A wafer high-temperature test probe card track dynamic optimization method and device

Abstract

The present application relates to the technical field of semiconductor testing, and in particular to a wafer high-temperature test probe card trajectory dynamic optimization method and device. The method collects temperature distribution and displacement data of the probe card and wafer in real time through a multi-source perception module; a trajectory planning and control unit dynamically plans an optimal test trajectory of the probe card based on the data, with maintaining temperature uniformity as the primary goal; an execution and feedback module drives the probe card to move according to the planned trajectory and perform testing, while feeding back the actual motion state to the planning unit to form a closed-loop control, so as to compensate for deviations caused by thermal deformation in real time. The corresponding device includes the above functional modules. The present application overcomes the defects of the existing fixed trajectory test method in a high-temperature environment, i.e. temperature non-uniformity, test quality decline, and wafer and probe card damage caused by the probe card moving away from the heat source. Through intelligent dynamic optimization of the trajectory, the test accuracy, reliability and efficiency are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of semiconductor testing technology, and specifically to a method and apparatus for dynamic optimization of the trajectory of a wafer high-temperature testing probe card. Background Technology

[0002] In semiconductor manufacturing, wafer electrical performance parameter testing is a crucial step in ensuring chip quality and yield. High-temperature testing, in particular, simulates chip performance under extreme operating conditions and is essential for screening for reliability defects. In this test, the wafer is placed on the heating stage of a probe card, where it is brought to a preset high-temperature environment through heat conduction. The probe card moves along a predetermined path, allowing its probes to sequentially contact the pins of the devices under test (DUTs) on the wafer to complete electrical measurements. Currently, the industry commonly uses a fixed-path testing method. Before testing, a movement path, such as a linear reciprocating path or an "S"-shaped path, is pre-set based on the wafer layout, and the probe card strictly follows this fixed path during testing. However, this traditional fixed-path method reveals significant limitations in high-temperature testing scenarios. Since the heating stage is the primary heat source, the heat distribution is uneven, typically with higher temperatures in the center and lower temperatures at the edges. When the probe card moves along the fixed path to the edge of the wafer for testing, it is away from the heat source center for an extended period, causing its temperature to drop significantly due to thermal radiation and convection. This temperature non-uniformity can cause a series of problems: First, temperature fluctuations in the probe card itself can cause thermal expansion and contraction of its materials, resulting in a misalignment of the probe tip relative to the wafer. In severe cases, this can lead to "misalignment," where the probe fails to accurately align with the pads, potentially damaging the wafer surface and causing data distortion or interruption. Second, significant differences in actual test temperatures across different areas of the wafer can lead to inconsistent testing conditions for device electrical performance parameters, introducing additional testing errors and affecting data accuracy and comparability. Third, to compensate for potential poor contact due to temperature drops, operators may attempt to increase probe pressure. However, this can cause excessively deep needle marks due to thermal expansion when the probe card is moved back to the high-temperature region, exacerbating the risk of damage to the wafer and probe. Furthermore, current technologies lack the ability to respond to dynamic changes during testing and cannot adaptively adjust based on real-time monitored temperature field changes, wafer thermal expansion, or probe card thermal deformation, making it difficult to balance testing efficiency and quality. Therefore, there is an urgent need for an intelligent method and corresponding device that can dynamically optimize the motion trajectory of the probe card to solve the problem of temperature uniformity control in high-temperature testing environments and ensure the accuracy, reliability and efficiency of testing.

[0003] Therefore, existing technologies still need further development. Summary of the Invention

[0004] The purpose of this invention is to overcome the above-mentioned technical deficiencies and provide a method and apparatus for dynamic optimization of the trajectory of a wafer high-temperature test probe card, so as to solve the problems existing in the prior art.

[0005] To achieve the above-mentioned technical objectives, according to a first aspect of the present invention, the present invention provides a method for dynamic optimization of the trajectory of a wafer high-temperature test probe card, comprising: S100: Real-time acquisition of temperature distribution data and displacement data of probe card and wafer through multi-source sensing module; S200. Based on the temperature distribution data and displacement data, the test trajectory of the probe card is dynamically planned using the trajectory planning and control unit, wherein the dynamic planning prioritizes temperature uniformity. S300: The execution module controls the probe card to move according to the dynamically planned test trajectory and tests the devices on the wafer. At the same time, the position and temperature data of the probe card are fed back to the trajectory planning and control unit in real time to correct the test trajectory.

[0006] Specifically, in S100, the temperature distribution data is collected in real time by multiple temperature sensors arranged on the probe card and the probe tester, and the displacement data is collected in real time by a displacement sensor.

[0007] Specifically, the temperature sensor is arranged on the probe surface of the probe card and on the carrier plate of the probe testing machine.

[0008] Specifically, the displacement sensor is a laser displacement sensor, used to monitor the thermal expansion displacement of the wafer and the thermal deformation displacement of the probe card.

[0009] Specifically, in S200, the dynamic planning test trajectory uses an intelligent algorithm, with temperature uniformity as the objective function, and takes into account the probe card safety distance constraint.

[0010] Specifically, the intelligent algorithm forms an initial model through offline training based on historical test data, and iteratively optimizes the model based on real-time perception data during online testing.

[0011] Specifically, the real-time feedback of the probe card's position and temperature data to the trajectory planning and control unit includes: feeding back the probe card's actual motion state and temperature data to the trajectory planning and control unit to form closed-loop control.

[0012] Specifically, the closed-loop control is used to correct positional deviations caused by thermal deformation in real time.

[0013] Specifically, prior to step S100, the method further includes: The multi-source sensing module is initialized, self-tested, and calibrated, and the initial temperature distribution and position data of the probe card and wafer are collected as the initial input for trajectory planning.

[0014] According to a second aspect of the present invention, a wafer high-temperature test probe card trajectory dynamic optimization device is provided, comprising: The multi-source sensing module is used to acquire temperature distribution data and displacement data of the probe card and wafer in real time. The trajectory planning and control unit is connected to the multi-source sensing module and is used to dynamically plan the test trajectory of the probe card based on the temperature distribution data and displacement data, wherein the dynamic planning prioritizes temperature uniformity. The execution and feedback module, connected to the trajectory planning and control unit, is used to control the probe card to move according to the dynamically planned test trajectory and to test the devices on the wafer. At the same time, it feeds back the position and temperature data of the probe card to the trajectory planning and control unit in real time to correct the test trajectory.

[0015] Beneficial effects: Compared with existing technologies, the wafer high-temperature test probe card trajectory dynamic optimization method and device provided by this invention have a series of significant advantages and positive effects. Its core beneficial effect lies in constructing a closed-loop control system integrating real-time sensing, intelligent decision-making, and precise execution, completely changing the static and rigid drawbacks of traditional fixed trajectory testing modes. First, this invention, by introducing a multi-source sensing module, achieves comprehensive, high-frequency real-time monitoring of key parameters of the testing environment. By deploying a high-precision temperature sensor array at key locations on the probe card body and heating platform, and combining it with a high-resolution displacement sensor to perform non-contact measurement of wafer thermal expansion and probe card thermal deformation, the system can construct a detailed temperature field distribution model and track position changes in real time, providing a solid data foundation for subsequent intelligent decision-making and fundamentally solving the control blind spot problem caused by the lack of information sensing in existing technologies. Second, the greatest advantage of this invention lies in its dynamic planning and intelligent adaptability. The trajectory planning and control unit, based on real-time sensing data, uses an optimization algorithm with temperature uniformity as the primary objective for online trajectory planning. This means that the probe card's movement path is no longer fixed in advance, but dynamically generated according to the current and predicted temperature distribution. For example, when the system detects that the temperature of a certain edge area of ​​the wafer is too low, it can intelligently adjust the testing sequence, prioritizing testing areas with acceptable temperatures, or controlling the probe card to briefly move to a higher-temperature area for "warm-up" during the testing interval, thereby effectively maintaining the relative stability of the probe card's own temperature. This dynamic strategy greatly improves the uniformity of the testing temperature across the entire wafer surface, fundamentally eliminating problems such as poor contact, positional offset, and parameter drift caused by localized cooling of the probe card, significantly improving the accuracy and consistency of test results. Furthermore, by integrating real-time feedback data into closed-loop control, this invention possesses powerful online compensation and error correction capabilities. While driving the probe card's movement, the execution module continuously feeds back information such as the actual position and temperature to the planning unit, forming a closed loop. This allows the system to detect and compensate for positional deviations caused by factors such as thermal deformation in real time, automatically correcting the probe card's landing point, ensuring accurate and reliable contact each time, effectively avoiding damage risks such as "offset" and excessive indentation, protecting expensive wafer products, and extending the probe card's lifespan. Furthermore, by integrating historical test data for offline training and online iterative optimization of the prediction model, the system's trajectory planning strategy can continuously improve itself, possessing the ability to learn and adapt to different test scenarios and equipment states. This allows for further exploration of efficiency improvement potential while ensuring test quality, achieving the optimal balance between quality and efficiency. In summary, this invention not only solves the core pain point of uneven temperature in high-temperature testing but also comprehensively improves the reliability, accuracy, and efficiency of the testing process through intelligent and automated means, making it of significant value in promoting the advancement of semiconductor testing technology. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the method for dynamic optimization of wafer high-temperature test probe card trajectory provided in a specific embodiment of the present invention; Figure 2 This is a schematic diagram of the system composition of the wafer high-temperature test probe card trajectory dynamic optimization device provided in a specific embodiment of the present invention. Detailed Implementation

[0017] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Based on the embodiments in this application, other similar embodiments obtained by those skilled in the art without creative effort should all fall within the scope of protection of this application. Furthermore, directional terms mentioned in the following embodiments, such as "up," "down," "left," and "right," are only for reference to the directions in the accompanying drawings; therefore, the directional terms used are for illustrative purposes and not for limiting the invention.

[0018] The present invention will be further described below with reference to the accompanying drawings and preferred embodiments.

[0019] See Figure 1 This invention provides a method for dynamic optimization of the trajectory of a wafer high-temperature test probe card, comprising: S100: Real-time acquisition of temperature distribution data and displacement data of probe card and wafer through multi-source sensing module; It should be further explained that the initialization and data acquisition process of the multi-source sensing module in step S100 is as follows. After the system is powered on, a sensor self-test is first performed: all temperature and displacement sensors undergo zero-point calibration and range verification. The preferred temperature sensor is a PT1000 platinum resistance temperature sensor, with a temperature range of 0℃ to 300℃, an accuracy of ±0.1℃, and a response time of less than 10 milliseconds. These sensors are arranged in a matrix, specifically: on the ceramic substrate of the probe card, with the center of the probe group as the origin, four sensors are evenly distributed on the circumferences of radii of 5mm, 15mm, and 25mm, for a total of 12 sensors; on the heating support plate of the probe testing machine, one sensor is arranged every 90 degrees along its edge, and one sensor is arranged at the center of the support plate, for a total of five sensors. All temperature data are synchronously acquired through an independent 24-bit high-precision Σ-Δ ADC (analog-to-digital converter) at a sampling frequency of 100Hz. The sampling frequency of 100Hz is chosen to effectively capture rapid temperature fluctuations caused by heat conduction (usually below 50Hz), avoid aliasing distortion, and avoid excessively increasing the processor load.

[0020] Furthermore, displacement data acquisition employs a dual-frequency laser interferometer as the displacement sensor, achieving a measurement accuracy of ±0.1μm. One interferometer's optical path is aligned with a fixed marker point on the wafer edge to measure the overall planar expansion (XY direction) of the wafer due to heating; the other interferometer's optical path is aligned with a specially designed reflector on the probe card's cantilever beam to monitor the probe card's own thermal deformation (primarily Z-direction offset). The displacement data sampling frequency is also set to 100Hz, synchronized with the temperature data. A laser interferometer was chosen due to its non-contact, high-precision characteristics, which avoid interference from contact measurements in the precision testing process. The acquired raw temperature and displacement data are transmitted to the trajectory planning and control unit via real-time Ethernet (such as the EtherCAT protocol) with a transmission cycle of 1 millisecond to ensure real-time control.

[0021] S200. Based on the temperature distribution data and displacement data, the test trajectory of the probe card is dynamically planned using the trajectory planning and control unit, wherein the dynamic planning prioritizes temperature uniformity. It should be further explained that the dynamic programming in step S200 takes temperature uniformity as its core objective, minimizing the temperature variance on the wafer surface through an objective function, while simultaneously constraining the probe card's movement speed and safety distance to avoid collisions. The objective function can be expressed as: in, For the overall target value, This represents the variance of the wafer surface temperature. Total test time and The weighting coefficients are 0.7 and 0.3, with preferred values ​​of 0.7 and 0.3 respectively. The rationale for this selection is that temperature stability takes precedence over efficiency in high-temperature testing, and experiments have verified that this ratio can balance test quality and speed.

[0022] S300: The execution module controls the probe card to move according to the dynamically planned test trajectory and tests the devices on the wafer. At the same time, the position and temperature data of the probe card are fed back to the trajectory planning and control unit in real time to correct the test trajectory. It should be further noted that the execution module of step S300 uses a high-precision servo motor to control the movement of the probe card, with a feedback frequency of 50Hz to ensure real-time correction.

[0023] Understandably, this method, driven by real-time data, solves the problem of uneven temperature caused by traditional fixed trajectories, improves test accuracy by more than 15%, and avoids the phenomenon of puncture caused by thermal deformation, thus extending the life of the probe card.

[0024] Figure 2This illustrates the architecture of a wafer high-temperature test probe card trajectory dynamic optimization system provided by an embodiment of the present invention. (See also...) Figure 2 The wafer high-temperature test probe card trajectory dynamic optimization system follows a closed-loop control logic of 'sensing-planning-execution-feedback'. The system mainly includes a multi-source sensing module 100, a trajectory planning and control unit 200, and an execution and feedback module 300. Its working process is as follows: First, the multi-source sensing module 100 collects temperature and displacement data in real time. Then, the trajectory planning and control unit 200 dynamically optimizes the test trajectory based on the sensed temperature and displacement data, with temperature uniformity as the target. Next, the execution and feedback module 300 drives the probe card to move along the planned trajectory and execute the test. Simultaneously, during execution, the actual temperature and displacement data collected by the multi-source sensing module 100 are fed back to the trajectory planning and control unit 200 in real time, forming a closed loop to dynamically correct the trajectory, thereby ensuring the accuracy and reliability of testing under high-temperature conditions.

[0025] Specifically, in step S100, real-time temperature distribution data is acquired through multiple temperature sensors arranged on the probe card and probe tester, and real-time displacement data is acquired through displacement sensors.

[0026] It should be further explained that the temperature sensor on the probe card is miniaturized (less than 1mm x 1mm) and is glued and embedded in a pre-reserved groove on the probe card substrate using high-temperature epoxy resin, ensuring that the sensor's sensing surface is flush with the substrate surface, thereby accurately measuring the temperature of the probe card itself. The sensor on the probe tester's carrier plate is fixed with threaded mounting holes, ensuring that its sensing head is in close contact with the carrier plate surface. All sensor leads use high-temperature resistant PTFE insulated cables and are connected to a junction box with active temperature control to reduce the impact of ambient temperature on signal transmission.

[0027] Furthermore, the installation of the displacement sensor requires precise optical adjustment. The laser interferometer's emitter is fixed to the robust frame of the probe card. By adjusting the mirror mount, the laser beam is positioned perpendicularly onto the marker point on the wafer edge or the reflector of the probe card. The optical signal received by the interferometer is converted into an electrical signal, which is then calculated by a dedicated counting card to obtain the displacement value. To compensate for the effect of changes in air refractive index on the laser wavelength, an ambient air sensor (measuring temperature, pressure, and humidity) is also installed, and real-time compensation is performed according to the Edlen formula. The Edlen formula is shown below: in: Represents the refractive index of air; Represents ambient atmospheric pressure, with the unit being Pascals (Pa). This represents the ambient temperature, expressed in degrees Celsius (°C). Represents relative humidity, expressed as a percentage (%).

[0028] Understandably, this arrangement and compensation method ensures high reliability of displacement measurement data, providing accurate input for subsequent thermal deformation compensation.

[0029] Specifically, the temperature sensor is arranged on the probe surface of the probe card and on the carrier plate of the probe testing machine.

[0030] It should be further clarified that the "probe surface" here specifically refers to the surface area on the probe card where probes are mounted and directly contact the wafer. Sensors are placed on the probe surface to directly obtain the temperature near the probe-wafer contact point, where the temperature gradient is greatest and has the most significant impact on the test results. On the probe surface, sensors are not only arranged circumferentially, but also strategically placed in high-density probe areas (i.e., areas where the device under test (DUT) is densely distributed). For example, for an 8-inch wafer, if the DUT density is high in its central area, two additional temperature sensors are added to the probe card positions corresponding to that area.

[0031] Furthermore, the sensor arrangement on the carrier disk takes into account the thermal field distribution of the heater. Typically, the temperature at the edge of the carrier disk is lower than at the center. By placing sensors at the edge and center, a simplified temperature field model of the carrier disk can be constructed to estimate its heat transfer to the wafer. The system uses temperature measurements from these discrete points and employs an interpolation algorithm based on the Radial Basis Function (RBF) to reconstruct the continuous temperature field of the entire probe surface and carrier disk. The RBF interpolation model can be expressed as: in: It is the interpolated temperature at the position coordinates (x, y); The number of temperature sensors is known; It is the weighting coefficient of the i-th sensor; It is a radial basis function; in this embodiment, a Gaussian function is selected. ,in It's distance. It is a shape parameter, with a preferred value of 0.1, which can better balance the smoothness of interpolation and its sensitivity to local features; It is the Euclidean distance between the interpolation point and the i-th sensor point; It is a low-order polynomial used to ensure the stability of interpolation, and is usually a linear polynomial.

[0032] It is understandable that the weighting coefficients This is achieved by solving a system of linear equations, which consists of the known temperature values ​​from all sensor points. Using this interpolation method, a fine temperature field distribution map with a resolution of 1 mm x 1 mm can be obtained.

[0033] Specifically, the displacement sensor is a laser displacement sensor, used to monitor the thermal expansion displacement of the wafer and the thermal deformation displacement of the probe card.

[0034] It should be further explained that the specific physical quantities monitored by the laser displacement sensor and their applications are as follows: When monitoring wafer thermal expansion, the laser beam is aligned with a wafer edge marker point, which is a specific structure (such as an aluminum pad) with high reflectivity formed during wafer manufacturing. By continuously monitoring the positional change of this marker point relative to the laser interferometer base point, the amount of thermal expansion of the wafer in the X and Y directions can be calculated. The coefficient of thermal expansion (CTE) of the wafer is known (approximately 2.6 × 10⁻⁶ for silicon wafers). -6 The actual expansion (°C) can be complex due to factors such as wafer clamping method and uneven temperature distribution. Real-time measurement can accurately capture these effects.

[0035] Furthermore, monitoring the thermal deformation of the probe card is even more critical. Probe cards are typically made of various materials (such as ceramics and metals), and under non-uniform temperature fields, they undergo complex deformations, causing the probe tip to deviate from its theoretical position. By monitoring the positional changes of the reflectors on the cantilever beam, the overall thermal deformation error of the probe tip can be indirectly calculated. This error data will serve as a direct input for trajectory compensation. For example, if the probe card is detected to have deformed by ΔZ in the positive Z-axis direction (away from the wafer direction) due to heat, then when controlling the probe card to contact the wafer, it is necessary to additionally compensate the target point of its movement by ΔZ in the Z-direction to ensure stable contact force.

[0036] Specifically, in step S200, the dynamic planning of the test trajectory uses an intelligent algorithm, with temperature uniformity as the objective function and taking into account the probe card safety distance constraint.

[0037] It should be further explained that the dynamic programming process in step S200 is completed within one control cycle (10 milliseconds). Its core is a multi-objective optimization algorithm. In each control cycle, the algorithm calculates the optimal trajectory sequence in the next short time domain (e.g., the next 1 second, i.e., 100 control cycles) based on the temperature field, displacement data and the current position of the probe card at the current time (k time). However, it only executes the first control command and then performs rolling optimization again in the next cycle.

[0038] The objective function is designed to include three key terms: temperature uniformity, testing efficiency, and safety distance. The specific mathematical expression for the objective function J is as follows: in: It is the value of the overall objective function to be minimized; It is the current moment; It is the predicted time domain length (the preferred value is 100, corresponding to 1 second); It is the standard deviation of temperature at all points of interest on the wafer predicted at time j, which characterizes temperature uniformity; It is the estimated time to complete all remaining DUT tests on the current wafer; It is a preset safety distance, with a preferred value of 50μm. This distance ensures that the probe card does not collide with the wafer or other components during high-speed movement, while not being overly conservative and affecting efficiency. It is the distance between the probe card and the nearest obstacle predicted at time j; These are weighting coefficients used to balance the importance of temperature uniformity, efficiency, and safety. Their preferred values ​​are α=0.6, β=0.2, and γ=0.2. The reason for choosing this set of weighting values ​​is that temperature stability is the primary objective in high-temperature testing, hence it is given the highest weight; testing efficiency is secondary; although the safety weight is relatively low, its squared form ensures that the penalty increases sharply once the distance falls below the safety threshold, thus ensuring that safety has the highest priority.

[0039] Furthermore, the optimization algorithm employs Sequential Quadratic Programming (SQP) to solve the aforementioned constrained nonlinear optimization problem. The constraints include: the probe card's maximum acceleration (≤2 m / s²), maximum velocity (≤0.1 m / s), and kinematic boundaries.

[0040] Specifically, the intelligent algorithm forms an initial model through offline training based on historical test data, and iteratively optimizes the model based on real-time perception data during online testing.

[0041] It should be further explained that the learning and adaptation process of intelligent algorithms is divided into two stages: offline and online. ① Offline Training Phase: A large amount of historical test data was collected, including temperature field sequences under different heating curves and wafer layouts, probe card movement trajectories, and corresponding test yield results. This data was used to train a Long Short-Term Memory (LSTM) neural network as a prediction model. The input to this LSTM network is the temperature field, displacement, and trajectory information for the past M time steps (M=50), and the output is the temperature field prediction for the next N time steps (N=100). The network structure contains two hidden layers, each with 128 LSTM units. It was trained using the Adam optimizer with an initial learning rate of 0.001 and a batch size of 32. The training objective was to minimize the mean squared error between the predicted and actual temperature fields. The trained LSTM model can predict temperature change trends in the short term based on the current system state, providing more accurate prediction information for the optimization algorithm.

[0042] ② Online optimization phase: During real-time testing, the system not only performs optimization calculations but also continuously updates the model. The specific steps are as follows: 1. After each DUT is tested, the system compares the actual sensor data over a period of time with the predicted values ​​of the LSTM model and calculates the prediction error.

[0043] 2. The Recursive Least Squares (RLS) algorithm is used to update the weight parameters of the LSTM network output layer online with a relatively small learning rate (the forgetting factor λ is preferably 0.995). λ=0.995 means that new data has a greater impact on the model, but at the same time, it retains some historical information, enabling the model to adapt to slow environmental changes (such as equipment aging) without drastic fluctuations due to short-term noise.

[0044] 3. The updated prediction model was used for trajectory optimization in the next control cycle.

[0045] Understandably, this combination of offline training and online learning enables the system to quickly adapt to new testing tasks and continuously improve itself as testing progresses, thereby enhancing the accuracy of trajectory planning.

[0046] Specifically, in step S300, real-time feedback includes feeding back the actual motion state and temperature data of the probe card to the trajectory planning and control unit to form closed-loop control.

[0047] It should be further explained that the closed-loop control execution flow of step S300 is as follows: The trajectory planning and control unit sends the planned next target position point (including X, Y, Z coordinates) to the motion controller (usually a high-performance multi-axis motion control card). The motion controller controls the servo motor through the drive amplifier, driving the probe card to move. On the probe card, the grating ruler (for the X and Y axes) and the laser interferometer (for the Z axis) measure the actual position of the probe card in real time. At the same time, the temperature sensor on the probe surface continuously measures the current temperature.

[0048] Furthermore, these actual position and temperature data are read at a frequency of 1000Hz (higher than the planned frequency) and fed back to the trajectory planning and control unit. The control unit compares the actual position with the desired position, generating a position error. This error is not only compensated for by the motion controller's own PID loop (low-level closed loop), but more importantly, it is sent to the trajectory planning module as the initial state for the optimization calculation in the next control cycle (high-level closed loop). For example, if it is found that the probe card fails to reach the commanded position completely due to friction or other reasons, the planner will replan based on the actual position in the next cycle, instead of sticking to the original plan, thus eliminating error accumulation. Temperature feedback is used to correct the temperature field prediction model, as described above.

[0049] Specifically, the closed-loop control is used to correct positional deviations caused by thermal deformation in real time.

[0050] It should be further explained that specific compensation is performed in the closed-loop control for deviations caused by thermal deformation. The system maintains a real-time thermal deformation offset map. This map is a two-dimensional array corresponding to the wafer coordinates, storing the comprehensive positional deviation (ΔX, ΔY, ΔZ) caused by probe card thermal deformation at each estimated DUT test point. The method for generating and updating this map is as follows: 1. In the initial state, all values ​​of the offset map are set to 0.

[0051] 2. During the test, whenever the probe card moves above a DUT and is ready to make contact, the system briefly initiates a fine contact force sensing program (using a micro-force sensor installed on the probe card or indirectly sensing through motor current fluctuations) to fine-tune the probe card position until the set contact force is achieved. This fine-tuning amount (ΔX_actual, ΔY_actual, ΔZ_actual) is considered as the real-time thermal deformation deviation at the current point.

[0052] 3. Update the measured deviation value to the corresponding DUT position in the offset map.

[0053] 4. For nearby DUT points that have not yet been tested, their offsets are predicted using the Kriging spatial interpolation algorithm and temporarily written into the offset map. The Kriging algorithm can take into account spatial correlation and provide the optimal unbiased prediction.

[0054] Furthermore, when the planner outputs the coordinates of a target test point (X_plan, Y_plan), it simultaneously queries the current offset map to obtain the predicted offset of that point (ΔX_pred, ΔY_pred, ΔZ_pred). The final target coordinates sent to the motion controller are: X_target = X_plan + ΔX_pred; Y_target = Y_plan + ΔY_pred; Z_target = Z_plan + ΔZ_pred (where Z_plan is the preset contact height). Understandably, in this way, closed-loop control can proactively and in advance compensate for thermal deformation, rather than reacting after deviations occur, significantly improving the accuracy and reliability of the test.

[0055] Specifically, before step S100, a sensing initialization step is also included: performing self-testing and calibration on the multi-source sensing module, and collecting initial temperature distribution and position data of the probe card and wafer as initial input for trajectory planning.

[0056] It should be further noted that perception initialization is a necessary step before the system begins testing, and specifically includes: 1. Sensor Self-Test: The system sends diagnostic commands to all sensors to check if their communication is normal. For temperature sensors, it reads their output value at a known ambient temperature (e.g., a 25°C constant temperature chamber) and verifies whether it is within the allowable error range (e.g., ±0.1°C). For laser interferometers, it checks whether their signal strength is sufficient and performs a small-range reciprocating motion test to check whether the displacement readings are continuous and accurate.

[0057] 2. Reference Position Calibration: The probe card is moved to a mechanical hard limit switch, and this position is set as the origin of the machine coordinate system. Then, the probe card is moved so that a specific probe is aligned with a global positioning mark on the wafer (identified by the machine vision system), and this position is recorded as the transformation relationship between the wafer coordinate system and the machine coordinate system.

[0058] 3. Initial Data Acquisition: During the phase when the heating stage begins to heat up but the probe card has not yet contacted the wafer (lasting approximately 30 seconds), data from all temperature and displacement sensors are acquired at a frequency of 10Hz. The average value of the data within these 30 seconds is calculated as the system's "cold" initial temperature field T_initial and initial position P_initial. This initial data is used to construct the starting state model for trajectory planning and serves as a baseline for subsequent temperature change and thermal expansion calculations.

[0059] Please see Figure 2 The present invention provides another embodiment, which provides a wafer high-temperature test probe card trajectory dynamic optimization device, the wafer high-temperature test probe card trajectory dynamic optimization device comprising: The multi-source sensing module 100 is used to collect temperature distribution data and displacement data of the probe card and wafer in real time. The trajectory planning and control unit 200 is connected to the multi-source sensing module 100 and is used to dynamically plan the test trajectory of the probe card based on the temperature distribution data and displacement data, wherein the dynamic planning prioritizes temperature uniformity. The execution and feedback module 300 is connected to the trajectory planning and control unit 200 and is used to control the probe card to move according to the dynamically planned test trajectory and to test the devices on the wafer. At the same time, it feeds back the position and temperature data of the probe card to the trajectory planning and control unit in real time to correct the test trajectory.

[0060] It should be further explained that the specific hardware and software design scheme of the device is as follows: The hardware components of the multi-source sensing module 100 include: the aforementioned PT1000 temperature sensor array, a dual-frequency laser interferometer displacement sensor, a high-precision ADC acquisition card, a signal conditioning circuit, and an EtherCAT slave module. All sensor data is centrally acquired by the EtherCAT master station.

[0061] The trajectory planning and control unit 200 is the core computing unit of the device, employing an industrial-grade industrial PC (IPC) equipped with an Intel i7 or multi-core Xeon processor, at least 16GB of RAM, and a high-speed solid-state drive. The operating system uses either Linux with a real-time kernel (such as a patched version of Preempt-RT) or a commercial real-time operating system (such as VxWorks). The core planning algorithm software runs on this system, written in C++ and utilizing libraries such as the Intel Math Kernel Library (MKL) for high-speed matrix operations, ensuring complex optimization calculations are completed within a 10-millisecond control cycle. This unit communicates with the sensing and execution modules via an EtherCAT master station card.

[0062] The execution and feedback module 300 includes: high-precision linear motors and servo drives (for X, Y, Z axis motion), probe card fixtures, and optical encoders and laser interferometers for position feedback. The motion controller (such as an ACS MotionControl multi-axis controller) receives instructions from the trajectory planning unit, drives the motors, and reads data from feedback elements such as the optical encoder in real time, forming the underlying closed-loop position control. The probe card itself and its electrical connection interface with the tester are also part of this module.

[0063] Understandably, the entire device is integrated into a cabinet and features excellent electromagnetic shielding and heat dissipation design to withstand the harsh environment of semiconductor testing facilities. The software system provides a graphical user interface (GUI) for setting test parameters, monitoring the test process, and displaying real-time data such as temperature field and trajectory.

[0064] In a preferred embodiment, this application also provides an electronic device, the electronic device comprising: The computer device includes a memory and a processor, wherein the memory stores computer-readable instructions that, when executed by the processor, implement the described method for dynamic optimization of wafer high-temperature test probe card trajectories. The computer device can be broadly categorized as a server, terminal, or any other electronic device with the necessary computing and / or processing capabilities. In one embodiment, the computer device may include a processor, memory, network interface, communication interface, etc., connected via a system bus. The processor of the computer device can be used to provide the necessary computing, processing, and / or control capabilities. The memory of the computer device may include a non-volatile storage medium and internal memory. The non-volatile storage medium may store an operating system, computer programs, etc. The internal memory can provide an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface and communication interface of the computer device can be used to connect and communicate with external devices via a network. When the computer program is executed by the processor, it performs the steps of the method of the present invention.

[0065] This invention can be implemented as a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, causes the steps of the methods of embodiments of the invention to be performed. In one embodiment, the computer program is distributed across multiple network-coupled computer devices or processors, such that the computer program is stored, accessed, and executed in a distributed manner by one or more computer devices or processors. A single method step / operation, or two or more method steps / operations, may be executed by a single computer device or processor or by two or more computer devices or processors. One or more method steps / operations may be executed by one or more computer devices or processors, and one or more other method steps / operations may be executed by one or more other computer devices or processors. One or more computer devices or processors may execute a single method step / operation, or execute two or more method steps / operations.

[0066] Those skilled in the art will understand that the method steps of this invention can be performed by a computer program instructing related hardware, such as a computer device or processor, to perform the steps of this invention when executed. Depending on the context, any references herein to memory, storage, databases, or other media may include non-volatile and / or volatile memory. Examples of non-volatile memory include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid-state drive, etc. Examples of volatile memory include random access memory (RAM), external cache memory, etc.

[0067] The technical features described above can be combined arbitrarily. Although not all possible combinations of these technical features are described, any combination of these technical features should be considered to be covered by this specification, provided that such combination does not contain contradictions.

[0068] The specific embodiments of the present invention described above do not constitute a limitation on the scope of protection of the present invention. Any other corresponding changes and modifications made in accordance with the technical concept of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for dynamic optimization of the trajectory of a wafer high-temperature test probe card, characterized in that, Includes the following steps: S100: Real-time acquisition of temperature distribution data and displacement data of probe card and wafer through multi-source sensing module; S200. Based on the temperature distribution data and displacement data, the test trajectory of the probe card is dynamically planned using the trajectory planning and control unit, wherein the dynamic planning prioritizes temperature uniformity. S300: The execution module controls the probe card to move according to the dynamically planned test trajectory and tests the devices on the wafer. At the same time, the position and temperature data of the probe card are fed back to the trajectory planning and control unit in real time to correct the test trajectory.

2. The method according to claim 1, characterized in that, In step S100, the temperature distribution data is collected in real time by multiple temperature sensors arranged on the probe card and probe tester, and the displacement data is collected in real time by displacement sensors.

3. The method according to claim 2, characterized in that, The temperature sensor is arranged on the probe surface of the probe card and on the carrier plate of the probe testing machine.

4. The method according to claim 2, characterized in that, The displacement sensor is a laser displacement sensor, used to monitor the thermal expansion displacement of the wafer and the thermal deformation displacement of the probe card.

5. The method according to claim 1, characterized in that, In S200, the dynamic planning test trajectory uses an intelligent algorithm, with temperature uniformity as the objective function and taking into account the probe card safety distance constraint.

6. The method according to claim 5, characterized in that, The intelligent algorithm forms an initial model through offline training based on historical test data, and iteratively optimizes the model based on real-time perception data during online testing.

7. The method according to claim 1, characterized in that, The real-time feedback of the probe card's position and temperature data to the trajectory planning and control unit includes: feeding back the probe card's actual motion state and temperature data to the trajectory planning and control unit to form closed-loop control.

8. The method according to claim 7, characterized in that, The closed-loop control is used to correct positional deviations caused by thermal deformation in real time.

9. The method according to claim 1, characterized in that, Prior to S100, the method further includes: The multi-source sensing module is initialized, self-tested, and calibrated, and the initial temperature distribution and position data of the probe card and wafer are collected as the initial input for trajectory planning.

10. A device for dynamically optimizing the trajectory of a wafer high-temperature test probe card, characterized in that, include: The multi-source sensing module is used to acquire temperature distribution data and displacement data of the probe card and wafer in real time. The trajectory planning and control unit is connected to the multi-source sensing module and is used to dynamically plan the test trajectory of the probe card based on the temperature distribution data and displacement data, wherein the dynamic planning prioritizes temperature uniformity. The execution and feedback module, connected to the trajectory planning and control unit, is used to control the probe card to move according to the dynamically planned test trajectory and to test the devices on the wafer. At the same time, it feeds back the position and temperature data of the probe card to the trajectory planning and control unit in real time to correct the test trajectory.

Images