WO2021176841A1 - 試料観察システム及び画像処理方法 - Google Patents

試料観察システム及び画像処理方法 Download PDF

Info

Publication number
WO2021176841A1
WO2021176841A1 PCT/JP2021/000484 JP2021000484W WO2021176841A1 WO 2021176841 A1 WO2021176841 A1 WO 2021176841A1 JP 2021000484 W JP2021000484 W JP 2021000484W WO 2021176841 A1 WO2021176841 A1 WO 2021176841A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
defect
learning
reference image
sample observation
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/JP2021/000484
Other languages
English (en)
French (fr)
Japanese (ja)
Inventor
直明 近藤
原田　実
陽平 嶺川
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hitachi High Tech Corp
Original Assignee
Hitachi High Tech Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hitachi High Tech Corp filed Critical Hitachi High Tech Corp
Priority to CN202180013541.6A priority Critical patent/CN115088061B/xx
Priority to KR1020227027317A priority patent/KR102808816B1/ko
Priority to US17/802,161 priority patent/US12333695B2/en
Publication of WO2021176841A1 publication Critical patent/WO2021176841A1/ja
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G06T7/001—Industrial image inspection using an image reference approach
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01J—ELECTRIC DISCHARGE TUBES OR DISCHARGE LAMPS
    • H01J37/00—Discharge tubes with provision for introducing objects or material to be exposed to the discharge, e.g. for the purpose of examination or processing thereof
    • H01J37/02—Details
    • H01J37/22—Optical, image processing or photographic arrangements associated with the tube
    • H01J37/222—Image processing arrangements associated with the tube
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B15/00—Measuring arrangements characterised by the use of electromagnetic waves or particle radiation, e.g. by the use of microwaves, X-rays, gamma rays or electrons
    • G01B15/04—Measuring arrangements characterised by the use of electromagnetic waves or particle radiation, e.g. by the use of microwaves, X-rays, gamma rays or electrons for measuring contours or curvatures
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
    • G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01J—ELECTRIC DISCHARGE TUBES OR DISCHARGE LAMPS
    • H01J37/00—Discharge tubes with provision for introducing objects or material to be exposed to the discharge, e.g. for the purpose of examination or processing thereof
    • H01J37/02—Details
    • H01J37/22—Optical, image processing or photographic arrangements associated with the tube
    • H—ELECTRICITY
    • H01—ELECTRIC ELEMENTS
    • H01J—ELECTRIC DISCHARGE TUBES OR DISCHARGE LAMPS
    • H01J37/00—Discharge tubes with provision for introducing objects or material to be exposed to the discharge, e.g. for the purpose of examination or processing thereof
    • H01J37/26—Electron or ion microscopes; Electron or ion diffraction tubes
    • H01J37/28—Electron or ion microscopes; Electron or ion diffraction tubes with scanning beams
    • H—ELECTRICITY
    • H10—SEMICONDUCTOR DEVICES; ELECTRIC SOLID-STATE DEVICES NOT OTHERWISE PROVIDED FOR
    • H10P—GENERIC PROCESSES OR APPARATUS FOR THE MANUFACTURE OR TREATMENT OF DEVICES COVERED BY CLASS H10
    • H10P74/00—Testing or measuring during manufacture or treatment of wafers, substrates or devices
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10056—Microscopic image
    • G06T2207/10061—Microscopic image from scanning electron microscope
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30141—Printed circuit board [PCB]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30148—Semiconductor; IC; Wafer

Definitions

  • the present invention relates to a sample observation system and an image processing method for observing circuit patterns and defects formed on a semiconductor wafer as a sample by using a charged particle microscope or the like.
  • a device for observing defects on a sample is a device for observing defects on a sample with high resolution based on the defect position coordinates (coordinate information indicating the position of the defect on the sample (wafer)) output by the defect inspection device.
  • a device that captures images and outputs an image, and a defect observation device (hereinafter referred to as a review SEM) using a scanning electron microscope (SEM) is widely used.
  • ADR Automatic Defect Review
  • ADC Automation Defect Classification
  • the ADR rediscovers the defect from the image captured by widening the field of view around the defect position coordinates output by the defect inspection device. , It has a function to obtain an image for observation by imaging the re-detected defect position at high magnification.
  • a defect detection method from the SEM image an image obtained by capturing an image of a region in which the same circuit pattern as the defect portion is formed is used as a reference image, and an image obtained by imaging the defect portion (hereinafter referred to as a defect image) is compared with the reference image.
  • a defect image an image obtained by imaging the defect portion
  • Patent Document 2 describes a method of generating a reference image from a database image according to design data and a captured image, comparing the captured image with the reference image, and detecting a defect candidate. ..
  • Non-Patent Document 1 discloses a method of learning the correspondence between an input image and an output image by using a neural network.
  • the system for observing defects on a sample according to the present invention (hereinafter referred to as a sample observation system) relates to a system for observing a sample such as a semiconductor wafer, acquiring an image, and observing the image.
  • Patent Document 1 describes a method of acquiring a reference image for each defect image and performing defect detection. However, if the reference image can be estimated from the defect image, the acquisition of the reference image can be omitted, so that the throughput of sample observation can be improved.
  • Patent Document 2 describes a method of generating a reference image from a database image according to design data and a captured image.
  • the design data is highly confidential information and cannot be taken out to a semiconductor manufacturing line, particularly a mass production line requiring high throughput, and it may be difficult to use the design data. As described above, when the design data cannot be used, it is difficult to estimate the reference image from the defective image, and none of the above-mentioned known examples mentions a method for solving this problem.
  • An object of the present invention is to solve the above-mentioned problems of the prior art, to make it possible to estimate a reference image from a defect image without using design data, and to improve the throughput of sample observation. There is.
  • Another object of the present invention is to solve the above-mentioned problems of the prior art, to make it possible to estimate the defect site from the defect image without using the design data, and to improve the throughput of sample observation. Is to enable.
  • the sample observation system and method having the scanning electron microscope and the computer of the present invention, or the computer included in the sample observation system (1) A plurality of images captured by the scanning electron microscope are acquired, and the images are acquired. (2) From the plurality of images, a learning defect image including the defect portion and a learning reference image not including the defect portion are acquired. (3) The estimation processing parameter is calculated by using the learning defect image and the learning reference image. (4) Obtain an inspection defect image including the defect part, (5) A pseudo reference image is estimated using the estimation processing parameters and the inspection defect image.
  • the sample observation system and method having the scanning electron microscope and the computer of the present invention, or the computer included in the sample observation system A plurality of images captured by the scanning electron microscope are acquired, and a plurality of images are acquired. From the plurality of images, a learning defect image including the defect site is acquired, and the defect image is obtained. Using the learning defect image, the estimation processing parameters are calculated. Obtain an inspection defect image including the defect site, The defect portion in the inspection defect image is estimated by using the estimation processing parameter and the inspection defect image.
  • the present invention it is possible to estimate a reference image from a defective image even when design data cannot be used in sample observation. Furthermore, by estimating the reference image, it is possible to omit the acquisition of the reference image, and it is possible to improve the throughput of sample observation.
  • the present invention it is possible to estimate the defect site from the defect image in the sample observation. Furthermore, by estimating the defect site from the defect image, it is possible to omit the acquisition of the reference image, and it is possible to improve the throughput of sample observation.
  • FIG. 1 It is a block diagram which shows the schematic structure of the sample observation system which concerns on Example 1.
  • FIG. It is a flow chart which shows the flow of the sample observation of the sample observation system which concerns on Example 1.
  • FIG. It is a figure which shows the example of the result of having specified the defect part in the sample observation of the sample observation system which concerns on Example 1.
  • FIG. It is a flow chart which shows the flow of the learning sequence of the sample observation system which concerns on Example 1.
  • FIG. It is a flow chart of the process which acquired the image pair for learning in the learning sequence of the sample observation system which concerns on Example 1.
  • FIG. It is a flow chart of the process which calculates the estimation process parameter in the learning sequence of the sample observation system which concerns on Example 1.
  • FIG. It is a figure which shows the process of aligning an image pair, and the process of cutting out an image in the calculation of the estimation processing parameter of the sample observation system which concerns on Example 1.
  • FIG. It is a block diagram which shows the structure of the neural network which estimates the pseudo reference image in the sample observation system which concerns on Example 1.
  • FIG. This is a GUI for setting a learning image size in the sample observation system according to the first embodiment. This is a GUI for setting learning end conditions in the sample observation system according to the first embodiment. This is a GUI for confirming the estimation error for each learning step in the sample observation system according to the first embodiment.
  • FIG. It is a timing chart of the process of acquiring the defect image for observation by the conventional sample observation system which compares with the sample observation system which concerns on Example 1.
  • FIG. It is a timing chart of the process of acquiring the defect image for observation by the sample observation system which concerns on Example 1.
  • FIG. It is a flow chart which shows the flow of the learning sequence of the sample observation system which concerns on Example 2.
  • FIG. It is a block diagram which shows the schematic structure of the computer in the sample observation system which concerns on Example 3.
  • FIG. It is a flow chart which shows the flow of the sample observation of the sample observation system which concerns on Example 3.
  • FIG. It is a flow chart of the process which calculates the estimation process parameter in the learning sequence of the sample observation system which concerns on Example 3.
  • FIG. It is a block diagram which shows the structure of the neural network which estimates the defect part in the sample observation system which concerns on Example 3.
  • FIG. It is a figure which shows the formula used when the pseudo reference image is estimated from the learning partial defect image of Examples 1 to 3. It is a figure which shows the formula used in the process of Example 3.
  • the sample observation system related to this embodiment will be described with reference to FIG.
  • a sample observation system equipped with a scanning electron microscope (SEM) as an imaging device for imaging a sample will be described.
  • the imaging device according to this embodiment may be an imaging device other than SEM, and may be an imaging device using charged particles such as an optical microscope or ions.
  • the image of the defect on the semiconductor wafer will be described as the image to be observed, an image of another sample such as a flat panel display or a biological sample may be used.
  • FIG. 1 shows the configuration of the sample observation device 100 according to this embodiment.
  • the sample observation device (also referred to as a sample observation system) 100 includes an SEM 101 that captures an image of a sample, a defect inspection device 102 that detects defects in the sample, and a computer 103.
  • the SEM 101 includes a stage 105 on which the sample wafer 104 to be observed is mounted and movable in the XY plane or the XYZ space, and an electron source 107 that generates an electron beam 106 to irradiate the sample wafer 104.
  • an electron lens (not shown) that converges the electron beam 106 on the sample wafer 104.
  • a deflector (not shown) for scanning the electron beam 106 on the sample wafer 104 is provided.
  • the defect inspection device 102 is a device that captures an optical image of the wafer surface and inspects the defect by comparing it with an image of a non-defective part (for example, an image of an adjacent chip).
  • a non-defective part for example, an image of an adjacent chip.
  • such an inspection device is affected by the illumination wavelength, and the resolution limit of the acquired image is about several hundred nanometers. Therefore, with respect to a defect on the order of several tens of nanometers on the wafer, the presence or absence of the defect can only be detected, and the defect coordinates on the wafer are output.
  • the computer 103 includes a user interface (denoted as user I / F in the figure) 111, a network interface (denoted as network I / F in the figure) 112, a control unit 113 for controlling the SEM 101, a storage unit 114 for storing information, and processing.
  • the unit 115 is provided.
  • the storage unit 114 is, for example, a magnetic disk device, a volatile and non-volatile semiconductor memory device, but may be another storage device. Further, the storage unit may be composed of a plurality of the above-mentioned devices.
  • the example of the processing unit may be any of CPU, GPU, FPGA, and LSI, and may be realized in combination.
  • the control unit 113 which will be described later, may be a subsystem (sometimes referred to as a control subsystem) different from the computer 103.
  • the computer 103 will be described by one example in the present specification, it may be a plurality of computers.
  • the processing for displaying the GUI is performed by a display computer such as a tablet or smartphone, and the other image processing is performed by another computer.
  • the computer may have a plurality of the above-mentioned components.
  • the user interface 111 is, for example, a touch panel, a display, a keyboard, a mouse, or the like, but may be another device as long as it can accept operations from a worker (user) and display information.
  • the network interface 112 is an interface for communicating with an external device such as a defect inspection device or SEM via a network.
  • control unit 113 The configuration of the control unit 113, the processing unit 115, and the storage unit 114 according to this embodiment will be described.
  • the control unit 113 includes a stage control unit 116, an electron beam control unit 117, and a detector control unit 118.
  • the stage control unit 116 controls the movement and stop of the stage 105.
  • the electron beam control unit 117 controls a deflector (not shown) so that the electron beam 106 is irradiated in a predetermined field of view, and controls the scan area of the electron beam 106 on the sample wafer 104.
  • the detector control unit 118 samples the signal from the detector 110 in synchronization with the scan of the electron beam 106 driven by a deflector (not shown), adjusts the gain, offset, and the like, and generates a digital image.
  • the control unit 113 may be realized by, for example, a circuit, or may be realized by a CPU, GPU, FPGA, or LSI.
  • the storage unit 114 has an image storage area 119 that stores digital images generated by the detector control unit 118 together with incidental information, a recipe storage area 120 that stores recipes including device and manufacturing process information, image imaging conditions, and the like.
  • the estimation processing parameter storage area 121 for storing the parameters related to the estimation processing of the pseudo reference image is included. It should be noted that each area does not necessarily have to be a separate area.
  • the data arrangement inside the storage unit 114 may be in any format as long as the recipe, parameters, and estimation processing parameters can be stored.
  • the processing unit 115 includes an estimation processing parameter calculation unit 122 that calculates an estimation processing parameter for estimating a pseudo reference image from a defect image, a pseudo reference image estimation unit 123 that estimates a pseudo reference image based on the estimation processing parameter, and a defect image. It is provided with a defect site identification unit 124 for identifying a defect site inside. If the processing unit 115 is a device such as a CPU or GPU that executes a predetermined process by a program, a program corresponding to the estimation processing parameter calculation unit 122, the pseudo reference image estimation unit 123, and the defect site identification unit 124 (collectively). An image processing program) is stored in the storage unit 114. Then, the processing unit 115 realizes these processes by reading the program.
  • the subsequent description of the processing according to the first to third embodiments is the processing performed by the computer 103. More specifically, the control of the SEM 101 is a process performed by the control unit 113, and the other processes are performed by the process unit 115, which is an example of sharing the process.
  • the processing unit 115 is a CPU or GPU
  • the program stored in the storage unit 114 (hereinafter, may be referred to as an image processing program) is read and realized.
  • the processing is performed by the estimation processing parameter calculation unit 122, the pseudo-reference image estimation unit 123, or the defect site identification unit 124, these units are included in the processing unit 115 and are therefore processed by the processing unit 115. You may think that.
  • the sample wafer 104 wait for the sample wafer 104 to be observed to be loaded on the stage 105 (S201).
  • the recipe corresponding to the sample wafer to be observed is read from the recipe storage area (S202).
  • the semiconductor pattern formed on the sample wafer 104 is manufactured through a number of manufacturing steps, and the appearance may be significantly different in each step. Further, the characteristics of the sample such as the ease of charging may differ. Therefore, it is common to adjust and store the imaging conditions for each device and manufacturing process. For the same reason, the estimation accuracy is improved by managing the estimation processing parameters of the pseudo-reference image for each process.
  • the defect coordinate information output by the defect inspection device 102 is received or read (S203).
  • all of the received or read defect coordinates may be the observation target, or the one sampled based on the user-specified condition may be the observation target.
  • it is confirmed whether or not the estimation processing parameters corresponding to the process in which the sample wafer 104 is processed are stored in the estimation processing parameter storage area 121 (S204), and if they are not stored (“None” in FIG. 2). In the case of), the estimation processing parameter is calculated by the learning sequence described later, and the result is stored in the estimation processing parameter storage area 121 (S205).
  • the defect coordinate information may be received from the defect inspection device described above via the network interface 112, or may be read from a portable storage medium such as a USB memory. Note that S205 may be realized by the estimation processing parameter calculation unit 122, or may be realized by the entire calculation unit 115 as described above.
  • the estimation processing parameters of the pseudo reference image are read from the estimation processing parameter storage area 121 (S206).
  • the defects to be observed on the sample wafer 104 are sequentially imaged using SEM101, and the following series of observations from S207 to S213 are performed.
  • the stage 105 is controlled and moved through the control unit 113 so that the defect to be observed on the sample wafer 104 is included in the imaging field of view of the SEM 101 (S207).
  • an inspection defect image is acquired (S208).
  • the electron beam 106 is irradiated on the sample wafer 104 by irradiating a relatively wide area including the observation target defect with the SEM 101 and scanned, and the generated secondary electrons 108 and backscattered electrons 109 are detected by the detector 110.
  • a relatively wide area including the defect to be observed is imaged, and the detection signal from the detector 110 obtained by the image is processed by the detector control unit 118 to process the relatively wide area including the defect to be observed. This is done by acquiring a low-magnification image of.
  • image preprocessing such as noise removal and brightness unevenness correction is applied to the defective image for inspection (S209), and the pseudo-reference image estimation unit 123 uses the estimation processing parameters read from the estimation processing parameter storage area 121.
  • a pseudo reference image is estimated from the inspection defect image that has undergone image preprocessing (S210).
  • This pseudo-reference image corresponds to an image in which a circuit pattern similar to that of the inspection defect image is observed and does not include defects. Design data is not required for estimating the pseudo-reference image, and the estimation processing parameters and the inspection defect image that has undergone image preprocessing are used.
  • the defect site identification unit 124 compares the inspection defect image with the pseudo-reference image, and identifies the defect site from the inspection defect image (S211).
  • the method described in Patent Document 1 or the like may be used.
  • the identified defect site is imaged at a high magnification with a narrow field of view, and an observation defect image (hereinafter referred to as an observation defect image) is acquired (S212), and an inspection defect image and a pseudo reference image are obtained. Then, the defective image for observation is stored in the image storage area 119 (S213).
  • an observation defect image hereinafter referred to as an observation defect image
  • the series of processes may be repeatedly executed for all the observation target defects of the sample wafer 104, but may be performed for some observation target defects based on other criteria.
  • the estimation processing parameter calculation unit 122 may be responsible for S205, or the processing unit 115 as a whole may be in charge of S205.
  • Example of the result of identifying the defective part An example of the result of identifying the defective portion in S211 will be described with reference to FIG.
  • the specific defect portion image 303 capable of distinguishing between the defect portion and the region other than the defect portion can be obtained.
  • the pixel value of the defect portion may be 1, and the pixel value of the region other than the defect portion may be 0.
  • the defect to be learned is set (S402). For this, all of the defect coordinates read in S203 may be used as a learning target, or those sampled based on a user-specified condition may be used as a learning target.
  • the set image including the defect to be learned hereinafter referred to as the defect image for learning
  • the region designed to form a circuit pattern similar to the vicinity of the defect position to be learned are included.
  • a pair of images hereinafter referred to as a learning reference image is acquired (S403).
  • the stage 105 is included in the imaging field of view of the SEM 101 so that a region (hereinafter referred to as a reference region) designed (or assumed) to form a circuit pattern similar to that around the defect position to be learned is included in the imaging field of view of the SEM 101.
  • a region hereinafter referred to as a reference region
  • information as accurate as the design data is not required to specify the reference area.
  • a semiconductor wafer a plurality of chips designed to form a similar circuit pattern are arranged on the wafer, so the simplest method is to use a region shifted by one chip from the defect coordinates as a reference region. Can be considered.
  • the reference area may be selected by other methods.
  • the reference region on the sample wafer 104 is irradiated with the electron beam 106 and scanned, and the generated secondary electrons 108 and backscattered electrons 109 are detected by the detector 110 to image the reference region, and the reference region is imaged.
  • the detection signal from the obtained detector 110 is processed by the detector control unit 118, and a learning reference image is acquired so as to be larger than the size acquired in S401 (S502).
  • the stage 105 is controlled and moved so that the region containing the defect to be learned (hereinafter referred to as the defect region) is included in the imaging field of view of the SEM 101 (S503).
  • the defect region on the sample wafer 104 is irradiated with the electron beam 106 and scanned, and the generated secondary electrons 108 and backscattered electrons 109 are detected by the detector 110 to image the defect region, and the defect region is imaged.
  • the detection signal from the obtained detector 110 is processed by the detector control unit 118, and a learning defect image is acquired so as to be larger than the size acquired in S401 (S504).
  • incidental information is added so that the learning defect image and the learning reference image are paired and stored in the image storage area 119 (S505).
  • either of the learning reference image and the learning defect image may be acquired first. That is, the processes of S501 to S502 may be executed after S503 to S504.
  • the learning defect image acquired in S403 is compared with the learning reference image, and the defect portion is specified in the same manner as in S211 (S404).
  • the identified defect site is imaged at a high magnification with a narrow field of view, an observation defect image is acquired, the observation defect image is stored in the image storage area 119, and the defects to be observed in S207 to S213 are used. Is excluded (S405).
  • the above processes of S403 to S405 are repeatedly executed for all or a part of the defects to be learned of the sample wafer 104.
  • the above is related to the parallel processing of the learning sequence and the defect observation processing.
  • the estimation processing parameter calculation unit 122 calculates the estimation processing parameter for estimating the pseudo reference image (S407: details will be described later).
  • S404 and S405 in addition to the above-mentioned merit of parallel processing, since it is not necessary to acquire an observation defect image in S212, there is a merit that sample observation can be performed efficiently. Note that S404 to S405 may be omitted, and if omitted, the defects to be learned are not excluded from the defects to be observed in S207 to S213.
  • image preprocessing such as noise removal and luminance unevenness correction is applied to all the learning defect images and the learning reference images acquired in S403 (S601).
  • image preprocessing such as noise removal and luminance unevenness correction is applied to all the learning defect images and the learning reference images acquired in S403 (S601).
  • the pair of the learning defect image and the learning reference image are aligned based on a predetermined evaluation value, and the alignment amount AX between the images is performed.
  • AY is obtained (S602).
  • a predetermined evaluation value a normalized intercorrelation coefficient, a mean square error, or the like may be used, and alignment may be performed based on the position where the evaluation value becomes the maximum or the minimum. If the image resolutions (the number of pixels per image in the same field of view) are different, the image resolutions shall be made uniform by linear interpolation or the like before alignment.
  • a learning partial defect image is cut out from the learning defect image, and a learning partial reference image is cut out from the learning reference image (S603).
  • FIG. 7 shows a learning defect image 701 and a learning reference image 702.
  • these two images are aligned based on a predetermined evaluation value, and an image alignment result 703 is obtained.
  • the image alignment result 703 does not show the defect portion included only in the learning defect image 701.
  • a learning partial defect image 704 is cut out from the learning defect image 701 as a region common to the learning defect image and the learning reference image based on the alignment amounts AX and AY, and the learning reference image 702 is used for learning.
  • a partial reference image 705 is cut out.
  • the previously learned estimation processing parameters may be read out from the estimation processing parameter storage area 121 and used as the initial values of the estimation processing parameters during learning.
  • the pseudo-defect image is estimated from the learning partial defect image (S605) based on the learning estimation processing parameter, the estimation error between the learning partial reference image and the pseudo-defect image is calculated (S606), and the estimation error is estimated.
  • the estimation processing parameter during learning is updated so that Next, it is confirmed whether or not the learning end condition acquired in S406 is satisfied (S608). If the learning end condition is satisfied, the estimation processing parameter during learning is stored in the estimation processing parameter storage area 121 as an estimation processing parameter (S609). If the learning end condition is not satisfied, the process returns to S605 again.
  • learning end condition 1 The estimation error is compared with the preset estimation error threshold value TH, and the estimation error is smaller than the estimation error threshold value TH.
  • Learning end condition 2 An operation to end learning was accepted from the user.
  • Learning end condition 3 The processes from S605 to S608 were repeated a predetermined number of times MR.
  • Non-Patent Document 1 the neural network described in Non-Patent Document 1 may be used.
  • This neural network is also used in S210 when estimating a pseudo reference image from the inspection defect image.
  • a U-shaped neural network called U-net as shown in FIG. 8 may be used.
  • Y indicates an input image.
  • F11 (Y), F12 (Y), F13 (Y), F21 (Y), F22 (Y), F23 (Y), F24 (Y), F31 (Y), F32 (Y) indicate intermediate data.
  • F (Y) indicates the estimation result of the pseudo reference image.
  • Equations 1 to 10 The intermediate data and the final result are calculated by Equations 1 to 10 in FIG.
  • "*" represents a convolution operation
  • DS represents an operation of applying a 2 ⁇ 2 max filter to an input image and reducing it in half in the space (XY) direction
  • US represents an input image.
  • CC represents an operation of combining two input images in the channel direction.
  • Equations 1 to 10 the meanings of the variables used in Equations 1 to 10 are as follows: W1 is c1 c0 ⁇ f1 ⁇ f1 size filter c0 is the number of channels of the input image f1 is the size of the spatial filter A c1 dimensional feature map is obtained by folding the c0 ⁇ f1 ⁇ f1 size filter c1 times into the input image. Be done.
  • B1 is a c1 dimensional vector (bias component corresponding to c1 filter)
  • W2 is a c1 ⁇ f2 ⁇ f2 size filter
  • B2 is a c2 dimensional vector
  • W3 is a c2 ⁇ f3 ⁇ f3 size filter
  • B3 is a c3 dimensional vector
  • W4 is a c3 ⁇ f4 ⁇ f4 size filter
  • B4 is a c2 dimensional vector
  • W5 is (C2 ⁇ 2)
  • B5 is c2 dimensional vector
  • W6 is c2 ⁇ f6 ⁇ f6 size filter
  • B6 is c1 dimensional vector W7 is (c1 ⁇ 2) ⁇ f7 ⁇ f7 size filter B7
  • the c4 dimensional vector W8 is a c4 ⁇ f8 ⁇ f8 size filter B8 is a c5
  • c0 and c5 are values determined by the number of channels of the learning partial defect image and the learning partial reference image.
  • the parameters calculated by the estimation processing parameter calculation processing (S405) are W1 to W8 and B1 to B8.
  • the estimation error calculation process (S607) is a process of evaluating the difference (error) between the estimation result F (Y) and the learning partial reference image, and the parameters are updated so that the estimation error obtained in this process becomes smaller. It is said.
  • a mean square error (Mean Squared Error) or the like may be used.
  • a general error back propagation method may be used in the learning of the neural network. Further, when calculating the estimation error, all the pairs of the acquired learning defect image and the learning reference image may be used, but a mini-batch method may be adopted. That is, a plurality of image pairs may be randomly extracted from the pair of the learning defect image and the learning reference image, and the parameters may be updated repeatedly. Further, a patch image may be randomly cut out from one image pair and used as an input image Y of the neural network. As a result, learning can be performed efficiently.
  • FIG. 9 shows the GUI 900 for setting the size of the learning image in S401.
  • the GUI 900 displays the inspection defect image size 901 and the imaging field of view 902 included in the recipe read in S202. Further, the GUI 900 displays an input unit 903 for setting the learning image size. After setting the learning image size through the GUI 900, pressing the "OK" button 904 executes the processing after S402. NS.
  • the input image is reduced to a size (XY direction) of 1 / (2 ⁇ d) at the depth d. Therefore, when using a neural network with a maximum depth D, the input image is required to have a size of (2 ⁇ D) ⁇ (2 ⁇ D) or more.
  • the sizes of the learning defective image and the learning partial reference image obtained by aligning the learning defective image and the learning reference image in S602 are the sizes of the learning defective image and the learning reference image, respectively. It becomes as follows. Specifically, the image size is reduced by the amount of alignment AX and AY.
  • the learning defect image and the learning reference image are acquired in S502 and S504 with the same size as the inspection defect image, and the learning defect image and the learning reference are acquired in S602.
  • the predetermined size is (2 ⁇ D) ⁇ (2 ⁇ D) when a neural network having a maximum depth D is used.
  • FIG. 10 shows the GUI 1000 for setting the learning end condition in S406.
  • an input unit 1001 for setting the number of times MR of repeating the processes of S605 to S608, an input unit 1002 for setting the estimation error threshold value TH, and an input unit 1003 for setting whether or not to accept the learning end operation by the user are displayed.
  • the estimation processing parameter of S407 is calculated by pressing the "learning start” button 1004. If the "Cancel" button 1006 is pressed during the calculation of the estimation processing parameters, the calculation of the estimation processing parameters can be interrupted.
  • the GUI 1100 is switched to confirm the progress of the learning estimation processing parameter update processing as shown in FIG.
  • the number of repetitions of the estimation parameter update and the transition of the estimation error are displayed on the graph 1101.
  • the "learning end” button 1103 of the GUI 1100 it is considered that the learning end operation has been accepted from the user, and the update of the estimation processing parameter during learning is finished, that is, it is determined in S608 that the learning end condition is satisfied.
  • the "estimated image confirmation" button 1102 of the GUI 1100 is pressed, the screen switches to the GUI 1200 as shown in FIG.
  • the image ID selection button 1201 On the GUI 1200, press the image ID selection button 1201 to specify the number of the image to be displayed, and the channel selection unit 1202 selects the type of image to be displayed such as the secondary electron image (SE) and the reflected electron image (BSE). Therefore, the pseudo-defect image estimation process for the specified image ID is executed using the in-learning estimation parameter, and the learning partial defect image 1203 of the specified image ID, the estimated pseudo-defect image 1204, and the learning The partial reference image 1205 is displayed.
  • the "OK" button 1206 of the GUI 1200 When the "OK" button 1206 of the GUI 1200 is pressed, the display switches to the original GUI 1100 display as shown in FIG.
  • a learning defect image and a learning reference image are acquired, estimation processing parameters are calculated using the learning defect image and the learning reference image, and in sample observation, a pseudo reference image is obtained from the inspection defect image.
  • FIG. 13 shows a sequence in which the defects (1) and (2) to be observed are sequentially observed in the conventional sample observation system.
  • the horizontal axis represents time, and the vertical axis represents the defect to be observed.
  • the sequence of step 1301 relating to the observation of the defect to be observed (1) includes: The stage is moved so that the reference region corresponding to the observation target defect (1) is included in the imaging field of view of the SEM101 (S).
  • the reference area is imaged with SEM101 to acquire a reference image for learning (RI).
  • the stage is moved so that the region including the observation target defect (1) is included in the imaging field of view of the SEM101 (S).
  • a relatively wide area including the defect to be observed (1) is imaged by SEM101 to acquire a defect image for learning (DI).
  • D defect image and the learning reference image
  • the defect site in the learning defect image is identified (D).
  • a relatively narrow area including the identified defect site is imaged with SEM101 to acquire an observation defect image (HI).
  • step 1302 A similar sequence is included in step 1302 relating to the observation of the next observation target defect (2).
  • stage movement (S) of the step 1302 is performed after the acquisition (HI) of the observation defect image of the observation target defect (1) is completed. This is because the observation target defect (1) remains within the imaging field of view of the SEM 101 until the acquisition (HI) of the observation defect image (HI) of the observation target defect (1) is completed in the step 1301.
  • FIG. 14 is a sequence relating to the processing of S207 to S213 of this embodiment.
  • the relationship between the horizontal axis and the vertical axis is the same as in the case of FIG.
  • the sequence of step 1401 relating to the observation of the defect to be observed (1) includes: The stage is moved so that the region including the observation target defect (1) is included in the imaging field of view of the SEM101 (S).
  • a relatively wide area including the defect to be observed (1) is imaged by SEM101 to acquire an inspection defect image (DI).
  • a pseudo-reference image is estimated from the inspection defect image based on the estimation processing parameters (P).
  • the defect site in the inspection defect image is identified by using the inspection defect image and the pseudo reference image (D).
  • a relatively narrow area including the identified defect site is imaged with SEM101 to acquire an observation defect image (HI).
  • step 1402 relating to the observation of the observation target defect (2), the same processing is performed on the observation target defect (2).
  • the sequence of FIG. 14 by estimating the pseudo reference image (P) from the defect image, the first stage movement (S) of the sequence described in FIG. 13 and the acquisition of the reference image (RI) are performed. It becomes unnecessary. As a result, the number of stage movements can be reduced by half, the imaging of the reference image can be omitted, and a plurality of defects to be observed on the sample wafer 104 can be detected by using the sample observation system. It is possible to improve the throughput when observing sequentially.
  • the defect region and the reference region are imaged by using SEM, the defect image for learning and the reference image for learning are acquired, and the estimation processing parameters are calculated using the defect image for learning and the reference image for learning.
  • the sample observation a method of improving the sample observation throughput by estimating a pseudo-reference image from the inspection defect image has been described.
  • the estimation processing parameters the larger the number of pairs of the learning defect image and the learning reference image, the more efficiently the learning becomes possible.
  • a learning defect image is generated by adding a pseudo defect to the learning reference image, and the learning reference image and the generated learning defect are generated.
  • a method of calculating the estimation processing parameter using a pair of images will be described.
  • the configuration of the sample observation system according to this embodiment is basically the same as the configuration shown in FIG. 1 described in Example 1. What is different is the processing flow of the learning sequence, and the sample observation flow other than the learning sequence includes a processing flow equivalent to the processing flow shown in FIG. 2 described in Example 1. Further, the GUI of the sample observation system according to the present embodiment has an interface equivalent to that shown in FIGS. 9 to 12 described in the first embodiment. Hereinafter, only the parts different from those of the first embodiment will be described.
  • set the area to be learned (S1501). This may be one or more regions specified by the user on the sample wafer, or one or more regions on the sample wafer 104 may be randomly set. However, the area to be learned does not include the defect coordinates output by the defect inspection device.
  • the stage 105 is controlled and moved (S1502) so that the set region to be learned is included in the imaging field of the SEM 101, and the electron beam 106 is irradiated to the region to be learned on the sample wafer 104.
  • the region to be learned is imaged by detecting the generated secondary electrons 108 and backscattered electrons 109 with the detector 110, and the detection signal from the detector 110 obtained by the image is captured by the detector control unit. Processing is performed in 118 to acquire a reference image for learning (S1503).
  • a learning defect image is acquired (S1504) so that the learning defect image and the learning reference image are paired.
  • Ancillary information is added and stored in the image storage area 119 (S1505).
  • the center position and size (width and height) of the region PR to which the pseudo defect is given may be randomly set in the plane of the learning reference image.
  • a certain offset may be added to the shade of the area PR as a pseudo defect.
  • the region PR may be set so as to include the edge of the circuit pattern, and the circuit pattern may be deformed based on the edge strength.
  • the above is a simulation of a minute defect, but a huge defect that covers the entire image may be generated.
  • the types of pseudo defects are not limited to these, and various defects may be modeled and generated.
  • an image including all the defect coordinates read in S203 or the defect coordinates sampled based on the user-specified condition may be used as the learning defect image. That is, in S1507, a pair of an image (first defect image) of an image containing a defect on the sample wafer 104 and a reference image (first reference image) corresponding to the first defect image, and a first The estimation processing parameter may be calculated by using the pair of the second reference image and the image containing the pseudo defect (second defect image) generated from the second reference image.
  • a learning defect image is generated by adding a pseudo defect to the learning reference image, and the learning reference image and the generated learning defect are generated. It is possible to calculate the estimation processing parameters using the pair of images, and in sample observation, by estimating the pseudo reference image from the defect image for inspection, it is possible to omit the acquisition of the reference image, and sample observation. It is possible to improve the throughput of.
  • a pseudo reference image is estimated from the inspection defect image in sample observation using the estimation processing parameters obtained by learning the correspondence between the learning defect image and the learning reference image.
  • a method of identifying a defect site in an inspection defect image by comparing an inspection defect image and a pseudo-reference image has been described.
  • a method of calculating the estimation processing parameters for estimating the defect portion in the defect image and estimating the defect portion in the inspection defect image based on the estimation processing parameter will be described.
  • the SEM 101 and the defect inspection device 102 are the same as the configuration shown in FIG. 1 described in the first embodiment, and the difference is the configuration of the computer 103.
  • the parts different from those of the first embodiment will be described.
  • the computer 103 of the sample observation system will be described with reference to FIG.
  • the storage unit 114 has an image storage area 119 that stores digital images generated by the detector control unit 118 together with incidental information, a recipe storage area 120 that stores recipes including device and manufacturing process information, image imaging conditions, and the like.
  • An estimation processing parameter storage area 1601 for storing parameters related to estimation processing of a defect portion in a defect image is provided.
  • the processing unit 115 includes an estimation processing parameter calculation unit 1602 that calculates an estimation processing parameter for estimating a defect portion in a defect image, and a defect portion estimation unit 1603 that estimates a defect portion based on the estimation processing parameter.
  • estimation processing parameters corresponding to the process in which the sample wafer 104 is processed are stored in the estimation processing parameter storage area 1601, and if they are not stored (when “none” in FIG. 17). Is calculated and stored in the estimation processing parameter calculation unit 1602 by the learning sequence described later (S1705). Next, the estimation processing parameter is read from the estimation processing parameter storage area 1601 (S1706).
  • the defects to be observed on the sample wafer 104 are sequentially imaged using SEM101, and a series of observations are performed. Since S1707 to S1709 are the same as S207 to S209 in FIG. 2, the description thereof will be omitted.
  • the defect site estimation unit 1603 estimates the defect site in the inspection defect image that has undergone image preprocessing using the estimation processing parameters (S1710). Design data is not required for estimating the defect portion, and the defect image for inspection to which the estimation processing parameters and the image preprocessing have been performed is used.
  • the estimated defect portion is imaged at a high magnification with a narrow field of view, an observation defect image is acquired (S1711), and the inspection defect image and the observation defect image are stored in the image storage area 119 (S1712). ..
  • the above processes from S1707 to S1712 are repeatedly executed for all the observed defects of the sample wafer 104.
  • a method of calculating the estimation processing parameter corresponding to S407 of FIG. 4 will be described with reference to FIG. Design data is not required for the calculation of the estimation processing parameters, and the learning defect image and the learning reference image are used.
  • S1801 to S1803 are the same as S601 to S603 in FIG. 6, the description thereof will be omitted.
  • the learning partial defect image and the learning partial reference image are compared, the defective portion in the learning partial defective image is specified, and the specific defective portion image showing the defective portion is obtained (S1804: S211 in FIG. 2). Equivalent).
  • the learning defect image is displayed on the GUI without specifying the defect portion using the learning defect image and the learning reference image, the user specifies the defect portion, and the designated defect portion is designated. An image in which a region other than the defective portion can be distinguished may be used as a specific defective portion image.
  • the estimation processing parameters during learning are initialized (S1805: corresponding to S604 in FIG. 6). At this time, the previously learned estimation processing parameters may be read out from the estimation processing parameter storage area 1601 and used as the initial values of the estimation processing parameters during learning.
  • the defect site estimation unit 1603 estimates the defect site from the learning partial defect image based on the learning estimation processing parameter (S1806), obtains the estimated defect site image, and obtains the estimated defect site image and the specific defect site image.
  • the estimation error is calculated using (S1807), and the estimation processing parameter during learning is updated so that the estimation error becomes small (S1808: corresponding to S607 in FIG. 6).
  • it is confirmed whether or not the learning end condition acquired in S406 is satisfied (S1809: corresponding to S608 in FIG. 6), and if the learning end condition is satisfied, the estimation processing parameter during learning is used as the estimation processing parameter for estimation processing. It is stored in the parameter storage area 1601 (S1810).
  • the learning end conditions are as described in Example 1. If the learning end condition is not satisfied, the process returns to S1806 again. This completes the flow of FIG.
  • the neural network shown in FIG. 19 may be used, the input image Y is the learning partial defect image, and F (Y) is the estimated defect portion image. ..
  • the estimated defect site image F (Y) is calculated by Equation 12 in FIG.
  • F (Y) (x, y, c) is the value of the pixel in which the coordinate value of the estimated defect site image F (Y) in the X direction is x, the coordinate value in the Y direction is y, and the channel is c.
  • F15 (Y) (x, y, c) is the same applies.
  • F11 (Y), F12 (Y), F21 (Y), F22 (Y), F23 (Y), F24 (Y), F31 (Y), and F32 (Y) are shown in FIG. 20 described in Example 1. It is calculated by Equations 1 to 8. Further, when calculating the estimation error of the defect portion in S1807, the specific defect portion image DD (1 channel binary image) is converted into the 2-channel image MSK by the formula 13 of FIG. 21 and estimated as the image MSK. The difference (error) from the defect site image F (Y) is calculated. As a method of quantifying the difference (error) between images, a mean square error (Mean Squared Error) or the like may be used as in S607.
  • a mean square error Mel Squared Error
  • W9 is two (c1 ⁇ 2) ⁇ f9 ⁇ f9 size filters
  • B9 is a two-dimensional vector.
  • the parameters calculated by the estimation processing parameter calculation processing are W1 to W6, W9, B1 to B6, and B9.
  • the estimation processing parameters for estimating the defect portion in the defect image are calculated using the learning defect image, and in the sample observation, the defect portion in the inspection defect image is calculated based on the estimation processing parameter.
  • the calculator is: (1) A plurality of images captured by the scanning electron microscope are acquired, and the images are acquired. (2) From the plurality of images, a learning defect image including the defect portion and a learning reference image not including the defect portion are acquired. (3) The estimation processing parameter is calculated by using the learning defect image and the learning reference image. (4) Obtain an inspection defect image including the defect part, (5) A pseudo reference image is estimated using the estimation processing parameter and the inspection defect image. Sample observation system.
  • ⁇ Viewpoint 2 The sample observation system according to the viewpoint 1.
  • the calculator is: (6) The pseudo reference image and the inspection defect image are compared to identify the defect portion of the inspection defect image. Sample observation system.
  • ⁇ Viewpoint 3 The sample observation system according to the viewpoint 1.
  • the computer is used as the process of (3): (3A) An alignment amount is obtained by aligning the learning defect image and the learning reference image based on a predetermined evaluation value. (3B) A learning partial defect image is cut out from the learning defect image based on the alignment amount. (3C) A learning partial reference image is cut out from the learning reference image based on the alignment amount. (3D) The estimation processing parameter is calculated using the learning partial defect image and the learning partial reference image. Sample observation system.
  • the sample observation system according to the viewpoint 1.
  • the estimation processing parameter is a neural network parameter, and is In the neural network, the minimum size of the image input to the input layer is the first size.
  • the computer is used as the process of (1): Each acquires the plurality of images of the first size or larger. Sample observation system.
  • the sample observation system according to the viewpoint 3.
  • the estimation processing parameter is a neural network parameter, and is In the neural network, the minimum size of the image input to the input layer is the first size.
  • the computer is used as the process of (3): (3E) Check that at least one of the size of the learning partial defect image and the learning partial reference image is equal to or larger than the first size. Sample observation system.
  • ⁇ Viewpoint 6 The sample observation system according to the viewpoint 1.
  • the computer acquires the end condition of the calculation process of the estimation process parameter, and obtains the end condition.
  • the computer is used as the process of (3): (3F) When it is detected that the end condition is satisfied, the update of the estimation processing parameter is completed. Sample observation system.
  • ⁇ Viewpoint 7 The sample observation system according to the viewpoint 6.
  • the computer In parallel with the calculation of the estimation processing parameters, the computer: (7) By comparing the learning defect image with the learning reference image, the defect portion of the learning defect image is specified. Sample observation system.
  • ⁇ Viewpoint 8 The sample observation system according to the first aspect.
  • the computer omits the acquisition of the reference image corresponding to the inspection defect image. Sample observation system.
  • the calculator is: A plurality of images captured by the scanning electron microscope are acquired, and a plurality of images are acquired. From the plurality of images, a learning defect image including the defect site is acquired, and the defect image is obtained. Using the learning defect image, the estimation processing parameters are calculated. Obtain an inspection defect image including the defect site, Using the estimation processing parameter and the inspection defect image, the defect portion in the inspection defect image is estimated. Sample observation system.
  • the processing shown above may be realized by an image processing program executed by the processing unit.
  • the image processing program may be distributed by a storage medium that can be read by a computer, or may be distributed by a distribution server computer.
  • the distribution server computer has a storage unit, a calculation unit, and a network interface 112. Specific examples of each part may be the same as in the case of the computer 103.
  • the image processing program is stored in the storage unit of the distribution server computer having such a configuration, and the processing unit reads the image processing program in response to the distribution request from the computer 103 and transmits the image processing program to the computer 103 via the network interface 112. You may.
  • Sample observation system 101 SEM 102: Defect inspection device 103: Computer

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Multimedia (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Quality & Reliability (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Electromagnetism (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Testing Or Measuring Of Semiconductors Or The Like (AREA)
  • Analysing Materials By The Use Of Radiation (AREA)
  • Length-Measuring Devices Using Wave Or Particle Radiation (AREA)
  • Image Analysis (AREA)
PCT/JP2021/000484 2020-03-06 2021-01-08 試料観察システム及び画像処理方法 Ceased WO2021176841A1 (ja)

Priority Applications (3)

Application Number Priority Date Filing Date Title
CN202180013541.6A CN115088061B (zh) 2020-03-06 2021-01-08 试样观察系统以及图像处理方法
KR1020227027317A KR102808816B1 (ko) 2020-03-06 2021-01-08 시료 관찰 시스템 및 화상 처리 방법
US17/802,161 US12333695B2 (en) 2020-03-06 2021-01-08 Sample observation system and image processing method

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2020-038745 2020-03-06
JP2020038745A JP7262409B2 (ja) 2020-03-06 2020-03-06 試料観察システム及び画像処理方法

Publications (1)

Publication Number Publication Date
WO2021176841A1 true WO2021176841A1 (ja) 2021-09-10

Family

ID=77613622

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2021/000484 Ceased WO2021176841A1 (ja) 2020-03-06 2021-01-08 試料観察システム及び画像処理方法

Country Status (5)

Country Link
US (1) US12333695B2 (https=)
JP (1) JP7262409B2 (https=)
KR (1) KR102808816B1 (https=)
TW (1) TWI793496B (https=)
WO (1) WO2021176841A1 (https=)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP4253943A4 (en) * 2020-11-30 2024-05-01 Konica Minolta, Inc. ANALYSIS DEVICE, INSPECTION SYSTEM AND LEARNING DEVICE
JP7847075B2 (ja) * 2022-01-27 2026-04-16 株式会社日立ハイテク 欠陥観察方法、装置、およびプログラム
JP2023177058A (ja) * 2022-06-01 2023-12-13 株式会社豊田中央研究所 顕微鏡像取得制御装置、顕微鏡像取得制御プログラム
JP2024092743A (ja) * 2022-12-26 2024-07-08 株式会社日立ハイテク マルチ荷電粒子ビーム装置と通信可能なプロセッサシステム、およびその方法
CN121112997B (zh) * 2025-11-17 2026-03-17 中国电建集团西北勘测设计研究院有限公司 海上光伏支架的应变监测方法、装置、程序产品与设备

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08136466A (ja) * 1994-11-10 1996-05-31 Dainippon Screen Mfg Co Ltd 画像パターン検査装置
JP2001085482A (ja) * 1999-09-14 2001-03-30 Toshiba Corp 半導体検査装置及び半導体検査方法
JP2004354251A (ja) * 2003-05-29 2004-12-16 Nidek Co Ltd 欠陥検査装置
JP2006269489A (ja) * 2005-03-22 2006-10-05 Hitachi High-Technologies Corp 欠陥観察装置及び欠陥観察装置を用いた欠陥観察方法
JP2007184565A (ja) * 2005-12-07 2007-07-19 Hitachi High-Technologies Corp 欠陥レビュー方法及びその装置
JP2010161247A (ja) * 2009-01-09 2010-07-22 Hitachi High-Technologies Corp Ｓｅｍを用いた欠陥観察方法及びその装置
JP2018137275A (ja) * 2017-02-20 2018-08-30 株式会社日立ハイテクノロジーズ 試料観察装置および試料観察方法
JP2018205163A (ja) * 2017-06-06 2018-12-27 株式会社デンソー 外観検査装置、変換データ生成装置、及びプログラム
US20190287230A1 (en) * 2018-03-19 2019-09-19 Kla-Tencor Corporation Semi-supervised anomaly detection in scanning electron microscope images

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3893825B2 (ja) 1999-12-28 2007-03-14 株式会社日立製作所 半導体ウェーハの欠陥観察方法及びその装置
JP2004349515A (ja) * 2003-05-23 2004-12-09 Hitachi High-Technologies Corp Ｓｅｍ式外観検査装置，レビュー装置、およびアライメント座標設定方法
US7333650B2 (en) 2003-05-29 2008-02-19 Nidek Co., Ltd. Defect inspection apparatus
US7764826B2 (en) * 2005-12-07 2010-07-27 Hitachi High-Technologies Corporation Method and apparatus of reviewing defects on a semiconductor device
JP5948262B2 (ja) * 2013-01-30 2016-07-06 株式会社日立ハイテクノロジーズ 欠陥観察方法および欠陥観察装置
WO2016092640A1 (ja) * 2014-12-10 2016-06-16 株式会社 日立ハイテクノロジーズ 欠陥観察装置および欠陥観察方法
JP2016148747A (ja) 2015-02-12 2016-08-18 株式会社ニューフレアテクノロジー 画像作成方法、検査方法および画像作成装置
JP6546826B2 (ja) * 2015-10-08 2019-07-17 株式会社日立パワーソリューションズ 欠陥検査方法、及びその装置
US10402688B2 (en) 2016-12-07 2019-09-03 Kla-Tencor Corporation Data augmentation for convolutional neural network-based defect inspection
WO2018167965A1 (ja) * 2017-03-17 2018-09-20 株式会社日立ハイテクノロジーズ 欠陥観察装置及び欠陥観察方法
EP3499459A1 (en) * 2017-12-18 2019-06-19 FEI Company Method, device and system for remote deep learning for microscopic image reconstruction and segmentation
JP7042118B2 (ja) * 2018-03-08 2022-03-25 株式会社東芝 検査装置、検査方法、及びプログラム
WO2019216303A1 (ja) 2018-05-09 2019-11-14 レーザーテック株式会社 検査装置、検査方法、学習方法、及びプログラム
CN110619618B (zh) * 2018-06-04 2023-04-07 杭州海康威视数字技术股份有限公司 一种表面缺陷检测方法、装置及电子设备
JP2020043266A (ja) * 2018-09-12 2020-03-19 株式会社日立ハイテクノロジーズ 半導体ウェハの欠陥観察システム及び欠陥観察方法
US11450012B2 (en) * 2019-10-31 2022-09-20 Kla Corporation BBP assisted defect detection flow for SEM images

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH08136466A (ja) * 1994-11-10 1996-05-31 Dainippon Screen Mfg Co Ltd 画像パターン検査装置
JP2001085482A (ja) * 1999-09-14 2001-03-30 Toshiba Corp 半導体検査装置及び半導体検査方法
JP2004354251A (ja) * 2003-05-29 2004-12-16 Nidek Co Ltd 欠陥検査装置
JP2006269489A (ja) * 2005-03-22 2006-10-05 Hitachi High-Technologies Corp 欠陥観察装置及び欠陥観察装置を用いた欠陥観察方法
JP2007184565A (ja) * 2005-12-07 2007-07-19 Hitachi High-Technologies Corp 欠陥レビュー方法及びその装置
JP2010161247A (ja) * 2009-01-09 2010-07-22 Hitachi High-Technologies Corp Ｓｅｍを用いた欠陥観察方法及びその装置
JP2018137275A (ja) * 2017-02-20 2018-08-30 株式会社日立ハイテクノロジーズ 試料観察装置および試料観察方法
JP2018205163A (ja) * 2017-06-06 2018-12-27 株式会社デンソー 外観検査装置、変換データ生成装置、及びプログラム
US20190287230A1 (en) * 2018-03-19 2019-09-19 Kla-Tencor Corporation Semi-supervised anomaly detection in scanning electron microscope images

Also Published As

Publication number Publication date
KR20220123708A (ko) 2022-09-08
TW202135119A (zh) 2021-09-16
US20230005123A1 (en) 2023-01-05
JP2021141231A (ja) 2021-09-16
US12333695B2 (en) 2025-06-17
CN115088061A (zh) 2022-09-20
JP7262409B2 (ja) 2023-04-21
KR102808816B1 (ko) 2025-05-19
TWI793496B (zh) 2023-02-21

Similar Documents

Publication Publication Date Title
KR102808816B1 (ko) 시료 관찰 시스템 및 화상 처리 방법
US11170483B2 (en) Sample observation device and sample observation method
JP4199939B2 (ja) 半導体検査システム
KR102350416B1 (ko) 결함 관찰 장치
JP5581286B2 (ja) 欠陥検査方法および欠陥検査装置
TWI515813B (zh) Charged particle - ray device
JP2020115131A (ja) 試験片の調査の方法およびそのシステム
JP2007218711A (ja) 電子顕微鏡装置を用いた計測対象パターンの計測方法
US20150146967A1 (en) Pattern evaluation device and pattern evaluation method
JP2017198589A (ja) パターン検査方法及びパターン検査装置
JPWO2011001635A1 (ja) 半導体検査装置及びそれを用いた半導体検査方法
US12154264B2 (en) Defect inspecting system and defect inspecting method
JP5415523B2 (ja) パターン検査装置及びその検査方法
JP2020113769A (ja) 画像推定方法およびシステム
TWI874699B (zh) 檢驗半導體樣品的系統及方法和包含指令之非暫時性電腦可讀媒體
TW202500985A (zh) 用於半導體試樣的量測偏差分析
CN115088061B (zh) 试样观察系统以及图像处理方法
JP4262269B2 (ja) パターンマッチング方法及び装置
JP6207893B2 (ja) 試料観察装置用のテンプレート作成装置
JP4262288B2 (ja) パターンマッチング方法及び装置
TW202011121A (zh) 用於較佳設計對準之目標選擇改良
WO2018167965A1 (ja) 欠陥観察装置及び欠陥観察方法
US20260026296A1 (en) Automatic creation of an imaging recipe
TW202544739A (zh) 用於提供一或多個半導體取樣的資訊的圖像之間的配準的系統和方法
JP2014130745A (ja) 荷電粒子線装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21765022

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 20227027317

Country of ref document: KR

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21765022

Country of ref document: EP

Kind code of ref document: A1

WWG Wipo information: grant in national office

Ref document number: 17802161

Country of ref document: US