Wafer detection probe variable speed impact contact method and related apparatus

Abstract

The application relates to a wafer detection probe variable-speed impact contact method and a related device, which comprises the following steps: controlling a probe to slowly approach a wafer pad at a first speed to a pause position; stopping at the pause position to eliminate residual oscillation, so that the probe enters an impact state; driving the probe to impact the wafer pad at a second speed greater than the first speed, so that the needle tip penetrates an oxide layer on the surface of the wafer pad and contacts an underlying metal layer with a movement momentum before the displacement reaches a static overdrive depth; and maintaining the contact pressure after the penetration. The application penetrates the oxide layer with the movement momentum instead of increasing the static overdrive depth, thereby stably establishing a low contact resistance while reducing wafer pad damage and needle tip aluminum accumulation, and having the advantages of considering contact reliability, detection beat and probe service life.

Description

Technical Field

[0001] This application relates to the field of semiconductor wafer testing, and in particular to a method and apparatus for variable speed impact contact of a wafer inspection probe. Background Technology

[0002] Wafer testing establishes an electrical connection between probes on a probe card and wafer pads to measure the chip's electrical parameters and filter out defective products. The pads are mostly aluminum pads, whose surfaces develop an oxide layer in the air. This oxide layer has high resistance, and the probes can only obtain stable, low contact resistance by penetrating the oxide layer and contacting the metal layer beneath it.

[0003] Existing probe contact methods press the probe against the pad at a constant low speed, relying on continuously increasing the probe's downward displacement relative to the pad, i.e., the static overdrive depth, to allow the probe tip to scrape and penetrate the oxide layer and then penetrate the metal layer. As the number of test batches accumulates, aluminum will adhere to the probe tip due to repeated scraping of the pad. After the probe tip is covered with aluminum, the contact resistance increases and the contact reliability decreases, requiring periodic shutdowns to clean the probe tip. Furthermore, the downward displacement applied to penetrate the oxide layer will leave deep probe marks on the pad.

[0004] Existing methods link the goal of penetrating the oxide layer with the means of increasing the static overdrive depth, thus creating a mutually restrictive contradiction. Because penetrating the oxide layer depends on continuous downward static overdrive, the greater the downward pressure, the more severe the mechanical damage to the pads and the smaller the available overdrive margin. Furthermore, downward scraping causes aluminum to accumulate on the probe tip, which increases the contact resistance with the number of tests. To control the contact resistance, the overdrive must be further increased or periodic probe cleaning must be scheduled. Increasing the overdrive will approach the damage limit of the pads more quickly, while periodic probe cleaning will both reduce the testing cycle and shorten the probe life due to tip abrasion. The probe array composed of multiple probes is also affected by the distribution of probe tip height and the distribution of pad flatness. Under the same nominal overdrive, the actual contact state of each probe is not consistent, further compressing the available overdrive window. To address the aforementioned contradictions, methods such as regularly cleaning the probe tip, monitoring contact resistance trends, and applying release material to the probe tip only alleviate one aspect of the problem. Cleaning the probe tip removes aluminum buildup but comes at the cost of downtime and probe grinding. Trend monitoring and degradation curves can only provide early warnings without changing the contact mechanism. Releasing material only provides remediation at the probe tip attachment level. None of these methods address the fundamental issue that the film breaking method relies on static overdrive. Therefore, it is difficult to stably establish low contact resistance without increasing pad damage, sacrificing detection cycle time, or shortening probe life. Summary of the Invention

[0005] In order to enable the probe to penetrate the oxide layer of the pad and stably establish a low contact resistance without increasing pad damage, sacrificing the detection cycle time, or shortening the probe life, and to solve the mutual constraint that existing methods bind oxide layer penetration with increasing static overdrive depth, this application provides a wafer inspection probe variable speed impact contact method and related apparatus.

[0006] Firstly, this application provides a variable-speed impact contact method for wafer inspection probes, which adopts the following technical solution: A variable-speed impact contact method for wafer inspection probes includes the following steps: S1. Controlling the probe to slowly approach the pad of the wafer to be inspected at a first speed until a preset pause position is reached; S2. Pausing at the pause position to eliminate residual oscillations generated by the slow approach, so that the probe enters the impact-ready state; S3. Driving the probe to impact the pad at a second speed greater than the first speed, so that the tip of the probe penetrates the oxide layer on the surface of the pad with kinetic momentum and contacts the metal layer below the oxide layer. The penetration is completed by the kinetic momentum of the tip before the displacement of the probe reaches the static overdrive depth; S4. After the tip penetrates the oxide layer, maintaining the contact pressure between the tip and the metal layer.

[0007] By adopting the above technical solution, the probe first approaches at a low speed and eliminates residual oscillations at the pause position, so that the probe enters a state of convergence of velocity and oscillation before impact, providing reproducible initial conditions for impact; then, the probe is driven at a second speed higher than the approach speed, so that the tip penetrates the oxide layer by means of motion momentum before the displacement reaches the static overdrive depth, thereby decoupling the goal of penetrating the oxide layer from the means of increasing the static overdrive depth; since the penetration is completed by momentum rather than continuous downward pressure, the mechanical damage to the pad and the aluminum accumulation caused by the tip scratching are reduced, and the contact pressure is maintained after the tip penetrates, ensuring the stability of the contact resistance during the measurement, so that the probe can still stably contact the metal layer with a small overdrive margin.

[0008] Optionally, S3 includes sub-steps S31-S33: S31. After the probe tip penetrates the oxide layer, measure the contact resistance between the probe tip and the metal layer. S32. When the contact resistance does not reach the preset contact resistance threshold, apply a supplementary impact to the pad without increasing the static overdrive depth of the probe. The energy of the supplementary impact increases gradually, and the contact resistance is remeasured after each supplementary impact. S33. When the contact resistance reaches the preset contact resistance threshold, stop applying the supplementary impact.

[0009] By adopting the above technical solution, and using whether the contact resistance meets the standard as the closed-loop criterion, the contact resistance is converged to the threshold without relying on increasing the downward pressure when the single impact does not fully penetrate the oxide layer by supplementing the impact with successively increasing energy under constant static overdrive depth.

[0010] Optionally, the method is applied to a probe array containing multiple probes; each probe in the probe array triggers S3 independently at the moment of its first contact with the corresponding pad under the same nominal overdrive conditions, and executes S31 to S33 independently. When the contact resistance of each probe reaches a preset contact resistance threshold, each probe stops applying supplementary impact to absorb the flatness distribution of the probe array.

[0011] By adopting the above technical solution, each probe in the array independently enters the impact and supplementary impact closed loop according to its first contact time. The probe that contacts first does not need to wait, and the probe that contacts later makes up for it with supplementary impact, thereby absorbing the dispersion of tip height and pad flatness under uniform nominal overdrive.

[0012] Optionally, the method further performs the following operations: monitoring the energy sequence taken from the supplementary impact and the rising slope as a function of the number of contacts; when the rising slope reaches a preset slope threshold, moving the probe to a cleaning station and applying a self-cleaning impact to the probe in the same manner as in S3, wherein the energy of the self-cleaning impact is higher than the energy required for the tip to penetrate the oxide layer; and after the self-cleaning impact is completed, recalibrating the energy baseline of the supplementary impact using the reference contact point.

[0013] By adopting the above technical solution, the slope of the impact energy increase with the number of contacts is used to characterize the accumulation of aluminum at the needle tip. When the slope reaches the threshold, the same impact method is used to apply a self-cleaning impact exceeding the film breaking energy to remove the aluminum accumulation, avoiding the need to set up an independent cleaning mechanism. After cleaning, the energy baseline is recalibrated to restore the closed-loop criterion.

[0014] Optionally, S2 includes sub-steps S21-S23: S21. Acquire the probe's position and velocity feedback. S22. When it is determined from the position and velocity feedback that the probe has reached zero velocity and the residual oscillation is below a preset oscillation threshold, confirm that the probe has entered the impact-ready state. S23. After confirming that the probe has entered the impact-ready state, trigger S3.

[0015] By adopting the above technical solution, zero velocity and residual oscillation below the threshold are used as the dual conditions for entering the impact state, so that each impact starts from a consistent velocity and oscillation level, thereby improving the reproducibility of impact penetration of the oxide layer.

[0016] Optionally, the second speed is adjusted according to the thickness of the oxide layer or the material of the pads, so that the momentum of the needle tip is matched with the energy required to penetrate the oxide layer.

[0017] By adopting the above technical solution, the second speed is pre-adjusted according to the oxide layer thickness or pad material, so that the momentum of the needle tip movement matches the energy required to penetrate different oxide layers, reducing insufficient penetration due to insufficient energy or pad damage due to excessive energy.

[0018] Optionally, the incremental energy of the supplementary impact is determined based on the residual between the current contact resistance and the preset contact resistance threshold, as well as the decrease in contact resistance caused by the previous application of the supplementary impact.

[0019] By adopting the above technical solution, the energy increment is adaptively adjusted according to the residual and the previous reduction, thereby reducing the number of supplementary impacts and energy overshoot.

[0020] Optionally, different actions can be taken based on the trajectory of the contact resistance as a result of the application of supplementary impact: apply a self-cleaning impact if the resistance does not decrease, reposition and retest or apply additional impact if the resistance decreases but fails to meet the standard, and increase the contact pressure when the resistance rises after meeting the standard.

[0021] By adopting the above technical solutions, aluminum accumulation, underpenetration, and contact drift can be distinguished and dealt with according to the change trajectory of contact resistance, thereby improving the contact success rate under abnormal conditions.

[0022] Optionally, a damping interval can be inserted between two consecutive applications of supplementary impact, and the next impact can be applied only after the residual oscillation is lower than the preset oscillation threshold.

[0023] By adopting the above technical solution, each supplementary impact starts from a low oscillation state, avoiding the superposition of residual oscillations from affecting the penetration consistency.

[0024] Optionally, the number of additional impacts required for the same contact point to reach a preset contact resistance threshold can be used as the characteristic quantity of aluminum accumulation at that contact point.

[0025] By adopting the above technical solution, the number of additional impacts required to meet the standard is used as an online indicator of the degree of aluminum accumulation, providing a criterion for determining the timing of maintenance.

[0026] Optionally, the energy of the supplementary impact is increased in preset steps, and when the preset energy limit is reached but the contact resistance still does not meet the standard, the corresponding contact point is marked as abnormal and transferred to maintenance.

[0027] By adopting the above technical solution, an energy limit is set for the supplementary impact, and contact points that continuously fail to meet the standard are transferred to maintenance, thus avoiding damage to the solder pads by the infinitely increasing energy.

[0028] Optionally, contact points marked as abnormal can be skipped in subsequent inspections or transferred to manual re-inspection.

[0029] By adopting the above technical solution, abnormal contact points are bypassed, avoiding single-point abnormalities from blocking the overall detection cycle of the array.

[0030] Optionally, the initial applied energy of the supplementary impact can be calculated based on the impact energy of the needle tip penetrating the oxide layer.

[0031] By adopting the above technical solution, the energy of the first supplementary impact is set based on the membrane rupture energy, thus shortening the iterative process of the energy approaching the target value.

[0032] Optionally, when multiple probes in the probe array apply supplementary impacts in parallel, the total power of the supplementary impacts is constrained by a preset power upper limit, and the supplementary impact energy quota of each probe is allocated according to the residual between the contact resistance of each probe and the preset contact resistance threshold.

[0033] By adopting the above technical solution, the energy quota of each needle is allocated according to the residual under the total power constraint, which takes into account both the efficiency of parallel supplementary impact and the power limit.

[0034] Optionally, the flatness distribution of the probe array can be determined based on the dispersion of the number of additional impacts required for each probe in the probe array to reach the preset contact resistance threshold.

[0035] By adopting the above technical solution, the array flatness distribution is inferred from the dispersion of the number of times each needle meets the standard, providing a quantitative basis for probe card leveling.

[0036] Optionally, probes in the probe array that first reach the preset contact resistance threshold maintain their current contact pressure without increasing the static overdrive depth, and wait for the remaining probes to reach the threshold before the probe array enters the test.

[0037] By adopting the above technical solution, the first calibration probe is kept under constant pressure and waits without increasing the overdrive, thus avoiding over-pressure on the first calibration probe to accommodate the later calibration probe.

[0038] Optionally, if the energy of the supplementary impact at the reference contact point after applying the self-cleaning impact is still higher than the preset reset target, an additional self-cleaning impact may be applied until it falls back to the preset reset target.

[0039] By adopting the above technical solution, and using the reset target as the closed-loop criterion to add self-cleaning impact, the cleaning effect can be verified rather than being executed in a single open-loop manner.

[0040] Optionally, the number of contacts required to reach a preset slope threshold can be predicted by extrapolation based on the rising slope, and the self-cleaning impact can be applied during the wafer change gap or the gap where the probe separates from the wafer.

[0041] By adopting the above technical solutions, maintenance timing can be predicted and self-cleaning can be scheduled to be performed during inspection intervals, thereby reducing the impact of self-cleaning on the inspection cycle.

[0042] Optionally, the energy of the self-cleaning impact can be determined by classifying the magnitude of the rise slope.

[0043] By adopting the above technical solution, self-cleaning energy is applied in stages according to the rate of aluminum accumulation, avoiding excessive impact on the needle tip when aluminum is only slightly accumulated.

[0044] Optionally, the energy baseline of the supplementary impact after each self-cleaning impact is monitored, and if the energy falls below the preset lower limit after a preset number of consecutive preset number of times, the probe is deemed to be faulty and an alarm is triggered or the probe is replaced.

[0045] By adopting the above technical solution, the attenuation of the energy baseline drop after self-cleaning is used as the criterion for the end of the probe's lifespan, and timely alarms are triggered to replace the probe.

[0046] Optionally, the energy of the self-cleaning impact is equal to the energy of the current supplementary impact plus a preset increment.

[0047] By adopting the above technical solution, the self-cleaning energy is set by adding the current supplementary impact energy to the incremental setting, so that the cleaning intensity adapts to the degree of aluminum accumulation.

[0048] Optionally, the upward slope is obtained by linearly fitting the supplementary impact energy corresponding to a preset number of consecutive contacts.

[0049] By adopting the above technical solution, the slope is linearly fitted using multiple energy data, thereby reducing the interference of a single energy fluctuation on the maintenance criteria.

[0050] Optionally, the triggering condition for self-cleaning shocks may also include the absolute value of the contact resistance exceeding a preset upper limit.

[0051] By adopting the above technical solution, the absolute value of contact resistance is used as an additional triggering condition for self-cleaning, covering the situation where the slope has not yet met the standard but the contact has deteriorated.

[0052] Optionally, the preset slope threshold can be recalibrated after the self-cleaning impact is completed.

[0053] By adopting the above technical solution, the slope threshold is recalibrated after cleaning, so that the subsequent maintenance criteria are adapted to the baseline after cleaning.

[0054] Optionally, the confirmation of the impact state includes comparing the probe's position and velocity feedback with a pre-stored reference state.

[0055] By adopting the above technical solution, the impact state is confirmed by comparing it with the pre-stored benchmark state, thereby improving the consistency of the initial conditions when entering the impact.

[0056] Optionally, the duration of the pause can be adaptively determined based on the initial speed or the approaching slow speed.

[0057] By adopting the above technical solution, the pause duration is adaptively adjusted according to the approach speed or stroke, achieving a balance between sufficient vibration damping and cycle time.

[0058] Optionally, the second velocity is configured with respect to the momentum required to penetrate the oxide layer according to a preset energy-velocity relationship.

[0059] By adopting the above technical solution, the second velocity is configured according to the energy-velocity relationship, so that the impact energy accurately corresponds to the momentum required to penetrate the oxide layer.

[0060] Optionally, the second velocity is not lower than the lower limit velocity that allows the needle tip's momentum to penetrate the oxide layer.

[0061] By adopting the above technical solution, a lower limit is set for the second velocity to ensure that the momentum of a single impact is sufficient to complete the penetration.

[0062] Optionally, the angle between the direction of the probe impacting the pad and the normal of the pad surface shall not exceed a preset angle, so as to limit the lateral displacement component when the probe tip penetrates the oxide layer.

[0063] By adopting the above technical solution, the impact incident angle is constrained to limit the lateral displacement component, thereby reducing the lateral scraping of the pin tip against the pad and the transfer of aluminum.

[0064] Optionally, the second speed and impact stroke are configured such that the depth of the pin mark formed by the pin tip on the pad does not exceed a preset pin mark depth threshold.

[0065] By adopting the above technical solution, the second speed and stroke are constrained to limit the needle mark depth, thus reserving usable pad area for subsequent wire bonding.

[0066] Secondly, the wafer inspection probe variable speed impact contact device provided in this application adopts the following technical solution: A wafer inspection probe variable speed impact contact device includes: a servo impact mechanism for driving the probe to move toward the pad of the wafer to be inspected at a controllable speed; a controller connected to the servo impact mechanism and configured to: control the probe to slowly approach the pad at a first speed to a preset pause position; pause at the pause position to eliminate residual oscillations generated by the slow approach; control the servo impact mechanism to drive the probe to impact the pad at a second speed greater than the first speed, so that the tip of the probe penetrates the oxide layer on the surface of the pad with kinetic momentum and contacts the metal layer below the oxide layer, the penetration is completed by the kinetic momentum of the tip before the displacement of the probe reaches the static overdrive depth; and maintain the contact pressure between the tip and the metal layer after the tip penetrates the oxide layer.

[0067] By adopting the above technical solution, the servo impact mechanism provides controllable approach and impact speed, and the controller coordinates slow approach, pause vibration damping and high-speed impact, so that the needle tip can penetrate the oxide layer before reaching the static overdrive depth by relying on the momentum of motion. This achieves decoupling of film breaking and static overdrive depth in hardware, and provides a device carrier for the execution of the method.

[0068] Optionally, the controller is also configured to: measure the contact resistance between the probe tip and the metal layer after the probe tip penetrates the oxide layer; when the contact resistance does not reach the preset contact resistance threshold, without increasing the static overdrive depth of the probe, control the servo impact mechanism to apply a supplementary impact with progressively increasing energy to the pad, and remeasure the contact resistance after each application; stop applying the supplementary impact when the contact resistance reaches the preset contact resistance threshold.

[0069] By adopting the above technical solution, the controller uses the contact resistance meeting the standard as a closed loop and drives the servo impact mechanism to apply a constant static pressure incremental supplementary impact, so that the device can make up for the penetration without increasing the overdrive.

[0070] Optionally, the probe is a probe array containing multiple probes; the controller is configured to enable each probe in the probe array to independently trigger an impact at the moment of its first contact with the corresponding pad under the same nominal overdrive conditions, and to independently perform contact resistance measurement and supplementary impact application, and to stop when the contact resistance of each probe reaches a preset contact resistance threshold, so as to absorb the flatness distribution of the probe array.

[0071] By adopting the above technical solution, the controller enables each pin of the array to independently close the loop according to its own contact time, absorbs the distribution of pin tip height and pad flatness, and ensures that multiple pins achieve standard contact under uniform nominal overdrive.

[0072] Optionally, the controller is also configured to: monitor the energy sequence taken from the supplementary impact and the rising slope as a function of the number of contacts; when the rising slope reaches a preset slope threshold, move the probe to the cleaning station and control the servo impact mechanism to apply a self-cleaning impact, the energy of which is higher than the energy required for the needle tip to penetrate the oxide layer; after the self-cleaning impact is completed, recalibrate the energy baseline of the supplementary impact using the reference contact point.

[0073] By adopting the above technical solution, the controller triggers the self-cleaning impact by supplementing the upward slope of the impact energy, reuses the same servo impact mechanism to apply the impact and recalibrate the baseline, so that the membrane breaking, supplementing the impact and maintenance share the same actuator.

[0074] In summary, this application includes at least one of the following beneficial technical effects: 1. By using kinetic momentum to penetrate the pad oxide layer before the probe displacement reaches the static overdrive depth, the penetration of the oxide layer is decoupled from the increase of the static overdrive depth. This reduces the mechanical damage to the pad and the aluminum buildup at the probe tip as the pressure decreases, and establishes a stable low contact resistance with a small overdrive margin.

[0075] 2. Using the contact resistance as a closed-loop criterion, supplementary impacts with progressively increasing energy are applied to compensate for penetration without increasing the static overdrive depth, and each pin in the probe array independently meets the standard according to its own contact time, thereby absorbing the dispersion of pin tip height and pad flatness, so that multiple pins can make stable contact under a uniform nominal overdrive.

[0076] 3. The slope of the increase in supplementary impact energy with the number of contacts is used to characterize the aluminum accumulation at the needle tip. The same impact method is used to apply self-cleaning impact and recalibrate the energy baseline after cleaning. The film breaking, supplementary impact and needle tip maintenance are unified into the same servo impact action, reducing the detection cycle time occupation and probe life loss caused by independent cleaning mechanism and periodic shutdown for needle cleaning. Attached Figure Description

[0077] Figure 1 This is a schematic flowchart of the variable speed impact contact method for wafer inspection probes disclosed in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of the variable speed impact contact device for wafer inspection probes disclosed in an embodiment of this application. Detailed Implementation

[0078] The present application will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the application and are not intended to limit the scope of the application.

[0079] This application discloses a variable-speed impact contact method for wafer inspection probes.

[0080] This method moves the probe towards the pad of the wafer under test at a controlled speed. It first approaches slowly to a preset pause position at a lower first speed and pauses there to eliminate residual oscillations generated during the approach. Then, it impacts the pad at a second speed higher than the first speed. This allows the probe tip to penetrate the oxide layer on the pad surface and contact the underlying metal layer before reaching the static overdrive depth, thanks to the momentum gained from the motion. This decouples the oxide layer penetration action from the static overdrive depth formed by the downward pressure. Furthermore, this method uses contact resistance feedback for closed-loop control of the supplementary impact, applies impact independently to each probe in the probe array, and performs self-cleaning of the probe tip using a homogeneous impact method, thus balancing the consistency of array contact and the lifespan maintenance of the probes. The following section first explains several technical terms involved in the embodiments of this application.

[0081] The probe tip is the end of the probe that comes into contact with the pad, and is used to establish an electrical connection with the conductive surface of the pad during testing.

[0082] The oxide layer of a solder pad is a non-conductive or weakly conductive thin film that naturally forms on the metal surface of the solder pad due to exposure to the environment. It covers the metal layer of the solder pad and prevents the formation of a low-resistance contact between the tip and the metal layer.

[0083] The metal layer is the conductive body beneath the oxide layer of the pads. After the needle tip penetrates the oxide layer, it comes into contact with the metal layer to form the conductive path required for the test.

[0084] Static overdrive depth is the amount of overtravel formed by the probe continuing to press down towards the pad after contacting it, which is a quasi-static displacement. It characterizes the depth to which the probe tip is pressed into the pad.

[0085] Momentum is the momentum that the probe gains due to its velocity during impact. The probe tip relies on this momentum to exert an impact on the oxide layer at the moment of contact.

[0086] The ready state is the state in which the probe remains in the paused position until its motion velocity converges to zero and the residual oscillation decays to a preset range. This state provides reproducible initial conditions for subsequent impacts.

[0087] The application environment of this method is described below. This method can be applied to wafer-level testing. For pads located on the wafer, multiple probes on a probe card make parallel contact with multiple pads to introduce test signals into the circuit under test on the wafer. The pads can be made of aluminum or aluminum alloy, and their surface is usually pre-oxided before testing.

[0088] The overall flow of this method is described in detail below with reference to the accompanying drawings. This method includes steps S1 to S4. The overall flow of this method is as follows: Figure 1 As shown.

[0089] S1. The probe slowly approaches the pad at a first speed to the preset pause position. The first speed is a lower approach speed than the subsequent impact speed, so that the movement of the probe is relatively gentle in the final stage of the approach, thereby reducing the impact and oscillation when it reaches the pad.

[0090] S2. The probe is brought to a pause position to eliminate residual oscillations generated during the slow approach, thus bringing the probe into an impact-ready state. During the pause, the probe's speed gradually converges to zero, and the mechanical oscillations excited during the approach decay over time. Once the residual oscillations decay to a preset range, the probe is in an impact-ready state where an impact can be applied.

[0091] In some embodiments, S2 includes sub-steps S21-S23.

[0092] S21. Acquire probe position and velocity feedback. Position feedback indicates the probe's current position relative to the pad, and velocity feedback indicates the probe's current speed. Both are used to determine whether the probe has come to a stop.

[0093] S22. When the velocity feedback characterizes the probe's motion velocity to zero and the amplitude of the residual oscillation is lower than the preset oscillation threshold, the probe is confirmed to have entered the impact-ready state. For example, the oscillation threshold can be a small amount of the probe tip displacement amplitude, and the probe is considered to have stopped when the measured residual oscillation amplitude is lower than this small amount.

[0094] S23. In response to the confirmation probe entering the impact-ready state, the impact action of S3 is triggered.

[0095] In some embodiments, a reference state representing a stopped state is pre-stored, which includes reference values ​​for position and velocity feedback. The position and velocity feedback collected in S21 are compared with this reference state. When the deviation between the two falls within a preset tolerance range, it is confirmed that the probe has entered the impact-ready state. Thus, the consistency of the impact-ready state determination is improved by using a reference comparison method.

[0096] In some embodiments, the dwell time at the pause position is dynamically adjusted based on the probe's approach speed or travel distance to the pad. For example, the higher the approach speed or the longer the approach travel, the greater the residual oscillations generated at the end of the approach, and the dwell time is correspondingly extended. The dwell time can be in the range of several milliseconds to tens of milliseconds so that the residual oscillations have decayed to a preset range by the end of the dwell.

[0097] S3. The probe impacts the pad at a second velocity greater than the first velocity, causing the probe tip to penetrate the oxide layer on the pad surface and contact the metal layer beneath it using the momentum gained from the motion. The penetration of the oxide layer by the probe tip is completed by the momentum of the probe tip before the probe's displacement reaches the static overdrive depth. When the probe tip impacts the oxide layer with its momentum, it applies a short-duration impact, causing the oxide layer to break instantly upon contact. The film breaking is completed by the momentum of the probe tip at the instant of contact, rather than by the gradual breaking through the static overdrive depth formed by continuous downward pressure from the probe. The probe tip completes the film breaking within a very shallow displacement into the pad using momentum, and the displacement required for film breaking is much smaller than the set static overdrive depth.

[0098] For example, if the oxide layer on the pad surface is approximately 10 nanometers thick, when the needle tip impacts at the second velocity, it breaks through the oxide layer and contacts the metal layer within a displacement of approximately 0.1 micrometers into the pad using kinetic momentum. The set static overdrive depth is approximately 15 micrometers, far greater than the approximately 0.1 micrometer displacement required to break through the oxide layer. Therefore, the oxide layer breaking action is completed before the displacement reaches the static overdrive depth. Thus, the film breaking action is no longer tied to increasing the static overdrive depth; the needle tip establishes a low-resistance contact with the metal layer with a shallower insertion, correspondingly reducing the overdrive depth accumulated on the pad.

[0099] The actuator that achieves a controllable second velocity can take various forms. In one implementation, a voice coil motor drives the probe to accelerate along the impact direction to obtain the second velocity; in another implementation, a piezoelectric stack drives the probe to generate high-speed micro-displacement to obtain the second velocity. Both methods enable the probe tip to impact the pad with controllable momentum, representing different implementations of impacting the pad with a second velocity.

[0100] In some embodiments, the second speed is adjusted according to the oxide layer thickness and pad material of the pad, so that the momentum of the probe tip is matched with the momentum required to penetrate the oxide layer. For example, the thicker the oxide layer or the harder the pad material, the second speed is increased accordingly to increase the momentum of the probe tip; for common aluminum pads, the second speed can be selected within a range that provides sufficient momentum for the probe tip to penetrate its surface oxide layer. Thus, different pad conditions are adapted in a speed-feedforward manner.

[0101] In some embodiments, the impact energy of the needle tip impacting the pad is correlated with a second velocity, and the impact energy increases with the increase of the second velocity. The impact energy is determined by both the equivalent mass of the needle tip and the second velocity; the greater the equivalent mass or the higher the second velocity, the greater the impact energy.

[0102] In some embodiments, the second speed is not lower than a lower limit speed that allows the movement momentum of the needle tip to penetrate the oxide layer on the surface of the pad, so as to ensure that the needle tip can reliably penetrate the oxide layer and contact the metal layer. This lower limit speed is determined according to the thickness and material of the oxide layer.

[0103] In some embodiments, the incident angle of the probe impacting the pad is limited to a preset angle threshold to reduce the lateral displacement component of the probe tip relative to the pad surface during the impact. A larger lateral displacement component makes it easier for the probe tip to plow the pad surface, carrying up metal and causing transfer and accumulation; limiting the incident angle reduces this lateral scratching, thereby reducing aluminum buildup. The incident angle threshold can be set to a small angle that makes the impact direction approximately perpendicular to the pad surface.

[0104] In some embodiments, the pin mark depth is the depth of the indentation formed by the pin tip on the pad, which increases with the increase of the second speed and the downward stroke; the second speed and the downward stroke are configured such that the pin mark depth does not exceed a preset depth threshold. Thus, the pin mark depth is controlled while ensuring oxide layer penetration, leaving sufficient usable area for the pad in subsequent wire bonding and other processes.

[0105] In some embodiments, S3 includes sub-steps S31-S33.

[0106] S31. After the needle tip penetrates the oxide layer, measure the contact resistance between the needle tip and the metal layer. The contact resistance characterizes the quality of the contact between the needle tip and the metal layer; a higher contact resistance indicates a poorer contact.

[0107] S32. When the contact resistance does not reach the preset contact resistance threshold, a supplementary impact is applied to the pad without increasing the static overdrive depth of the probe. The energy of the supplementary impact increases progressively, and the contact resistance is remeasured after each supplementary impact. After a single impact, there may still be insufficient contact between the probe tip and the metal layer. By applying supplementary impacts with progressively increasing energy while maintaining a constant static overdrive depth, the contact is gradually improved without increasing the overdrive depth formed by the downward pressure. For example, if the contact resistance threshold is set to 0.5 ohms, and the contact resistance is measured to be 2.0 ohms after the first impact, which is higher than the threshold, a first supplementary impact is applied. If the energy of the first supplementary impact is higher than the energy of the first impact, the contact resistance drops to 1.0 ohms after the impact, still higher than the threshold, then a second supplementary impact with even higher energy is applied. After the second impact, the contact resistance drops to 0.4 ohms, lower than the threshold. Throughout the entire process, the static overdrive depth of the probe remains constant.

[0108] S33. When the contact resistance reaches the preset contact resistance threshold, stop the supplementary impact.

[0109] There are several ways to progressively increase the energy of the supplementary impact. In one implementation, the energy of each supplementary impact increases by a fixed step size based on the previous energy; in another implementation, the energy of each supplementary impact increases by a fixed multiple based on the previous energy. Both methods result in a progressive increase in the energy of the supplementary impact, representing different implementations of progressive energy increase.

[0110] In some embodiments, the energy increment for the next supplementary impact is determined based on the deviation of the current contact resistance from a threshold and the magnitude of the decrease in contact resistance caused by the previous supplementary impact. A larger current deviation results in a larger energy increment, accelerating the convergence of the contact resistance to the threshold; a smaller decrease in the previous impact indicates slower contact improvement, thus requiring a correspondingly larger energy increment. This adaptive energy increment reduces the number of supplementary impacts needed to reach the threshold contact resistance.

[0111] In some embodiments, the contact state is classified and appropriate measures are taken based on the trajectory of the contact resistance changing with the number of impacts during the supplementary impact process, such as the downward slope or inflection point of the trajectory. The trajectory of the contact resistance includes the following three types of patterns.

[0112] Firstly, the contact resistance does not decrease significantly after repeated impacts. This indicates that there may be a barrier between the probe tip and the metal layer that is difficult to break through due to factors such as aluminum buildup. In this case, the probe should be moved to a cleaning station to clean the probe tip before making contact again.

[0113] Secondly, the contact resistance decreases with each additional impact but never reaches the threshold. This indicates insufficient penetration of the probe tip into the oxide layer. In this case, without increasing the static overdrive depth, the energy of the additional impact should be increased, or the probe should be moved to another contact point on the pad for re-contact.

[0114] Third, the contact resistance initially reaches the threshold and then rises again. This pattern indicates that the contact between the probe tip and the metal layer has drifted, at which point the contact pressure between the probe tip and the pad is maintained to stabilize the contact.

[0115] Therefore, targeted measures can be taken for different forms of contact resistance trajectories to improve the success rate of establishing reliable contact under abnormal conditions.

[0116] In some embodiments, a vibration damping interval is inserted between two adjacent supplementary impacts. After the residual oscillations excited by the previous impact decay to a preset range, the next supplementary impact is applied to avoid the residual oscillations of the previous impact being superimposed on the next impact and affecting the consistency of penetration.

[0117] In some embodiments, the number of additional impacts required to bring the contact resistance to a threshold is used as a measure of the degree of aluminum buildup. A higher number of additional impacts indicates greater obstruction between the needle tip and the metal layer, and consequently, a higher degree of aluminum buildup. This number can be used to determine when the needle tip needs maintenance.

[0118] In some embodiments, an upper limit is set for the energy of the supplementary impact. When the energy of the supplementary impact reaches this upper limit but the contact resistance still does not reach the threshold, the supplementary impact on the current pad is stopped and the process switches to the tip maintenance procedure. Setting an energy upper limit can prevent the energy of the supplementary impact from increasing indefinitely and causing excessive damage to the pad.

[0119] In some embodiments, pads that consistently fail to reach the contact resistance threshold are marked and skipped for subsequent manual re-inspection, in order to prevent a single abnormal pad from blocking the overall testing process.

[0120] In some embodiments, the energy of the first supplementary impact is estimated based on the energy required to break down the oxide layer of the pad, so that the energy of the first supplementary impact is comparable to the energy required to break down the oxide layer, and the supplementary impact starts from an energy close to the target, reducing the number of supplementary impacts required for the contact resistance to converge to the threshold.

[0121] In some embodiments, the method is applied to a probe array including multiple probes, each probe in the array is set with the same nominal overdrive depth; each probe independently triggers its own S3 when it first contacts the corresponding pad, and each performs contact resistance closed-loop supplementary impact from S31 to S33; each probe stops supplementary impact when its contact resistance reaches a threshold, thereby absorbing the flatness dispersion between the pins in the array.

[0122] The nominal overdrive depth is a target value for the overdrive depth uniformly set for all probes in the array. Due to the dispersion of probe tip height and pad plane caused by manufacturing and assembly, the actual contact time of each probe with its respective pad is not the same when the array is pressed down to the same nominal overdrive depth.

[0123] For example, a probe array includes a first probe, a second probe, and a third probe, all with the same nominal overdrive depth. Due to the dispersion of probe tip height and pad plane, the three probes do not contact their respective pads simultaneously; the first probe contacts first, followed by the second, and the third probe contacts last. Each probe independently triggers its own impact and supplementary impact closed loop at the moment of its first contact with the pad. The first probe achieves the required contact resistance after one supplementary impact, the second probe after two supplementary impacts, and the third probe after three supplementary impacts. Once each probe achieves its target, it stops supplementary impacting independently, without affecting the others. Thus, even with the dispersion of probe tip height and pad flatness, each probe still achieves the required contact state, absorbing the impact of the flatness dispersion on the contact of each probe in the probe-by-probe closed loop.

[0124] The moment when each probe first contacts the pad can be determined in several ways. One implementation method is to determine contact based on a sudden change in the contact resistance of each probe from an open-circuit state; another implementation method is to determine contact based on the probe's displacement reaching a preset contact position. Both methods can determine the initial contact time of each probe and represent different implementations of determining the initial contact time. It should be understood that the number of probes in the probe array is not limited to the three methods mentioned above, and this method is also applicable to arrays with more probes and probe arrays with different probe pitches.

[0125] In some embodiments, when supplementary impacts are applied in parallel to each pin of the array, the power quota for the supplementary impact is allocated to each pin based on the deviation of its contact resistance from a threshold value, constrained by the upper limit of the total power supply. The probe with the larger deviation receives a higher power quota, prioritizing the improvement of probes with poor contact. The sum of the power quotas for all pins does not exceed the upper limit of the total power supply. For example, when three probes apply supplementary impacts in parallel, the third probe with the largest contact resistance deviation receives the highest power quota, while the first probe with the smaller deviation receives a lower power quota. Thus, the efficiency of parallel supplementary impacts and the need for contact improvement of each pin are balanced without exceeding the total power supply limit.

[0126] There are several ways to allocate power quotas. One implementation is to divide the total power equally among all pins; another implementation is to prioritize the allocation of power to each pin based on the magnitude of their contact resistance deviation, allocating power first to the pins with larger deviations. Both methods allocate supplementary impact power to each pin within the total power limit, representing different implementations of power quota allocation.

[0127] In some embodiments, the dispersion of the number of additional impacts required for each pin to reach the contact resistance threshold reflects the flatness distribution between the probe array and the pad plane. The more dispersed the number of impacts required for each pin to reach the threshold, the greater the difference in contact conditions between the pins, and the greater the corresponding flatness distribution. This dispersion can be used to evaluate the flatness condition between the probe card and the wafer.

[0128] In other embodiments, for probes in the array that have already reached the contact resistance threshold, their contact pressure on the pads is maintained while waiting for the remaining probes to reach the threshold. During this waiting period, the static overdrive depth of the probe is not increased. This ensures that the probes that have reached the threshold first maintain contact without being over-pressed into the pads due to continued pressure.

[0129] In some embodiments, the method also monitors the rising slope of the energy sequence of the supplementary impact with the number of contacts. When the rising slope reaches a preset slope threshold, the probe is moved to the cleaning station and a self-cleaning impact is applied to the probe in the same way as in S3. The energy of the self-cleaning impact is higher than the energy required for the needle tip to penetrate the oxide layer. After self-cleaning, the energy baseline of the supplementary impact is recalibrated at the reference contact point.

[0130] As the probe repeatedly contacts the pads, aluminum and other pad materials gradually accumulate on the probe tip, worsening the contact between the probe tip and the metal layer. The supplementary impact energy required to reach the contact resistance threshold increases with each contact, and the slope of the supplementary impact energy reflects the rate of material accumulation on the probe tip. By applying a higher-energy self-cleaning impact to the probe tip using the same impact method as for film breaking, the material accumulated on the probe tip is removed by the impact action, eliminating the need for a separate cleaning mechanism for cleaning the probe tip.

[0131] When monitoring the rising slope of the supplementary impact energy sequence, the supplementary impact energy at which each of several consecutive contacts reaches the contact resistance threshold is taken. The rising slope is determined based on the rate at which this energy increases with the number of contacts. For example, the slope threshold corresponds to a certain increase in supplementary impact energy after a certain number of contacts. When the measured rate of energy increase reaches this slope threshold, it is determined that the needle tip adhesion is relatively severe, triggering self-cleaning. The energy of the self-cleaning impact is higher than the energy required to break through the oxide layer, ensuring that the material deposited on the needle tip is removed by impact.

[0132] After self-cleaning, the probe re-contacts the pad at the reference contact point. The additional impact energy required to reach the contact resistance threshold at this reference contact point is used as the new energy baseline, and subsequent rise rate monitoring starts from this new baseline. This ensures that the monitoring of additional impact energy returns to the baseline corresponding to the clean state after self-cleaning, avoiding the miscalculation of energy accumulation before self-cleaning into the subsequent rise rate.

[0133] Cleaning stations can take various forms. In one implementation, a dedicated cleaning block is provided at the cleaning station, and a probe applies a self-cleaning impact to this block to remove deposited material. In another implementation, a waste pad or test pad pre-reserved on the wafer is used at the cleaning station, and the probe applies a self-cleaning impact to this pad. Both methods involve the probe applying a self-cleaning impact at the cleaning station using the same impact method as S3, representing different implementations of self-cleaning.

[0134] In some embodiments, the cleaning effect of the probe tip is retested after the self-cleaning impact, thus forming a closed loop of self-cleaning. Specifically, after self-cleaning, the probe re-contacts the pad at a reference contact point, and the supplementary impact energy required to reach the contact resistance threshold at this reference contact point is retested. The retested energy is compared with a preset reset target. If the retested energy is higher than the reset target, it indicates that the deposits on the probe tip have not been sufficiently removed, so an additional self-cleaning impact is applied until the retested energy falls below the reset target. For example, the reset target is set as the supplementary impact energy corresponding to the reference contact point when the probe tip is in a clean state. If the retested energy after one self-cleaning is still higher than the reset target, an additional self-cleaning impact is applied and the retest is repeated until the retested energy is lower than the reset target. This allows the removal effect of self-cleaning to be verified, rather than ending with a single open-loop self-cleaning impact.

[0135] Several criteria can be used to determine whether the needle tip has returned to a clean state. One implementation uses whether the additional impact energy required to reach the contact resistance threshold at the reference contact point has fallen back to the reset target. Another implementation uses whether the contact resistance at the reference contact point after the first impact has fallen back to a preset value. Both criteria reflect whether the needle tip has returned to a clean state and represent different implementations of the reset determination.

[0136] In some embodiments, self-cleaning is scheduled during the intervals of the test flow to reduce its impact on the test cycle time. Specifically, the number of contacts required for the supplemental impact energy to reach a slope threshold is predicted by extrapolating the rise rate of the supplemental impact energy, and self-cleaning is scheduled to be performed during the die change interval or the interval when the probe separates from the pad before that number of contacts. For example, if the current rise rate predicts that the slope threshold will be reached after several more contacts, self-cleaning is performed in advance during the most recent die change interval. This allows self-cleaning to avoid the period when the probe is in contact with the pad for testing, reducing its impact on the overall test cycle time.

[0137] Self-cleaning can be triggered in several ways. One implementation involves periodically performing self-cleaning at fixed contact cycles; another implementation involves self-cleaning being driven by the event that the rising slope of the replenishing impact energy reaches a certain threshold. Both methods can trigger self-cleaning and represent different implementations of self-cleaning triggering.

[0138] In some embodiments, the energy of the self-cleaning impact is divided into different levels based on the magnitude of the upward slope of the supplementary impact energy. The greater the upward slope, the more severe the needle tip adhesion, and thus a higher energy level of self-cleaning impact is selected. Therefore, a lower energy self-cleaning impact is selected when the needle tip has only slight adhesion, avoiding unnecessary impact on the needle tip.

[0139] In other embodiments, the amount of drop in the baseline of the replenished impact energy after each self-cleaning relative to the amount before self-cleaning is monitored. When the amount of drop after multiple self-cleanings is consistently insufficient, it is determined that the needle tip can no longer be recovered through self-cleaning, and a needle tip replacement prompt is issued so that the needle tip can be replaced in a timely manner when self-cleaning is no longer effective.

[0140] As an alternative implementation, the energy of the self-cleaning impact is increased by a certain increment based on the current supplementary impact energy, so that the energy of the self-cleaning impact is higher than the energy of the current supplementary impact and sufficient to remove the material deposited on the needle tip.

[0141] The rising slope of the supplementary impact energy can optionally be determined by linearly fitting the supplementary impact energy and the number of contacts for each of several consecutive contacts, and using the slope of the fitted line.

[0142] In addition to triggering self-cleaning based on the rising slope of the supplementary impact energy, in some embodiments self-cleaning is also triggered when the absolute value of the contact resistance after the supplementary impact exceeds a preset upper limit of resistance, in order to cover the situation where the rising slope has not yet reached the slope threshold, but the contact has been significantly deteriorated.

[0143] In some other embodiments, after self-cleaning, while recalibrating the supplementary impact energy baseline, the slope threshold for triggering self-cleaning is recalibrated according to the new energy baseline, so that the slope threshold is adapted to the energy baseline after self-cleaning, thereby ensuring that the triggering criterion for self-cleaning remains consistent with the current tip state after each self-cleaning.

[0144] S4. After the needle tip penetrates the oxide layer and contacts the metal layer, maintain the contact pressure between the needle tip and the metal layer to maintain stable contact between the needle tip and the metal layer during measurement.

[0145] This application also discloses a wafer inspection probe variable speed impact contact device, which is used to perform the above-described wafer inspection probe variable speed impact contact method. The device includes a servo impact mechanism and a controller. The servo impact mechanism drives the probe to move towards the pads of the wafer to be inspected at a controllable speed; the controller is connected to the servo impact mechanism and controls the servo impact mechanism to drive the probe to complete the actions of slow approach, pause for vibration damping, and impact film breaking. The structure of the device is as follows: Figure 2 As shown.

[0146] Specifically, the controller controls the probe to slowly approach the pad at a first speed to a preset pause position; the controller controls the probe to stay at the pause position to eliminate residual oscillations and enter the impact-ready state; the controller controls the servo impact mechanism to drive the probe to impact the pad at a second speed greater than the first speed, so that the tip penetrates the oxide layer on the surface of the pad with kinetic momentum and contacts the metal layer below the oxide layer, wherein the action of the tip penetrating the oxide layer is completed by the kinetic momentum of the tip before the displacement of the probe reaches the static overdrive depth; and the contact pressure between the tip and the metal layer is maintained after the tip penetrates the oxide layer.

[0147] Therefore, a servo-driven impact mechanism provides controllable approach and impact speeds, while a controller coordinates slow approach, pause for vibration damping, and high-speed impact. This allows the probe tip to penetrate the oxide layer before reaching the static overdrive depth, thus decoupling the film-breaking action from the static overdrive depth in hardware. The servo-driven impact mechanism can be an actuator such as a voice coil motor or piezoelectric stack that can drive the probe to generate controllable impacts.

[0148] In some embodiments, the controller is also configured to perform the operations defined in the subordinate steps of the aforementioned method, including closed-loop control of the supplementary impact with contact resistance feedback without increasing the static overdrive depth, applying the impact and supplementary impact closed loop independently to each probe of the probe array at its first contact time, and applying a self-cleaning impact of the same source to the probe based on the rising slope of the supplementary impact energy with the number of contacts and recalibrating the energy baseline; the specific limitations of these operations can be found in the corresponding descriptions in the above method, and will not be repeated here.

[0149] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0150] 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, and should all be included within the protection scope of the present invention.

Claims

1. A method for variable-speed impact contact of a wafer inspection probe, characterized in that, Includes the following steps: S1. Control the probe to slowly approach the pads of the wafer to be inspected at a first speed until it reaches the preset pause position; S2. Pause at the pause position to eliminate residual oscillations caused by the slow approach, so that the probe enters the impact-ready state; S3. Drive the probe to impact the pad at a second speed greater than the first speed, so that the tip of the probe penetrates the oxide layer on the surface of the pad with kinetic momentum and contacts the metal layer below the oxide layer. The penetration is completed by the kinetic momentum of the tip before the displacement of the probe reaches the static overdrive depth. S4. After the probe tip penetrates the oxide layer, maintain the contact pressure between the probe tip and the metal layer; S3 includes the following sub-steps: S31. After the probe tip penetrates the oxide layer, measure the contact resistance between the probe tip and the metal layer; S32. When the contact resistance is greater than a preset contact resistance threshold, apply a supplementary impact to the pad without increasing the static overdrive depth of the probe, the energy of the supplementary impact increases progressively, and the contact resistance is remeasured after each application of the supplementary impact; S33. When the contact resistance is less than or equal to the preset contact resistance threshold, stop applying the supplementary impact.

2. The wafer inspection probe variable speed impact contact method according to claim 1, characterized in that, The method is applied to a probe array containing multiple probes; each probe in the probe array independently triggers S3 at the first contact time between its tip and the corresponding pad under the same nominal overdrive conditions, and independently executes S31 to S33. When the contact resistance of each probe is less than or equal to the preset contact resistance threshold, the supplementary impact is stopped to absorb the flatness distribution of the probe array.

3. The wafer inspection probe variable speed impact contact method according to claim 1, characterized in that, The method further performs the following operations: monitoring the energy sequence from the supplementary impact and the rising slope as a function of the number of contacts; when the rising slope reaches a preset slope threshold, moving the probe to a cleaning station and applying a self-cleaning impact to the probe in the same manner as in S3, wherein the energy of the self-cleaning impact is higher than the energy required for the tip to penetrate the oxide layer; and recalibrating the energy baseline of the supplementary impact using a reference contact point after the self-cleaning impact is completed.

4. The wafer inspection probe variable speed impact contact method according to claim 1, characterized in that, S2 includes the following sub-steps: S21. Acquire the position feedback and velocity feedback of the probe; S22. When it is determined from the position feedback and velocity feedback that the probe has reached zero velocity and the residual oscillation is lower than a preset oscillation threshold, confirm that the probe has entered the impact-ready state; S23. After confirming that the probe has entered the impact-ready state, trigger S3.

5. The wafer inspection probe variable speed impact contact method according to claim 1, characterized in that, The second speed is adjusted according to the thickness of the oxide layer or the material of the pad, so that the momentum of the needle tip is matched with the energy required to penetrate the oxide layer.

6. A variable-speed impact contact device for wafer inspection probes, characterized in that, include: A servo impact mechanism is used to drive a probe to move towards the pad of the wafer under test at a controllable speed. A controller is connected to the servo impact mechanism and configured to: control the probe to slowly approach the pad at a first speed to a preset pause position; pause at the pause position to eliminate residual oscillations generated by the slow approach; control the servo impact mechanism to drive the probe to impact the pad at a second speed greater than the first speed, causing the probe tip to penetrate the oxide layer on the surface of the pad with kinetic momentum and contact the metal layer below the oxide layer, the penetration occurring when the probe displacement reaches the static overdrive depth. Previously, the motion momentum of the needle tip was used; and after the needle tip penetrated the oxide layer, the contact pressure between the needle tip and the metal layer was maintained; the controller is also configured to: measure the contact resistance between the needle tip and the metal layer after the needle tip penetrates the oxide layer; when the contact resistance is greater than a preset contact resistance threshold, without increasing the static overdrive depth of the probe, control the servo impact mechanism to apply a supplementary impact with progressively increasing energy to the pad, and remeasure the contact resistance after each application; when the contact resistance is less than or equal to the preset contact resistance threshold, stop applying the supplementary impact.

7. The wafer inspection probe variable speed impact contact device according to claim 6, characterized in that, The probe is a probe array containing multiple probes; the controller is configured to enable each probe in the probe array to independently trigger an impact at the moment of its first contact with the corresponding pad under the same nominal overdrive conditions, and to independently perform contact resistance measurement and supplementary impact application, and to stop when the contact resistance of each probe is less than or equal to the preset contact resistance threshold, so as to absorb the flatness distribution of the probe array.

8. The wafer inspection probe variable speed impact contact device according to claim 6, characterized in that, The controller is further configured to: monitor the energy sequence from the supplementary impact and the rising slope as the number of contacts changes; when the rising slope reaches a preset slope threshold, move the probe to the cleaning station and control the servo impact mechanism to apply a self-cleaning impact, wherein the energy of the self-cleaning impact is higher than the energy required for the needle tip to penetrate the oxide layer; after the self-cleaning impact is completed, recalibrate the energy baseline of the supplementary impact using the reference contact point.

Images