JP2004177139A - Support program for preparation of inspection condition data, inspection device, and method of preparing inspection condition data - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an inspection condition data preparing support program for preparing in a short time an inspection condition data for an inspection device for detecting the position of a defect such as a foreign matter and a pattern abnormality in a substrate. <P>SOLUTION: This inspection condition data preparing support program is provided with following steps. Defect candidates are reviewed in a reviewed result display field 5, and a decided result is input as to a defect or a pseudo-defect. Symbols are displayed, while discriminating the defect and the a pseudo-defect, on a wafer map 2, a chip map and a defect feature distribution display field 7. A modification recommended parameter is calculated based on a distribution of feature amounts of a defect group and a pseudo-defect group. A modification recommended value is also calculated. Thereby, the preparation time for the inspection condition data is shortened, the operation time for the inspection device is increased, and the inspection frequency is improved. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method for inspecting a substrate having a fine circuit pattern such as a semiconductor integrated circuit, and more particularly to a technique for inspecting a substrate for foreign matter and pattern abnormalities in a manufacturing process.
[0002]
[Prior art]
An inspection of a semiconductor integrated circuit will be described as an example.
[0003]
A semiconductor integrated circuit is manufactured by repeating a process of transferring a pattern formed on a photomask onto a semiconductor wafer by film formation, lithography, and etching. Defects such as circuit pattern abnormalities and foreign matter occurring during the manufacturing process greatly affect the yield of semiconductor integrated circuits. Therefore, in a semiconductor integrated circuit manufacturing factory, an inspection process is provided in the middle of a manufacturing process in order to detect a defect early and take a measure.
[0004]
2. Description of the Related Art As an apparatus for inspecting a foreign substance present on a pattern of a semiconductor wafer and a pattern abnormality, there is a foreign substance inspection apparatus that irradiates a semiconductor wafer with a laser beam and finds the foreign substance based on a difference in intensity of scattered light. There is also a visual inspection device that irradiates a semiconductor wafer with visible light, ultraviolet light, or an electron beam, captures an image, compares the captured image with an image of a circuit pattern of an adjacent chip, and finds a foreign substance or a pattern abnormality.
[0005]
In order to use a foreign substance inspection device or a visual inspection device, inspection condition data corresponding to a circuit pattern and an inspection process is required, but it takes several hours to create the inspection condition data.
[0006]
In the following, the inspection condition data will be described using the appearance inspection apparatus as an example. The visual inspection apparatus detects a defect by comparing with a circuit pattern of an adjacent chip. In order to compare with the circuit pattern of the adjacent chip, the chip arrangement information in the wafer is registered in the inspection condition in advance. In addition, the visual inspection device reads the alignment marks on the wafer to accurately grasp the position and direction of the wafer. The position information of the alignment mark and the image of the alignment mark for this purpose are registered. In addition, the appearance inspection apparatus may detect color unevenness on a wafer, grains of metal wiring, and the like as defects. These are called pseudo defects because they do not affect the performance of the semiconductor integrated circuit. An area where many false defects are detected is usually set as a non-inspection area. Further, in the process of manufacturing a semiconductor, since the material and circuit pattern on the wafer surface differ depending on the inspection process and product type, the optimal intensity of visible light, ultraviolet light and electron beam in the inspection differs. In addition, a threshold value is provided for the luminance of the image to identify a defect. This threshold value also differs depending on the type and the inspection process.
[0007]
Among these inspection condition data, the threshold value that requires the longest time for setting will be described below. Representative threshold values used in the visual inspection device include a difference luminance threshold value set for a luminance difference between an image of the chip to be inspected (target image) and an image of the same coordinates of the adjacent chip (target image). And a size threshold set for the size of the defect. Further, another difference luminance threshold value may be provided only in the peripheral portion of the wafer where the variation in film thickness is large and a pseudo defect is easily detected.
[0008]
Usually, the threshold value is set as follows. First, a trial inspection is performed using the standard inspection condition data. The user observes the coordinates of the defect candidate detected by the visual inspection device with a microscope, and determines whether the defect candidate is a defect or a pseudo defect. Hereinafter, this work is called a review. If many pseudo defects are detected by the review, it is necessary to correct the inspection condition data. Therefore, one of these threshold values is corrected, and the test inspection is performed again. As a result, if the number of pseudo defects does not decrease, another threshold value is corrected. When the number of pseudo defects decreases, the threshold value is corrected again. In this way, the trial inspection and the adjustment are repeated several times, and the optimum threshold value is set. However, finding an optimum value for a plurality of threshold values by such trial and error is a time-consuming operation.
[0009]
In Patent Document 1 below, in order to reduce the number of reviews, a test is performed using a plurality of test condition data, a logical sum of the detected objects output is obtained, and a review is performed on the logical sum. A method is described in which a plurality of inspection conditions are evaluated in a single review to reduce the time required to generate the inspection conditions.
[0010]
Japanese Patent Application Laid-Open Publication No. 2003-189686 discloses an optical appearance inspection apparatus that uses polarized light for irradiation light, performs an inspection using a plurality of inspection condition data having different polarization conditions, and discriminates a defect from a pseudo result among the inspection results. A method for selecting an inspection condition with the highest value is described.
[0011]
[Patent Document 1]
JP 2001-250852 A
[Patent Document 2]
JP-A-2000-155509
[0012]
[Problems to be solved by the invention]
According to the method described in Patent Document 1 of the related art, the number of reviews can be reduced, but a plurality of trial inspections must be performed at a time, which does not lead to a reduction in inspection condition data creation time.
[0013]
In addition, the method described in Patent Document 2 also requires a trial inspection to be performed several times, and similarly does not reduce the inspection condition data creation time. Accordingly, it is an object of the present invention to provide a means for reducing the number of trial inspections for obtaining the optimum condition and shortening the time for generating inspection condition data.
[0014]
[Means for Solving the Problems]
A means for achieving the above object is an inspection condition data creation support program for an inspection apparatus for detecting a foreign substance or a pattern abnormality included in a wafer, wherein the inspection apparatus reads coordinate data and feature amount data of a defect candidate detected by the inspection apparatus. And a step of inputting a result of determining whether the defect candidate is a defect or a pseudo defect by a review, and classifying the defect candidate into a defect group and a pseudo defect group based on the defect determination result input in the step. And a step of outputting the classified defect group and the pseudo defect group with different symbols in a two-dimensional scatter diagram using a defect feature amount on one axis.
[0015]
More specifically, it is as shown in the claims.
[0016]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, an embodiment of the present invention will be described using an appearance inspection apparatus used in a semiconductor manufacturing process as an example.
[0017]
First, a defect detection algorithm for detecting a defect by the visual inspection device equipped with the inspection condition data creation support program of the present application will be described with reference to FIG. In FIG. 4, (1a) is a target chip image. The wiring 201a and the defect 202a exist in the target chip image of (1a). (1b) is a reference chip image which is an image adjacent to the target chip and having the same intra-chip coordinates as the target chip image. (1b) The wiring 201b exists in the reference chip image. (1c) is a difference image between the target chip image and the reference chip image. The operation for obtaining the difference image is as follows. (2a) is a luminance distribution on the line AA 'of the reference chip image. Here, the luminance is an electric signal obtained by A / D converting the brightness of the wafer surface. (2a) has a peak corresponding to the defect 202a and the wiring 201a. (2b) is a luminance distribution on the BB 'line segment of the reference chip image, and a peak corresponding to the wiring 201b exists. When the luminance distribution of (2b) is subtracted from the luminance distribution of (2a), the peak corresponding to the wiring 201 is canceled, and the peak corresponding to the defect remains, and the luminance distribution shown in (2c) is obtained. By performing this operation on the entire image, a difference image shown in (1c) is obtained.
[0018]
As shown in (2c), the size of the defect is obtained by measuring the range where the difference luminance continuously exceeds the threshold value.
[0019]
However, in practice, the difference luminance does not become 0 even at coordinates where no defect is present due to variations in the thickness of the wafer surface film and the like. Therefore, a difference luminance threshold value is provided as shown in (2c), and a portion where the peak of the difference luminance exceeds the difference luminance threshold value is recognized as a defect. Further, at the edge of the wiring or the like, a large difference in luminance is generated due to a slight shift of the coordinates of the target chip image and the reference chip image. In such a case, the range in which the difference luminance exceeds the difference luminance threshold is narrow. Therefore, erroneous detection is suppressed by providing a size threshold value. In addition, a luminance threshold is provided for luminance in order to remove the influence of electrical noise.
[0020]
The feature amounts of defects that are important in the above algorithm are the peak of the difference luminance and the peak and magnitude of the luminance. Hereinafter, the peak of the difference luminance is defined as the difference luminance of the defect, and the peak of the luminance is defined as the luminance of the defect.
[0021]
The defect detection device detects a pseudo defect when setting the threshold value incorrectly. We predicted that there was a difference in the distribution of these feature amounts between the defect group and the pseudo defect group, and thought that an optimal threshold could be set by analyzing this difference.
[0022]
Therefore, we examined a method of visually recognizing the difference in the feature distribution between the defect group and the pseudo defect group. As a result, it was found that a two-dimensional scatter plot in which the vertical axis is the feature amount of the defect and the horizontal axis is the space coordinate axis is optimal for visually recognizing the difference between the pseudo defect and the defect distribution. In some inspection processes, the occurrence rate of pseudo defects is high on the outer peripheral side of the wafer. However, when the square of the distance from the center of the wafer is selected as the horizontal axis, this tendency is clearly displayed. Based on the above analysis, an inspection condition data creation support program having a user interface for displaying a two-dimensional scatter diagram using a defect feature amount as one axis and a square of a distance from a wafer center as another axis. Was invented.
[0023]
Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
[0024]
FIG. 2 shows an example of the appearance inspection device 30 of the present invention.
[0025]
The visual inspection device 30 includes a defect inspection unit 31, a control unit 32, a calculation unit 33, a main storage device 34, a secondary storage device 35, a network interface 36, a keyboard 37, a mouse 38, and a monitor 39. It is composed of The secondary storage device 35 stores an inspection result 100, a defect determination result file 101, an inspection condition data 102, a t-test result file 103, a program 104 of the present application, and a standard inspection condition data file.
[0026]
FIG. 3 shows an example of the defect inspection unit 31. The defect inspection unit 31 includes a stage 152, a lamp 155, a sensor 156, an A / D (Analog / Digital) converter 157, a temporary image memory 158, an image processing unit 159, a spectral plate 160, and a lens 161.
[0027]
The wafer 151 is mounted on a stage 152, and is moved by the stage 152 to the front, rear, left and right. In the wafer 151, a reference chip 154 is a chip to be compared in the inspection of the target chip 153. Similarly, the image capturing location corresponding to the target chip image capturing location 153a on the target chip 153 is the reference chip image capturing location 154a.
[0028]
The operation of the defect inspection unit 31 is as follows. First, the light emitted from the lamp 155 is reflected in the 90 ° direction by the spectral plate 160 and is irradiated on the wafer 151 via the lens 161. The sensor 156 captures an image of the reference chip image capturing location 154a via the lens 161 and the spectral plate 160. The image of the reference chip is converted into luminance by the A / D converter 157 and stored in the temporary image memory 158. Subsequently, the stage 152 moves to capture an image of the target chip image capturing location 153a. The target chip image is converted into luminance by the A / D converter 157. Subsequently, the image processing unit 159 processes the luminance of the target chip image and the luminance of the reference chip image stored in the temporary image memory 158 to detect a defect. The coordinates and feature amount of the detected defect are stored in the inspection result file 100 inside the secondary storage device 35.
[0029]
Next, FIG. 7 shows an example of the inspection result file 100. The inspection result file 100 describes the defect candidate number of the defect candidate detected by the defect inspection unit 31, the defect candidate coordinates on the wafer surface, the difference luminance, the luminance, and the size of the defect candidate.
[0030]
Next, FIG. 8 shows an example of the defect determination result file 101. In the defect determination result file 101, each defect candidate is uniquely assigned a defect number. For each defect candidate, 1 is described as a determination result when the defect is a defect, 2 when the defect is a pseudo defect, and 0 when no determination is performed.
[0031]
FIG. 9 shows an example of the inspection condition data file 102. The inspection condition data file 102 describes a wafer size, the number of chips on the wafer, and a chip pitch which is a distance from an adjacent chip. In addition, the position of the alignment chip, the coordinates of the alignment mark, and the alignment image name are described. Further, a non-inspection area is described. Further, as detection conditions, an irradiation light amount, a luminance threshold, a difference luminance threshold, a peripheral difference luminance threshold, and a magnitude threshold are described.
[0032]
Next, an example of a GUI (Graphic User Interface) of the program of the present application will be described with reference to FIG.
[0033]
The GUI 1 is displayed on the monitor 39 of the inspection device 30. The GUI 1 includes a wafer map 2, a legend display column 3, a chip map 4, a review column 5, a pseudo defect ratio display column 6, a defect feature selection column 7, a defect feature distribution display column 8, a target pseudo defect number input column 11, and a calculation. It comprises a button 12, a change recommended parameter display field 13, a recommended value display field 14, and a fully automatic button 22.
[0034]
The review column 5 is used for a review in which the user determines whether or not the defect candidate detected by the defect inspection device 30 is a defect. A review image 16 is displayed in the review column 5, and a defect button 17 and a pseudo defect button 18 are arranged.
[0035]
In the wafer map 2, the chip map 4, and the defect feature amount distribution display column 8, the defect, the pseudo defect, and the defect candidate which has not been reviewed are displayed by different symbols. As shown in the legend display column 3, each symbol is displayed according to the defect symbol 19, the pseudo defect is displayed according to the pseudo defect symbol 20, and the defect candidate which has not been reviewed is displayed according to the non-review symbol 21.
[0036]
The user first creates an inspection condition data file 102a using the standard inspection condition data file 105, and performs an inspection using the inspection condition data file 102a. When the inspection is completed, the inspection result is displayed on GUI1. At this time, the display of all the defect candidates is the no-review symbol 21.
[0037]
Next, the cursor 15 is moved to a defect candidate displayed on the wafer map 2 by operating the mouse 38, and the defect is selected by clicking. Then, the review image 16 of the selected defect candidate is displayed in the review column 5. The user determines from the review image 16 whether the defect candidate is a defect or a pseudo defect. The determined result is input by clicking the defect button 17 or the pseudo defect button 18. The user reviews the defect candidates in order and inputs the determination result. The input judgment result is described in a defect judgment result file 101 in the secondary storage device 35.
[0038]
The displays in the wafer map 2, the chip map 4, and the defect feature distribution display column 8 are updated in accordance with the input of the determination result. In the defect feature amount distribution display column 8, the difference brightness is displayed as a defect feature amount on the vertical axis and the squared distance from the wafer center is displayed on the horizontal axis by default. From the display of the two-dimensional scatter diagram, the user visually judges whether or not the difference between the distribution of the defect group and the distribution of the pseudo defect group is significant. The user selects a defect feature to be displayed and a horizontal axis from the defect feature selection column to determine whether or not the difference between the defect group and the pseudo defect group is significant for other defect features. .
[0039]
Next, the user inputs the maximum allowable number of pseudo defects in the target number of pseudo defects input field 11. When the calculation button 12 is selected with the cursor 15 and clicked with the mouse 38, the recommended change parameters are obtained according to the algorithm shown in FIG. 3, and the recommended change parameters are displayed in the recommended change parameter display column 13. In addition, a recommended value is obtained in accordance with the algorithm shown in FIG. 4, and a recommended parameter value is displayed in a recommended value display column 14. The user can set the optimum threshold value by rewriting the inspection condition data file 102a according to the calculation result. The above is the method of creating inspection condition data using GUI1.
[0040]
Next, an algorithm for calculating the recommended change parameter shown in FIG. 3 will be described using the inspection result 100 of FIG. 5, the defect determination result file 101 of FIG. 6, and the t-test result of FIG.
[0041]
In step 51, the inspection result file 100 is read. In step 52, the defect determination result file 101 is read.
[0042]
In step 53, a t-test is performed on the size of the defect group and the pseudo defect group to calculate a one-sided probability. The t-test is a general technique for testing whether the average values of two normal distributions are different. The one-sided probability is a probability that the average values of two normal distributions match. The one-sided probability obtained by the t test is substituted for the variable Ts. In the present embodiment, as shown in the t test result 108, Ts is 1.4 × 10 ^-6 It becomes. In step 54, a t test is performed on the luminance with the defect group and the pseudo defect group, and the one-sided probability is substituted for the variable Ti. In the present embodiment, as shown in the t test result 108, Ti is 8.7 × 10 ^-2 It becomes. In step 55, a t-test is performed on the difference luminance between the defect group and the pseudo defect group, and the one-sided probability is substituted for a variable Td. In this embodiment, Td = 2.3 × 10 ^-6 It becomes. In step 56, a t test is performed on the difference luminance outside the 80 mm from the center of the wafer using the defect group and the pseudo defect group. The one-sided probability obtained by the t test is substituted for the variable Tp. In this embodiment, Tp = 3.1 × 10 ^-1 It becomes. The t-test results of steps 53 to 56, the average value and the standard deviation of the pseudo defect group are recorded in the t-test result file.
[0043]
Next, in step 58, the one-sided probability of the minimum t test among Ts, Ti, Td, and Tp is obtained. In the present embodiment, as shown in FIG. 10, Td is minimum. In step 59, a recommended change parameter is selected based on the minimum one-sided probability obtained in step 58. In the present embodiment, the difference luminance threshold value is the recommended change parameter. In step 60, the recommended change parameters are displayed in the recommended change parameter display column 13 of the GUI 1 in FIG.
[0044]
Next, an algorithm for calculating the recommended value of the change recommended parameter shown in FIG. 4 will be described using the inspection result 100 of FIG. 5, the defect determination result file 101 of FIG. 6, and the t-test result of FIG.
[0045]
In step 71, the target number of pseudo defects inputted from the target number of pseudo defects input column 11 in FIG. In step 72, the inspection result file 100 is read. In step 73, the total number of defect candidates is obtained from the inspection result file 100 and substituted into the variable N. In step 74, the variable D is divided by the variable N to calculate a target false alarm rate, which is substituted for the variable IR. In this embodiment, since D = 1 and N = 40, IR = 0.025.
[0046]
In step 75, a feature value with the smallest one-sided probability is obtained, and the average value of the pseudo defects of the feature value is substituted into a variable M. Similarly, the standard deviation is substituted for the variable S. In step 76, a value of a feature amount in which the cumulative probability density is 1−IR when the pseudo defect group follows the normal distribution of the average value M and the standard deviation S is calculated and substituted into the variable A.
[0047]
In step 77, the variable A is displayed in the recommended value display field 14 in FIG.
[0048]
The above is an example in which a person reviews a defect image to classify a defect and a pseudo defect.
[0049]
Next, a method of classifying defects and pseudo defects by an ADC (Automatic Defect Classification) and automatically correcting the inspection condition data will be described with reference to FIG.
[0050]
The user inputs the maximum allowable number of pseudo defects in the target pseudo defect number input field 11 in FIG. 1 and operates and clicks the fully automatic button 22 with the mouse 38, thereby starting the algorithm in FIG.
[0051]
In FIG. 12, at step 301, the number of target pseudo defects is read. In step 302, the standard inspection condition data file 105 is read, and based on this, a new inspection condition data file 102a is created. In step 303, an inspection is performed. In step 304, the coordinates of the defect candidate and the feature amount of the defect are stored in the inspection result file 100. In step 305, an image of a defect candidate is captured. In step 305, all the defect candidates are imaged and the images are stored. In step 306, the image is classified into a defect and a pseudo defect by the ADC using the captured image. In step 307, the number of pseudo defects is counted. In step 308, it is determined whether the number of pseudo defects is larger than the target number of pseudo defects. If the number of pseudo defects is large, the process proceeds to step 309. In step 309, a recommended change parameter is calculated according to the algorithm of FIG. In step 310, the recommended value of the change recommended parameter is calculated according to the algorithm of FIG. In step 311, the change recommended parameter of the inspection condition data file 102a is changed to a recommended value. After step 311, the process returns to step 303. If the number of pseudo defects is smaller than the target number of pseudo defects in step 308, the process ends.
[0052]
As described above, according to the inspection condition data creation method of the present invention, the user reviews the defect image and classifies the defect image into a defect group and a pseudo defect group, and displays the result on the GUI. The difference in the feature amount of the defect group can be easily visually recognized. Further, a threshold value to be corrected and its correction value can be calculated from the difference between the feature amounts of the defect group and the pseudo defect group.
[0053]
In this embodiment, the defect feature display area uses the defect feature on the vertical axis and the coordinate axis on the horizontal axis. In some inspection devices, the difference luminance threshold is changed depending on the brightness of the reference image in order to improve the sensitivity in a dark area. In such a case, as shown in FIG. 13, the defect feature amount display area 8b of the two-dimensional scatter diagram using the luminance on the horizontal axis and the differential luminance on the vertical axis may be used.
[0054]
In this embodiment, a two-dimensional scatter diagram is used for the defect feature amount display area. However, in order to easily make a quantitative difference between the number of defect groups and the number of pseudo defect groups, as shown in FIG. The quantity display area 8c may be used.
[0055]
Although the present embodiment has been described with reference to the appearance inspection apparatus, the present invention is not limited to this, and a foreign matter inspection apparatus or the like may be used as long as the apparatus detects a defect on a substrate.
[0056]
In the present embodiment, the defect candidate to be reviewed is selected from the wafer map 2, but the defect candidate to be reviewed may be selected from the chip map 4 or the defect feature distribution display column 8.
[0057]
In this embodiment, the difference luminance, the luminance, and the magnitude are used as the feature amounts to be displayed, but the present invention is not limited to these. In this embodiment, the horizontal axis to be displayed is selected from the distance from the center of the wafer, the square of the distance from the center of the wafer, the X and Y coordinates in the wafer coordinate system, and the X and Y coordinates in the chip coordinate system. However, the present invention is not limited to these.
[0058]
In the present embodiment, the algorithm shown in FIG. 5 is used as an algorithm for selecting the recommended change parameter using the difference in the feature amount between the defect group and the pseudo defect group. However, the present invention is not limited to this.
[0059]
Further, in the present embodiment, the algorithm shown in FIG. 6 is used as an algorithm for calculating the recommended value of the change recommended parameter using the difference between the feature amounts of the defect group and the pseudo defect group, but the present invention is not limited to this. .
[0060]
According to the embodiment of the present invention, the following effects can be obtained.
[0061]
According to the present invention, it is possible to visually judge the result of reviewing defect candidates output by inspection under standard inspection conditions and discriminating between a defect group and a pseudo defect group. Further, it is possible to calculate a threshold value to be corrected and its correction value using the feature amount data of the defect group and the pseudo defect group. By setting this value as new inspection condition data, the user can perform inspection under optimum inspection conditions. Since no further trial inspection and adjustment are required, it is possible to shorten the time for creating inspection condition data.
[0062]
As described above, the embodiment of the present invention has been specifically described based on the embodiment. However, the present invention is not limited to this, and can be changed without departing from the gist thereof.
[0063]
【The invention's effect】
According to the present invention, inspection condition data can be easily created.
[Brief description of the drawings]
FIG. 1 is a diagram showing an example of a system screen of the present invention.
FIG. 2 is a diagram showing an example of the configuration of the apparatus of the present invention.
FIG. 3 is an example showing a configuration of a defect inspection unit of the present invention.
FIG. 4 is a diagram illustrating a defect detection algorithm of the apparatus of the present invention.
FIG. 5 is a diagram illustrating an example of an algorithm for calculating a change recommended parameter according to the present invention;
FIG. 6 is a diagram illustrating an example of an algorithm for calculating an optimum value according to the present invention.
FIG. 7 is a diagram showing an example of an inspection result file output by the apparatus of the present invention.
FIG. 8 is a diagram illustrating an example of a defect determination result file according to the present invention.
FIG. 9 is a diagram illustrating an example of an inspection condition data file according to the present invention.
FIG. 10 is an example of a t-test result of a defect group and a pseudo defect group for each feature amount calculated by the program of the present invention.
FIG. 11 is a diagram showing an example of a t-test result file of the present invention.
FIG. 12 is an example of an algorithm for automatically creating optimum inspection condition data according to the present invention.
FIG. 13 is another example of a defect feature amount distribution display column used in the GU11.
FIG. 14 is another example of a defect feature amount distribution display column used in the GU11.
[Explanation of symbols]
1 ... GUI of this program, 2 ... wafer map, 3 ... legend display field, 4 ... chip map, 5 ... review field, 6 ... pseudo defect ratio display field, 7 ... defect feature quantity selection field, 8 ... defect feature quantity distribution Display field, 8b ... defect feature quantity distribution display field using defect feature quantities on the vertical and horizontal axes, 8c ... defect feature quantity distribution display field using histogram, 11 ... target pseudo defect number input field, 12 ... calculation button, 13 ... recommended change parameter display field, 14 ... recommended value display field, 15 ... cursor, 16 ... review image, 17 ... defect button, 18 ... pseudo defect button, 19 ... defect symbol, 20 ... pseudo defect symbol, 21 ... no review Symbol, 22: Fully automatic button, 30: Appearance inspection device, 31: Defect inspection unit, 32: Control unit, 33: Operation unit, 34: Main storage device, 35: Secondary storage device, 36: Network interface 37, a keyboard, 38, a mouse, 39, a monitor, 51, a step of reading an inspection result, 52, a step of reading a review determination result, 53, a step of performing a t-test on size, and 54, a t-test of luminance. Step 55: Step of performing a t-test on the difference brightness, 56: Step of performing a t-test on the difference brightness around the wafer, 58: Step of obtaining the minimum one-sided probability from the t-test result obtained in Steps 53 to 56 59: Step of selecting recommended change parameters, 60: Step of displaying change recommended parameters in the change recommended parameter display column, 71: Step of reading target number of false reports, 72: Step of reading inspection results, 73: Calculating total defect candidate number , 74 ..., calculating the target false alarm rate, 75 ... one side Substituting the average value of the pseudo defects into the variable M and substituting the standard deviation into the variable S for the feature amount having the minimum rate. 76: If the distribution of the pseudo defects follows the normal distribution of the average value M and the standard deviation S, In this case, a step of calculating a feature quantity having a cumulative probability density of 1-IR and substituting it for a variable A, a step of displaying the variable A in a recommended value display column, a step of displaying an inspection result file, a step of detecting a defect Judgment result file, 102: inspection condition data file, 103: t test result file, 104: program, 105: standard inspection condition data file, 151: wafer, 152: stage, 153: target chip, 153a: target chip photographing location, 154: Reference chip, 154: Reference chip shooting location, 155: Lamp, 156: Sensor, 158: A / D converter, 158: Temporary image memo Reference numeral 159: Image processing unit, 160: Lamp lens, 161: Sensor lens, 201a: Wiring on the target chip image, 201b: Wiring on the reference chip image 202a: Defect on the reference chip image, 202c: Difference image Upper defect, 301: a step of inputting the target number of pseudo defects, 302: a step of creating an inspection condition data file 102a from a standard inspection condition data file, 303: an inspection step, 304: an inspection result in the inspection result file 100 Storing, 305: capturing and storing images of all defect candidates, 306: classifying defects and pseudo defects by ADC using the defect candidate images, 307: counting the number of pseudo defects, 308 ... Step of comparing the number of pseudo defects with the target number of pseudo defects, 309: Recommended change parameter Calculating a, 310 ... calculating a correction value, the step of modifying the 311 ... detection condition data file 102a.

Claims (16)

A program that is executed to create detection condition data of an inspection device that detects positions of a plurality of defects such as a foreign substance or a pattern abnormality that the substrate has,
Inputting coordinates of each defect candidate detected in the inspection and one or more feature amounts corresponding to the coordinates;
Inputting a determination result as to whether the defect is a defect or a pseudo defect for each defect candidate, and classifying the defect candidate group into a defect group and a pseudo defect group;
Plotting the defect or the pseudo defect for each of the defect candidates with a different symbol on a two-dimensional scatter diagram in which one item of the feature amount is set as one axis and the other as a coordinate axis, and outputs the plotted result. Detection condition data creation support program. A program that is executed to create detection condition data of an inspection device that detects positions of a plurality of defects such as a foreign substance or a pattern abnormality that the substrate has,
Inputting coordinates of each defect candidate detected in the inspection and one or more feature amounts corresponding to the coordinates;
Inputting a determination result as to whether the defect is a defect or a pseudo defect for each defect candidate, and classifying the defect candidate group into a defect group and a pseudo defect group;
Plotting the defect or the pseudo defect with a different symbol for each defect candidate on a two-dimensional scatter diagram in which each axis is one item of the feature amount and outputting the same. Support program. A program that is executed to create detection condition data of an inspection device that detects positions of a plurality of defects such as a foreign substance or a pattern abnormality that the substrate has,
Inputting coordinates of each defect candidate detected in the inspection and one or more feature amounts corresponding to the coordinates;
Inputting image data captured for each coordinate of the defect candidate;
For each of the defect candidates, determining whether the defect is a defect or a pseudo defect using the image data, classifying the defect candidate group into a defect group and a pseudo defect group,
Plotting the defect or the pseudo defect for each of the defect candidates with a different symbol on a two-dimensional scatter diagram in which one item of the feature amount is set as one axis and the other as a coordinate axis, and outputs the plotted result. Detection condition data creation support program. A program that is executed to create detection condition data of an inspection device that detects positions of a plurality of defects such as a foreign substance or a pattern abnormality that the substrate has,
Inputting coordinates of each defect candidate detected in the inspection and one or more feature amounts corresponding to the coordinates;
Inputting image data captured for each coordinate of the defect candidate;
For each of the defect candidates, determining whether the defect is a defect or a pseudo defect using the image data, classifying the defect candidate group into a defect group and a pseudo defect group,
Plotting the defect or the pseudo-defect with a different symbol for each defect candidate on a two-dimensional scatter diagram in which each axis is one item of the feature amount, and outputting the plotted image. Detection condition data creation support program. 4. The method according to claim 1, wherein the characteristic amount includes one of a size of the defect candidate, a luminance of the defect candidate, and a difference luminance of the defect candidate. 4. A detection condition data creation support program according to 4. As the coordinate axes of the two-dimensional scatter diagram, the distance from the center of the substrate, the square of the distance from the center of the substrate, or the X or Y coordinate of the orthogonal coordinate system defined on the substrate, or the orthogonal coordinates defined for each chip of the substrate The detection condition data creating program according to claim 1, wherein the system is set to one of X and Y coordinates of the system. A program that is executed to create detection condition data of an inspection device that detects a position of a plurality of defects such as a foreign substance or a pattern abnormality that the substrate has,
Inputting the coordinates and the feature amount of each defect candidate detected by the inspection;
For each defect candidate, inputting a determination result of whether the defect is a defect or a pseudo defect, classifying the defect candidate group into a defect group and a pseudo defect group,
Dividing the size of the defect feature into a plurality of classes;
Calculating the number of defects in the defect group and the number of pseudo defects in the pseudo defect group for each class;
A step of outputting a histogram indicating the number of defects and the number of pseudo defects for each class on one axis as a class of the defect feature quantity, and the other axis. An inspection device that detects the position of multiple defects such as foreign substances and pattern abnormalities on the substrate,
Means for detecting defect candidates;
Means for storing coordinates of each defect candidate detected by the means, and one or more feature amounts corresponding to the coordinates;
Means for capturing an image of the defect candidate;
Means for storing the defect candidate image;
Means for outputting the defect candidate image;
Means for inputting and storing the result of determining whether the defect candidate is a defect or a pseudo defect,
Means for plotting and outputting the defect or the pseudo defect with a different symbol for each defect candidate on a two-dimensional scatter diagram in which one item of the feature amount is set as one axis and the other as a coordinate axis. Inspection equipment. A program that is executed to create detection condition data of an inspection device that detects the position of a plurality of defects such as a foreign substance or a pattern abnormality that the substrate has,
Inputting coordinates of each defect candidate detected in the inspection and one or more feature amounts corresponding to the coordinates;
Inputting a determination result as to whether the defect is a defect or a pseudo defect for each defect candidate, and classifying the defect candidate group into a defect group and a pseudo defect group;
Classifying the defect candidates into a defect group and a pseudo defect group based on the determination result input in the step;
A step of selecting detection condition data to be corrected by using a difference in feature amount between the defect group and the pseudo defect group. In the step of selecting the detection condition data to be corrected, the difference between the average value of the feature values of the defect group and the pseudo defect group is tested, and the most different feature value between the defect group and the pseudo defect group is selected.
10. The detection condition data creation support program according to claim 9, wherein detection condition data strongly related to the selected feature amount name is selected as detection condition data to be corrected. An inspection device that detects the position of defects such as foreign substances or pattern abnormalities that the substrate has,
Means for detecting defect candidates;
Means for storing coordinates of the defect candidate detected by the means, and one or more feature amounts corresponding to the coordinates;
Means for capturing an image of the defect candidate;
Means for storing the defect candidate image;
Means for outputting the defect candidate image;
Means for inputting a determination result of whether the defect candidate is a defect or a pseudo defect,
Means for classifying the defect candidate group into a defect group and a pseudo defect group based on the determination result input by the means,
Means for selecting detection condition data to be corrected by using a difference in feature amount between the defect group and the pseudo defect group. A program that is executed to create detection condition data of an inspection device that detects a position of a defect such as a foreign substance or a pattern abnormality that the substrate has,
Inputting the coordinates of the defect candidate detected by the inspection and one or more feature data corresponding to the coordinates;
Inputting a determination result as to whether the defect candidate is a defect or a pseudo defect, and classifying the defect candidate group into a defect group and a pseudo defect group;
Calculating a correction value of the detection condition data to be corrected by using a difference in the feature amount between the defect group and the pseudo defect group. An inspection device that detects the position of defects such as foreign substances or pattern abnormalities that the substrate has,
Means for detecting defect candidates;
Means for storing the coordinates of the defect candidate detected by the means, and one or more feature amounts corresponding to the coordinates;
Means for capturing an image of the defect candidate;
Means for storing the defect candidate image;
Means for outputting the defect candidate image;
Means for inputting a determination result of whether the defect candidate is a defect or a pseudo defect,
Means for classifying the defect candidate group into a defect group and a pseudo defect group according to the determination result input by the input means;
Means for calculating a correction value of the detection condition data to be corrected using the difference between the feature amounts of the defect group and the pseudo defect group. A program that is executed to create detection condition data of an inspection device that detects a position of a defect such as a foreign substance or a pattern abnormality that the substrate has,
Inputting the coordinates of the detected defect candidate and a feature amount corresponding to the coordinates;
For each of the defect candidates, inputting a determination result as to whether it is a defect or a pseudo defect, and classifying the defect candidate group into a defect group and a pseudo defect group;
Selecting detection condition data to be corrected by using a difference in feature amount between the defect group and the pseudo defect group;
Calculating a correction value of the detection condition data to be corrected using the difference between the feature amount data of the defect group and the pseudo defect group;
Writing the calculated correction value in the detection condition data. An inspection device that detects the position of defects such as foreign substances or pattern abnormalities that the substrate has,
Means for detecting defect candidates;
Means for storing coordinates of each defect candidate detected by the means, and one or more feature amounts corresponding to the coordinates;
Means for capturing an image of the defect candidate;
Means for storing the defect candidate image;
Means for outputting the defect candidate image;
Means for inputting a determination result as to whether the defect candidate is a defect or a pseudo defect;
Means for classifying the defect candidate group into a defect group and a pseudo defect group based on the determination result input by the input means;
Means for selecting detection condition data to be corrected using the difference between the feature amounts of the defect group and the pseudo defect group;
Means for calculating a correction value of the detection condition data to be corrected using the difference between the feature data of the defect group and the feature amount data of the pseudo defect group;
Means for describing the calculated correction value in the detection condition data. A detection condition data creation method of an inspection apparatus for detecting a position of a defect such as a foreign substance or a pattern abnormality having a substrate,
Determine whether each defect candidate detected by the inspection is a defect or a pseudo defect,
According to the determination result, a defect group and a pseudo defect group are classified,
The detection condition data to be corrected is selected using the difference between the feature amounts of the defect group and the pseudo defect group,
The correction value of the threshold to be corrected is calculated using the difference between the feature amounts of the defect group and the pseudo pseudo defect group,
A method for creating detection condition data, wherein the calculated correction value is set as detection condition data.

Images