Probe station and positioning precision compensation method thereof
Through the cooperation of bilinear interpolation algorithm and machine vision system, the problems of complex calibration and environmental deformation of probe tables are solved, and high-precision probe table positioning and testing are achieved, which improves production efficiency and system stability.
Patent Information
- Application Number
- CN202411252914.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-09
- Publication Date
- 2024-11-26
AI Technical Summary
During the calibration process, the existing wafer probe table has complex instrument installation, time-consuming and functional limitations, and changes in ambient temperature and humidity lead to deformation of the moving platform, affecting the accuracy and accuracy of the probe insertion into the pad.
The mapping relationship table is refined by a bilinear interpolation algorithm, combined with the machine vision system to detect and compensate for the needle marks, and convert the motion command coordinates through the inverse bilinear interpolation algorithm to realize high-precision positioning and testing of the probe table.
It significantly improves the positioning accuracy and testing accuracy of the probe table, ensures high-precision operation under different environmental conditions, improves the stability and production efficiency of the system, and reduces equipment maintenance costs.
Smart Images

Figure CN119024142A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of probe station positioning, and in particular to a probe station and a positioning accuracy compensation method thereof. Background Art
[0002] The wafer probe station is one of the indispensable key equipment in the semiconductor manufacturing process. It is mainly used to test the electrical characteristics of semiconductor wafers. The test principle is that the probe station guides the probe to contact the test point on the wafer, and obtains the chip electrical performance data by controlling the integrated tester. This test process is used to batch screen the chips before they are cut and packaged, which greatly affects the quality and efficiency of chip production. With the rapid development of semiconductor technology, the wafer size and circuit density continue to increase. Wafer probe station technology will evolve towards finer detection accuracy and faster test speed to meet the challenges brought by advanced packaging technologies such as 3D packaging and heterogeneous integration.
[0003] At present, in the research and development process of the fully automatic wafer probe station, by benchmarking the international advanced probe station competitors, a high-precision linear motor four-axis motion platform is adopted to achieve the micron-level positioning accuracy target. However, it still has the following shortcomings when applied: ① During the calibration process of the probe station, it was found that the existing calibration method relying on laser interferometer has defects such as complex instrument installation process, long calibration time, and limited functions, which cannot fully meet the complete calibration requirements. Moreover, when the probe station is in use, the motion platform has a certain degree of thermal sensitivity, which causes its calibration data to need to be corrected and updated according to the current situation of the equipment. ② During the probe test pad process, it was found that changes in ambient temperature and humidity will cause deformation of many components such as the motion platform shaft, optical components and their tooling structure, grating scale base, needle card chuck, etc., thereby affecting the accuracy and precision of the probe piercing the pad.
[0004] Therefore, how to improve the positioning accuracy of the probe station and achieve accurate probe testing of the electrical performance of the pad has become a technical problem that needs to be solved urgently. In view of this, the present invention is proposed. Summary of the invention
[0005] In order to overcome the above-mentioned defects, the present invention provides a probe station and a positioning accuracy compensation method thereof. The method is easy to operate, short time consumption, high positioning accuracy and probe test accuracy, strong operation reliability, high production efficiency, low production cost, and has broad application prospects and important significance.
[0006] The technical solution adopted by the present invention to solve the technical problem is: a positioning accuracy compensation method of a probe station, comprising the following steps:
[0007] S1: Calibrate the motion platform of the probe station;
[0008] Performing alignment operations on multiple feature points on the calibration object through the probe station to obtain multiple motion instruction coordinates of the motion platform of the probe station;
[0009] Based on the multiple motion command coordinates of the motion platform and the MAP map of the calibration object, a mapping relationship table is generated; wherein the MAP map is an ideal coordinate set of multiple feature points on the calibration object;
[0010] Refining the mapping relationship table by using a bilinear interpolation algorithm to perform incremental updates of the mapping relationship table; and converting the current motion instruction coordinate of the motion platform into an ideal coordinate in the MAP diagram by using an inverse bilinear interpolation algorithm to ensure the accuracy of the mapping relationship table;
[0011] S2: After the probe of the probe station is inserted into the test object, the needle mark on the test object is detected by the machine vision system of the probe station to obtain the position coordinates of the needle mark and its deviation value relative to the set center point coordinates on the test object;
[0012] S3: Based on the detection result of the needle mark, a linear interpolation algorithm is used to perform error compensation on the position coordinates of the needle mark.
[0013] As a further improvement of the present invention, a probe station performs an alignment operation on a plurality of feature points on a calibration object to obtain a plurality of motion instruction coordinates of a motion platform of the probe station; comprising:
[0014] A high-precision wafer with a diameter not greater than 2 mm is selected as the calibration object, and correspondingly, a plurality of grains arranged in rows and columns on the high-precision wafer are characteristic points of the calibration object;
[0015] The high-precision wafer is placed on the execution end of the motion platform, and the machine vision system is first used to perform preliminary alignment and position confirmation on the high-precision wafer. The machine vision system is then used to perform alignment operations on the grains on the high-precision wafer row by row and column by column, so as to obtain multiple motion instruction coordinates of the motion platform, and the multiple motion instruction coordinates of the motion platform correspond one-to-one to several of the grains respectively.
[0016] As a further improvement of the present invention, the machine vision system uses automatic optical alignment technology to perform preliminary alignment and position confirmation on the high-precision wafer, and performs alignment operations on the grains on the high-precision wafer;
[0017] The motion command coordinates of the motion platform are fed back in real time in a closed loop through the grating ruler feedback system of the probe station.
[0018] As a further improvement of the present invention, a gradient analysis is performed on the mapping relationship table to determine whether the data in the mapping relationship table changes continuously:
[0019] If the data in the mapping relationship table changes continuously, incremental update of the mapping relationship table is allowed;
[0020] If the data in the mapping relationship table changes, incremental update of the mapping relationship table is not allowed.
[0021] As a further improvement of the present invention, a bilinear interpolation algorithm is used to refine the MAP diagram in the mapping relationship table, that is, a bilinear interpolation algorithm is used to calculate the planning value of the motion instruction coordinates of the motion platform.
[0022] As a further improvement of the present invention, the bilinear interpolation algorithm uses four points in the MAP to perform interpolation, that is, according to Q 11 (x 1 ,y 1 ), Q 12 (x 1 ,y 2 ), Q 21 (x 2 ,y 1 ) and Q 22 (x 2 ,y 2 ) to calculate the value of the unknown function at point p = (x, y), including:
[0023] The first step is to perform two linear interpolation operations in the x direction to obtain:
[0024]
[0025] Among them, R 1 =(x,y 1 ), R 2 =(x,y 2 );
[0026] The second step is to perform a linear interpolation operation in the y direction, that is, through R 1 Point and R 2 The points are interpolated in the y direction to calculate point p:
[0027] Where p = (x, y).
[0028] As a further improvement of the present invention, the inverse bilinear interpolation algorithm uses the four current motion command coordinates of the motion platform for interpolation, that is, according to and Calculate the unknown function at point p * = the value at (x, y), including:
[0029] The first step is to and Perform interpolation operation according to coefficient u and get point; and and Perform interpolation according to the coefficient v and get point;
[0030] Where u∈[0,1];
[0031] Where v∈[0,1];
[0032] The second step is based on Dot and Calculate p * point:
[0033] Among them, v∈[0,1]; p * =(x,y).
[0034] As a further improvement of the present invention, the test object is a test wafer; accordingly, after the probe of the probe station penetrates into the pad of the grain on the test wafer, the machine vision system is used to obtain a picture of the test wafer, and the obtained picture is subjected to ROI slicing, binomial filtering, thresholding, connected domain, pad screening and needle mark processing in turn to obtain the position coordinates of the needle mark, the center point coordinates of the pad, and the deviation value between the two.
[0035] As a further improvement of the present invention, a linear interpolation algorithm is used to perform error compensation on the position coordinates of the needle mark, specifically including:
[0036] Select two reference dies on the test wafer that are in the same row or column as the designated die, and the sequence values of the two reference dies are x 0 and x 1 , and the deviation values of the needle mark position coordinates on the two reference grains relative to the center point coordinates of the pad are y 0 and 1 ;
[0037] The deviation value y of the needle mark position coordinates on the specified die relative to the center point coordinates of the pad is calculated by the following formula (1):
[0038]
[0039] Wherein, x is the sequence value of the specified grain.
[0040] The present invention further provides a probe station, which performs positioning accuracy compensation by using the positioning accuracy compensation method of the probe station as described in the present invention.
[0041] The beneficial effects of the present invention are as follows: on the one hand, the present invention innovates the calibration method of the motion platform of the probe station, especially in the calibration, the bilinear interpolation algorithm is used to refine the mapping relationship table, and the inverse bilinear interpolation algorithm is used to convert the current motion command coordinates of the motion platform into an ideal coordinate in the MAP diagram, so that the positioning accuracy of the motion platform can be significantly improved, which satisfies the high-precision positioning measurement requirements well; on the other hand, the present invention is based on the image processing algorithm of machine vision to efficiently and accurately detect the needle mark data, and the needle mark detection data is error compensated by using a linear interpolation algorithm, which can realize the electrical properties of the probe accurately testing the test object (such as the pad), and also ensure that the probe station can still maintain high-precision probe test operation under different environmental conditions or in long-term operation, thereby improving the test stability and accuracy of the entire system of the probe station. In short, the positioning accuracy compensation method of the probe station described in the present application is simple to operate, short in time, high in operating accuracy (including positioning accuracy and probe test accuracy), strong in reliability, high in production efficiency, and reduces equipment downtime and maintenance costs, and has a wide range of application prospects and important significance. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Figure 1 It is a schematic flow chart of the positioning accuracy compensation method of the probe station described in Example 1 of the present invention;
[0043] Figure 2 Schematic diagram of the process of step S1 in embodiment 1 of the present invention;
[0044] Figure 3 The schematic diagram of the point relationship of the mapping relationship table is a diagram of using a bilinear interpolation algorithm to refine the point relationship of the mapping relationship table in step S13 of embodiment 1 of the present invention;
[0045] Figure 4 Schematic diagram of the process of step S2 in Example 1 of the present invention;
[0046] Figure 5 This is a schematic diagram of the relationship between the pad position and the needle mark position obtained by using the minimum enclosing rectangle tool in step S2 of Example 1 of the present invention, wherein the blue box encloses the pad position and the red box encloses the needle mark position. DETAILED DESCRIPTION
[0047] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.
[0048] Embodiment 1:
[0049] Please see attached Figure 1 As shown, it is a flow chart of the positioning accuracy compensation method of the probe station of the present invention.
[0050] In the present application, the probe station preferably adopts a fully automatic probe station, which mainly includes a motion platform, a probe card, a machine vision system and a grating ruler feedback system, wherein the motion platform is used to drive the object to be tested to move, and its positioning accuracy directly affects the accuracy and production efficiency of the test results, so the motion platform usually has an XYZR four-axis motion function to drive the object to be tested to be able to perform linear motion in the XYZ three-axis direction and to rotate; in particular, in order to achieve precise position control and reliable repeatability, the XY axes of the motion platform are also preferably linear motors. The probe card is connected to the test instrument and is used to perform performance testing on the object to be tested. The positioning accuracy of the probe card inserted into the object to be tested also directly affects the accuracy of the test results. The machine vision system is used to collect the state information of the object to be measured, and it includes a hardware part and a software part, wherein the preferred configuration of the hardware part is: including a TUH100 series high-performance industrial microscope lens with a 10-fold magnification and a 1.3-megapixel CMOS industrial camera, so that the hardware part has a safer working distance and a pixel accuracy of 0.4um / pixel; the software part includes an image function library Halcon, whose main operator is shape-based template matching, which can provide pixel-level matching results, and the operation time is in milliseconds. The grating ruler feedback system is used to perform real-time closed-loop feedback on the motion command coordinates of the motion platform. It is understandable that the encoder pulse position in the grating ruler feedback system is the real-time motion command coordinates of the motion platform; in addition, for the grating ruler feedback system, the grating ruler material is generally selected to be the same or similar to the machine material of the probe station, so as to avoid the difference in the thermal expansion coefficients of the grating ruler and the machine, which leads to different expansion amplitudes of the two when the temperature changes, thereby causing relative motion between the two to affect the measurement accuracy.
[0051] Based on the above structure / functional configuration of the probe station, the positioning accuracy compensation method of the probe station described in the present application includes the following steps:
[0052] S1: Calibrate the motion platform of the probe station. For details, please refer to the attached Figure 2 As shown,
[0053] S11: performing an alignment operation on a plurality of feature points on the calibration object by the probe station to obtain a plurality of motion instruction coordinates of a motion platform of the probe station.
[0054] Specifically, the preferred operation method for realizing the alignment operation of the probe station on multiple feature points on the calibration object in the present application is:
[0055] S111: A high-precision wafer with a diameter not greater than 2 mm is selected as the calibration object. Accordingly, a plurality of rows and columns of dies on the high-precision wafer are characteristic points of the calibration object.
[0056] As is well known, the wafer is made of high-purity single crystal silicon, and its processing process includes a series of precise photolithography processes, such as: by coating a layer of photoresist on the surface of the wafer, and then forming a complex microelectronic circuit on the wafer after exposure, development, etching and other processing. This application is based on the characteristics of the wafer itself, and then by optimizing the wafer size, such as: requiring the diameter of the wafer to be no more than 2mm, it can be ensured that the dimensional accuracy of the high-precision wafer is much higher than the micron-level positioning accuracy requirement of the fully automatic probe station. In addition, according to the distribution of the grains on the wafer (the high-precision wafer) with equal row spacing and equal column spacing, it can be further ensured that the high-precision wafer is well applied to the fully automatic probe station for calibration.
[0057] S112: placing the high-precision wafer on the execution end of the motion platform, wherein the execution end of the motion platform can be understood as a loading platform connected to the output end of the R-axis (such as a rotating motor); first using the machine vision system to perform preliminary alignment and position confirmation on the high-precision wafer, and then using the machine vision system to perform alignment operations on the grains on the high-precision wafer row by row and column by column, so as to obtain multiple motion command coordinates of the motion platform. It can be understood that the multiple motion command coordinates of the motion platform correspond one-to-one to several of the grains.
[0058] Furthermore, in the present application, the machine vision system uses automatic optical alignment technology (Auto Optical Alignment) to perform preliminary alignment and position confirmation (Alignment) on the high-precision wafer, and to perform alignment operations on the grains on the high-precision wafer. Herein, the preliminary alignment and position confirmation of the high-precision wafer is used as an example for explanation: the machine vision system uses automatic optical alignment technology to capture (including edge finding, visual effects, matching, etc.) the surface features of the high-precision wafer, and then combines the MAP map of the high-precision wafer (which is stored in the control system of the fully automatic probe station) to achieve preliminary alignment and position confirmation of the high-precision wafer.
[0059] Note: Automatic optical alignment technology is a conventional technical means in the field of optical imaging technology, so it will not be described in detail here.
[0060] Furthermore, in the present application, the motion command coordinates of the motion platform are fed back in real time in a closed loop through the grating scale feedback system of the probe station, that is, after the machine vision system matches and aligns the unique features of each grain on the high-precision wafer one by one, the grating reading under the corresponding closed-loop range is recorded.
[0061] In addition, it can be seen from the above that after step S112 is completed to align the grains on the high-precision wafer, the straightness, linearity, and orthogonality of the XY axes of the motion platform are also detected and corrected, ensuring the high precision and reliability of the motion platform.
[0062] S12: Generate a mapping relationship table based on the multiple motion command coordinates of the motion platform and the MAP map of the calibration object; wherein the MAP map is an ideal coordinate set of multiple feature points on the calibration object.
[0063] Specifically, based on the fact that the calibration object adopts the above-mentioned high-precision wafer, the essence of this step S12 is to establish a one-to-one correspondence between the multiple motion instruction coordinates of the motion platform and the ideal coordinates of several grains on the high-precision wafer, and generate a table (i.e., the mapping relationship table). It can be understood that based on the mapping relationship table, basic data can be provided for subsequent correction and compensation operations.
[0064] In addition, after the mapping relationship table is generated, the present application also performs a gradient analysis on the mapping relationship table to determine whether the data in the mapping relationship table is continuously changing: ① If the data in the mapping relationship table is continuously changing, incremental updating of the mapping relationship table is allowed, that is, the following step S13 is allowed; ② If the data in the mapping relationship table jumps, incremental updating of the mapping relationship table is not allowed, that is, the following step S13 is not allowed.
[0065] S13: Refine the mapping relationship table by using a bilinear interpolation algorithm to perform incremental updates of the mapping relationship table, i.e., generate more mapping points to make the mapping relationship table more precise, so as to improve positioning accuracy; and, convert the current motion instruction coordinates of the motion platform into an ideal coordinate in the MAP diagram by using an inverse bilinear interpolation algorithm, so as to ensure the accuracy of the mapping relationship table and the effectiveness of error compensation.
[0066] Specifically, based on the data type in the mapping relationship table, the present application uses a bilinear interpolation algorithm to refine the MAP diagram in the mapping relationship table to generate more ideal coordinates. It can also be understood that: the present application uses a bilinear interpolation algorithm to calculate more planning values of the motion instruction coordinates of the motion platform.
[0067] Furthermore, the present application adopts a bilinear interpolation algorithm to refine the MAP diagram in the mapping relationship table. The bilinear interpolation algorithm uses four points in the MAP diagram for interpolation, that is, according to Q 11 (x 1 ,y 1 ), Q 12 (x 1 ,y 2 ), Q 21 (x 2 ,y 1 ) and Q 22 (x 2 ,y 2 ) to calculate the value of the unknown function at point p = (x, y), including: Figure 3 As shown,
[0068] The first step is to perform two linear interpolation operations in the x direction to obtain:
[0069]
[0070] Among them, R 1 =(x,y 1 ), R 2 =(x,y 2 );
[0071] The second step is to perform a linear interpolation operation in the y direction, that is, through R 1 Point and R 2 The points are interpolated in the y direction to calculate point p:
[0072] Where p = (x, y).
[0073] It can be understood that point p is the ideal coordinate that is subdivided, and is also the motion command coordinate planned for the motion platform. Based on the MAP diagram in the above-mentioned mapping relationship table, the motion platform can be better guided to accurately locate during the actual motion process, that is, the positioning accuracy of the motion platform during the actual motion process is greatly improved.
[0074] In addition, the method of using the inverse bilinear interpolation algorithm to calculate the current motion command coordinates of the motion platform is as follows: the inverse bilinear interpolation algorithm uses the four current motion command coordinates of the motion platform for interpolation, that is, according to and Calculate the unknown function at point p * = the value at (x, y), including:
[0075] The first step is to and Perform interpolation operation according to coefficient u and get point; and and Perform interpolation according to the coefficient v and get point;
[0076] Where u∈[0,1];
[0077] Where v∈[0,1];
[0078] The second step is based on Dot and Calculate p * point:
[0079] Among them, v∈[0,1]; p * =(x,y).
[0080] Understandable, * The point belongs to the ideal coordinate of a grain in the MAP. Based on the above inverse bilinear interpolation operation, more accurate error compensation can be achieved.
[0081] S2: After the probe of the probe station is inserted into the test object, the machine vision system of the probe station is used to detect the needle mark on the test object to obtain the position coordinates of the needle mark and its deviation value relative to the set center point coordinates on the test object.
[0082] Specifically, the test object described in the present application is a test wafer. Accordingly, after the probe of the probe station penetrates into the pad of the grain on the test wafer, the machine vision system is used to obtain a picture of the test wafer, and the obtained picture is sequentially subjected to ROI slicing, binomial filtering, thresholding, connected domain, pad screening and needle mark processing, and the position coordinates of the needle mark, the center point coordinates of the pad, and the deviation value between the two are obtained; the specific workflow can be found in the attached Figure 4 shown.
[0083] Note: ① Since the test object is a test wafer, when using the machine vision system to detect needle marks, it actually includes two tasks, namely: detecting the pad and detecting the needle mark; among them, by detecting the pad, the center point coordinates of the pad can be obtained, and by detecting the needle mark, the position coordinates of the needle mark can be obtained, and then the deviation value between the center point of the pad and the needle mark can be obtained.
[0084] ② The above-mentioned "ROI slicing, binomial filtering, thresholding, connected domain, pad and needle mark screening" and other processing are all conventional methods in image processing technology, so they will not be described in detail here, but only briefly explained as follows:
[0085] The ROI slicing is to segment the image of the test wafer to reduce the amount of calculation and effectively avoid the interference of complex background images outside the ROI.
[0086] The binomial filter is a filter that performs processing on the image after ROI slicing to effectively filter out noise in the image, reduce false detection and misdetection caused by noise, and facilitate subsequent feature screening and image processing. The binomial filter provides an effect similar to that of the Gaussian filter, but only uses integer operations, so the processing speed is much faster than that of the Gaussian filter.
[0087] The thresholding (also called binarization) is to perform thresholding on the image of the feature area to better detect the position of the pad and the needle mark. In addition, in this application, since the pad and the needle mark are adjacent to each other, a dividing grayscale value needs to be set to distinguish the pad and the solder point, that is, the needle mark threshold range is set to [0, threshold), and the pad threshold range is set to [threshold, 255].
[0088] The connected domain is to connect the valid adjacent pixels after binarization into components (Object) and return the component list.
[0089] The screening of pads and needle marks: It is characterized by the pixel area size, and uses the feature histogram tool to select valid areas from the component list. These valid areas include pad areas and needle mark areas. Since the pad area and the needle mark area overlap, this application preferably uses the minimum enclosing rectangle tool to obtain the pad location and the needle mark location (see the attached Figure 5 and then calculate the relative relationship between the two.
[0090] Note: The minimum enclosing rectangle tool can effectively avoid false detection caused by the overlap of the pad area and the needle mark area, and can more conveniently output the distance information from the needle mark to the pad boundary, so as to effectively determine whether the needle mark deviates from the center position of the pad, and the processing is very efficient.
[0091] S3: Based on the detection result of the needle mark, a linear interpolation algorithm is used to perform error compensation on the position coordinates of the needle mark.
[0092] Specifically, the method for using the linear interpolation algorithm to perform error compensation on the position coordinates of the needle mark in the present application is: select two reference grains (generally the two outermost grains) on the test wafer and located in the same row or column as the specified grain (which can be understood as the grain to be compensated for the error), and the sequence values of the two reference grains are x 0 and x 1, and the deviation values of the needle mark position coordinates on the two reference grains relative to the center point coordinates of the pad are y 0 and 1 ; Then, the deviation value y of the needle mark position coordinates on the specified die relative to the center point coordinates of the pad is calculated by the following formula (1):
[0093]
[0094] Wherein, x is the sequence value of the specified grain (a known quantity).
[0095] From the above, it can be seen that, on the one hand, the present application innovates the calibration method of the motion platform of the probe station, especially during calibration, a bilinear interpolation algorithm is used to refine the mapping relationship table, and an inverse bilinear interpolation algorithm is used to convert the current motion instruction coordinates of the motion platform into an ideal coordinate in the MAP diagram, so that the positioning accuracy of the motion platform can be significantly improved, and the high-precision positioning measurement requirements are well met; on the other hand, the present application uses an image processing algorithm based on machine vision to efficiently and accurately detect needle mark data, and uses a linear interpolation algorithm to perform error compensation on the needle mark detection data, which can enable the probe to accurately test the electrical properties of the test object (such as a pad), while also ensuring that the probe station can maintain high-precision probe testing operations under different environmental conditions or in long-term operation, thereby improving the test stability and accuracy of the entire probe station system.
[0096] Embodiment 2:
[0097] This embodiment 2 provides a probe station, and the probe station performs positioning accuracy compensation using the positioning accuracy compensation method of the probe station described in embodiment 1. It is understandable that the present application does not impose any restrictions on the specific structure, type, etc. of the probe station.
[0098] In summary, the positioning accuracy compensation method of the probe station described in the present invention is simple to operate, short time consumption, high positioning accuracy and probe test accuracy, strong operation reliability, high production efficiency, low production cost, and has broad application prospects and important significance.
[0099] Many specific details are described in the above description to facilitate a full understanding of the present invention. However, the above description is only a preferred embodiment of the present invention. The present invention can be implemented in many other ways different from those described herein, so the present invention is not limited to the specific implementation disclosed above. At the same time, any person familiar with the art can make many possible changes and modifications to the technical solution of the present invention using the methods and technical contents disclosed above without departing from the scope of the technical solution of the present invention, or modify it into an equivalent embodiment of equivalent changes. Any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of protection of the technical solution of the present invention.
Claims
1. A positioning accuracy compensation method for a probe station, characterized in that: The following steps are involved: S1: Calibrate the motion platform of the probe station; Performing alignment operations on multiple feature points on the calibration object through the probe station to obtain multiple motion instruction coordinates of the motion platform of the probe station; Based on the multiple motion command coordinates of the motion platform and the MAP map of the calibration object, a mapping relationship table is generated; wherein the MAP map is an ideal coordinate set of multiple feature points on the calibration object; Refining the mapping relationship table by using a bilinear interpolation algorithm to perform incremental updates of the mapping relationship table; and converting the current motion instruction coordinate of the motion platform into an ideal coordinate in the MAP diagram by using an inverse bilinear interpolation algorithm to ensure the accuracy of the mapping relationship table; S2: After the probe of the probe station is inserted into the test object, the needle mark on the test object is detected by the machine vision system of the probe station to obtain the position coordinates of the needle mark and its deviation value relative to the set center point coordinates on the test object; S3: Based on the detection result of the needle mark, a linear interpolation algorithm is used to perform error compensation on the position coordinates of the needle mark.
2. The positioning accuracy compensation method of the probe station according to claim 1, characterized in that: Performing alignment operations on multiple feature points on the calibration object through a probe station to obtain multiple motion instruction coordinates of a motion platform of the probe station; including: A high-precision wafer with a diameter not greater than 2 mm is selected as the calibration object, and correspondingly, a plurality of grains arranged in rows and columns on the high-precision wafer are characteristic points of the calibration object; The high-precision wafer is placed on the execution end of the motion platform, and the machine vision system is first used to perform preliminary alignment and position confirmation on the high-precision wafer. The machine vision system is then used to perform alignment operations on the grains on the high-precision wafer row by row and column by column, so as to obtain multiple motion instruction coordinates of the motion platform, and the multiple motion instruction coordinates of the motion platform correspond one-to-one to several of the grains respectively.
3. The positioning accuracy compensation method of the probe station according to claim 2, characterized in that: The machine vision system uses automatic optical alignment technology to perform preliminary alignment and position confirmation on the high-precision wafer, and performs alignment operations on the grains on the high-precision wafer; The motion command coordinates of the motion platform are fed back in real time in a closed loop through the grating ruler feedback system of the probe station.
4. The positioning accuracy compensation method of the probe station according to claim 1, characterized in that: Performing gradient analysis on the mapping relationship table to determine whether the data in the mapping relationship table changes continuously: If the data in the mapping relationship table changes continuously, incremental update of the mapping relationship table is allowed; If the data in the mapping relationship table changes, incremental update of the mapping relationship table is not allowed.
5. The positioning accuracy compensation method of the probe station according to claim 1, characterized in that: The MAP diagram in the mapping relationship table is refined by using a bilinear interpolation algorithm, that is, the planning value of the motion instruction coordinates of the motion platform is calculated by using a bilinear interpolation algorithm.
6. The positioning accuracy compensation method of the probe station according to claim 5, characterized in that: The bilinear interpolation algorithm uses the four points in the MAP to interpolate, that is, according to Q 11 (x1,y1),Q 12 (x1,y2),Q 21 (x2,y1) and Q 22 The function values at the four points (x2, y2) calculate the value of the unknown function at the point p = (x, y), including: The first step is to perform two linear interpolation operations in the x direction to obtain: Where, R1=(x,y1), R2=(x,y2); The second step is to perform a linear interpolation operation in the y direction, that is, to interpolate points R1 and R2 in the y direction to calculate point p: Where p = (x, y).
7. The positioning accuracy compensation method of the probe station according to claim 1, characterized in that: The inverse bilinear interpolation algorithm uses the four current motion command coordinates of the motion platform to interpolate, that is, according to and Calculate the unknown function at point p * = the value at (x, y), including: The first step is to and Perform interpolation operation according to coefficient u and get point; and and Perform interpolation according to the coefficient v and get point; Where u∈[0,1]; Where v∈[0,1]; The second step is based on Dot and Calculate p * point: Among them, v∈[0,1]; p * =(x,y).
8. The positioning accuracy compensation method of the probe station according to claim 1, characterized in that: The test object is a test wafer; accordingly, after the probe of the probe station penetrates into the pad of the grain on the test wafer, the machine vision system is used to obtain a picture of the test wafer, and the obtained picture is subjected to ROI slicing, binomial filtering, thresholding, connected domain, pad screening and needle mark processing in turn, to obtain the position coordinates of the needle mark, the center point coordinates of the pad, and the deviation value between the two.
9. The positioning accuracy compensation method of the probe station according to claim 8, characterized in that: A linear interpolation algorithm is used to compensate the position coordinates of the needle mark for errors, specifically including: Select two reference die on the test wafer and located in the same row or column as the designated die, the sequence values of the two reference die are x0 and x1 respectively, and the deviation values of the needle mark position coordinates on the two reference die relative to the center point coordinates of the pad are y0 and y1 respectively; The deviation value y of the needle mark position coordinates on the specified die relative to the center point coordinates of the pad is calculated by the following formula (1): Wherein, x is the sequence value of the specified grain.
10. A probe station, characterized in that: The probe station performs positioning accuracy compensation using the positioning accuracy compensation method of the probe station according to any one of claims 1 to 9.
Citation Information
Cited By
Method for calibrating position between die bond camera and die bond head
CN119515964A
Wafer probe station real-time error compensation control system based on multi-sensor fusion
CN120010384A
Method, apparatus and system for controlling prober
KR103049665B1
