Defect automatic detection apparatus and focusing method for defect detection

Abstract

The application provides a focusing method for defect detection, comprising the following steps: placing a to-be-detected substrate on a workpiece table and making a center point of the to-be-detected substrate be at an optimal focal plane of a detection objective lens; acquiring vertical heights of a plurality of sampling points by a vertical measurement sensor; performing polynomial fitting on all sampling points of the to-be-detected substrate, and outputting all valid sampling point data of the measured fitting substrate surface type; fitting the substrate surface type according to the output valid sampling point data, and adjusting a vertical height parameter of a substrate area where a detection light spot is located in real time, or adjusting an X-direction tilt value, a Y-direction tilt value and the vertical height parameter in real time, so that the to-be-detected substrate as a whole is placed at the optimal focal plane of the detection objective lens. The vertical control of a multilayer high-order complex substrate surface type is realized, and the vertical control precision is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of lithography equipment measurement, in particular to a defect automatic detection device and a focusing method for defect detection. BACKGROUND

[0002] The lithography technology is widely used in the integrated circuit manufacturing process. The technology exposes the mask pattern to the photoresist through the lithography system. The lithography system finally determines the feature size of the integrated circuit, and its accuracy is extremely important to the lithography process. In order to obtain the best imaging effect, the substrate (such as a silicon wafer) coated with photoresist needs to be placed at the height of the best focal plane during exposure. However, due to the processing technology and other reasons, the surface of the substrate is not an ideal plane, and the surface fluctuation is sufficient to affect the quality of the exposure imaging of the lithography machine.

[0003] The optical automatic detection equipment is an important detection equipment widely used for semiconductor production line applications, which is used for defect detection in key process steps such as lithography, electroplating, etching, etc. It can provide defect identification, statistics, report and other functions, and can improve the control ability of the production line to product defects, which is of great significance to improve the process stability of the production line and improve the product yield.

[0004] The line array scanning defect detection equipment uses a line array time delay integration (TDI) scanning detection method, and realizes ultra-high resolution and yield through a high-speed large area array TDI camera. In the alignment link, the scanning alignment method is used to save time. In the review stage, the review method adopts two types of area array camera review and line array TDI camera review, and the optimal method is selected according to the specific use condition. In all processes, the closed loop method is used to complete the real-time focusing. The closed loop real-time focusing adjusts the vertical parameters of the workpiece table in real time according to the measurement results of the vertical measurement sensor, and realizes the real-time closed loop control of the best focal plane.

[0005] However, due to the large change of the substrate surface type, the height of different parts of the light spot at the same time in different areas on the substrate is not the same, or the height of the same part of the light spot at different times in different areas on the substrate is not the same, which leads to that only according to the best focal plane of the detection process determined by a certain area of the substrate at a certain time, the detection picture taken is blurred. On the other hand, if all the sampling points are directly used for substrate surface type fitting, the invalid data with too large or too small Z-direction height value in the sampling points is also used, especially when the substrate is checked in the later multi-layer process, which leads to the difference of the surface type of the layer to be fitted. SUMMARY

[0006] The purpose of the present application is to provide a defect automatic detection device and a focusing method for defect detection, to solve the problems of blurred detection picture of the measured substrate and the difference of the surface type of the layer to be fitted.

[0007] To solve the above technical problems, the application provides a focusing method for defect detection, comprising:

[0008] placing a substrate to be measured on a plane at a predetermined distance from a detection objective lens;

[0009] moving the substrate to be measured along an X direction or a Y direction by a predetermined distance, measuring the entire substrate surface profile of the substrate to be measured by a vertical measurement sensor, and obtaining the vertical height of a plurality of sampling points;

[0010] performing polynomial fitting on all sampling points of the substrate to be measured, calculating the fitting residual of the vertical height of the sampling points and the vertical height at the same position of the fitted substrate surface profile, and determining whether the fitting residual is less than a preset threshold value; if the fitting residual is less than or equal to the preset threshold value, determining that the measurement data of the sampling point is valid; if the fitting residual is greater than the preset threshold value, performing filtering processing and iterative determination until all the measurement data of the sampling points used for fitting the substrate surface profile are valid;

[0011] fitting the substrate surface profile according to the output valid sampling point data, and adjusting the vertical height parameter of the substrate area where the detection light spot is located in real time, or adjusting the X direction tilt value, or the Y direction tilt value and the vertical height parameter in real time, so that the real-time measured substrate area of the substrate to be measured is placed on the best focal plane of the detection objective lens.

[0012] Optionally, the filtering processing comprises:

[0013] if the fitting residual is less than or equal to the filtering preset threshold value, determining that the sampling point is valid, and recording and saving the sampling point data;

[0014] if the fitting residual is greater than the filtering preset threshold value, determining that the sampling point is invalid, and deleting the sampling point data.

[0015] Optionally, the recorded valid sampling points are re-fitted and filtered, and the iteration number is determined; if the iteration number is greater than a preset number, an error is reminded, and the measurement sampling points are increased; if the iteration number is less than or equal to the preset number, the loss percentage of the valid sampling points relative to all the sampling points is calculated, and when the loss percentage exceeds a constraint percentage, an error is reminded, and the measurement sampling points are increased.

[0016] Optionally, in the step of placing the substrate to be measured on the plane at a predetermined distance from the detection objective lens, the substrate to be measured is placed on a workpiece table, and the center point of the substrate to be measured is placed on the best focal plane of the detection objective lens.

[0017] Optionally, the workpiece table drives the substrate to be measured to move step by step along the X direction or the Y direction by a predetermined distance.

[0018] Optionally, the base surface type is fitted according to the output effective sampling points, the X-direction tilt amount, the Y-direction tilt amount and the waveform distribution of the fitting residual of the base surface type are calculated by the measurement model, and the vertical height parameter is calculated according to the fitting result.

[0019] Optionally, the difference between the set value and the real-time measurement value of the vertical measurement sensor is sent to a worktable vertical control mechanism, and the vertical height of the worktable is adjusted in real time by the worktable vertical control mechanism.

[0020] Optionally, after the measurement data for fitting the base surface type is all effective, N-order polynomial fitting is performed on all effective sampling points to obtain the vertical height fitting value of the all effective sampling points, wherein N is an integer greater than or equal to 1.

[0021] Optionally, the difference between the vertical height fitting value of the all effective sampling points and the initial optimal focal plane height value of the worktable is sent to a worktable vertical control mechanism, and the vertical height of the worktable is adjusted by the worktable vertical control mechanism.

[0022] Optionally, the difference between the vertical height fitting value of the sampling points and the initial optimal focal plane height value is sent to an optical system worktable vertical control mechanism, and the vertical height of the optical system is adjusted by the optical system vertical control mechanism.

[0023] Optionally, in the step of judging whether the fitting residual is less than a preset threshold, the three times variance of the fitting residual is used for judgment.

[0024] Based on the same inventive concept, the present application also provides a defect automatic detection device for executing the above-mentioned focusing method for defect detection, comprising:

[0025] A worktable for carrying a to-be-measured substrate;

[0026] An optical system comprising a detection objective lens for emitting a light spot;

[0027] A vertical measurement sensor for measuring the vertical height of the to-be-measured substrate;

[0028] An adjustment structure for adjusting the movement of the worktable or the detection objective lens;

[0029] The control unit is used for receiving the sampling point data measured by the vertical measurement sensor, performing polynomial fitting on all sampling points of the to-be-measured substrate, calculating fitting residuals of the vertical height of the sampling point and the vertical height at the same position of the fitted substrate surface type, and judging whether the fitting residuals are less than a preset threshold value; if the fitting residuals are less than or equal to the preset threshold value, it is determined that the measurement data of the sampling point is valid; if the fitting residuals are greater than the preset threshold value, filtering processing is performed and iterative judgment is performed until all the measurement data of the sampling points used for fitting the substrate surface type are valid; the substrate surface type is fitted according to the output valid sampling point data, the vertical height parameter of the substrate area where the detection light spot is located is adjusted in real time, or the X-direction tilt value, or the Y-direction tilt value and the vertical height parameter are adjusted in real time, so that the real-time measured substrate area of the to-be-measured substrate is placed on the best focal plane of the detection objective lens.

[0030] Optionally, the adjustment structure comprises:

[0031] The Rz rotating table is located below the workpiece table and is used for adjusting the tilt of the workpiece table.

[0032] The workpiece table horizontal control mechanism is located below the Rz rotating table and is used for controlling the X-direction movement and Y-direction movement of the workpiece table.

[0033] The vertical control mechanism is used for controlling the vertical movement of the workpiece table or the optical system.

[0034] Optionally, the vertical control mechanism comprises a workpiece table vertical control mechanism located between the Rz rotating table and the workpiece table horizontal control mechanism, and the workpiece table vertical control mechanism is used for controlling the vertical movement or Rx-direction movement and Ry-direction movement of the workpiece table.

[0035] Optionally, the vertical control mechanism comprises an optical system vertical control mechanism located below the vertical measurement sensor, and the optical system vertical control mechanism is used for controlling the vertical movement of the optical system.

[0036] Compared with the prior art, the present application has the following advantages:

[0037] The application provides a kind of defect automatic detection device and focusing method for defect detection, by vertical measurement sensor to the whole to be measured substrate surface type measurement, obtain the vertical height of different horizontal position of workpiece table, then calculate the fitting residual of the vertical height of sampling point and the vertical height at the same position of fitted substrate surface type, and judge whether the fitting residual is less than the preset threshold value, if the fitting residual is less than the preset threshold value, the measurement data of sampling point is valid, if the fitting residual is greater than the preset threshold value, filtering processing is carried out and iterative judgment is carried out until the sampling point data for fitting substrate surface type measurement is all valid;During online scanning detection process, according to the valid sampling point fitting substrate surface type output, the vertical actuator of workpiece table is adjusted in real time, so that the whole to be measured substrate is placed on the best focal plane of the detection objective lens;Realize the vertical control of multilayer high-order complex substrate surface type, improve the vertical control precision, improve the image quality, which is conducive to strengthening the final detection effect, so as to solve the problem of to be measured substrate detection picture blur and fitting layer surface type difference. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 is the defect detection device structure schematic diagram of the embodiment of the application;

[0039] Figure 2 is another defect detection device structure schematic diagram of the embodiment of the application;

[0040] Figure 3 is the focusing method flow chart for defect detection of the embodiment of the application;

[0041] Figure 4 is another focusing method flow chart for defect detection of the embodiment of the application;

[0042] In the figure,

[0043] 1-optical system;2-cross beam;3-vertical measurement sensor;4-to be measured substrate;5-workpiece table;6-Rz rotary table;7-workpiece table vertical control mechanism;7a-optical system vertical control mechanism;8-workpiece table horizontal control mechanism. DETAILED DESCRIPTION

[0044] The application provides a kind of defect automatic detection device and focusing method for defect detection, by vertical measurement sensor to the whole to be measured substrate surface type measurement, obtain the vertical height of different horizontal position of workpiece table, then calculate the fitting residual of the vertical height of sampling point and the vertical height at the same position of fitted substrate surface type, and judge whether the fitting residual is less than the preset threshold value, if the fitting residual is less than the preset threshold value, the measurement data of sampling point is valid, if the fitting residual is greater than the preset threshold value, filtering processing is carried out and iterative judgment is carried out until the sampling point data for fitting substrate surface type measurement is all valid;During online scanning detection process, according to the valid sampling point fitting substrate surface type output, the vertical actuator of workpiece table is adjusted in real time, so that the whole to be measured substrate is placed on the best focal plane of the detection objective lens;Realize the vertical control of multilayer high-order complex substrate surface type, improve the vertical control precision, improve the image quality, which is conducive to strengthening the final detection effect, so as to solve the problem of to be measured substrate detection picture blur and fitting layer surface type difference.

[0045] Figure 1 is the defect detection device structure schematic diagram of the embodiment of the application; Figure 1As shown, the embodiment provides a defect automatic detection device, which comprises an optical system 1, a vertical measurement sensor 3, a workpiece table 5, an adjusting structure and a control unit (not shown in the figure). The optical system 1 is used to emit a light spot, and the optical system 1 comprises a detection objective lens (not shown in the figure). The workpiece table 5 is used to carry a substrate 4 to be measured. The vertical measurement sensor 3 is fixed on a crossbeam 2, for example, and is used to measure the vertical height of the substrate 4 to be measured. The adjusting structure is used to adjust the movement of the workpiece table or the detection objective lens, and the adjusting structure comprises an Rz rotary table 6, a workpiece table vertical control mechanism 7 and a workpiece table horizontal control mechanism 8. The Rz rotary table 6 is located below the workpiece table 5, for example, and is used to realize the rotation of the workpiece table 5 around the Z axis. The workpiece table horizontal control mechanism 8 is located below the Rz rotary table 6, for example, and is used to control the movement of the workpiece table 5 in the X direction and the Y direction. The workpiece table vertical control mechanism 7 is located between the Rz rotary table 6 and the workpiece table horizontal control mechanism 8, for example, and is used to control the vertical movement of the workpiece table 5 to adjust the substrate 4 to be measured to be placed on the best focal plane. The control unit is used to receive the sampling point data measured by the vertical measurement sensor 3, perform polynomial fitting on all the sampling points of the substrate to be measured, calculate the fitting residual of the vertical height of the sampling point and the vertical height of the same position of the fitted substrate surface type, and judge whether the fitting residual is less than a preset threshold value. If the fitting residual is less than or equal to the preset threshold value, it is determined that the measurement data of the sampling point is valid. If the fitting residual is greater than the preset threshold value, filtering processing is performed and iterative judgment is performed until all the measurement data of the sampling points used for fitting the substrate surface type is valid. The substrate surface type is fitted according to the output valid sampling point data, and the vertical height parameter of the substrate area where the detection light spot is located is adjusted in real time, or the X direction tilt value, or the Y direction tilt value and the vertical height parameter are adjusted in real time, so that the real-time measured substrate area of the substrate to be measured is placed on the best focal plane of the detection objective lens.

[0046] The embodiment adopts the triangulation method to measure the vertical height of the substrate to be measured. Specifically, during vertical measurement, one side (for example, the left side) of the vertical measurement sensor 3 emits a measurement light spot, the light spot is incident on the upper surface of the substrate 4 to form a reflection, and the other side (for example, the right side) receives the measurement light spot reflected by the upper surface of the substrate 4. The vertical height of the substrate 4 to be measured is calculated by using the triangulation principle known to those skilled in the art.

[0047] During the whole vertical measurement process, the workpiece table horizontal control mechanism 8 performs horizontal x direction and y direction movement, and the vertical measurement sensor 3 measures the vertical height z of the corresponding workpiece table (x, y). During the measurement process, the workpiece table 5 does not need to move vertically. After the vertical measurement is completed, the scanning detection process is started.

[0048] During the scanning and detection process, the difference between the set value and the real-time measured value of the workpiece stage 5 in the vertical measurement sensor 3 is sent to the workpiece stage vertical control mechanism 7. Then, the workpiece stage vertical control mechanism 7 sends the difference between the set value and the real-time measured value of the workpiece stage 5 to the control unit, thereby realizing the closed-loop control of the workpiece stage vertical control mechanism 7.

[0049] Figure 2 This is a schematic diagram of another defect detection device according to an embodiment of the present invention. Figure 2 As shown, the workpiece stage vertical control mechanism 7 in the automatic defect detection device can also be set below the vertical measurement sensor 3. The workpiece stage vertical control mechanism 7 is used to control the optical system 1 to perform vertical movement, so as to adjust the optical system 1 so that the substrate 4 to be measured is placed in the optimal focal plane.

[0050] Figure 3 This is a flowchart of a focusing method for defect detection according to an embodiment of the present invention. Figure 3 As shown, combined with Figure 1 As shown, based on the same inventive concept, this embodiment also provides a focusing method for defect detection, including:

[0051] Step S10: Place the substrate to be tested in a plane at a predetermined distance from the detection objective lens; the substrate to be tested moves in steps along the X or Y direction at predetermined distances, and the vertical measurement sensor measures the entire surface shape of the substrate to be tested, acquiring multiple sampling points (X... i Y i The vertical height Z i .

[0052] Step S20, sampling all points (X) of the substrate under test. i Y i Z i Polynomial fitting is performed.

[0053] Step S30: Calculate the fitting residual between the vertical height of the sampling point and the vertical height at the same position as the fitted base surface shape, and determine whether the fitting residual is less than a preset threshold. If the fitting residual is less than the preset threshold, the measurement data of the sampling point is determined to be valid. If the fitting residual is greater than the preset threshold, filtering is performed and iterative judgment is performed until all sampling point data used for fitting the base surface shape measurement are valid.

[0054] Step S40: Fit the substrate surface shape based on the output effective sampling point data, and adjust the parameters of the X-direction tilt value Rx, Y-direction tilt value Ry, and vertical height Z of the substrate region where the detection spot is located in real time, so that the entire substrate under test is placed at the optimal focal plane of the detection objective lens.

[0055] Specifically, in step S10, placing the to-be-tested substrate at a plane at a predetermined distance from the detection objective lens includes placing the to-be-tested substrate on the workpiece table and making the center point of the to-be-tested substrate be at the best focal plane of the detection objective lens, or, when the to-be-tested substrate is not located at the best focal plane of the detection objective lens, adjusting the workpiece table so that the center point of the to-be-tested substrate is at the best focal plane of the detection objective lens. After the to-be-tested substrate 4 is loaded onto the workpiece table 5, the positional relationship between the substrate coordinate system wcs and the workpiece table coordinate system wscs is calculated according to the alignment model, and the correspondence between a point (x_wcs, y_wcs) in the substrate coordinate system and a point (x_wscs, y_wscs) in the workpiece table coordinate system satisfies formula (1):

[0056]

[0057] wherein Rw is a substrate upper sheet rotation angle; C wx is a substrate upper sheet x direction translation amount; C wy is a substrate upper sheet y direction translation amount. Thus, the coordinates of each point in the substrate coordinate system can correspond to different motion positions of the workpiece table. During the sampling point measurement process, the workpiece table 5 is stepped by nx and ny steps in the x direction and the y direction, respectively, at a certain step pitch, corresponding to a series of sampling points on the surface of the to-be-tested substrate 4. At each sampling point, the workpiece table vertical control mechanism 7 of the workpiece table 5 brings the process substrate 4 into the best focal plane of the center coordinate point of the imaging objective lens in the optical system 1, and keeps the height Z of the workpiece table 5 corresponding to the best focal plane unchanged, Rx=0, Ry=0, and the vertical measurement sensor 3 records the vertical height z i corresponding to the workpiece table (x i , y i ).

[0058] In step S20, according to the overall continuous and gradual change characteristics of the surface topography of the to-be-tested substrate 4, a series of sampling points (x i , y i , z i ) recorded by the vertical measurement sensor 3 are calculated by using formula (2), that is, the following polynomial surface fitting formula:

[0059] n order fitting formula:

[0060]

[0061] wherein z i is the fitted vertical height, a kl is the kth and lth order coefficients of the surface topography of the to-be-tested substrate, and n is the order of the model, k+l≤n. According to a series of sampling points (x i , y i , z i ) recorded by the vertical measurement sensor 3 during the measurement process, the polynomial coefficients akl .

[0062] First order surface type fitting expansion:

[0063] z i = a 00 +a 10 *x i +a 01 *y i (2_1) Second order surface type fitting:

[0064] z i = a 00 +a 20 *x i 2 +a 10 *x i +a 11 *x i *y i +a 02 *y i 2 +a 01 *y i (2_2) Third order surface type fitting:

[0065] z i = a 00 +a 30 *x i 3 +a 21 *x i 2 *y i +a 20 *x i 2 +a 10 *x i +a 12 *x i *y i 2 +a 11 *x i x*y i +a 03 *y i 3 +a 02 *y i 2 +a 01 *y i (2_3)

[0066] The fitting residual is calculated from the vertical height of the measured base fitted in formula (2) and the measured vertical height, and the fitting residual result is as follows:

[0067]

[0068] wherein z is the measured vertical height, is the surface fitting result, and ss is the actual fitting residual.

[0069] Taking the first-order fitting as an example,

[0070] ss i =z i -a 00 -a 10 *x i -a 01 *y i (3_1)

[0071] The step S30 further comprises a sub-step of:

[0072] In step S31, it is judged whether the three times variance of the fitting residual is less than or equal to a preset threshold value. If yes, it is considered that all the recorded measurement points are valid, and the measurement model result is output. If no, step S22 is executed.

[0073] In step S31, when making a specific judgment, the three times variance of the fitting residual is used for judgment.

[0074] In step S32, the recorded sampling points are subjected to fitting filtering. If the fitting residual is less than or equal to a filtering preset threshold value, it is judged that the recorded measurement points are valid, and the sampling point data is recorded and saved. Otherwise, it is judged that the recorded measurement sampling points are invalid, and are deleted.

[0075] In step S33, the iteration number is judged. If the iteration number exceeds a preset iteration number, an error is reminded, and the number of measurement points is increased. If the iteration number is less than or equal to the preset iteration number, step S24 is executed.

[0076] In step S34, the loss percentage p% of the valid sampling points relative to all the sampling points is counted. When the loss percentage exceeds a constrained percentage, an error is reminded, and the number of sampling points is increased. When the loss percentage is less than or equal to the constrained percentage, the recorded valid sampling points are subjected to fitting filtering until all the measurement sampling points for fitting the base surface are valid.

[0077] In the embodiment, taking the first order as an example, the three times variance ss1 of the residual is counted. If ss1 (ss i )≤d, wherein d is a preset threshold value set, it is considered that all the recorded measurement sampling points are valid, and the measurement model result is output. Otherwise, the first-order fitting residual is subjected to threshold filtering of d', and if ss iIf d'≤d, it is determined that the recorded measurement sampling point is valid, the sampling point is saved, otherwise, it is determined that the recorded measurement sampling point is invalid and is deleted. Iteration number is determined, and finally the loss percentage p% of the valid points relative to all measurement sampling points is calculated. When the loss percentage exceeds the constraint percentage, an error is reported and the number of sampling points is increased. When the loss percentage is less than or equal to the constraint percentage, the valid sampling points are fitted and filtered until all the measurement sampling points used for fitting the base surface type are valid.

[0078] In step S40, the first-order filter fitting is completed, the measurement model output is performed, and the calculated x-direction tilt value Rx and y-direction tilt value Ry of the measured base 4 and the fitting residual ss are output i The waveform distribution diagram of the fitting residual ss is output, and the vertical focus plane compensation value is set as the setting value of the vertical measurement sensor 3. During online operation, the difference between the setting value and the real-time measurement value of the workpiece table in the vertical measurement sensor 3 is sent to the workpiece table vertical control mechanism 7 for real-time closed-loop adjustment, and the x-direction tilt value Rx and the y-direction tilt value Ry are sent to the workpiece table vertical control mechanism 7 for real-time adjustment as the setting value of the vertical measurement sensor 3 according to the fitting and filtering results.

[0079] The vertical height of the measured base 4 is obtained by full-surface measurement, the waveform distribution diagram of the fitting residual ss is output according to the results of the first-order fitting and filtering model, and the height difference h of different process layers is determined according to the waveform distribution diagram of the fitting residual ss i The waveform distribution diagram of the fitting residual ss is output, and the height difference h of different process layers is determined according to the waveform distribution diagram of the fitting residual ss i The waveform distribution diagram of the fitting residual ss is output, and the height difference h of different process layers is determined according to the waveform distribution diagram of the fitting residual ss

[0080] In another embodiment, the workpiece table vertical control mechanism 7 is located between the workpiece table 5 and the workpiece table horizontal control mechanism 8, and the workpiece table vertical control mechanism 7 is used for focus adjustment and controls the vertical movement of the workpiece table. At this time, the workpiece table vertical control mechanism 7 cannot control the workpiece table to adjust the x-direction tilt value Rx and the y-direction tilt value Ry.

[0081] Figure 4 is another focus adjustment method flow chart for defect detection according to an embodiment of the present application. As shown in Figure 4 , the present embodiment also provides a focus adjustment method for defect detection, comprising:

[0082] In step S10, the measured base is placed on a plane at a predetermined distance from the detection objective lens; the measured base is moved in a predetermined distance step along the X direction or the Y direction, and the vertical measurement sensor measures the entire base surface of the measured base to obtain a series of sampling points (X i , Y i ) of the vertical height Z i .

[0083] In step S10, placing the to-be-measured substrate at a plane at a predetermined distance from the detection objective lens includes placing the to-be-measured substrate on the workpiece table and making the center point of the to-be-measured substrate be at the best focal plane of the detection objective lens, or, when the to-be-measured substrate is not located at the best focal plane of the detection objective lens, adjusting the workpiece table so that the center point of the to-be-measured substrate is at the best focal plane of the detection objective lens.

[0084] In step S20, polynomial fitting is performed on all sampling points (X i , Y i , Z i ) of the to-be-measured substrate.

[0085] In step S30, the fitting residual of the vertical height of the sampling point and the vertical height at the same position of the fitted substrate surface is calculated, and it is determined whether the fitting residual is less than a preset threshold value. If the fitting residual is less than the preset threshold value, the measurement data of the sampling point is valid. If the fitting residual is greater than the preset threshold value, filtering processing is performed and iterative determination is performed until all the measurement data of the sampling points used for fitting the substrate surface is valid.

[0086] In step S30, when the specific determination is performed, the three times variance of the fitting residual is used for determination.

[0087] In step S40a, polynomial fitting is performed according to the output valid sampling point data, and the parameter of the vertical height Z of the substrate area where the detection light spot is located is adjusted in real time, so that the to-be-measured substrate as a whole is placed at the best focal plane of the detection objective lens.

[0088] In step S40a, in this embodiment, taking the first order as an example, the first order filtering fitting is completed, and all the valid sampling points are counted, and the N-order polynomial fitting is performed on all the valid sampling points as follows:

[0089]

[0090] Wherein, C i,(m_i) is the coefficient, the coefficients C of the N-order polynomial are solved, and m and N are the orders of the polynomial.

[0091] Any control sampling point horizontal coordinate (x0, y0) is substituted into the above fitted polynomial, and the vertical height fitting value of the control sampling point is obtained as follows:

[0092]

[0093] The height setting value of the workpiece table vertical control mechanism 7 of the workpiece table 5 is as follows:

[0094] Z_set0= Zfit0 _BF_Z (6)

[0095] Wherein, BF_Z is the optimal focal plane height value, Z_set0 is the setting value sent to the vertical actuator 7 of the workpiece stage 5, and Zfit0 is the vertical height fitting value of the control sampling point.

[0096] In another embodiment, the optical system vertical control mechanism 7 is located below the vertical measurement sensor 3 and is used for focal plane adjustment to control the vertical movement of the autofocus optical system 1. That is, the optical system vertical control mechanism 7 is used to control the Z-axis movement of the optical system.

[0097] like Figure 2 and Figure 4 As shown, in this embodiment of the defect detection device using the imaging lens in optical system 1 as the vertical motion actuator, a focusing method for defect detection is also provided, including:

[0098] Step S10: Place the substrate to be tested in a plane at a predetermined distance from the detection objective lens; the substrate to be tested moves in steps along the X or Y direction at predetermined distances, and the vertical measurement sensor measures the entire surface shape of the substrate to be tested, acquiring a series of sampling points (X... i Y i The vertical height Z i .

[0099] In step S10, placing the substrate to be tested in a plane at a predetermined distance from the testing objective includes placing the substrate to be tested on the workpiece stage and placing the center point of the substrate to be tested in the optimal focal plane of the testing objective, or, when the substrate to be tested is not located in the optimal focal plane of the testing objective, adjusting the workpiece stage so that the center point of the substrate to be tested is in the optimal focal plane of the testing objective.

[0100] Step S20, sampling all points (X) of the substrate under test. i Y i Z i Polynomial fitting is performed.

[0101] Step S30: Calculate the fitting residual between the vertical height of the sampling point and the vertical height at the same position as the fitted base surface shape, and determine whether the fitting residual is less than a preset threshold. If the fitting residual is less than the preset threshold, the measurement data of the sampling point is valid. If the fitting residual is greater than the preset threshold, filter processing is performed and iterative judgment is performed until all sampling point data used for fitting the base surface shape measurement are valid.

[0102] In step S30, when making a specific judgment, three times the variance of the fitted residual is used for the judgment.

[0103] Step S40b: Perform polynomial fitting based on the output effective sampling point data, and adjust the vertical height Z parameter of the optical system in real time so that the substrate under test is placed at the optimal focal plane of the detection objective lens.

[0104] In step S40b, the first-order filter fitting is completed, and all effective sampling points are counted, and the N-order polynomial fitting is performed on all effective sampling points as follows:

[0105]

[0106] Wherein, C i,(m_i) is a coefficient, and the coefficients C, m, and N of the N-order polynomial are solved, and m and N are orders of the polynomial.

[0107] Any control sampling point horizontal coordinate (x0, y0) is substituted into the above-mentioned fitted polynomial, and the vertical height fitting value of the control sampling point is obtained as follows:

[0108]

[0109] The height setting value of the worktable vertical control mechanism 7 of the optical system 1 is issued as follows:

[0110] Z_set0_lens = Zfit0_BF_Z_lens (7)

[0111] Wherein, BF_Z_lens is the height value defined at the best focal plane of the vertical execution mechanism 7 of the optical system 1, and Z_set0_lens is the setting value issued to the vertical execution mechanism 7 of the optical system 1.

[0112] As can be seen from the above, in the defect automatic detection device and the focusing method for defect detection provided in the embodiment of the present application, the entire to-be-measured substrate surface type is measured by the vertical measurement sensor, and the vertical height of the worktable at different horizontal positions is obtained; the fitting residual of the vertical height of the sampling point and the vertical height at the same position of the fitted substrate surface type is calculated, and it is judged whether the fitting residual is less than a preset threshold value; if the fitting residual is less than the preset threshold value, the measurement data of the sampling point is valid; if the fitting residual is greater than the preset threshold value, the filtering processing is performed, and the iterative judgment is performed until all the sampling point data used for measuring the fitted substrate surface type are all valid; in the online scanning detection process, the substrate surface type is fitted according to the output valid sampling point data, and the vertical execution mechanism of the worktable is adjusted in real time, so that the to-be-measured substrate is placed on the best focal plane of the detection objective lens; the vertical control of the multi-layer high-order complex substrate surface type is realized, the vertical control precision is improved, the image quality is improved, and the final detection effect is improved; thus, the problems of blurred detection picture of the to-be-measured substrate and the difference in the surface type of the fitted layer can be solved.

[0113] The above description is only a description of the preferred embodiments of the present application, and does not limit the scope of the present application in any way. Any modification or modification of the present application made by a person skilled in the art according to the above disclosure is within the protection scope of the claims.

Claims

1. A focusing method for defect detection, characterized in that, include: Place the substrate to be tested in a plane at a predetermined distance from the detection objective lens; The substrate under test moves along the X or Y direction at a predetermined distance, and the vertical measurement sensor measures the entire surface shape of the substrate under test to obtain the vertical height of multiple sampling points; Polynomial fitting is performed on all sampling points of the substrate to be tested, the fitting residual between the vertical height of the sampling point and the vertical height at the same position of the fitted substrate surface is calculated, and it is determined whether the fitting residual is less than a preset threshold. If the fitting residual is less than or equal to a preset threshold, the measurement data of the sampling point is determined to be valid; if the fitting residual is greater than the preset threshold, filtering is performed and iterative judgment is made until all sampling point data used for fitting the base surface shape measurement are valid. The filtering process includes determining that the sampling point is invalid and deleting the sampling point data if the fitting residual is greater than the preset filtering threshold. The substrate surface shape is fitted based on the output effective sampling point data. The waveform distribution of the X-direction tilt value, Y-direction tilt value and fitting residual of the substrate surface shape is calculated and output through the measurement model. The vertical height parameter is statistically analyzed based on the fitting results. The vertical height parameter of the substrate area where the detection spot is located is adjusted in real time, or the X-direction tilt value, or the Y-direction tilt value and vertical height parameter are adjusted in real time, so that the real-time measured substrate area of ​​the substrate under test is placed at the optimal focal plane of the detection objective lens.

2. The focusing method for defect detection as described in claim 1, characterized in that, The filtering process further includes: If the fitting residual is less than or equal to the preset filtering threshold, the sampling point is determined to be valid, and the sampling point data is recorded and saved.

3. The focusing method for defect detection as described in claim 2, characterized in that, Re-fit and filter the valid sampling points that have been saved and determine the number of iterations; If the number of iterations exceeds the preset number, an error message will be displayed, and an additional measurement sampling point will be added. If the number of iterations is less than or equal to the preset number, the percentage loss of the valid sampling points relative to all sampling points is calculated. When the percentage loss exceeds the constraint percentage, an error message is displayed and an additional measurement sampling point is added.

4. The focusing method for defect detection as described in claim 1, characterized in that, The step of placing the substrate to be tested on a plane at a predetermined distance from the testing objective includes placing the substrate to be tested on the workpiece stage and placing the center point of the substrate to be tested at the optimal focal plane of the testing objective.

5. The focusing method for defect detection as described in claim 4, characterized in that, The workpiece stage drives the substrate under test to move in steps along the X or Y direction at a predetermined distance.

6. The focusing method for defect detection as described in claim 1, characterized in that, The difference between the set value and the real-time measured value of the vertical measurement sensor is sent to a workpiece stage vertical control mechanism, which then adjusts the vertical height of the workpiece stage in real-time through a closed-loop mechanism.

7. The focusing method for defect detection as described in claim 6, characterized in that, The difference between the fitted vertical height values ​​of all valid sampling points and the initial optimal focal plane height value of the workpiece stage is sent to the vertical control mechanism of the workpiece stage, and the vertical height of the workpiece stage is adjusted by the vertical control mechanism of the workpiece stage.

8. The focusing method for defect detection as described in claim 6, characterized in that, The difference between the vertical height fitting value of the sampling point and the initial optimal focal plane height value is sent to the vertical control mechanism of the optical system, and the vertical height of the optical system is adjusted by the vertical control mechanism of the optical system.

9. The focusing method for defect detection as described in claim 1, characterized in that, After all the measurement data used to fit the base surface shape are valid, an Nth-order polynomial fit is performed on all valid sampling points to obtain the vertical height fitting value of all valid sampling points, where N is an integer ≥1.

10. The focusing method for defect detection as described in claim 1, characterized in that, In the step of determining whether the fitting residual is less than a preset threshold, three times the variance of the fitting residual is used for the determination.

11. An automatic defect detection device, characterized in that, A focusing method for defect detection as described in any one of claims 1-10, comprising: The workpiece stage is used to support the substrate to be tested. An optical system, comprising a detection objective lens for emitting a light spot; A vertical measurement sensor is used to measure the vertical height of the substrate under test. An adjustment structure is used to adjust the movement of the workpiece stage or the detection objective lens; The control unit is configured to receive sampling point data measured by the vertical measurement sensor and perform polynomial fitting on all sampling points of the substrate under test. It calculates the fitting residual between the vertical height of the sampling point and the vertical height at the same position as the fitted substrate surface shape, and determines whether the fitting residual is less than a preset threshold. If the fitting residual is less than or equal to the preset threshold, the measurement data of the sampling point is deemed valid. If the fitting residual is greater than the preset threshold, filtering is performed and iterative judgment is made until all sampling point data used for fitting the substrate surface shape is valid. The filtering process includes determining that if the fitting residual is greater than the preset filtering threshold, the sampling point is invalid and the sampling point data is deleted. The control unit fits the substrate surface shape based on the output valid sampling point data, calculates and outputs the waveform distribution of the X-direction tilt value, Y-direction tilt value, and fitting residual of the substrate surface shape through a measurement model, and statistically analyzes the vertical height parameter based on the fitting result. It then adjusts the vertical height parameter of the substrate region where the detection spot is located in real time, or adjusts the X-direction tilt value, or the Y-direction tilt value, and the vertical height parameter in real time, so that the real-time measured substrate region of the substrate under test is placed at the optimal focal plane of the detection objective lens.

12. The automatic defect detection device as described in claim 11, characterized in that, The adjustment structure includes: The Rz rotary table, located below the workpiece stage, is used to enable the workpiece stage to rotate around the Z-axis. The workpiece stage horizontal control mechanism is located below the Rz rotary table and is used to control the movement of the workpiece stage in the X and Y directions. A vertical control mechanism is provided to control the workpiece stage or optical system to perform vertical movement, Rx direction movement, and Ry direction movement.

13. The automatic defect detection device as described in claim 12, characterized in that, The vertical control mechanism includes a workpiece stage vertical control mechanism, located between the Rz rotary table and the workpiece stage horizontal control mechanism. The workpiece stage vertical control mechanism is used to control the workpiece stage to perform vertical movement.

14. The automatic defect detection device as described in claim 12, characterized in that, The vertical control mechanism includes an optical system vertical control mechanism located below the vertical measurement sensor. The optical system vertical control mechanism is used to control the vertical movement of the optical system.

Images