JP2008153572A - Contamination inspecting device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To register a template image to be used for an alignment in an optimum method to reduce a time required for error and evaluation in pattern matching. <P>SOLUTION: The selection, identification, and comparison of an optimum template with alighnment marks are determined with the operation function of a relative value disposed at a contamination inspecting device. Namely, the contamination inspecting device includes a device that registers features of the alighnment mark formed on the surface of an inspected object, a device that samples image data of the alighnment mark formed on the surface of the inspected object, and a data operation processing device that calculates the relative values from both features by extracting the features from the image data, thereby registering the image data of the alignment marks based on thresholds of the relative values. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a foreign matter inspection apparatus for detecting foreign matter, scratches, defects, dirt, etc. (hereinafter collectively referred to as foreign matter) present on the surface of an inspection object such as a glass substrate or a semiconductor wafer. The present invention relates to a foreign matter inspection apparatus that corrects the position of an inspection object with high accuracy and detects foreign matter with high accuracy and high sensitivity.

  As a foreign matter inspection device that detects foreign matter existing on the surface of an inspection object such as a glass substrate or a semiconductor wafer, the surface of the inspection object is irradiated with a light beam such as a laser beam, and reflected light or scattered light generated on the surface. It is known that the presence or absence of a foreign object is detected by detecting (for example, see Patent Document 1). In this foreign matter detection device, when a large number of IC chips having essentially the same pattern are formed on a semiconductor wafer, an image signal is created from the intensity of reflected or scattered light detected from each chip and obtained from adjacent chips. Compared with a prepared image signal or an image signal of a good chip prepared in advance, a case where the deviation signal of both is equal to or greater than a threshold value is determined as a foreign object.

  When collecting the image signal, it is necessary that the chips arranged in parallel in the lateral direction on the surface of the object to be inspected are placed parallel to the scanning direction of the light beam. Since deviation signals are collected compared to adjacent chips, deviation signals vary due to differences in the types and densities of patterns included in the detection area, such as the wiring layout within the chip and the scribe lines between chips. It is because it affects.

  As an alignment method for placing the object to be inspected in parallel, the coordinates (X, Y) of two points in the object to be inspected are based on the alignment mark formed in the chip on the surface of the object to be inspected. Correction is performed by moving the inspection stage based on the amount of deviation of the inspection object that is collected and calculated from the coordinates.

  With the recent high integration and miniaturization of semiconductors, it is necessary to detect finer foreign substances. To suppress variations in deviation signals and improve detection accuracy and reproducibility, further alignment accuracy is required. Although desired, alignment has become difficult due to miniaturization of the alignment mark and a decrease in contrast caused by the manufacturing process.

  Regarding the alignment method of the inspection position, for example, the projection waveform of the reticle substrate is prepared as reference image data (hereinafter referred to as a template), and the amount of deviation is determined by pattern matching with the projection waveform actually obtained from the reticle substrate for position adjustment. What is desired is known (see, for example, Patent Document 2).

  As for the pattern matching method, an image signal is detected from an object to be inspected, a predetermined feature amount is extracted from the image signal, an abstract pattern is formed, and an abstract pattern obtained from this and a reference image (template) Are known (for example, see Patent Document 3).

Japanese Patent Laid-Open No. 5-47901 Japanese Patent Laid-Open No. 10-106941 JP 11-340115 A

  The technique described in Patent Document 2 corrects the positional deviation of the forward path and the backward path collectively when scanning the inspection object by reciprocating with a light beam. It cannot be applied to a foreign substance inspection apparatus that requires alignment. In addition, because it is intended for reticle substrates created as jigs dedicated to positional misalignment, the effects of various thin films such as semiconductor films, metal films, and insulating films formed in the manufacturing process of semiconductor devices, etc. Is not considered, and problems such as unrecognition due to a decrease in the contrast of the alignment mark and misrecognition (mismatching) with other alignment marks are inevitable.

  Conventionally, an operator selects an alignment mark while looking at an observation screen of an inspection object, and registers the image data as a template. However, when the inspection process of the semiconductor device product is a manufacturing process in which the contrast decreases, it is difficult to visually check the alignment mark, and the selection of the alignment mark is repeatedly performed based on past experience and feeling. However, since the recognition accuracy is lowered, the operator needs a long time to find out the complicated template evaluation conditions and to collect the template. Further, when the angle (θ) is shifted in the object to be inspected, it is difficult to use the template as it is, and it is necessary to correct the angle. There are problems such as a decrease in alignment accuracy and variations due to errors in the correction processing.

  An object of the present invention is to perform pattern matching so as not to be unrecognizable or erroneously recognized even when the alignment mark of the inspection object is in a condition that makes pattern matching difficult.

  Another object of the present invention is to equip a function for analyzing and registering whether or not an image of a pattern collected from an object to be inspected is appropriate as a template, thereby facilitating selection of an appropriate template and determination of evaluation conditions. is there.

  The present invention irradiates a surface of an inspection object with a light beam, acquires an image signal from the received light intensity of reflected light or scattered light, and compares the surface of the inspection object with an image signal obtained from an adjacent inspection object In the foreign matter inspection apparatus for inspecting the presence or absence of foreign matter present in the apparatus, an apparatus for registering feature points of an alignment mark formed on the surface of the inspection object, and an apparatus for reading the alignment mark on the surface of the inspection object And performing alignment by detecting an alignment mark formed on the surface of the inspection object based on the feature point.

  In the foreign matter inspection apparatus for irradiating the inspection object with a light beam and inspecting the presence or absence of the foreign matter existing on the surface of the inspection object based on the received light intensity of the light reflected or scattered from the inspection object, A device for inputting feature points of an alignment mark formed on the surface of the object, a device for displaying the feature points of the input alignment mark, and a device for registering the feature points of the input alignment mark. Alignment is performed by detecting alignment marks formed on the surface of the inspection object based on the points.

  In addition, in the foreign matter inspection apparatus for inspecting the presence or absence of foreign matter existing on the surface of the inspection object by irradiating the inspection object with a light beam and receiving or receiving light of reflected or scattered light, it is formed on the surface of the inspection object. A device for registering feature points of an alignment mark, a device for collecting image data of alignment marks formed on the surface of the object to be inspected, and extracting feature points from the image data and correlating them from both feature points A data operation processing device for calculating a value, and registering image data of the alignment mark based on a threshold value of the correlation value.

  In addition, in the foreign matter inspection apparatus for inspecting the presence or absence of foreign matter existing on the surface of the inspection object by irradiating the inspection object with a light beam and receiving or receiving light of reflected or scattered light, it is formed on the surface of the inspection object. A device for registering feature points of an alignment mark, a device for collecting image data of alignment marks formed on the surface of the object to be inspected, and extracting feature points from the image data and correlating them from both feature points A data calculation processing device for calculating a value and a display device for displaying the calculation result of the correlation value are displayed, and the determination result of the pass / fail of the alignment mark is displayed based on a threshold value of the correlation value. In this foreign matter inspection apparatus, it is desirable that the calculation results of the correlation values are arranged in order of the correlation values and displayed on the display device in a list format.

  In addition, the surface of the object to be inspected is irradiated with a light beam, the image signal is obtained from the received light intensity of reflected light or scattered light, and is present on the surface of the object to be inspected compared with the image signal obtained from the adjacent object to be inspected. In the foreign matter inspection apparatus for inspecting the presence or absence of foreign matter, image processing means for extracting feature points from image data of alignment marks formed on the surface of the object to be inspected and pattern matching the feature points of the template, and matching mutual An arithmetic processing means for calculating a probability or a score (score) from a feature point; a judgment processing means for identifying a case where the feature points match with a predetermined probability or score (score) as a reference alignment mark; Arithmetic processing means for calculating a deviation amount of the object to be inspected from coordinates obtained by recognizing at least two of the alignment marks, and the deviation amount Characterized in that a drive mechanism for aligning by moving the inspection stage based.

  In the present invention, in order to automatically select an optimum alignment mark, an alignment mark evaluation function as to whether or not it is suitable as a template for a foreign substance inspection apparatus, a display function of the evaluation result, an alignment mark selection function, and the image data A registration function for storing as a template can be provided.

  Further, in order to automatically extract alignment marks as template candidates, data input means for inputting feature points of the template, data registration means for storing data, and designation in the chip based on the feature point data of the template Detection means for collecting image data of a section, data registration means for storing the collected image data, image processing means for extracting feature points from the image data, and data calculation processing for comparing with feature points of a registered template A display means for displaying the coordinates and appearance shape of the extracted alignment mark candidate and the analysis value such as the suitability degree and suitability rank calculated by the data calculation processing section, and automatically selecting from the alignment mark candidate as an optimum template or Registration means for selecting and registering image data. Can.

  In addition, in order to evaluate in advance the possibility of misrecognition of a template candidate, detection means for collecting image data of a designated section in the chip based on the extracted feature point data of the template candidate and the sample Data registration means for storing image data, image processing means for extracting feature points from the image data, a data calculation processing section for comparing with the extracted feature points of the template, coordinates and appearance shape of the collected image data, and data calculation A display means for displaying an analysis value such as the aptitude degree calculated by the processing unit and the aptitude rank can be provided.

  According to the present invention, it is possible to perform pattern matching in which alignment marks cannot be recognized and erroneous recognition is suppressed.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. The foreign matter inspection apparatus of the present invention can inspect foreign matter existing on the surface of an inspection object such as a glass substrate used for a semiconductor wafer, an ALTIC substrate, a TFT-LCD, etc. In this embodiment, the semiconductor wafer This will be described as an example of an inspection object.

  FIG. 1 is a plan view showing a schematic configuration of a foreign substance inspection apparatus according to an embodiment of the present invention. The foreign matter inspection apparatus includes one or more load ports 10, a transport unit 20, a pre-alignment unit 30, an inspection unit 40, and a data processing unit 50. One or more wafer cassettes 11 containing a plurality of semiconductor wafers 1 to be inspected are placed on the load port 10. The wafer cassette 11 can also be operated separately for exclusive use for collecting the semiconductor wafer 1 to be inspected and the semiconductor wafer 1 determined to be defective as a result of the inspection.

  FIG. 2 is a block diagram showing an outline of control of the data processing unit 50. The data processing unit 50 includes a host computer 100, an input device 110 including a keyboard, a touch panel, or a mouse, a display device 120 including a CRT or a flat panel display, an output device 130 such as a printer, a floppy (registered trademark) disk (FD), or a compact. An external storage device 140 that controls external media such as a disc (CD) is configured. The host computer 100 also includes a data arithmetic processing device 150 and a storage device 160 such as a hard disk drive (HDD). Based on the command from the input device 110, the whole foreign matter inspection apparatus is controlled. Control and analysis results related to alignment, control of inspection conditions, analysis of collected data, and operation state of the foreign matter inspection apparatus are displayed on the display device 120, and further output to an output device 130 such as a printer. Various condition settings are input from the input device 110 and stored in the storage device 160 as recipe data or the like.

  FIG. 14 is a flowchart showing the processing procedure of the semiconductor wafer 1 in the first embodiment. In the foreign matter inspection apparatus of FIG. 1, the semiconductor wafer 1 is transported by a signal transmitted from the data processing unit 50 by executing the transport program (S101), driving a servo motor via a pulse control board and a drive circuit (not shown), This is done by controlling the handling arm 22 disposed in the transport assembly 21. The handling arm 22 carries out the semiconductor wafer 1 from the shelf of the wafer cassette 11 instructed by the data processing unit 50, and carries it from the load port 10 to the pre-alignment unit 30 (S102). When unloading the semiconductor wafer 1 from the wafer cassette 11, the U-shaped contact portion 22 a is controlled so that the approximate center thereof coincides with the approximate center of the semiconductor wafer 1, and vacuum suction is performed at the suction hole 23 of the contact portion 22 a, The semiconductor wafer 1 is held.

  The handling arm 22 moves forward above the pre-alignment chuck 31 while holding the semiconductor wafer 1, descends at a position where the respective centers of the semiconductor wafer 1 and the pre-alignment chuck 31 coincide with each other, and the semiconductor on the pre-alignment chuck 31. A wafer 1 is mounted. Next, the pre-alignment chuck 31 vacuum-sucks the back surface of the semiconductor wafer 1 and holds the semiconductor wafer 1.

  The pre-alignment chuck 31 is configured to be movable and rotatable in the X, Y, and θ directions. The detection device 32 includes a light emitting unit such as a laser beam and a light receiving unit such as a CCD line sensor, detects the position and intensity of light reaching the light receiving unit from the light emitting unit, and is a semiconductor placed on the pre-alignment chuck 31. The outer periphery of the wafer 1 and the position of the V notch (or orientation flat) are detected. Based on the detection result of the detection device 32, the data processing unit 50 drives the pulse motor through a pulse control circuit and a drive circuit (not shown), moves and rotates the pre-alignment chuck 31, and adjusts the rough position of the semiconductor wafer 1. (Pre-alignment) is performed (S103).

  After the pre-alignment is completed, the pre-alignment chuck 31 releases the vacuum suction of the semiconductor wafer 1. The handling arm 22 penetrates below the semiconductor wafer 1 mounted on the pre-alignment chuck 31 and rises at a position where the center of the U-shaped contact portion 22 a substantially coincides with the center of the semiconductor wafer 1. 1 is lifted from the pre-alignment chuck 31. Then, the semiconductor wafer 1 is transferred from the pre-alignment unit 30 to the inspection stage 41 of the inspection unit 40 (S104).

  In the inspection unit 40, an inspection stage chuck 42 is disposed on an inspection stage 41. The inspection stage chuck 42 includes elevating pins 43a, 43b, and 43c that can be raised and lowered. With the lifting pins 43a, 43b, and 43c raised, the handling arm 22 moves upward above the inspection stage chuck 42 with the semiconductor wafer 1 lifted, and the approximate centers of the semiconductor wafer 1 and the inspection stage chuck 42 coincide with each other. The semiconductor wafer 1 is transferred to the lift pins 43a, 43b and 43c. Next, after the handling arm 22 is retracted from above the inspection stage chuck 42, the lift pins 43 a, 43 b, 43 c are lowered, and the semiconductor wafer 1 is mounted on the inspection stage chuck 42. The inspection stage chuck 42 vacuum-sucks the back surface of the semiconductor wafer 1 and fixes the semiconductor wafer 1.

  When the semiconductor wafer 1 is placed on the inspection stage chuck 42, the data processing unit 50 reads the chip size from the recipe data registered in the storage device 160 via the input device 110, and the data arithmetic processing unit 150 reads the semiconductor wafer. The coordinates of the chips arranged in parallel to 1 are calculated. The inspection stage chuck 42 is configured to be able to move and rotate in the X, Y, and θ directions, and detects the position (coordinates) by a position detector (not shown) such as a laser scale or the like, while detecting the position (coordinates). Control the inspection position etc. above. The data processing unit 50 moves and rotates the inspection stage chuck 42 by driving a servo motor (not shown) via a pulse control board and a drive circuit (not shown), for example, the first chip 310 of the first chip 310 shown in FIG. Move to one chip designated coordinate 311 (S105).

A CCD camera (not shown) is placed on the inspection stage chuck 42 of the inspection unit 40. The designated range in the vicinity of the first chip designated coordinates 311 is searched (S106), and for example, the first alignment mark 312 of the first chip 310 is collected as image data from the CCD camera (S107). From the collected image data of the first alignment mark 312, the average lightness of the entire image is calculated by the data processing unit 150 based on the gradation data for each pixel of the CCD. Next, a multiplier is obtained in which the average brightness and the average brightness of the template are substantially the same, and a difference between the brightness obtained by multiplying the brightness for each pixel and the average brightness of the entire image is calculated for each pixel. By this normalization process, changes between pixels are extracted as normalized brightness waveforms (projection waveforms) and stored in the storage device 160.
In the present embodiment, the reference brightness is used as the average brightness of the template. However, the same effect can be obtained by obtaining a multiplier for adjusting the brightness based on the preset brightness.

  The data calculation processing device 150 performs calculation processing of the projection waveform, and calculates an edge intensity waveform 410 representing the change in light and shade (gradation) shown in FIG. 4 by, for example, differentiation processing. Next, a maximum edge strength value 411 indicating the maximum value is extracted from the obtained edge strength waveform 410, and a multiplier that is substantially the same as the preset feature amount setting value 420 is obtained. The entire edge intensity waveform 410 is corrected based on this multiplier so that the maximum edge intensity value 411 decreases when the contrast is high so that the intensity value 411 increases. This correction suppresses the influence on the pattern matching accuracy due to a decrease in contrast such as a film on the surface of the semiconductor wafer 1 and fluctuations in illuminance.

  Based on the feature amount threshold value 430 registered in the storage device 160 in advance, the edge position 440 is detected from the peak exceeding the threshold value and the position of the CCD pixel, the intensity value of each edge position 440, the position of the CCD pixel, etc. Are stored in the storage device 160 in association with position information (coordinate information) from the position detector.

  The template data serving as a comparison reference is subjected to image processing similar to that of the first alignment mark 312 described above, and feature points such as the intensity value of each edge position 440 and the CCD pixel position are registered in the storage device 160 in advance. Yes. The feature points of the first alignment mark 312 are subjected to pattern matching based on the feature points of the template, and the similarity is determined (S108). If the first alignment mark 312 is different from the template pattern, the search within the designated range is continued. Steps S106 to S108 are repeated to search for alignment marks that can be correlated while repeating pattern matching. A pattern detected with a predetermined correlation is recognized as the first alignment mark 312, the coordinates (X, Y) of the first alignment mark 312 are detected by the position detector, and the coordinates are registered in the storage device 160 ( S109). Since analogy determination is performed based on feature points that have been normalized to each other, even when the contrast of the first alignment mark 312 is low due to the influence of the process processing step, high-precision pattern matching is ensured and erroneous recognition can be suppressed. it can.

  For the pattern recognized as the first alignment mark 312 by pattern matching, an index value (score value) representing the matching state is calculated by the data arithmetic processing device 150 based on the feature point and displayed on the screen of the display device 120. . The score value is calculated by, for example, equation (1) based on correlation data between the first alignment mark 312 and the feature points in the template. As long as the degree of pattern matching can be converted into a numerical value and the degree can be displayed, the present invention can be used without being limited to the expression, and the same effect can be obtained.

  In the equation (1), S is the size of the score value, Mp is the number of matching edges in pattern matching, Tt is the total number of edges at the time of template collection, and Ta is the total number of edges of the first alignment mark 312.

  The score value S is stored in the storage device 160 for each pattern matching of the semiconductor wafer 1 and displayed on the screen of the display device 120. By checking the pattern matching state via the score value S, it is possible to determine whether the template is good or bad. Further, by comparing with the past score value S data stored in the storage device 160, the state of the alignment mark, that is, the processing state of the semiconductor device and the state of the foreign matter inspection apparatus can be diagnosed. Based on a preset threshold value of the score value S, it is also possible to transmit an abnormality display, an alarm, or a device abnormality to a remote diagnosis means. It can be set via

  After detecting the coordinates of the first alignment mark 312, the inspection stage chuck 42 is moved to the second chip designated coordinates 321 on the chip matrix of the semiconductor wafer 1 (S110). The second alignment mark 322 of the second chip 320 is also subjected to image processing and pattern matching in the same manner as the first alignment mark 312 (S111 to S113), and the coordinates (X, Y) of the detected second alignment mark 322 are detected. Is stored in the storage device 160 together with the calculated score value (S114). As the management reference value of the score value, a setting value (threshold value) can be set from a setting field provided on the display device 120 screen. When at least one of the first alignment mark 312 and the second alignment mark 322 does not satisfy the threshold condition, the data processing unit 50 can display an alarm.

  Note that score value data processing is calculated according to the situation in which the foreign object inspection device is used, such as cases where the occurrence of erroneous recognition is small and pattern matching status monitoring is not required, or when the throughput of the foreign object inspection device is required. The processing steps can be skipped, and the data processing method and the display method can be changed by setting from the input device 110.

  From the two (X, Y) coordinates of the collected first alignment mark 312 and second alignment mark 322, the data arithmetic processing unit 150 calculates the amount of angular deviation, and based on the command from the data processing unit 50, the inspection is performed. The angle of the stage chuck 42 is corrected (S115), and chips arranged in parallel on the semiconductor wafer 1 are placed in parallel with high accuracy with respect to the scanning direction of the light beam.

  Above the inspection stage chuck 42, a light projecting system device and a light receiving system device (not shown) are arranged, and a light beam such as a laser beam is irradiated from the light projecting system device onto the surface of the semiconductor wafer 1. The inspection stage chuck 42 is moved in the Y direction and the X direction by driving a servo motor to scan the surface of the semiconductor wafer 1 with the light beam.

  The light receiving system device detects reflected light or scattered light generated from the surface of the semiconductor wafer 1 and is processed by the host computer 100 of the data processing unit 50 based on the detection result of the light receiving system device, and is present on the surface of the semiconductor wafer 1. A foreign object to be detected is detected (S116).

  When the foreign matter inspection on the surface of the semiconductor wafer 1 is completed, the semiconductor wafer 1 is transferred from the inspection unit 40 to the load port 10 by the reverse procedure of placing the semiconductor wafer 1 from the wafer cassette 11 onto the inspection stage chuck 42, and the same. The wafer cassette 11 is stored on the same shelf. When the number of foreign matters on the surface of the semiconductor wafer 1 exceeds a set value, the semiconductor wafer 1 can be separated and transferred to the wafer cassette 11 of another load port 10. The classification setting of the semiconductor wafer 1 at the time of unloading can be set via the input device 110 from the setting screen of the display device 120.

  In this embodiment, correction is performed using two (X, Y) coordinates. However, although correction processing takes time, if three or more (X, Y) coordinates positioned vertically and horizontally are used, the correction is further performed. Highly accurate correction is possible. In this embodiment, correction is performed by using the same row of chips arranged in the horizontal direction with respect to the notches of the semiconductor wafer 1, but it is sufficient if the chip arrangement can be corrected in parallel to the scanning direction of the light beam. The same performance can be obtained even with correction in the direction, oblique direction, or arbitrary chip arrangement.

  As the distance between the first chip 310 and the second chip 320 used for position correction increases, the accuracy of angle correction can be improved. However, when the predetermined accuracy cannot be obtained for angle correction of the pre-alignment chuck 31, the alignment mark is searched. May fall outside the specified range or cause recognition failure. Therefore, the designated coordinates of the intermediate correction chip 330 are provided between the first chip 310 and the second chip 320, and after the coarse position correction is performed by the intermediate correction chip 330, the first chip 310 and the second chip are corrected. It is also possible to finely adjust the positional deviation at 320. Whether or not the intermediate correction chip 330 is used and various condition settings are set via the input device 110 from a setting screen provided on the screen of the display device 120. The alignment mark can be stably recognized even in the semiconductor wafer 1 having a large positional deviation by the selection means of the positional deviation correction method.

  FIG. 5 shows the above-described condition setting screen. The condition setting screen includes a chip condition setting unit 510 for setting the first chip 310, the second chip 320, and the intermediate correction chip 330, the first chip specifying coordinates 311, the second chip specifying coordinates 321, and the intermediate correction chip specifying. Coordinate condition setting means 520 for setting the coordinates of alignment marks such as coordinates, search condition setting means 530 for setting a search area, template setting means 540 for setting a template used for pattern matching, angle and position correction conditions, and score It is constituted by a matching condition setting means 550 for setting a value data processing method, a threshold value and the like. Among the setting conditions of each item, for example, a condition or numerical value setting that influences matching is displayed on the setting value display means 560 at the top of the setting screen. When each condition setting button is selected via the input device 110, a setting screen related to the condition setting button is opened, and detailed data can be set. In the present embodiment, the setting screen uses buttons, but the setting screen can be configured using icons, input spaces, or other means as long as the screen and setting conditions can be displayed and set. Good.

  When the alignment mark is determined by pattern matching, a template to be compared is necessary. Next, a method for collecting an optimum template from patterns formed on the semiconductor wafer 1 will be described mainly with reference to FIGS. 6, 7, and 15. Parts that overlap with those of the first embodiment are omitted, and FIGS. 1 and 2 are cited as necessary.

  FIG. 15 is a flowchart showing a processing procedure of the semiconductor wafer 1 in the second embodiment, and FIG. 6 shows a CAD (Computer Aided Design) screen 720 for drawing a template candidate 710 for an alignment mark. The CAD screen 720 is provided as a function of the foreign substance inspection device and is displayed on the screen of the display device 120. The CAD screen 720 is subjected to various processes by the host computer 100 based on instructions from the input device 110. A grid and a scale can be displayed on the CAD screen 720, and the approximate size of the template candidate 710 can be confirmed. In addition, by selecting a figure created on the screen of the display device 120 with a position display means such as a mouse pointer 730, shape information such as the width, height, and angle of the template candidate 710 is displayed on the appearance display means 750. be able to. On the CAD screen 720, for example, a drafting means 740 for drawing and editing the template candidate 710, such as a solid line, a broken line or a chain line, a circle or a rectangle, and a graphic inversion or rotation, size reduction, or enlargement, is an icon or mark. Alternatively, it is displayed as an icon such as a button. The template candidate 710 can be read out, edited, and saved from the external storage device 140 via the storage device 160 or a storage medium. In addition, the design data of the alignment mark can be input via the external storage device 140. Further, based on the image on the screen of the display device 120 drawn freehand through the position display means such as the mouse pointer 730, the template candidate 710 that matches the feature point of the image may be read from the storage device 160. it can.

  FIG. 7 is an example of a template candidate 710 used in this embodiment. (A) provides a rectangle inside the rectangular outer ring, and a cross-shaped figure at the center. (B) is an L-shaped outer ring in which L-shaped figures are arranged side by side. (C) is the figure which chamfered the corner | angular part of the pattern of (b). (D) is a figure in which a rectangle is arranged in the vicinity of (c).

  By executing the program (S201), the figures shown in (a) to (d) are created and registered in the storage device 160 as template candidates 710 for searching the surface of the semiconductor wafer 1. When searching for similar alignment marks present on the semiconductor wafer 1, one or more template candidates 710 are selected from the registered template candidates 710 and entered via the input device 110 (S202). It is also possible to read and use graphics created and registered in the past.

  In order to increase the matching accuracy, it is generally preferable that the template candidate 710 has many corners and the figure itself is independent. However, in the case where the angle deviation of the semiconductor wafer 1 is large or in a wafer having a low contrast process, recognition failure may occur. Therefore, it is desirable to select a template figure according to the situation.

  After selecting the template candidate 710, the conveyance program is executed. The designated semiconductor wafer 1 taken out from the wafer cassette 11 of the load port 10 is transported to the inspection unit 40 and fixed on the inspection stage chuck 42 by the same method as in the first embodiment (S203 to S205). The template candidate 710 can be selected after the semiconductor wafer is fixed on the inspection stage chuck 42, and only the inspection and evaluation processes for the template candidate 710 are automatically executed by setting the inspection mode. You can switch to

  After fixing the semiconductor wafer 1 on the inspection stage chuck 42, the servo motor of the stage chuck 42 is driven in the same manner as in the first embodiment to move to the designated template evaluation designated coordinates 810 (S206). The template evaluation designated coordinates 810 are provided on the display device 120 as a template candidate 710 evaluation screen shown in FIG. 9, and one or more coordinates can be set via the input device 110. A plurality of coordinates can be set, and an average evaluation result of the template candidate 710 can be obtained.

  The inspection stage chuck 42 is moved with the template evaluation designated coordinate 810 as a starting point, and a pattern similar to the template candidate 710 is searched from the pattern to be inspected 830 formed in the designated range 820 of the semiconductor wafer 1 by the CCD camera ( S207).

  Template candidates 710 registered in the storage device 160 are stored as line segment image data. The image of the pattern to be inspected 830 collected via the CCD camera (S208) is subjected to image processing into line-divided image data by the data processing unit 150, and pattern matching is performed with the entered template candidate 710 as needed. (S209). When the size of the template candidate 710 and the pattern to be inspected 830 is different or when the semiconductor wafer 1 is misaligned, the accuracy of pattern matching is reduced, so that the angle variable range setting 920 and the size variable range 930 are designated. Within the range (shape correction setting means), pattern matching is performed while changing the image data of the template candidate 710 in a matrix (image correction means) according to the size variable amount 991 and the angle variable amount 992 (variable amount setting means). Done. In the pattern matching, when a preset correlation coefficient 910 is satisfied, the coordinates of the inspection pattern 830 are detected by a position detector such as a laser scale provided in the inspection stage 41.

  Also, based on the feature points of the template candidate 710 and the feature points of the pattern to be inspected 830, a score value is calculated by, for example, equation (2) (S210). This score value can be used without being limited to this formula as long as the degree of pattern matching is digitized and the degree can be ranked. As the feature points of the pattern to be inspected 830 and the template candidate 710, the feature points at the time of pattern matching are used. However, the feature points are extracted by the same or different feature point extraction means by separately performing image processing on the mutual image data. It is also possible to use. By changing the pattern matching and the feature point for calculating the score value, the analysis accuracy of the difference between the template candidate 710 and the pattern to be inspected 830 can be improved.

  In equation (2), Ss is the size of the score value between the pattern-matched template candidate 710 and the pattern to be inspected 830, Ms is the number of matching feature points between the pattern to be inspected 830 and the template candidate 710, and Ts. Is the total number of feature points of the collected pattern 830 to be inspected, and Td is the total number of feature points of the template candidate 710.

  The calculated score values are calculated based on the size variable amount 991 and the angle variable amount 992 obtained by changing the template candidate 710 image at the time of pattern matching (correction value calculation means), the angle correction value 1050 and the size correction value. 1060 and the template candidate 710 and the image of the pattern to be inspected 830 are registered in the recording device 160 (S211). If the pattern to be inspected 830 is different from the template candidate 710 to be obtained, the search within the designated range 820 is continued. While repeating the above steps (S207 to 209), the pattern to be inspected 830 that satisfies the correlation coefficient 910 is registered in the same manner, and continues until the search within the designated range 820 is completed.

  The pattern matching is performed using, for example, geometric shape pattern matching. The feature points of the image data are extracted from the image data of the pattern to be inspected 830 by feature point extraction means such as Moravec operator, SUSAN (Smallest Univalue Segment Assimilating Nucleus), and MIC (Minimum Intensity Charge). The feature point extraction means can be set by the feature point extraction setting means 921 such as an input space, icon, button provided on the template evaluation setting screen. It is also possible to extract the position coordinates and the number of corners of the corner portion (such as a corner of the wiring pattern) of the pattern to be inspected 830 as feature points from the line-divided image data and perform pattern matching based on the feature points. . Since the degree of matching of the feature points varies depending on the template candidate 710 to be searched and the manufacturing process of the semiconductor wafer 1, it is desirable to select an appropriate method according to the use situation.

  FIG. 10 is a display screen showing the inspection result of the pattern to be inspected 830 detected from the template evaluation designated coordinates 810. The pattern to be inspected 830 that satisfies the correlation coefficient 910 by pattern matching has a score value rank 1010, template candidate 710 images, score values 1020, acquired images 1030, coordinates 1070, determination results 1040, angle correction values 1050, and size correction values 1060. A list is displayed (analysis result display means) on the display device 120 in association with various evaluation data such as (S212). This display can be changed depending on the setting in descending order of score value 1020 or in ascending order. When the setting of the template process selection means 960 is automatic setting 970, the one with the highest score value 1020 is automatically adopted as the template screen. In this way, by selecting the optimal template candidate 710 based on the analysis means of the foreign substance inspection apparatus, mismatching between artificial judgment and mechanical judgment is suppressed, and the optimal template can be selected. This template candidate 710 can be selected manually 980, and an appropriate image can be selected as a template from the list display while referring to the acquired image 1030, the score value 1020, and the like.

  Next, the inspection stage chuck 42 is moved to the coordinates of the inspection pattern 830 adopted as a template (S213). The angle of the inspection stage chuck 42 is changed for each angular resolution 940 (S214), the image position of the inspection pattern 830 is corrected (S215), and image data of the inspection pattern 830 is captured by the CCD camera for each angle. However, it is registered as a template in the storage device 160 (S217), and continues until imaging within the range set in the angle correction range 941 is completed (S214 to S218). When the imaging of the template image data group with the shifted angle is completed, the semiconductor wafer 1 is unloaded from the inspection stage chuck 42 and collected in the wafer cassette 11 (S219). Pattern identification accuracy can be improved by preparing a plurality of template images whose angles are shifted in advance and performing pattern matching using a group of image data whose angles are shifted. It is possible to detect and predict the angle shift amount from the image data obtained by pattern matching of the image data group. Based on the shift amount, pattern matching of one alignment mark (first alignment mark 312 or second alignment mark 322) is possible. Θ can be corrected by. Coarse θ correction is possible when the angle deviation of the semiconductor wafer 1 is large, and correction processing can be speeded up and throughput can be improved.

  In the present embodiment, after being selected as a template, an image data group in which the angle serving as the basic image data for angle correction is shifted is picked up. It can also be performed when the pattern 830 is detected, automating the template search and improving operability. However, since image data to be collected increases, it is desirable to collect the image data group after selecting it as a template in order to suppress the enlargement of the system.

  Note that when the angle deviation of the semiconductor wafer 1 is small and the image data group and the coarse θ correction are unnecessary, or when the search processing speed of the template candidate 710 is important, the processing step can be skipped. The data processing method can be changed by setting from the input device 110.

  In addition, the edge position 440 is detected from the collected image data in the same manner as in the first embodiment, and is similarly registered in the storage device 160, and the image data and the edge position 440 are used as a template. The semiconductor wafer 1 after the template evaluation is collected in the wafer cassette by the same method as in the first embodiment.

  Next, a method for detecting the alignment mark formed on the semiconductor wafer 1 will be described mainly with reference to FIGS. Those that overlap with the first and second embodiments are omitted, and FIGS. 1 and 2 are cited as necessary.

  FIG. 16 is a flowchart showing a processing procedure of the semiconductor wafer 1 in the third embodiment, and FIG. 11 is a setting screen for detecting alignment marks formed on the semiconductor wafer 1. Along with the execution of the transfer program (S301), the semiconductor wafer 1 to be measured is unloaded from the wafer cassette 11 and placed on the inspection stage chuck 42 by the same method as in the first embodiment (S302 to S304). Thereafter, the inspection stage chuck 42 moves to the position of the designated chip 1210 according to the setting on the setting screen (S305). After alignment with the designated coordinates 1220 of the designated chip 1210, the search range 1230 is scanned by the CCD camera (S306), and the average position waveform is sampled and the edge position 440 is obtained by the same method as in the first embodiment while imaging. It detects (S307-S308). When an object satisfying the edge position number 1250 is detected in the set image range 1240 (S309), the image data and coordinates of the object and feature points such as the calculated edge number are stored in the storage device 160. Register (S310). The object is registered while searching within the designated range 1220 (S306 to S310), and the imaging is completed upon completion of scanning of the designated range.

  The registered image data is displayed as a list on the display device 120 in ascending order or descending order of the number of edge positions 1250 on the display screen of the sampled object image shown in FIG. 12 (S311). By selecting the image data of the object displayed on the display device 120 and selecting the movement instruction means 1310 button via the input device 110, the inspection stage chuck 42 moves to the designated coordinates and visually confirms the object. By selecting the registration instruction means 1320 button, the object can be directly registered as a template (S312). If the reliability is insufficient, an image of the object can be collected by the same method as in the second embodiment and registered as a template. Note that the above buttons may be input means capable of inputting signals, and are not particularly limited thereto. Other input means such as icons and keyboards can also be employed.

  Next, a template reliability evaluation method, such as whether the selected template candidate 710 is valid as a template or whether it is misrecognized as another pattern formed on the semiconductor wafer 1, is mainly shown in FIGS. It explains using. A duplicated description of the first embodiment, the second embodiment, and the third embodiment is omitted, and FIGS. 1 and 2 are cited as necessary.

  FIG. 13 shows an operation screen used for the template reliability evaluation operation, and FIG. 17 shows a flowchart of the processing procedure. The operation screen for template reliability evaluation shown in FIG. 13 includes a registration number input unit 1410 for inputting a template registration number, a display instruction unit 1413 for displaying an image of the registration number, and a registered image for displaying template images in a thumbnail. Display means 1411; selection means 1412 for determining template candidates 710 for reliability evaluation; correlation coefficient setting means 1460 for pattern matching; score threshold setting means 1450 for setting score value threshold; semiconductor wafer 1 to be inspected As a setting screen on the display device 120, a chip setting unit 1430 for designating a chip within the chip, a coordinate setting unit 1431 for designating an in-chip coordinate as a starting point, an in-chip search range setting unit 1440 for setting a search area, and the like are included. The setting value can be changed. Also, the registration data related to the template candidate 710 at the time of registration includes image display means 1420 for displaying an image, in-chip coordinate display means 1421 for displaying a relative position in the chip, and the number of feature points for displaying the number of feature points such as edges. Displayed via display means 1422, score value display means 1423 for displaying the matching state of feature points, correlation coefficient display means 1425 for displaying the pattern matching status, manufacturing process display means 1424 for displaying the processing process of the semiconductor wafer 1, etc. Display function. By displaying the reference data when the template candidate 710 is collected, artificial setting mistakes are reduced and the condition setting for reliability evaluation is facilitated. In the present embodiment, these inputs and display means are composed of input and display spaces and buttons. However, the present invention can be applied as long as it can input, transmit and display signals. An icon, a keyboard, or other signal input transmission means or display means may be used.

  By selecting a button provided on the screen of the display device 120, the screen is displayed (S401). When the registered image display unit 1411 is selected, the template candidates 710 stored in the storage device 160 are displayed as thumbnails on a display screen (not shown) set separately. By selecting an arbitrary template candidate 710 from the template candidate 710 group using candidate selection means (not shown), data of the template candidate 710 is read from the storage device 160. When the registration number of the template candidate 710 is known, for example, the registration number is directly input to the registration number input unit 1410 by the input device 110, the data is read by the display instruction unit 1413, and the template candidate 710 is displayed as the image display unit. It may be performed while confirming at 1420. Next, by selecting the selection means 1412, it is adopted as a template candidate 710 to be subjected to reliability evaluation (S402). In this embodiment, reliability evaluation is performed based on one template candidate 710. However, it is also possible to set two or more template candidates 710 and perform reliability evaluation at the same time. The evaluation time required for adoption can be shortened.

  The semiconductor wafer 1 to be measured is unloaded from the wafer cassette 11 and placed on the inspection stage chuck 42 by the same method as in the first embodiment together with the execution instruction means 1470 of the reliability evaluation program (S403 to S405). Thereafter, the inspection stage chuck 42 moves to the setting position of the first chip setting means 1430 according to the setting on the setting screen (S406). Next, after adjusting the setting position of the chip coordinate setting means 1431 to the target position (pattern) formed on the semiconductor wafer 1 while scanning the setting range of the in-chip search range setting means 1440 with a CCD camera (S407). ) Is collected (S408). While performing pattern matching, it is determined whether or not the predetermined value of the correlation coefficient threshold value setting means 1460 is satisfied (S409), and the object (pattern) that satisfies the predetermined correlation coefficient is the edge shown in the first embodiment. A feature point is extracted by detection or the like, and a score value is calculated based on the feature point (S410). The coordinates of the object (pattern) satisfying the score value of the score threshold setting unit 1450 are collected and registered in the storage device 160 together with the feature points such as image data, correlation coefficient, and score value (S411). Since the degree of pattern matching can be adjusted by the correlation coefficient threshold value setting means 1460, it is possible to suppress the influence of contrast that changes in the manufacturing process. Further, the score threshold setting means 1450 can control the number of objects (patterns) to be sampled, and can suppress an increase in inspection time. Targets (patterns) that satisfy the conditions are registered as needed while searching within the in-chip search range 1440 (S407 to S411), and move to the set position of the next chip setting means 1430 upon completion of scanning of the in-chip search range 1440. (S406). The same object (pattern) search is repeated, and all the chip inspections set by the chip setting unit 1430 are completed and the imaging is ended.

  After inspecting all the chips set by the chip setting means 1430, the data of the object stored in the storage device 160 is displayed in a list (not shown) on the setting screen provided on the display device 120 screen. (S412). The list is configured by associating the template candidate 710 with the image data and feature points of the sampled object (pattern), and based on numerical values representing a pseudo degree such as a score value and a correlation coefficient, the chip setting unit 1430. Displayed in descending or descending order for each chip set in. It is possible to judge whether it is appropriate as a template from the numerical value of the score value and correlation coefficient, and the possibility of misrecognition from the magnitude of the score value and correlation coefficient between each target object (pattern) and the magnitude of the difference Can be judged. Further, the stability of the pattern matching can be determined from the order of the objects (patterns) between the chips set by the chip setting unit 1430 and the numerical value. The template is registered as a template by selecting a template candidate 710 and selecting a template selection means (not shown) provided on the display device 120 screen while referring to the result of the list (S412).

  When evaluating a plurality of template candidates 710, by selecting an automatic selection means, the data processing unit 50 determines and adopts an appropriate template candidate 710 from the above relationship. The template candidate 710 can also be selected by checking the state. At the time of this manual setting, for example, when the evaluation result of the list display is not satisfactory, the reliability can be re-evaluated by selecting the re-execution instruction unit 1471 after changing various settings. Needless to say, it can be re-evaluated even when various settings are not changed. This plurality of evaluations is effective for reducing the evaluation time and adopting a highly reliable template because, for example, the θ angle deviation of the semiconductor wafer 1 can be compared and evaluated under the same conditions. However, this template registration step (S412) is not indispensable, and the step may be skipped as long as the template candidate 710 is only evaluated for reliability. Along with the end of the evaluation, the wafer is carried out by the same method as in the first embodiment, collected in the original wafer cassette 11, and the program is ended (S413 to S414). The reliability evaluation function of the template candidate 710 can check in advance the possibility of erroneous recognition of the target pattern and the selection of the template candidate 710 and the threshold setting value for avoiding erroneous recognition. The operating rate can be dramatically improved.

  If the reliability evaluation program freezes due to misrecognition during the evaluation, the evaluation can be interrupted by the interruption instructing means 1480 or the evaluation can be stopped by the stop instructing means 1490. Further, it is possible to select the recovery means 1491 to cancel the error and restore the inspection apparatus to the initial state.

  The foreign substance inspection apparatus related to the manufacture of the semiconductor integrated circuit has been described above as an example. However, the technique related to the alignment of the present invention is manufactured from a flat panel display substrate, a glass substrate for TFT, a flat substrate such as an ALTIC substrate. The present invention can be widely applied to various processes and apparatuses necessary for manufacturing discs, flat panel display devices, masks and the like.

  In the present invention, the embodiment of the foreign matter inspection apparatus is shown. However, the present invention can be applied regardless of the semiconductor inspection apparatus and the semiconductor manufacturing apparatus. For example, the alignment mark formed on the inspection object is used. As long as the position correction processing of the inspection object is performed, it can be applied to a length measuring device such as a CD-SEM and an exposure device.

It is a top view which shows schematic structure of the foreign material inspection apparatus concerning this invention. It is a block diagram which shows the outline of control of a data processing part. It is a figure explaining the chip | tip used for alignment in a semiconductor wafer, and designated coordinates. It is a figure explaining the processing and edge position of an edge strength waveform. It is a figure explaining the setting screen of alignment. It is a figure explaining the CAD screen for drawing a template candidate. It is a figure explaining one Example of a template candidate. It is a figure explaining the designated coordinate and designated range for evaluating a template candidate. It is a figure explaining the setting screen for evaluating a template candidate. It is a figure explaining the screen which displays the evaluation result of a template candidate. It is a figure explaining the setting screen for searching for the alignment mark currently formed in the chip | tip of a semiconductor wafer. It is a figure explaining the screen which displays the search result of the alignment mark currently formed in the chip | tip of a semiconductor wafer. It is a figure explaining the setting screen for evaluating the reliability of a template candidate. It is a flowchart which shows the method of angle correction of the semiconductor wafer using a template. It is a flowchart which shows the method for extract | collecting an optimal template. 6 is a flowchart illustrating a method for searching for template candidates in a semiconductor wafer. It is a flowchart which shows the method for reliability evaluation of a template candidate, and selecting an appropriate template.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Semiconductor wafer, 10 ... Load port, 11 ... Wafer cassette, 20 ... Transfer part,
DESCRIPTION OF SYMBOLS 21 ... Transport assembly, 22 ... Handling arm, 22a ... Contact part, 23 ... Suction hole, 30 ... Pre-alignment part, 31 ... Pre-alignment chuck, 32 ... Detection apparatus, 40 ... Inspection part, 41 ... Inspection stage, 42 ... Inspection Stage chuck, 43a, 43b, 43c ... Elevating pins, 50 ... Data processing unit, 100 ... Host computer, 110 ... Input device, 120 ... Display device, 130 ... Output device, 140 ... External storage device, 150 ... Data arithmetic processing device , 160 ... storage device, 310 ... first chip, 320 ... second chip,
330 ... Intermediate correction chip, 311 ... First chip designated coordinate, 312 ... First alignment mark, 321 ... Second chip designated coordinate, 322 ... Second alignment mark, 410 ... Edge strength waveform, 411 ... Maximum edge strength value, 420 ... feature value setting value, 430 ... feature value threshold value, 440 ... edge position, 510 ... chip condition setting means, 520 ... coordinate condition setting means, 530 ... search condition setting means, 540 ... template setting means, 550 ... matching condition setting Means 560 ... Setting value display means 710 ... Template candidates 720 ... CAD screen 730 ... Mouse pointer 740 ... Drawing means 750 ... Appearance display means 810 ... Designated coordinates for template evaluation, 820 ... Designated range, 830 ... Pattern to be inspected, 910 ... correlation coefficient, 920 ... angle variable range setting, 921 ... Marking point extraction setting means, 930 ... size variable range, 940 ... angle resolution, 941 ... angle correction range, 960 ... processing selection means, 970 ... automatic setting, 980 ... manual setting, 991 ... size variable amount, 992 ... angle variable amount 1010: Ranking, 1020 ... Score value, 1030 ... Acquired image, 1040 ... Determination result, 1050 ... Angle correction value, 1060 ... Size correction value, 1070 ... Coordinate, 1210 ... Designated chip, 1220 ... Designated coordinate, 1230 ... Search range 1240 ... Image range, 1250 ... Number of edge positions, 1310 ... Movement instruction means, 1320 ... Registration instruction means, 1410 ... Registration number input means, 1411 ... Registration image display means, 1412 ... Selection means, 1413 ... Display instruction means, 1420 ... Image display means, 1421 ... In-chip coordinate display means, 1422 ... Feature point number display means 1423 ... Score value display means, 1424 ... Manufacturing process display means, 1425 ... Correlation coefficient display means, 1430 ... Chip setting means, 1431 ... Coordinate setting means, 1440 ... In-chip search range setting means, 1450 ... Score threshold setting means, 1460 ... correlation coefficient setting means, 1470 ... execution instruction means, 1471 ... re-execution instruction means, 1480 ... interruption instruction means, 1490 ... stop instruction means, 1491 ... recovery instruction means.

Claims (6)

  Foreign matter present on the surface of the inspection object by irradiating the surface of the inspection object with a light beam, obtaining an image signal from the received light intensity of the reflected or scattered light, and comparing with the image signal obtained from the adjacent inspection object A foreign matter inspection apparatus for inspecting the presence or absence of an inspection apparatus, comprising: an apparatus for registering feature points of an alignment mark formed on a surface of the inspection object; and an apparatus for reading the alignment mark on the surface of the inspection object, A foreign matter inspection apparatus characterized in that alignment is performed by detecting an alignment mark formed on the surface of the inspection object based on a feature point.   In a foreign matter inspection apparatus that irradiates a light beam to an inspection object and inspects for the presence or absence of foreign matter existing on the surface of the inspection object based on the received light intensity of light reflected or scattered from the inspection object, A device for inputting feature points of an alignment mark formed on the surface; a device for displaying the feature points of the input alignment mark; and a device for registering the feature points of the input alignment mark. A foreign matter inspection apparatus characterized in that an alignment mark formed on the surface of the object to be inspected is detected for alignment.   In a foreign matter inspection apparatus that irradiates a light beam to an inspection object and inspects for the presence or absence of foreign matter existing on the surface of the inspection object based on the received light intensity of light reflected or scattered from the inspection object, A device for registering feature points of alignment marks formed on the surface, a device for collecting image data of alignment marks formed on the surface of the object to be inspected, and extracting both feature points from the image data A foreign matter inspection apparatus comprising: a data calculation processing device that calculates a correlation value from a point, and registering image data of the alignment mark based on a threshold value of the correlation value.   In a foreign matter inspection apparatus that irradiates a light beam to an inspection object and inspects for the presence or absence of foreign matter existing on the surface of the inspection object based on the received light intensity of light reflected or scattered from the inspection object, A device for registering feature points of alignment marks formed on the surface, a device for collecting image data of alignment marks formed on the surface of the object to be inspected, and extracting both feature points from the image data A data calculation processing device that calculates a correlation value from a point, and a display device that displays the calculation result of the correlation value, and displays a determination result of pass / fail of the alignment mark based on a threshold value of the correlation value Foreign matter inspection apparatus characterized by the above.   5. The foreign matter inspection apparatus according to claim 4, wherein the calculation result of the correlation value is arranged in order of the correlation value and displayed in a list format on the display device.   Foreign matter present on the surface of the inspection object by irradiating the surface of the inspection object with a light beam, obtaining an image signal from the received light intensity of the reflected or scattered light, and comparing with the image signal obtained from the adjacent inspection object Image processing means for extracting feature points from image data of alignment marks formed on the surface of the object to be inspected and pattern matching with template feature points, and matching mutual feature points A calculation processing means for calculating a probability or a score (score), a determination processing means for identifying a case where the feature points match with a predetermined probability or a score (score) as a reference alignment mark, Arithmetic processing means for calculating a deviation amount of the object to be inspected from coordinates collected by recognizing at least two alignment marks, and based on the deviation amount; Particle inspection apparatus being characterized in that a drive mechanism for aligning by moving the inspection stage.

Images