, characterized in that, The lower surface of the capping layer is further provided with an upper detection electrode; the upper detection electrode is disposed above the free end of the cantilever beam; the upper surface of the substrate layer is further provided with a lower detection electrode; the lower detection electrode is disposed below the free end of the cantilever beam; the step of controlling the cantilever beam to bend upward to contact the capping layer, and testing whether the cantilever beam can de-adhere to the capping layer under the restoring force corresponding to the first bending length includes: The upper detection electrode is controlled to apply electrostatic force to drive the cantilever beam to bend upwards until it contacts the cover layer; Remove the electrostatic force applied by the upper detection electrode and test whether the cantilever beam can de-adhere to the capping layer under the restoring force corresponding to the first bending length; Accordingly, after testing whether the cantilever beam can de-adhere to the cap layer under the restoring force corresponding to the first bending length, the method further includes: If the cantilever beam cannot de-adhere to the capping layer, control the lower detection electrode to apply electrostatic force to drive the cantilever beam to de-adhere to the capping layer, and obtain the electrostatic force value at the time of de-adhesion; The magnitude of the adhesion force is determined based on the electrostatic force value during the disadhesion.

10. The testing method for MEMS wafers as described in claim 8, characterized in that, There are multiple cantilever beams, and the lengths of each cantilever beam are different; the distance from the protrusion of any cantilever beam to its free end is not equal to the length of other cantilever beams, nor is it equal to the distance from the protrusion of other cantilever beams to its free end. Before controlling the cantilever beam to bend upwards to contact the cap layer, the method further includes: The first bending length and the second bending length of each cantilever beam are arranged in ascending order to obtain the test sequence. Accordingly, after controlling the cantilever beam to bend upwards to contact the cap layer and testing whether the cantilever beam can de-adhere to the cap layer under the restoring force corresponding to the first bending length, the method further includes: The bending lengths of other cantilever beams in the screening test sequence between the first and second bending lengths of the cantilever beam; In the selected test sequence, the other cantilever beams are tested one by one in ascending order to see if they can adhere to the delaminations they are in contact with under the restoring force at the corresponding bending length.

Images