WO2012144183A1 - 欠陥分類方法及び欠陥分類システム - Google Patents
欠陥分類方法及び欠陥分類システム Download PDFInfo
- Publication number
- WO2012144183A1 WO2012144183A1 PCT/JP2012/002609 JP2012002609W WO2012144183A1 WO 2012144183 A1 WO2012144183 A1 WO 2012144183A1 JP 2012002609 W JP2012002609 W JP 2012002609W WO 2012144183 A1 WO2012144183 A1 WO 2012144183A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- classification
- defect
- recipe
- class
- Prior art date
Links
- 230000007547 defect Effects 0.000 title claims abstract description 392
- 238000000034 method Methods 0.000 title claims description 327
- 238000006243 chemical reaction Methods 0.000 claims abstract description 39
- 230000008569 process Effects 0.000 claims description 249
- 238000003384 imaging method Methods 0.000 claims description 38
- 238000004519 manufacturing process Methods 0.000 claims description 12
- 238000004891 communication Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 abstract description 137
- 238000003860 storage Methods 0.000 abstract description 41
- 235000012431 wafers Nutrition 0.000 description 32
- 239000000523 sample Substances 0.000 description 26
- 230000006870 function Effects 0.000 description 19
- 238000010586 diagram Methods 0.000 description 10
- 238000012552 review Methods 0.000 description 10
- 239000004065 semiconductor Substances 0.000 description 9
- 238000010894 electron beam technology Methods 0.000 description 8
- 230000002950 deficient Effects 0.000 description 6
- 238000001514 detection method Methods 0.000 description 6
- 230000006872 improvement Effects 0.000 description 6
- 238000007689 inspection Methods 0.000 description 6
- 230000008859 change Effects 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 238000012986 modification Methods 0.000 description 5
- 230000001133 acceleration Effects 0.000 description 4
- 238000001878 scanning electron micrograph Methods 0.000 description 4
- 230000010365 information processing Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 125000003821 2-(trimethylsilyl)ethoxymethyl group Chemical group [H]C([H])([H])[Si](C([H])([H])[H])(C([H])([H])[H])C([H])([H])C(OC([H])([H])[*])([H])[H] 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000011946 reduction process Methods 0.000 description 2
- 238000004626 scanning electron microscopy Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000005672 electromagnetic field Effects 0.000 description 1
- 230000001678 irradiating effect Effects 0.000 description 1
- 238000003825 pressing Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
- G01N23/22—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by measuring secondary emission from the material
- G01N23/225—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by measuring secondary emission from the material using electron or ion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
- G01N21/9501—Semiconductor wafers
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2223/00—Investigating materials by wave or particle radiation
- G01N2223/40—Imaging
- G01N2223/421—Imaging digitised image, analysed in real time (recognition algorithms)
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2223/00—Investigating materials by wave or particle radiation
- G01N2223/60—Specific applications or type of materials
- G01N2223/611—Specific applications or type of materials patterned objects; electronic devices
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2223/00—Investigating materials by wave or particle radiation
- G01N2223/60—Specific applications or type of materials
- G01N2223/646—Specific applications or type of materials flaws, defects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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; CALCULATING OR 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
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L22/00—Testing or measuring during manufacture or treatment; Reliability measurements, i.e. testing of parts without further processing to modify the parts as such; Structural arrangements therefor
- H01L22/10—Measuring as part of the manufacturing process
- H01L22/12—Measuring as part of the manufacturing process for structural parameters, e.g. thickness, line width, refractive index, temperature, warp, bond strength, defects, optical inspection, electrical measurement of structural dimensions, metallurgic measurement of diffusions
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L22/00—Testing or measuring during manufacture or treatment; Reliability measurements, i.e. testing of parts without further processing to modify the parts as such; Structural arrangements therefor
- H01L22/20—Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L2924/00—Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
- H01L2924/0001—Technical content checked by a classifier
- H01L2924/0002—Not covered by any one of groups H01L24/00, H01L24/00 and H01L2224/00
Definitions
- the present invention relates to a defect classification method and a defect classification system for classifying defects on a semiconductor wafer.
- defect analysis is performed using a defect inspection apparatus and a defect observation apparatus.
- the defect inspection apparatus is an apparatus that observes a wafer using optical means or an electron beam and outputs the position coordinates of the detected defect. Since it is important for defect inspection equipment to process a wide area at high speed, the amount of image data is reduced by increasing the pixel size of the acquired image as much as possible (that is, lowering the resolution). Although the existence of a defect can be confirmed from the detected low-resolution image, it is difficult to determine the type of the defect (defect type) in detail.
- the defect observation apparatus is an apparatus that uses the output information of the defect inspection apparatus, images defect coordinates of a wafer with high resolution, and outputs an image.
- the semiconductor manufacturing process has been miniaturized, and accordingly the defect size has reached the order of several tens of nanometers. In order to observe the defects in detail, a resolution of the order of several nanometers is required.
- a defect observation apparatus using a scanning electron microscope (SEM: Scanning Electron Microscope) has been widely used.
- the review SEM has an ADR (Automatic Defect Review) function that automatically collects high-resolution images (defect images) of defects on the wafer using the defect coordinates output by the defect inspection apparatus.
- the review SEM has an ADC (Automatic Defect Classification) function that automatically identifies and classifies defect types from defect images.
- Patent Document 1 describes a method of performing image processing on a defect image to quantify an appearance feature amount of a defective part and classifying the defect image using a neural network. Further, as a method that can be easily dealt with even when there are many types of defects (defect types) to be classified, for example, Patent Document 2 describes a method of classifying by combining a rule-based classification method and a teaching classification method. Yes.
- classification is performed based on the classification recipe.
- the classification recipe includes various parameters such as image processing parameters, information on defect types to be classified (classification class), defect images (teaching images) belonging to each classification class, and the like.
- the classification recipe is updated by adding a classification class for a new defect when a new defect type occurs due to a change in the manufacturing process.
- Patent Document 3 describes a method of determining a new defect as a defect whose classification class is unknown (unknown defect) and automatically adding a new classification class to the classification recipe when automatically classifying defect images.
- the unknown defect includes a defect outside the classification class defined in the classification recipe, which occurs due to a mistake in teaching by the user.
- defect images are classified manually by a person in front of a defect observation device, and the defect observation device generally has an automatic defect image classification function as part of its function. It is.
- a plurality of defect observation apparatuses have been introduced in a semiconductor wafer production line, and an increase in costs for managing classification recipes has become a problem.
- a method of connecting a plurality of image detection devices and an information processing device over a network, transferring a captured image to the information processing device, and determining whether there is an abnormality appearing in the image by the information processing device Is described in Patent Document 4.
- JP-A-8-21803 Japanese Patent Laid-Open No. 2007-225531 JP 2000-57349 A JP 2004-226328 A
- the ADC function (automatic defect classification function) described above calculates various features such as the size and shape of the defective part as feature amounts from the captured SEM image, and the defects are defined in advance from the calculated feature amounts. This is a function for classifying into a plurality of defect classes.
- review SEMs have been put on the market by several manufacturers, and each company provides this ADC function by installing it in a defect classification system (defect classification device) that is sold together with its own review SEM.
- this defect classification system not only the above-described automatic classification function of defect images, but also a display function for presenting the classification result to the user, a function for accepting input from the user and correcting the result of automatic classification, Alternatively, it also has a function of transferring the classification result to a yield management database server or the like installed in the production line.
- the manufacture of a semiconductor wafer consists of a plurality of processes (hereinafter referred to as “processes”), and the types of defects that occur due to differences in the processes differ, so that a classification recipe suitable for each process is generally created.
- processes processes
- parameters appropriate for classification are often different because of different performance and characteristics.
- machine difference performance difference
- FIG. 1 shows a conventional system configuration example, in which the image capturing apparatuses 101 and 101 ′ are connected to the corresponding classification modules 103 and 103 ′, and the classification modules 103 and 103 ′ correspond to the corresponding classification recipes 104 and 104, respectively. 'Is connected.
- the classification modules 103 and 103 ' are connected to the yield management database server 105 through communication means 106 such as a network.
- the classification procedure is as follows. First, with respect to images obtained from a plurality of image capturing apparatuses (two image capturing apparatuses 1 and 2 in FIG.
- the classification modules 103 and 103 ′ corresponding to the respective apparatuses are used.
- the defect classification is performed respectively.
- the classification modules 103 and 103 ′ perform classification based on the respective classification recipes 104 and 104 ′, and the classification result is transmitted to the yield management database server 105 via the communication unit 106 and stored.
- the image pickup apparatuses 101 and 101 ' correspond to the defect observation apparatus described above, and the classification modules 103 and 103' indicate apparatuses capable of executing ADC.
- the classification classes of the classification recipe 1 and the classification recipe 2 need to be the same, and images in which the same type of defects are imaged (hereinafter referred to as the same type of defect images) must be registered in each classification class.
- the classification classes of a plurality of classification recipes are the same and the defect images registered in all the classification classes are the same kind of defect images in the same classification class, the classification definitions are the same.
- the same classification class means that the defect type to be classified in the classification class is the same, and if the defect type to be classified is the same, even if the name of the classification class is different, it is the same classification class. Call.
- the classification recipe itself exists separately for each image capturing apparatus. Therefore, when the classification recipe is set individually, the classification definition may not be kept the same for each classification recipe 104. There is a problem that there is.
- FIG. 1 as described in the description of Patent Document 3, a case where a new defect occurs and the classification recipe 104 in the classification module 103 corresponding to one image capturing apparatus 101 is updated.
- the classification recipe 104 ′ in the classification module 103 ′ corresponding to the other image capturing apparatus 101 ′ is independent from the classification recipe 104 and is not updated.
- the classification recipe is set individually as described above. As in the case, there may be a difference in the classification definition for each classification recipe.
- Patent Documents 3 and 4 As described above, in Patent Documents 3 and 4, the above-mentioned problem that occurs when a plurality of manufacturers and different types of defect observation apparatuses are operated in the same process is not recognized, and the classification definition is the same to solve this. There was no description of how to keep it.
- the present invention solves the above problems and improves the reliability of statistical process management by maintaining the same classification definition even when a plurality of different defect observation apparatuses are operated in the same process.
- a classification system and a defect classification method are provided.
- An apparatus for imaging a sample and a defect classification method for classifying a defect image using a classification recipe corresponding to the process of manufacturing the sample, the classification class defined in the classification recipe of the first image imaging apparatus The same classification class as defined in the classification recipe of the second image pickup device corresponding to the same process as the classification recipe of the first image pickup device, and the second image pickup device A step of identifying a defect image of the same type as a teaching image registered in a classification class defined in a classification recipe of the first imaging device from among the defect images; and Registering in the same classification class as the classification class of the first image capturing apparatus in which the teaching image is registered, among the classification classes defined in the classification recipe of the image capturing apparatus of A defect classification method comprising and.
- a defect classification system connected to a plurality of image pickup devices via communication means, wherein the defect classification images picked up by the plurality of image pickup devices are classified, and information for classification is stored Classification recipe management means for managing the classified recipe, and the classification recipe management means is registered in the classification class of the classification recipe in the first image imaging device which is one of the plurality of image imaging devices.
- a corresponding defect identifying unit that identifies a defect image of the same type as the teaching image from a defect image captured by a second image capturing apparatus that is one of the plurality of image capturing apparatuses installed in the same process;
- An image conversion unit that converts the teaching image so as to be similar to the defect image obtained from the second image capturing device by image conversion.
- a defect classification method and defect classification that solve the above-described problems and improve the reliability of statistical process management by maintaining the same classification definition in a plurality of classification recipes corresponding to the same process.
- a system can be provided.
- FIG. 1 is a diagram illustrating a configuration example of an image capturing apparatus according to a first embodiment. It is a figure which shows the processing flow of the classification process of the defect classification system of Example 1.
- FIG. It is a figure which shows the processing flow of the classification recipe creation process of the defect classification system of Example 1.
- FIG. It is a figure which shows the processing flow of the classification class setting process of the defect classification system of Example 1.
- FIG. It is a figure which shows an example of GUI which specifies the same kind of defect image of the defect classification system of Example 1.
- FIG. 3 is a diagram illustrating an example of an acquired image obtained by the image capturing apparatus according to the first embodiment.
- FIG. 3 is a diagram illustrating an example of a defect cross section and the arrangement of detectors in the image capturing apparatus according to the first embodiment.
- FIG. 5 is a diagram for explaining the arrangement of detectors and the detection shadow direction in the image pickup apparatus of Embodiment 1.
- FIG. 3 is a diagram illustrating an example of a shadow detection image in the image capturing apparatus according to the first embodiment. It is a figure which shows the processing flow of the process which specifies the defect image of the same kind of the defect classification system of Example 1.
- FIG. It is a figure which shows the processing flow of the classification class setting process of the defect classification system of Example 2.
- FIG. It is a figure which shows the structural example of the defect classification system of Example 2.
- FIG. It is a figure which shows the processing flow of the classification process of the defect classification system of Example 2.
- FIG. It is a figure which shows the processing flow of the classification recipe update process of the defect classification system of Example 2.
- FIG. It is a figure which shows the processing flow of the classification class update process of the defect classification system of Example 2.
- FIG. It is a figure which shows an example of GUI which selects the classification
- FIG. It is a figure which shows the processing flow of the classification class setting process of the defect classification system of Example 3.
- FIG. It is a figure which shows the processing flow of the classification recipe creation process of the defect classification system of Example 3.
- an input of the defect classification system according to the present invention may be other than an SEM image, and may be an optical type.
- a defect image captured using a means or an ion microscope may be used.
- the defect classification system 201 is configured to be connected to N (N ⁇ 2) image capturing apparatuses 200-1 to 200-n via a communication unit 204 such as a network.
- the image capturing device 200 (200-1 to 200-n) is a device that acquires an image of a corresponding part, and a detailed configuration thereof will be described later.
- the defect classification system 201 receives input of defect images obtained by a plurality of image capturing devices, classifies them, and displays a classification result as a keyboard for receiving display of data to the operator and input from the operator. It has a function of outputting to an input / output unit 217 configured using a mouse display device or the like. Details of the first embodiment of the defect classification system 201 will be described below.
- the defect classification system 201 creates and updates a classification recipe, stores the classification recipe, the defect image, and information associated with the defect image, and classifies the defect image input from each image capturing apparatus.
- the recipe management unit 202 includes a processing unit 207 that executes processing related to the classification recipe, and a storage unit 208 that stores information.
- the storage unit 208 stores a defect image captured by the image capturing apparatus 200.
- the image storage unit 213, the image capturing device 200, the classification recipe storage unit 214 that stores the classification recipe created for each process, and the accompanying information such as the process obtained from the image capturing device together with the defect image are stored for each defect image.
- the auxiliary information storage unit 215 is used as appropriate.
- the processing unit 207 includes a corresponding defect specifying unit 209 that performs processing for specifying a defect image of the same type for the defect image obtained from each image capturing apparatus 200, and a device and a process that are captured for each classification recipe and defect image.
- An information specifying unit 210 that specifies information, a recipe updating unit 211 that creates a classification recipe, updates a classification class, and the like, and an image conversion unit 212 that converts an image by image processing are used as appropriate.
- the information specifying unit 210 specifies a classification recipe for the same process based on the process information for each defect image stored in the classification recipe specifying unit 214 and the accompanying information storage unit 215, or stores it in the image storage unit 213.
- the process information of the defect image being taken and the information of the imaged device are specified.
- the processing procedure and method of the processing unit 207 will be described later.
- the classification module 103 includes a classification processing unit 216 that classifies defect images based on a classification recipe. Details of the processing of the classification processing unit 216 will be described later.
- the example of the defect classification system 201 shown in FIG. 2 may be operated in one arithmetic device (hereinafter referred to as a PC) or may be operated separately in a plurality of PCs. When operating with a plurality of PCs, for example, a method of operating the recipe management unit 202 as a recipe server with a single PC is also conceivable.
- the classification module 203 is shown, but a configuration using a plurality of classification modules may be used.
- FIG. 3 is a diagram illustrating a detailed configuration example of the image capturing apparatus 200 described above.
- the image capturing apparatus 200 includes an SEM column 301, an SEM control unit 308, an input / output I / F 309, a storage unit 311, and an accompanying information creation unit 314, which are appropriately connected via a communication unit 315. Are connected to each other.
- An input / output unit 310 is connected to the input / output I / F 309 to input / output data to / from an operator.
- the SEM column 301 includes an electron source 302, a stage 306 on which the sample wafer 307 is placed, and secondary electrons generated from the sample wafer 307 and backscattering as a result of irradiating the sample wafer 307 with the primary electron beam from the electron source 302.
- a plurality of detectors 303, 304, and 305 for detecting electrons are appropriately used.
- the SEM column 301 also includes a deflector for scanning the observation region of the sample wafer 307 with a primary electron beam and an image for generating a digital image by digitally converting the intensity of detected electrons.
- a generation unit and the like are also included as appropriate.
- the storage unit 311 is used for processing that reduces the influence of shot noise by acquiring a plurality of images at the same location and creating an average image thereof, which are SEM imaging conditions, such as acceleration voltage, probe current, and frame addition number
- SEM imaging conditions such as acceleration voltage, probe current, and frame addition number
- the imaging recipe storage unit 312 for storing the number of images), the visual field size, and the like, and the image memory 313 for storing the acquired image data are appropriately included.
- the accompanying information generation unit 314 includes information accompanying each image data, for example, imaging conditions such as acceleration voltage, probe current, and frame addition number at the time of imaging, ID information for specifying the imaging device, and detection used for generating the image. It has a function of creating accompanying information such as the types and properties of the devices 303 to 305, the wafer ID and process, and the date and time when the image was captured.
- the wafer ID and process information may be input by the user from the input / output unit 310 or the like, or may be read from the surface of the wafer or from a box (not shown) in which the wafer is stored. .
- the generated accompanying information is transferred together with the image data when the image data is transferred via the input / output I / F 309.
- the SEM control unit 308 is a part that controls all processes performed by the image capturing apparatus 200 such as image acquisition.
- the stage 306 is moved to bring a predetermined observation site on the sample wafer 307 into the imaging field of view, the primary electron beam is irradiated onto the sample wafer 307, and electrons generated from the sample wafer 307 are emitted.
- Detection by the detectors 303 to 305, imaging of the detected electrons, storage in the image memory 313, creation of accompanying information for the captured image by the accompanying information creation unit 314, and the like are performed.
- Various instructions from an operator, designation of imaging conditions, and the like are executed through an input / output unit 310 including a keyboard, a mouse, a display, and the like.
- the configuration of the image pickup apparatus 200 shown in FIG. 3 is an example, and when the manufacturer and model number are different, the configurations and the numbers of the detectors 303 to 305 may be different.
- manufacturers and model numbers are different, even when the same defect is imaged, a difference occurs such as a difference in the captured image itself or a difference in the image quality of the captured image due to a difference in detector configuration or characteristics. Due to these differences, it is difficult to compare defect images obtained from different apparatuses in the state where the images are taken.
- the problem in order to make the classification definition of the classification recipe in each image capturing device the same, it is necessary to compare defect images captured by different image capturing devices and specify the same type of defect image. .
- a processing flow for classifying an input defect image in the defect classification system according to the present invention will be described with reference to FIG.
- the process of classifying the defect image is executed by the classification processing unit 216 of the classification module 203.
- a defect image to be classified is read from the image storage unit 213 (S401).
- the accompanying information of the defect image is read from the accompanying information storage unit 215 (S402).
- the accompanying information is a condition at the time of image capturing, and includes at least an ID for identifying the image capturing device that captured the defect image and an ID for identifying the process of the captured wafer.
- acceleration voltage and probe current at the time of imaging, imaging field size, imaging date and time, imaging coordinates, and the like may be stored and used as information at the time of classification.
- the information specifying unit 210 specifies the image pickup device that picked up the defect image and the process of the picked-up wafer (S403).
- the image capturing apparatus ID and the process ID included in the accompanying information of the defect image read out in the process S402 may be used.
- the image storage unit 213 has a hierarchical structure (directory structure), and the defect image transmitted from the image capturing device is specified by being stored in a hierarchy (directory) for each image capturing device and the captured process.
- the classification recipe corresponding to the imaging device and the process that captured the defect image to be classified is read from the classification recipe storage unit 214 among the classification recipes for each of the imaging device and the process (S404). A method for generating the classification recipe will be described later with reference to FIG.
- the classification recipe here refers to classification parameters including classification class information in classification processing, teaching images belonging to each classification class, classification identification plane information for classification into each classification class, and defect images. It includes parameters for processing for extracting regions and processing for calculating feature values.
- a defect area is extracted from the read defect image (S405).
- a value (feature amount) obtained by quantifying the feature relating to the defect is calculated for the extracted defect area (S406).
- the image is classified using the calculated feature amount and the classification identification surface included in the classification recipe (S407).
- a defect classification method a neural network, SVM (Support Machine Machine), or the like may be used, or a rule type classifier and a teaching type classifier may be used in combination as described in Patent Document 2. good.
- the processing flow when classifying the input defect image of one defect has been described above. However, in order to classify a plurality of defect images, the processing of S401 to S407 may be repeatedly executed for the number of defect images.
- the classification recipe is information that defines the defect image classification method, and includes information such as defect classification class (defect type), image processing parameters, and classification identification surface for classification into each classification class. Contains parameters. As described above, it is necessary to create a classification recipe for each image capturing device and each process. As a premise for creating a classification recipe, a defect image captured by a combination of a process and a device for which a classification recipe is to be created is stored in the image storage unit 213. Must be preserved.
- FIG. 5 (a) shows a conventional classification recipe creation (single classification recipe creation) method for creating a classification recipe for each apparatus and process
- FIG. 5 (b) is a classification recipe creation method according to the present invention.
- a classification class is set by registering a classification class definition and a teaching image (S501).
- classification classes are defined, and teaching images are registered in the respective classification classes.
- image processing parameters are set in image processing for recognizing a defective area or a wiring pattern in the defect image (S502), the classification parameters are adjusted (S503), and the classification recipe created in this way is set. It is stored in the classification recipe storage unit (S504).
- image processing parameters are set so that an appropriate image processing result is obtained for the teaching image registered in the processing S501.
- the adjustment of the classification parameter in the processing S503 may be performed by a method such as teaching the teaching image registered in the processing S501 to the classification processing unit in the classification module and creating a classification identification surface. If an image exists, it can be performed automatically.
- FIG. 5B The process of FIG. 5B is processed by the recipe update unit 211.
- the method for creating a classification recipe according to the present invention (FIG. 5B) it is possible to create a classification recipe for a plurality of image pickup devices in the same process at once, and in addition, classification definitions for all classification recipes to be created Can be the same.
- a common classification class is set in a classification recipe of a plurality of image pickup devices (S511).
- S511 Details of the processing S511 in the present invention will be described later with reference to FIGS.
- step S512 the following steps S513 to S515 are executed for the classification recipe corresponding to the N imaging devices (N ⁇ 2) in the same process for which the common classification class setting is performed in step S511.
- An image imaging apparatus in the same process corresponding to the classification recipe for which a common classification class setting has been performed in step S511 is assumed to be an apparatus i (1 ⁇ i ⁇ N).
- the image processing parameters of the classification recipe of the device i are set (S513), the classification parameters of the classification recipe of the device i are adjusted (S514), and the created classification recipe of the device i is stored in the classification recipe storage unit 214. (S515).
- Processes S513 to S515 are classification recipes for the device i to be processed, but may be executed by the same method as the processes S502 to S504 described above.
- step S502 and step S513 may be read from a pre-defined table or the like, or may be manually defined by the user.
- the user when reading image processing parameters from a predefined table, the user can create N classification recipes in the same process only by executing step S511.
- the value may be converted by using a conversion table or the like.
- the image processing parameter conversion table refers to a value corresponding to the image conversion processing parameter or a conversion value for each combination of the device i to be converted from the image processing parameter of the image processing parameter conversion device (here, device 1). This is a table in which the calculation formula is defined.
- the user can create classification recipes for N apparatuses in the same process only by performing the process S511 and the process S513 on the apparatus 1.
- FIG. 6 is an example of a processing flow at the time of setting a common classification class in the classification recipe creation in the defect classification system according to the present invention, and is a diagram showing details of the processing S511 in FIG.
- a common classification class is defined in a plurality of classification recipes corresponding to a plurality of imaging devices in the same process, and the same type of defect image is registered as a teaching image in each classification class.
- the classification definition can be made the same for all classification recipes.
- the image capturing apparatus is described as an apparatus, and the captured wafer process is appropriately abbreviated as a process.
- a classification recipe for classifying the defect image captured for the process A of the apparatus 1 is classified into a classification recipe for the apparatus 1 and the process A, and a defect image captured from the process A of the apparatus 1 is a defect of the apparatus 1 and the process A. It is abbreviated as an image (or an image of the apparatus 1 and the process A), and the same abbreviation is used even if the apparatus and the process such as the apparatus 2 and the process B are different.
- FIG. 6 illustrates a case where the classification recipe of process A is created when there are two devices (device 1 and device 2).
- a defect image captured by a combination of a process (process A) and a device (device 1, device 2) for which a classification class is to be set is stored in the image storage unit 213.
- the device imaged by the information identification unit 210 is identified (S601).
- the defect image of the process A as a method for identifying the imaged device, it may be determined from the accompanying information for each defect image stored in the accompanying information storage unit 215 or the like, or specified by the user from the input / output unit 217 You may do it.
- a classification recipe is created in each device (device 1, device 2) in step A, and a common classification class is defined for each device (S602).
- the classification recipe of the classification recipe of the apparatus 1 / process A and the classification recipe of the apparatus 2 / process A can be made the same definition.
- a part or all of the defect images of the apparatus 1 / process A stored in the image storage unit 213 are registered as teaching images in each classification class of the classification recipe of the apparatus 1 / process A (S603).
- the registration of the classification class and the teaching image in the processing S602 and the processing S603 may be specified by the user at the input / output unit 217, or the classification class defined in the file or the like and information on the defect image to be registered are read.
- step S606 the defect image of the apparatus 2 / process A stored in the image storage unit 213 is registered for each classification class of the classification recipe of the apparatus 2 / process A.
- the corresponding defect identification unit 209 is the same type as the defect image registered as the teaching image in the classification class of the classification recipe of the apparatus 1 / process A among the defect images of the apparatus 2 / process A in the process S603.
- the image (defect image of the same type) in which the defect is captured is specified (S604).
- the defect image registered as the teaching image in each classification class of the apparatus 1 and process A is converted into an image captured from the apparatus 2, and the feature amounts of these images are compared.
- a defect image of the same type as the teaching image is specified from the defect images of the apparatus 2 and the process A.
- the defect image of the apparatus 2 / process A identified in the process S604 is registered in the corresponding classification class of the classification recipe of the apparatus 2 / process A (S605).
- the device 1 and the device 2 has been described as an example. However, even when there are three or more devices, this processing flow can be applied by executing the process S606 by the number of devices. is there.
- FIG. 7 is an example of a GUI for executing the processing S603 to S605 at the time of setting the classification class described in FIG. 6 in the defect classification system according to the present invention.
- FIG. 7 as in the description of FIG. 6, a case where the classification recipe of the process A is created with two devices (device 1 and device 2) will be described as an example.
- reference numeral 701 indicates the name of the device where the defect image of the process A is captured and information on the process
- reference numeral 702 displays a plurality of defect images captured by each device in a display area corresponding to each device. is doing.
- Reference numeral 703 denotes a combo box for selecting a defect image to be displayed. For example, a secondary electron image or a backscattered electron image captured by the detectors 303 to 305 can be selected.
- Reference numeral 704 denotes a defect image selected by the user as a defect of the same type for the defects in the apparatus 1 and process A, and these defect images can be determined by highlighting a frame or an image on the image.
- the input / output unit 217 may use a mouse, a keyboard, a pen tablet, or the like, or information such as a defect ID for specifying an image is described in a file, and these are read and selected. It is also good to do.
- Reference numeral 705 denotes a button for registering the image selected in the apparatus 1 / step A in the classification class defined in step S602 (S603).
- a method of registering an image in a classification class is not limited to a mode in which a button on the screen is pressed, and a method of registering by a drag and drop operation with a mouse after selection is also conceivable.
- a button 706 is used to specify a defect image of the same type as the defect image registered as the teaching image in each classification class in the apparatus 1 / process A from the image of the apparatus 2 / process A (S604).
- Reference numeral 707 denotes a mark indicating the same type of defect image identified in step S604.
- the image may be distinguished from other images by surrounding the image with a frame or highlighting the image. It is also possible to display the classification class name or symbol in the image.
- the defect image of the same type may be specified for each classification class, or may be performed for a plurality of classification classes.
- the classification class of the apparatus 2 / process A is set. This is a button for correcting / changing a defect image to be registered.
- Reference numeral 709 denotes a button for registering the defect image of the apparatus 2 / process A identified in step S604 in the corresponding classification class of the classification recipe of the apparatus 2 / process A (S605).
- process S604 in order to compare defect images obtained from devices of different manufacturers and types, and to identify the same type of defect image, the difference in the captured image itself due to the difference in the configuration and characteristics of the detector, Consider the difference in the image quality of the captured image. This will be described below with reference to FIGS.
- the image capturing apparatus 200 shown in FIG. 3 includes three detectors, and the image capturing apparatus 200 can simultaneously acquire three images of the observation location on the sample wafer.
- FIG. 8 is an example of three captured images acquired for the foreign matter on the surface of the sample wafer.
- FIG. 8A is an image obtained by detecting secondary electrons generated from the sample wafer by the detector 303
- FIGS. 8B and 8C show backscattered electrons generated from the sample wafer.
- the images are acquired by the two detectors 304 and 305, respectively.
- 8A is referred to as an upper image
- FIGS. 8B and 8C are referred to as a left image and a right image, respectively.
- the upper image in FIG. 8A is an image in which the circuit pattern and the outline of the defective part are clearly observed.
- the left and right images in FIGS. 8 (b) and 8 (c) are images in which the shadows generated due to the uneven state of the surface can be observed.
- Such a difference in image properties is caused by the arrangement of the detector, the energy band of the detected electrons of the detector, the electromagnetic field applied to the column that affects the trajectory of the generated electrons from the sample, and the like.
- the image quality also changes depending on the imaging conditions, for example, the electron acceleration voltage, the probe current amount, the frame addition number, and the like.
- FIG. 9 shows the positional relationship between the cross section of the sample wafer and the backscattered electron detectors 304 and 305 in the case where the protruding defect 901 and the recessed defect 902 exist on the sample wafer 307, respectively.
- FIG. 9 shows the positional relationship between the cross section of the sample wafer and the backscattered electron detectors 304 and 305 in the case where the protruding defect 901 and the recessed defect 902 exist on the sample wafer 307, respectively.
- (b) schematically.
- two detectors of backscattered electrons are arranged at opposing positions diagonally above the sample wafer 307. The primary electron beam is incident from directly above.
- Backscattered electrons generated from the observation site have a characteristic that their energy is strong and directional, so most of the backscattered electrons generated in the direction of one detector reach the detector on the opposite side. do not do. As a result, as shown in FIGS. 8B and 8C, an image capable of observing the shadow according to the uneven state of the observation site can be acquired.
- FIG. 10 is a diagram schematically showing the direction of the detector and the direction of the shadow of the acquired image.
- FIG. 10A shows an example in which detectors are arranged along the X direction of the coordinate system. Images (a-1) and (a-2) schematically show images obtained by the detectors 304 and 305, respectively.
- the positions of the bright and dark areas on the images (a-1) and (a-2) obtained by the detectors 304 and 305 are shaded in the X direction as shown in the figure. .
- the bright area is an area having high brightness on the image.
- the bright region means that many backscattered electrons generated at the site are detected by the detector, while the dark region is a region where the backscattered electrons generated at the site are not detected by the detector. It is. Light and darkness appear in this way because backscattered electrons have directionality, so depending on the generation direction of backscattered electrons in each part and the position and direction of the detector that detects backscattered electrons, This is because the brightness is determined.
- FIG. 10 (b) shows a case where the direction of the detector is rotated 45 degrees clockwise with respect to FIG. 10 (a). The direction of the shadow of the images (b-1) and (b-2) obtained by the detector having the arrangement shown in FIG. 10 (b) rotates corresponding to the rotation of the detector. Similarly, FIG.
- FIG. 10C shows a case where the detector is arranged at a position rotated 45 degrees counterclockwise with respect to FIG. Similarly, the shadow directions of the images (c-1) and (c-2) obtained by the detectors having the arrangement shown in FIG. 10 (c) are rotated corresponding to the rotation of the detectors. Thus, if the direction of the detector changes, the direction of the shadow changes.
- FIGS. 11 (a) and 11 (b) when images are obtained by the detectors 304 and 305, it is determined whether the observation object is convex or concave. It cannot be determined that there is no information about the configuration.
- FIG. 11 (a) is an image obtained by acquiring a convex defect with the configuration of the detector of FIG. 10 (b)
- FIG. 11 (b) is an image showing a concave defect in FIG. 10 (c).
- a plurality of image capturing devices 200 are connected, but the types of the image capturing devices may be different.
- the manufacturers of the devices are different, or even a plurality of products with different detector configurations are provided even by the same manufacturer.
- the case where the number of detectors of the image pickup apparatus is three and the detector for detecting backscattered electrons faces the case has been described as an example in which the relative position with respect to the sample changes.
- the other conditions such as the number of detectors, the direction of each detector, the detected electronic energy band, and the like may be different for each apparatus.
- the energy of the generated sample can change even under imaging conditions, the obtained image may also change under these conditions.
- images obtained from devices of different manufacturers and models can be compared as they are due to differences in captured images themselves and differences in image quality of captured images due to differences in detector configuration and characteristics. It is difficult to identify the same type of defect image. Therefore, in the present invention, the image to be compared is subjected to image conversion by the image conversion unit 212 to eliminate the difference in the picked-up image itself and the difference in the image quality of the picked-up image due to the difference in the configuration and characteristics of the detector. Convert to a possible image.
- the image conversion process performed by the image conversion unit 212 will be described.
- the image conversion process means a series of processes that take an image set as an input, read corresponding incidental information from the incidental information storage unit 215, and output an image set obtained by processing them. Specifically, image quality improvement processing, shadow direction conversion processing, image mixing processing, and the like are included.
- an image quality improvement process for example, there is a noise reduction process.
- the SEM when the probe current at the time of image capturing is low or the frame addend is small, an image with a reduced S / N is easily obtained. Further, even under the same imaging conditions, when the imaging devices are different, images with different S / Ns may be obtained due to different electron detection yields at the detectors. Even in the same type of device, if the degree of adjustment is different, there may be a difference in S / N due to machine differences between devices.
- Specific examples of the noise reduction process include various noise filter processes.
- noise filter processing is executed on the image of the apparatus 1. Samples of images taken by the apparatus 2 are prepared, and the variance of the luminance values in the flat portion of the image after the noise filter processing of the apparatus 1 and the image of the apparatus 2 are compared. The above processing is repeated until the variance difference reaches a value exceeding a predetermined threshold value.
- the above processing is an example, but by these processing, an image similar to the image of the device 2 can be created from the image of the device 1.
- image quality improvement processing is sharpness conversion processing for reducing the difference in sharpness due to image blurring caused by the beam diameter of the primary electron beam.
- SEM SEM
- an observation site is scanned with an electron beam focused on a diameter of several nanometers.
- This beam diameter affects the sharpness of an image. That is, when the beam is thick, blurring occurs and an image with reduced sharpness is obtained. That is, in a plurality of apparatuses having different primary electron beam focusing performances, images with different sharpness can be obtained.
- deconvolution processing is effective, and conversely, to obtain an image with lower sharpness from the obtained image, a low-pass filter is effective.
- deconvolution processing is performed on the image of the apparatus 1.
- Samples of the image of the apparatus 2 are prepared, and the frequency intensity is calculated by processing such as Fourier transform on the image of the apparatus 1 and the image of the apparatus 2 that have been subjected to the deconvolution process, and the intensity of the high-frequency component is approximately the same ( For example, the above processing is repeated (until the difference in intensity between the two high-frequency components exceeds a predetermined threshold value).
- the above processing is an example, but by these processing, an image similar to the image obtained by the device 2 can be created from the image of the device 1.
- the image quality improvement process there is a contrast conversion process.
- this process when the image brightness changes slowly over the entire observation field due to the charging phenomenon of the sample surface, this brightness change is corrected, and the brightness of the circuit pattern part and defect part is corrected to improve visibility. Including processing to obtain a high image.
- the SEM when the imaging conditions are different, or when the imaging models are different even under the same imaging conditions, the light / dark relationship between the circuit pattern and the non-pattern part may be reversed.
- This contrast conversion process can unify the appearance of images captured between different devices or under different conditions by correcting the inverted brightness.
- contrast conversion processing is performed on the image of the apparatus 1. Samples of the image of the device 2 are prepared, and the luminance value average and variance of the image of the device 1 and the image of the device 2 on which the contrast conversion processing has been executed reach the degree of identification (for example, the difference between the luminance value average and the variance of both) The above process is repeated (until a predetermined threshold is exceeded).
- the above processing is an example, but by these processing, an image similar to the device 2 can be created from the image of the device 1.
- image conversion processing include shadow information conversion processing.
- the shadow information obtained by detecting backscattered electrons is strongly influenced by the arrangement of detectors in the apparatus.
- FIG. 11 when images having different detector arrangement forms coexist, there is a possibility that the determination of the concavo-convex state may be erroneous. Therefore, in order to prevent this, an image in which the direction of the shadow is converted is created.
- geometric transformation processing such as rotation processing and mirror image inversion processing is performed on the image in order to convert the shadow direction.
- rotation processing and inversion processing since the entire image is the processing target, it is not possible to change only the shadow direction. Therefore, when the rotation / reversal processing is performed, the captured circuit pattern and the like are similarly converted. However, this is not a problem in the process of determining the unevenness by analyzing the shadow. Because, in the determination of unevenness, the unevenness etc.
- FIG. 8 shows an example in which secondary images and backscattered electrons are separated and detected by the three detectors of the image pickup device shown in FIG. 3 to obtain three images. If they are different, it is assumed that the number of detectors and the types of detected electrons are also different. Therefore, a plurality of different images are created by mixing a plurality of detected images. For example, one device 1 can obtain an image in which a secondary electron image and a backscattered image are completely separated, while another device 2 is completely separated if a mixed image is detected. An image similar to the image obtained by the device 2 can be created by generating a plurality of images obtained by mixing the images from the detected image of the device 1. It should be noted that the various image conversion processes exemplified above can be executed in combination rather than independently.
- the conversion parameter table refers to the processing contents of the image conversion process for each combination of the target device (device 2) for which a similar image is to be created from the device (device 1) from which the original image was captured, It is a table in which processing procedures, parameters used in each processing, and the like are described.
- FIG. 12 in the case of two devices (device 1 and device 2), the teaching image is already registered in device 1 and process A, and device 2 is a defect image taken from the same process by device 2.
- device 1 and process A A case where a defect image of the same type as the teaching image is specified from the images of the process A will be described as an example.
- the defect image of the apparatus 2 and process A is stored in the image storage unit 213.
- the teaching image registered in each classification class in the apparatus 1 and step A is converted into an image similar to the image captured by the apparatus 2 by the image conversion unit 212 (S1201).
- the defect image of the apparatus 2 / process A stored in the image storage unit 213 in the process S1201 is compared with the teaching image of the apparatus 1 / process A after image conversion for each classification class (S1202). Based on the comparison result, a defect image of the same type as the teaching image registered in each classification class of the apparatus 1 / process A is specified for each classification class from the defect images of the apparatus 2 / process A (S1203).
- a feature amount such as the degree of unevenness of the defect portion and the size of the defect is calculated from each image, and it is determined that the defect image is the same type when the feature amount is close. Or the like.
- a method of determining the same type of defect image using the classification processing unit 216 is conceivable. In that case, the image converted for each classification class is taught to the classification processing unit 216, and the same type of defect image of the image converted image among the images of the apparatus 2 and step A is classified into each classification class.
- step S1201 the teaching image of the apparatus 1 / process A is not converted into an image of the apparatus 2, but the defect image of the apparatus 2 / process A is converted into an image of the apparatus 1, and the apparatus 1 / process is converted. You may compare with the teaching image of A.
- the teaching image of the apparatus 1 and the process A and the defect image of the apparatus 2 and the process A are converted into images captured by the apparatus 3 different from the apparatuses 1 and 2, respectively, and the converted image is processed in step S1202. You may compare each other.
- the apparatus 3 may be an apparatus installed in the same process, or may be an apparatus installed in a different process as long as the apparatus can perform image conversion.
- step S605 a part or all of the image converted in step S1201 may be registered in each classification class of the classification recipe of apparatus 2 and process A.
- the processing S1202 and the processing S1203 are not executed, and the teaching image of the device 1 is converted into an image. Or all of them may be registered in each classification class of the classification recipe of the apparatus 2 and the process A. If the machine difference between the apparatus 1 and the apparatus 2 is sufficiently small, the image conversion by the processing S604 is not executed, and part or all of the teaching image of the apparatus 1 is classified into each classification class of the classification recipe of the apparatus 2 and the process A. You may register for.
- the process S604 is appropriately executed according to the number of devices. Needless to say, this processing flow can be applied.
- a classification recipe in another device in the same process as the created classification recipe may already exist.
- devices are additionally arranged for the purpose of improving yield management efficiency by increasing the number of captured images with respect to a process in which devices and classification systems are already installed.
- the process of operating the device is step A
- the first installed device is device 1
- the added device is device 2 as an example. explain.
- the classification recipe of the apparatus 1 / process A exists, but the classification recipe of the apparatus 2 / process A does not exist.
- the classification recipe of the apparatus 1 / process A is determined by the processing flow of FIG. 5B and FIG. It is more efficient to create only the classification recipe of the apparatus 2 and the process A than to create both the classification recipe of the apparatus 2 and the process A.
- a process for setting the classification class of the apparatus 2 / process A when such a classification recipe of the apparatus 1 / process A exists will be described with reference to FIG.
- the processing flow for creating the classification recipe is the same as the processing flow described with reference to FIG. 5B.
- the previously existing classification recipe in FIG. It is not necessary to execute S515, and it may be performed only for a newly created classification recipe.
- it is assumed that the image of the apparatus 2 / process A is stored in the image storage unit 213 and the classification recipe of the apparatus 1 / process A exists, but the classification recipe of the apparatus 2 / process A does not exist.
- a reference classification recipe is selected (S1301).
- a classification recipe that already exists is selected, and in the following, an example in which the classification recipe of the apparatus 1 and the process A is selected as a reference classification recipe in the processing S1301 will be described.
- the information specifying unit 210 reads the apparatus and process information of the reference classification recipe (apparatus 1 / process A recipe) (S1302).
- the process information may be read from the classification recipe, may be determined from the accompanying information for each image stored in the accompanying information storage unit 215, or may be input by the user from the input / output unit 203. .
- a new classification recipe (apparatus 2 / process A classification recipe) of apparatus (apparatus 2) other than apparatus 1 is created with the apparatus that has captured the same process A image as the reference classification recipe. And set the classification class.
- an apparatus (apparatus 2) other than the apparatus 1 is specified by an apparatus that has captured an image of the same process A as the reference classification recipe (S1303).
- the classification recipe of the apparatus 2 / process A is created, and the same classification class as the classification class of the reference classification recipe (classification recipe of the apparatus 1 / process A) is defined (S1304).
- the defect image of the same kind as the teaching image registered in each classification class of the reference classification recipe (classification recipe of apparatus 1 and process A) is stored in the image storage unit 213 and is the image of apparatus 2 and process A. (S1305).
- the method of specifying is as described in the description of FIG.
- the image specified in step S1305 is registered in the classification class defined in step S1304 in the classification recipe newly created in step S1303 (classification recipe of apparatus 2 and step A) (S1306).
- process S601 When the classification recipe of process A is specified in process S601, if both the classification recipe of apparatus 1 and process A and the classification recipe of apparatus 2 and process A do not exist in classification recipe storage unit 214, the process flow of FIG. Is executed, and if any of the classification recipes of the apparatus 1 and the process A and the classification recipe of the apparatus 2 and the process A exists, a method of performing a conditional branch such as executing the processing flow of FIG.
- the classification definition in the classification recipe is kept the same in the same process. Describes the method and, as a method for identifying the same type of defect image for images of different manufacturers and types, converts the image to be compared into a similar image by image conversion, and converts the converted image and the feature amount of the converted image.
- a second embodiment of the defect classification system according to the present invention will be described with reference to FIGS.
- the second embodiment is a defect classification system that creates a recipe by the same processing flow as that of the first embodiment.
- the difference from Example 1 is that when performing defect classification, it has a function of determining a new defect as an unknown defect and a function of updating a classification recipe using an undetermined intelligent defect.
- a defect classification method and a classification recipe update method will be described.
- a case where an image captured by an observation apparatus equipped with an SEM is classified as in the first embodiment will be described as an example.
- the input of the defect classification system according to the present embodiment may be other than the SEM image. An image captured using the above means or an ion microscope may be used.
- FIG. 14 shows a configuration diagram of an embodiment of the defect classification system according to the second embodiment. The description of the same configuration as that of the defect classification system according to the first embodiment is omitted. The difference from the defect classification system according to the first embodiment is that an unknown defect determination unit 1402 that determines an unknown defect in the classification module 203 ′ and an unknown defect determined in the storage unit 208 ′ in the recipe management unit 202 ′. Is an unknown defect storage unit 1401 for storing.
- Processes S401 to S406 in FIG. 15 are the same as the process flow (FIG. 4) of the classification process of the first embodiment, and a description thereof is omitted.
- the unknown defect determination unit 1402 after calculating the feature amount (S406), the unknown defect determination unit 1402 performs processing for classifying the defect and determining the unknown defect (S1501), and the determined unknown defect is stored in the unknown defect storage unit 1401. The portion for performing the process of saving (S1502) is different.
- a method of determining based on the feature value distribution of the teaching image registered in the classification recipe and the Euclidean distance of the feature point to be classified may be used.
- the Euclidean distance is a set value.
- a method for determining an unknown defect in the above case can be considered.
- FIG. 16 is a flow when the same new classification class is defined in a classification recipe corresponding to a plurality of devices in the same process due to an unknown defect, and the classification recipe is updated.
- Processes S512 to S513 are the same as those described with reference to FIG. 5B of the first embodiment, and a description thereof is omitted.
- a part including a process of adding a new common classification class in step S1601 and registering an unknown defect in the new classification class is different. Details of the processing S1601 will be described later with reference to FIG.
- the image processing parameters and the classification parameters before the update of each classification recipe may be set as they are without being changed. In that case, the user can update a plurality of classification recipes with the same classification definition by executing only step S1601.
- Fig. 17 shows the processing flow for updating classification classes.
- classification recipes in a plurality of apparatuses in the same process can be updated collectively with the same classification definition.
- FIG. 17 a case where two devices (device 1 and device 2) are operated in step A will be described as an example.
- the classification recipe of apparatus 1 / process A and the classification recipe of apparatus 2 / process A exist and the classification recipe of apparatus 1 / process A is updated.
- a classification recipe in which a new class is defined by an unknown defect and the classification class is updated is specified (S1701).
- the classification recipe update process described in FIG. 16 is executed, and when the classification recipe in step S504 is stored, an update flag is set for each classification recipe, and the classification recipe in which the update flag is set. May be specified, or may be selected by the user using a GUI described later with reference to FIG.
- FIG. 17 a case where the classification recipe of the apparatus 1 and the process A is updated and specified in the process S1701 will be described as an example.
- step S1707 the classification class in the classification recipe in another apparatus is updated for the process specified in step S1701.
- the classification class of the classification recipe of the apparatus 2 / process A is updated.
- the information identification unit 210 first reads the updated classification recipe device and process information from the classification recipe storage unit 214 (S1702), and performs the same process as the updated classification recipe and the classification in another device.
- a recipe is specified (S1703).
- the classification recipe of the apparatus 2 and process A is specified.
- the classification recipe (the classification recipe of the apparatus 2 / process A in the example of FIG.
- the unknown defect identified in the process S1705 is registered in the new classification class defined in the classification recipe of the apparatus 2 / process A in the process S1704 (S1706).
- the number of images to be compared can be reduced.
- the unknown defect of the same type has not occurred before the time when the unknown defect registered in the new classification class of the classification recipe of the apparatus 1 / process A was imaged.
- the unknown defect imaged before the above time does not include the same type of defect image.
- the same kind is used from the time when the unknown defect registered in the new classification class of the apparatus 1 / process A is captured after the time when the unknown defect is imaged. It is good to specify a defect image.
- FIG. 18 shows an example of a GUI that displays a list of classification recipes for the same process and updates the classification class.
- Reference numeral 1801 denotes an area for displaying a list of classification recipes in the same process.
- Reference numeral 1802 denotes a process name displayed as a list in 1801.
- the classification recipes in the process A of the apparatuses 1 to 4 are listed and displayed.
- information such as classification class name (1803), device name (1804), process name (1805), classification class number (1806), number of taught images (1807), and the like are also displayed. May be.
- a new class is defined in the classification recipe of step A shown by 1801, and the updated classification recipe is not updated by a method such as surrounding with a line, changing a background color, or adding a mark as in 1808. Be able to distinguish from the classification recipe.
- Reference numeral 1809 denotes a button for instructing to update the classification recipe of the same process on the basis of the classification recipe updated by the processing flow of FIG. 17, and the classification recipe of the same process of another apparatus is updated by pressing the button. To do. At this time, there may be a classification list that does not want to be updated due to reasons such as performing another classification test, although the classification recipe is in the same process. In that case, when a check box (1810) corresponding to each classification recipe is arranged and the button 1809 is pressed, only the classification recipe checked in 1810 may be updated.
- a third embodiment of the defect classification system according to the present invention will be described with reference to FIGS.
- the third embodiment is a defect classification system that creates and updates a classification recipe by the same processing flow as that of the second embodiment.
- the difference from the second embodiment is that a plurality of defect classification systems 201 are installed.
- a classification recipe creating method and an updating method according to the third embodiment will be described.
- the input of the defect classification system according to the present embodiment may be other than the SEM image, and optical. It may be an image captured using an expression means or an ion microscope.
- FIG. 19 shows a configuration diagram of an embodiment of the defect classification system according to the present embodiment.
- the description of the same configuration as that of the defect classification system according to the first and second embodiments is omitted.
- the difference from the other embodiments is that a plurality of defect classification systems 201 (M ⁇ 2) are arranged via the communication means 204.
- FIG. 20 shows a processing flow for creating a classification recipe in the third embodiment.
- a classification recipe for process A is first created in the defect classification system 1.
- the 20 includes a processing flow (S2001) executed in the defect classification system and a processing flow (S2002) executed in the defect classification system 2. First, the processing flow S2001 is executed.
- a process for creating a classification recipe is specified (S2003).
- S2003 a process for creating a classification recipe
- a classification recipe corresponding to each device is created in the defect image of the process A stored in the defect classification system 1 (S2004).
- the process S2004 may be performed by executing the process flow for creating a classification recipe for a plurality of apparatuses in the same process described with reference to FIG. 5B and the process flow for setting a common classification class described with reference to FIG.
- the creation of a classification recipe is instructed in each device of step A in the defect classification system 2 (S2005).
- a classification recipe creation instruction signal, a part or all of the classification recipe of the process A of the defect classification system 1 created in the process S2004, information on the process to be created, and the like are sent to the defect classification system 2.
- a classification recipe in each device of the process A is created based on the classification recipe and process information transmitted in the process S2005 (S2006).
- these classification recipes are used as the standard classification recipe in FIG. 13 in the processing flow described in FIG. 5B and FIG. Then, a classification recipe is created and a classification class is set (S2006).
- each defect classification system is instructed to create a classification recipe in S2005, and the process flow S2002 is performed for each defect. Applicable by running in a classification system.
- the processes S2003 and S2004 in FIG. 20 may be replaced with the process flows described in FIGS. 16 and 17, and the process S2006 may be replaced with the process flows in FIGS.
- the embodiment described here takes a function (ADC) for automatically classifying a defect image captured by a review SEM as an example, and the creation of a classification recipe having the same classification class and the same classification, which are specific processing contents thereof
- ADC function for automatically classifying a defect image captured by a review SEM
- An update method for adding classes has been described, but similar images that can be compared by image conversion can also be generated for other defect observation and inspection devices that have a classification function and need to have the same classification class. If so, the present invention is applicable.
Landscapes
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Pathology (AREA)
- Health & Medical Sciences (AREA)
- Immunology (AREA)
- General Health & Medical Sciences (AREA)
- Biochemistry (AREA)
- Analytical Chemistry (AREA)
- Chemical & Material Sciences (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Testing Or Measuring Of Semiconductors Or The Like (AREA)
- Analysing Materials By The Use Of Radiation (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
Description
(1)試料を撮像する装置及び前記試料を製造する工程に対応する分類レシピを用いて欠陥画像を分類する欠陥分類方法であって、第一の画像撮像装置の分類レシピで定義された分類クラスと同一の分類クラスを、前記第一の画像撮像装置の分類レシピと同一の工程に対応する第二の画像撮像装置の分類レシピで定義するステップと、前記第二の画像撮像装置により撮像された欠陥画像の中から、前記第一の画像撮像装置の分類レシピで定義された分類クラスに登録された教示画像と同種の欠陥画像を特定するステップと、前記特定された欠陥画像を、前記第二の画像撮像装置の分類レシピで定義された分類クラスのうち、前記教示画像が登録された前記第一の画像撮像装置の分類クラスと同一の分類クラスに登録するステップと、を有することを特徴とする欠陥分類方法である。
(2)複数の画像撮像装置と通信手段を介して接続された欠陥分類システムであって、前記複数の画像撮像装置により撮像された欠陥画像を分類する分類手段と、分類のための情報を格納した分類レシピを管理する分類レシピ管理手段と、を有し、前記分類レシピ管理手段は、前記複数の画像撮像装置の一つである第一の画像撮像装置における分類レシピの分類クラスに登録された教示画像と同種の欠陥画像を、同一の工程に設置された前記複数の画像撮像装置の一つである第二の画像撮像装置より撮像された欠陥画像の中から特定する対応欠陥特定部と、前記教示画像を、画像変換によって前記第二の画像撮像装置から得られた欠陥画像と類似するように変換する画像変換部と、を有することを特徴とする欠陥分類システムである。
なお、図2で示した欠陥分類システム201の例は、一台の演算装置(以下、PC)内で運用しても良いし、複数台のPC内で分けて運用しても良い。複数のPCで運用する場合は、例えば、レシピ管理部202を一台のPCにてレシピサーバとして運用する方法も考えられる。また、図2で示した欠陥分類システム201の例では、分類モジュール203が一つの例を示したが、複数の分類モジュールを用いて構成しても構わない。複数の分類モジュールを有する場合はそれぞれ異なるPCで運用し、画像撮像装置毎に撮像した欠陥画像を処理する分類モジュールを割り当てるなどの方法も考えられる。なお、ここで示した変形例は後述の実施形態でも適用可能である。
まず、図5(a)を用いて、従来手法である単体の分類レシピ作成方法を説明する。まず、分類クラスの定義および教示画像を登録することによって分類クラスの設定を行う(S501)。ここでは、分類クラスを定義し、それぞれの分類クラスに教示画像を登録する。次に、欠陥画像中の欠陥領域や配線パターンなどを認識するための画像処理における画像処理パラメータの設定を行い(S502)、分類パラメータを調整し(S503)、このようにして作成した分類レシピを分類レシピ記憶部に保存する(S504)。画像処理パラメータ設定処理S502では、処理S501にて登録された教示画像に対して適切な画像処理結果が得られるように、画像処理パラメータの設定を行う。処理S503の分類パラメータの調整には、例えば、処理S501で登録した教示画像を分類モジュール内の分類処理部に教示し、分類識別面を作成するなどの手法によって行えば良く、分類クラス毎に教示画像が存在すれば自動で行うことが可能である。
Claims (15)
- 試料を撮像する装置及び前記試料を製造する工程に対応する分類レシピを用いて欠陥画像を分類する欠陥分類方法であって、
第一の画像撮像装置の分類レシピで定義された分類クラスと同一の分類クラスを、前記第一の画像撮像装置の分類レシピと同一の工程に対応する第二の画像撮像装置の分類レシピで定義するステップと、
前記第二の画像撮像装置により撮像された欠陥画像の中から、前記第一の画像撮像装置の分類レシピで定義された分類クラスに登録された教示画像と同種の欠陥画像を特定するステップと、
前記特定された欠陥画像を、前記第二の画像撮像装置の分類レシピで定義された分類クラスのうち、前記教示画像が登録された前記第一の画像撮像装置の分類クラスと同一の分類クラスに登録するステップと、
を有することを特徴とする欠陥分類方法。 - 請求項1記載の欠陥分類方法であって、
前記特定するステップは、前記第一の画像撮像装置の分類レシピで定義された分類クラスに登録された教示画像を、前記第二の画像撮像装置で撮像された画像と類似するように画像変換するステップを有することを特徴とする欠陥分類方法。 - 請求項2記載の欠陥分類方法であって、
前記特定するステップでは、前記画像変換された画像と前記第二の画像撮像装置により撮像された欠陥画像とを比較することにより、前記同種の欠陥画像を特定することを特徴とする欠陥分類方法。 - 請求項3記載の欠陥分類方法であって、
前記特定するステップでは、前記画像変換された画像から算出された特徴量と、前記第二の画像撮像装置により撮像された欠陥画像から算出された特徴量とを比較することにより、前記同種の欠陥画像を特定することを特徴とする欠陥分類方法。 - 請求項1記載の欠陥画像の分類方法であって、
前記特定するステップでは、前記第一の画像撮像装置の分類レシピで定義された分類クラスに登録された教示画像と、前記第二の画像撮像装置で撮像された欠陥画像とを、第三の画像撮像装置により撮像された画像と類似するように画像変換するステップを有することを特徴とする欠陥分類方法。 - 請求項5記載の欠陥画像の分類方法であって、
前記特定するステップでは、前記教示画像を画像変換した画像と、前記欠陥画像を画像変換した画像と、を比較することによって、前記同種の欠陥画像を特定することを特徴とする欠陥分類方法。 - 請求項6記載の欠陥分類方法であって、
前記特定するステップでは、前記教示画像を画像変換した画像から算出された特徴量と、前記欠陥画像を画像変換した画像から算出された特徴量とを比較することにより、前記同種の欠陥画像を特定することを特徴とする欠陥分類方法。 - 請求項1乃至7のいずれかに記載の欠陥分類方法であって、
前記特定するステップでは、
前記第二の画像撮像装置により撮像された欠陥画像のうち未知欠陥と判定された未知欠陥画像の中から、前記第一の画像撮像装置の分類レシピで定義された分類クラスに登録された教示画像と同種の欠陥画像を特定することを特徴とする欠陥分類方法。 - 請求項1乃至8のいずれかに記載の欠陥分類方法であって、
前記第二の画像撮像装置により撮像された欠陥画像の中から特定された画像に対してマークを表示するステップを有することを特徴とする欠陥分類方法。 - 請求項1乃至9のいずれかに記載の欠陥分類方法であって、
前記第二の画像撮像装置により撮像された欠陥画像の中から特定された画像を、表示画面上の一つのウィンドウ内にまとめて表示するステップを有することを特徴とする欠陥分類方法。 - 請求項2乃至4のいずれかに記載の欠陥分類方法であって、
前記登録するステップでは、
前記画像変換した画像を、前記第二の画像撮像装置の分類レシピ内で定義された分類クラスのうち、前記教示画像が登録された前記第一の画像撮像装置の分類クラスと同一の分類クラスに登録することを特徴とする欠陥分類方法。 - 複数の画像撮像装置と通信手段を介して接続された欠陥分類システムであって、
前記複数の画像撮像装置により撮像された欠陥画像を分類する分類手段と、
分類のための情報を格納した分類レシピを管理する分類レシピ管理手段と、
を有し、
前記分類レシピ管理手段は、
前記複数の画像撮像装置の一つである第一の画像撮像装置における分類レシピの分類クラスに登録された教示画像と同種の欠陥画像を、同一の工程に設置された前記複数の画像撮像装置の一つである第二の画像撮像装置より撮像された欠陥画像の中から特定する対応欠陥特定部と、
前記教示画像を、画像変換によって前記第二の画像撮像装置から得られた欠陥画像と類似するように変換する画像変換部と、
を有することを特徴とする欠陥分類システム。 - 請求項12記載の欠陥分類システムであって、
前記分類手段は、分類レシピに分類クラスが定義されていない未知欠陥を判定する未知欠陥判定手段を有し、
前記分類レシピ管理手段は、さらに、
前記教示画像と同種の欠陥画像を、前記第二の画像撮像装置により撮像された欠陥画像の中で分類レシピに分類クラスが定義されていない未知欠陥であると前記未知欠陥判定手段により判定された欠陥画像の中から特定することを特徴とする欠陥分類システム。 - 請求項12又は13記載の欠陥分類システムであって、
前記分類レシピ管理手段は、さらに、
前記第二の画像撮像装置により撮像された欠陥画像の中から特定された画像がマーク表示されるように、表示画面を有する入出力部にデータを出力することを特徴とする欠陥分類システム。 - 請求項12あるいは13に記載の欠陥分類システムであって、
前記分類レシピ管理手段は、さらに、
前記第二の画像撮像装置から撮像された画像の中から特定された画像を、表示画面上の一つのウィンドウ内にまとめて表示することを特徴とする欠陥分類システム。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201280019470.1A CN103502801B (zh) | 2011-04-20 | 2012-04-16 | 缺陷分类方法以及缺陷分类系统 |
KR1020137027296A KR101614592B1 (ko) | 2011-04-20 | 2012-04-16 | 결함 분류 방법 및 결함 분류 시스템 |
US14/112,105 US9401015B2 (en) | 2011-04-20 | 2012-04-16 | Defect classification method, and defect classification system |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2011-093603 | 2011-04-20 | ||
JP2011093603A JP5715873B2 (ja) | 2011-04-20 | 2011-04-20 | 欠陥分類方法及び欠陥分類システム |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2012144183A1 true WO2012144183A1 (ja) | 2012-10-26 |
Family
ID=47041317
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2012/002609 WO2012144183A1 (ja) | 2011-04-20 | 2012-04-16 | 欠陥分類方法及び欠陥分類システム |
Country Status (5)
Country | Link |
---|---|
US (1) | US9401015B2 (ja) |
JP (1) | JP5715873B2 (ja) |
KR (1) | KR101614592B1 (ja) |
CN (1) | CN103502801B (ja) |
WO (1) | WO2012144183A1 (ja) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2016194210A1 (ja) * | 2015-06-04 | 2016-12-08 | 株式会社日立ハイテクノロジーズ | 欠陥画像分類装置および欠陥画像分類方法 |
CN107408209A (zh) * | 2014-12-03 | 2017-11-28 | 科磊股份有限公司 | 无需取样及特征选择的自动缺陷分类 |
US11177111B2 (en) | 2019-04-19 | 2021-11-16 | Hitachi High-Tech Corporation | Defect observation device |
Families Citing this family (35)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7570796B2 (en) | 2005-11-18 | 2009-08-04 | Kla-Tencor Technologies Corp. | Methods and systems for utilizing design data in combination with inspection data |
US9659670B2 (en) | 2008-07-28 | 2017-05-23 | Kla-Tencor Corp. | Computer-implemented methods, computer-readable media, and systems for classifying defects detected in a memory device area on a wafer |
JP5608575B2 (ja) * | 2011-01-19 | 2014-10-15 | 株式会社日立ハイテクノロジーズ | 画像分類方法および画像分類装置 |
US9170211B2 (en) | 2011-03-25 | 2015-10-27 | Kla-Tencor Corp. | Design-based inspection using repeating structures |
JP5715873B2 (ja) * | 2011-04-20 | 2015-05-13 | 株式会社日立ハイテクノロジーズ | 欠陥分類方法及び欠陥分類システム |
US9087367B2 (en) | 2011-09-13 | 2015-07-21 | Kla-Tencor Corp. | Determining design coordinates for wafer defects |
KR101851241B1 (ko) * | 2011-12-06 | 2018-04-24 | 삼성전자 주식회사 | 휴대 단말기의 컨텐츠 통합 관리 방법 및 장치 |
US9189844B2 (en) | 2012-10-15 | 2015-11-17 | Kla-Tencor Corp. | Detecting defects on a wafer using defect-specific information |
US9053527B2 (en) | 2013-01-02 | 2015-06-09 | Kla-Tencor Corp. | Detecting defects on a wafer |
US9134254B2 (en) | 2013-01-07 | 2015-09-15 | Kla-Tencor Corp. | Determining a position of inspection system output in design data space |
US9311698B2 (en) | 2013-01-09 | 2016-04-12 | Kla-Tencor Corp. | Detecting defects on a wafer using template image matching |
WO2014149197A1 (en) | 2013-02-01 | 2014-09-25 | Kla-Tencor Corporation | Detecting defects on a wafer using defect-specific and multi-channel information |
US9865512B2 (en) * | 2013-04-08 | 2018-01-09 | Kla-Tencor Corp. | Dynamic design attributes for wafer inspection |
US9310320B2 (en) | 2013-04-15 | 2016-04-12 | Kla-Tencor Corp. | Based sampling and binning for yield critical defects |
CN103904002B (zh) * | 2014-03-20 | 2016-08-17 | 上海华力微电子有限公司 | 一种验证缺陷检测程序灵敏度的方法 |
US9830421B2 (en) * | 2014-12-31 | 2017-11-28 | Kla-Tencor Corp. | Alignment of inspection to design using built in targets |
JP6244329B2 (ja) * | 2015-05-12 | 2017-12-06 | 東京エレクトロン株式会社 | 基板の検査方法、基板処理システム及びコンピュータ記憶媒体 |
US9594021B2 (en) * | 2015-08-12 | 2017-03-14 | National Applied Research Laboratories | Apparatus of detecting transmittance of trench on infrared-transmittable material and method thereof |
CN108886007B (zh) * | 2016-04-13 | 2020-03-27 | 科磊股份有限公司 | 用于基于电设计意图的缺陷分类的系统及方法 |
US10810733B2 (en) | 2016-05-24 | 2020-10-20 | Hitachi High-Tech Corporation | Defect classification apparatus and defect classification method |
KR101744194B1 (ko) * | 2016-08-19 | 2017-06-09 | 인하대학교 산학협력단 | 반도체 fab 제조공정에서 유클리드 거리를 활용한 웨이퍼 자동 불량 검사 분류 예측 장치 및 방법 |
CN106290378B (zh) * | 2016-08-23 | 2019-03-19 | 东方晶源微电子科技(北京)有限公司 | 缺陷分类方法和缺陷检查系统 |
JP2018091771A (ja) * | 2016-12-06 | 2018-06-14 | 大日本印刷株式会社 | 検査方法、事前画像選別装置及び検査システム |
US10790114B2 (en) * | 2017-06-29 | 2020-09-29 | Kla-Tencor Corporation | Scanning electron microscope objective lens calibration using X-Y voltages iteratively determined from images obtained using said voltages |
JP7004145B2 (ja) * | 2017-11-15 | 2022-01-21 | オムロン株式会社 | 欠陥検査装置、欠陥検査方法、及びそのプログラム |
KR20190073756A (ko) * | 2017-12-19 | 2019-06-27 | 삼성전자주식회사 | 반도체 결함 분류 장치, 반도체의 결함을 분류하는 방법, 그리고 반도체 결함 분류 시스템 |
KR102549196B1 (ko) | 2018-02-07 | 2023-06-30 | 어플라이드 머티리얼즈 이스라엘 리미티드 | 반도체 시편의 심층 학습 기반 검사 방법 및 그의 시스템 |
KR102666072B1 (ko) * | 2018-06-05 | 2024-05-17 | 삼성전자주식회사 | 기계 학습 장치 및 기계 학습 장치의 학습 방법 |
JP7289658B2 (ja) * | 2018-09-07 | 2023-06-12 | 株式会社フジクラ | 分類装置、分類方法、分類プログラム、及び検査装置 |
KR102324622B1 (ko) * | 2018-12-12 | 2021-11-12 | 어플라이드 머티리얼즈 이스라엘 리미티드 | 프로세스 모니터링 |
US11087449B2 (en) * | 2019-10-24 | 2021-08-10 | KLA Corp. | Deep learning networks for nuisance filtering |
JP7496702B2 (ja) | 2020-03-27 | 2024-06-07 | 日鉄ステンレス株式会社 | 疵検査装置及び疵検査方法 |
JP7379303B2 (ja) * | 2020-09-14 | 2023-11-14 | 株式会社東芝 | 類似不良検索/表示システム、装置及び方法 |
CN113012128B (zh) * | 2021-03-18 | 2022-04-15 | 长鑫存储技术有限公司 | 缺陷表征方法和装置 |
US20220301147A1 (en) * | 2021-03-18 | 2022-09-22 | Changxin Memory Technologies, Inc. | Defect characterization method and apparatus |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH11282822A (ja) * | 1998-03-26 | 1999-10-15 | Toshiba Corp | 画像処理プログラム合成方法及びその装置並びに最適解探索方法及びその装置 |
JP2004047939A (ja) * | 2002-05-23 | 2004-02-12 | Hitachi High-Technologies Corp | 欠陥分類器の生成方法および欠陥自動分類方法 |
JP2005274285A (ja) * | 2004-03-24 | 2005-10-06 | Olympus Corp | 欠陥分類辞書教示装置 |
WO2012046431A1 (ja) * | 2010-10-08 | 2012-04-12 | 株式会社日立ハイテクノロジーズ | 欠陥分類システム及び欠陥分類装置及び画像撮像装置 |
Family Cites Families (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP3476913B2 (ja) * | 1994-07-08 | 2003-12-10 | オリンパス株式会社 | 欠陥種別判定装置及びプロセス管理システム |
US6542830B1 (en) * | 1996-03-19 | 2003-04-01 | Hitachi, Ltd. | Process control system |
US6246787B1 (en) * | 1996-05-31 | 2001-06-12 | Texas Instruments Incorporated | System and method for knowledgebase generation and management |
JP4220595B2 (ja) * | 1998-08-10 | 2009-02-04 | 株式会社日立製作所 | 欠陥の分類方法並びに教示用データ作成方法 |
WO2003028089A1 (fr) * | 2001-09-19 | 2003-04-03 | Olympus Optical Co., Ltd. | Systeme de controle de tranches en semiconducteur |
JP2004226328A (ja) * | 2003-01-24 | 2004-08-12 | Hitachi Ltd | 外観検査システムおよびそれらを用いた品質評価システムおよび品質評価情報提供システム |
US7602962B2 (en) | 2003-02-25 | 2009-10-13 | Hitachi High-Technologies Corporation | Method of classifying defects using multiple inspection machines |
US20050075841A1 (en) * | 2003-08-05 | 2005-04-07 | Netanel Peles | Automated defect classification system and method |
JP4176041B2 (ja) * | 2004-04-14 | 2008-11-05 | オリンパス株式会社 | 分類装置及び分類方法 |
WO2006046236A1 (en) * | 2004-10-26 | 2006-05-04 | May High-Tech Solutions, Ltd. | Method and apparatus for residue detection on a polished wafer |
WO2007074770A1 (ja) * | 2005-12-26 | 2007-07-05 | Nikon Corporation | 画像解析によって欠陥検査を行う欠陥検査装置 |
JP4644613B2 (ja) | 2006-02-27 | 2011-03-02 | 株式会社日立ハイテクノロジーズ | 欠陥観察方法及びその装置 |
JP5271491B2 (ja) * | 2006-10-26 | 2013-08-21 | 株式会社日立ハイテクノロジーズ | 電子線応用装置および試料検査方法 |
WO2008086282A2 (en) * | 2007-01-05 | 2008-07-17 | Kla-Tencor Corporation | Methods and systems for using electrical information for a device being fabricated on a wafer to perform one or more defect-related functions |
JP5327973B2 (ja) * | 2007-12-05 | 2013-10-30 | 芝浦メカトロニクス株式会社 | 特徴解析装置 |
SG188094A1 (en) * | 2008-01-30 | 2013-03-28 | Rudolph Technologies Inc | High resolution edge inspection |
US8175373B2 (en) * | 2009-02-16 | 2012-05-08 | Kla-Tencor Corporation | Use of design information and defect image information in defect classification |
JP5715873B2 (ja) * | 2011-04-20 | 2015-05-13 | 株式会社日立ハイテクノロジーズ | 欠陥分類方法及び欠陥分類システム |
-
2011
- 2011-04-20 JP JP2011093603A patent/JP5715873B2/ja active Active
-
2012
- 2012-04-16 US US14/112,105 patent/US9401015B2/en active Active
- 2012-04-16 CN CN201280019470.1A patent/CN103502801B/zh active Active
- 2012-04-16 KR KR1020137027296A patent/KR101614592B1/ko active IP Right Grant
- 2012-04-16 WO PCT/JP2012/002609 patent/WO2012144183A1/ja active Application Filing
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH11282822A (ja) * | 1998-03-26 | 1999-10-15 | Toshiba Corp | 画像処理プログラム合成方法及びその装置並びに最適解探索方法及びその装置 |
JP2004047939A (ja) * | 2002-05-23 | 2004-02-12 | Hitachi High-Technologies Corp | 欠陥分類器の生成方法および欠陥自動分類方法 |
JP2005274285A (ja) * | 2004-03-24 | 2005-10-06 | Olympus Corp | 欠陥分類辞書教示装置 |
WO2012046431A1 (ja) * | 2010-10-08 | 2012-04-12 | 株式会社日立ハイテクノロジーズ | 欠陥分類システム及び欠陥分類装置及び画像撮像装置 |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107408209A (zh) * | 2014-12-03 | 2017-11-28 | 科磊股份有限公司 | 无需取样及特征选择的自动缺陷分类 |
CN107408209B (zh) * | 2014-12-03 | 2021-07-06 | 科磊股份有限公司 | 用于在半导体工艺中进行缺陷分类的系统和方法 |
WO2016194210A1 (ja) * | 2015-06-04 | 2016-12-08 | 株式会社日立ハイテクノロジーズ | 欠陥画像分類装置および欠陥画像分類方法 |
KR20170141255A (ko) * | 2015-06-04 | 2017-12-22 | 가부시키가이샤 히다치 하이테크놀로지즈 | 결함 화상 분류 장치 및 결함 화상 분류 방법 |
KR101978995B1 (ko) | 2015-06-04 | 2019-05-16 | 가부시키가이샤 히다치 하이테크놀로지즈 | 결함 화상 분류 장치 및 결함 화상 분류 방법 |
US11177111B2 (en) | 2019-04-19 | 2021-11-16 | Hitachi High-Tech Corporation | Defect observation device |
Also Published As
Publication number | Publication date |
---|---|
CN103502801A (zh) | 2014-01-08 |
KR101614592B1 (ko) | 2016-04-21 |
JP2012225768A (ja) | 2012-11-15 |
KR20130135962A (ko) | 2013-12-11 |
JP5715873B2 (ja) | 2015-05-13 |
US9401015B2 (en) | 2016-07-26 |
CN103502801B (zh) | 2015-09-16 |
US20140072204A1 (en) | 2014-03-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5715873B2 (ja) | 欠陥分類方法及び欠陥分類システム | |
JP4644613B2 (ja) | 欠陥観察方法及びその装置 | |
KR102195029B1 (ko) | 결함 분류 장치 및 결함 분류 방법 | |
JP2022172401A (ja) | パターン検査システム | |
JP5103058B2 (ja) | 欠陥観察装置及び欠陥観察方法 | |
US9390490B2 (en) | Method and device for testing defect using SEM | |
JP5957378B2 (ja) | 欠陥観察方法および欠陥観察装置 | |
WO2012046431A1 (ja) | 欠陥分類システム及び欠陥分類装置及び画像撮像装置 | |
US9040937B2 (en) | Charged particle beam apparatus | |
JP5444092B2 (ja) | 検査方法およびその装置 | |
JP5948138B2 (ja) | 欠陥解析支援装置、欠陥解析支援装置で実行されるプログラム、および欠陥解析システム | |
JP5325580B2 (ja) | Semを用いた欠陥観察方法及びその装置 | |
JP2007225351A (ja) | 欠陥表示方法およびその装置 | |
JP2006179889A (ja) | 回路パターンの検査方法及び検査装置 | |
JP4262288B2 (ja) | パターンマッチング方法及び装置 | |
JP2015008059A (ja) | 荷電粒子線装置および画像蓄積方法 | |
WO2015064399A1 (ja) | 荷電粒子線装置およびプログラム記録媒体 |
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: 12774882 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 20137027296 Country of ref document: KR Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14112105 Country of ref document: US |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 12774882 Country of ref document: EP Kind code of ref document: A1 |