Measurement system, computer, and measurement method

Abstract

Provided is a technique capable of measuring an overlap offset amount and the like with high accuracy. In a measurement system, a processor acquires an image (309) obtained by capturing an image of a structure of a semiconductor device with a microscope, acquires a measurement region generation rule (308) relating to the structure, generates a measurement region for disposing the structure on the basis of the image and the measurement region generation rule, and disposes the measurement region on the structure of the image. Measurements relating to the structure are performed using a portion in the measurement region of the image (313).

Description

Technical Field

[0001] The present disclosure relates to a technique for measuring the size, overlap, etc. of a sample such as a semiconductor. Background Art

[0002] Conventional semiconductor pattern measurement devices, such as charged particle beam devices, include technologies that use images captured by a scanning electron microscope (SEM) to measure pattern dimensions, center coordinates, and overlap offset. Overlap offset, for example, refers to the amount of overlap between lower and upper patterns.

[0003] With the recent advancement in miniaturization and multi-layered structures of patterns produced by semiconductor processes, there is a growing demand for reducing the misalignment of patterns between multiple layers in the exposure system. Consequently, the importance of accurately measuring this misalignment and providing feedback to the exposure system is increasing.

[0004] Measuring methods for such overlay offsets include measurement devices using the aforementioned SEM. The SEM generates and outputs a captured image by detecting particles such as secondary electrons and backscattered electrons generated when a charged particle beam is irradiated onto a sample, such as a semiconductor wafer. The measurement device uses this captured image as the measured image, performs appropriate image processing, and calculates the positions of the patterns of the multiple layers being measured, such as the overlay offset. This makes it possible to measure overlay offsets and other parameters.

[0005] Examples of the prior art include International Publication No. 2021 / 038815 (Patent Document 1) and Japanese Patent Application Laid-Open No. 2020-187876 (Patent Document 2).

[0006] Prior art literature

[0007] Patent Literature

[0008] Patent Document 1: International Publication No. 2021 / 038815

[0009] Patent Document 2: Japanese Patent Application Laid-Open No. 2020-187876 Summary of the Invention

[0010] Problems to be solved by the invention

[0011] Patent Document 1 describes a method for measuring the amount of overlap offset by generating a segmented image from an input image (measurement target) of a semiconductor having a predetermined structure, using a learning model generated based on supervisory data and sample images. The supervised data is an image in which a label representing the structure of the semiconductor in the sample image is assigned to each pixel, and the learning model includes parameters for inferring the supervisory data from the sample image.

[0012] Patent document 2 describes the following: it comprises: a charged particle beam irradiation unit that irradiates a charged particle beam to a sample; a first detector that detects secondary electrons from the sample; a second detector that detects reflected electrons from the sample; and an image processing unit that generates a first image including an image of a first pattern located on the surface of the sample based on the output of the first detector, and generates a second image including an image of a second pattern located at a lower layer than the surface of the sample based on the output of the second detector, wherein a control unit adjusts the position of a measurement area in the first image based on a first template image regarding the first image, and adjusts the position of a measurement area in the second image based on a second template image regarding the second image, thereby measuring an overlapping offset.

[0013] As semiconductor device patterns become increasingly smaller, the outline of the pattern reflected in the measured image may become unclear. In particular, the boundary between the overlapping upper and lower patterns, or the boundary between the lower pattern and the background without a pattern, may become unclear. This can reduce the accuracy of measurements such as the overlap offset.

[0014] For example, the technology in Patent Document 1 is difficult to generate accurate supervisory data at the pixel level. If the learning model learns incorrect supervisory data, the boundaries of the generated segmented image will differ from the boundaries of the actual pattern. In this case, the measurement accuracy of the overlap offset and other parameters will be reduced.

[0015] Furthermore, as semiconductor device patterns become increasingly miniaturized and processes become more complex, process variations in individual patterns that deviate from each other on the measured image may become larger relative to the pattern size, potentially reducing measurement accuracy for overlay offsets and the like.

[0016] For example, the technique in Patent Document 2 places a measurement area defined by a template image on the image being measured. Consequently, if the pattern in the image being measured deviates or fluctuates in size or position compared to the pattern in the template image, the measurement area may not be located at the position of the pattern being measured. This reduces the accuracy of overlay measurement.

[0017] Furthermore, the aforementioned process variations are as follows. If any disruptions or variations occur during the semiconductor device manufacturing process (in other words, the process), these variations or variations are reflected in the pattern structure of the manufactured semiconductor device, resulting in variations or variations in the dimensions, position, and other aspects of the actual pattern structure. Furthermore, these variations in the actual structure also manifest as variations in the pattern structure in the measured image.

[0018] An object of the present disclosure is to provide a technique for measuring the above-mentioned overlay shift amount and the like, which can stably, in other words, measure the overlay shift amount and the like with higher accuracy.

[0019] Means for solving problems

[0020] A representative embodiment of the present disclosure has the following structure. The embodiment is a semiconductor device measurement system comprising a microscope and a processor, wherein the processor obtains an image of a structure of the semiconductor device captured by the microscope, obtains a measurement region generation rule related to the structure, generates a measurement region for arranging the structure based on the image and the measurement region generation rule, arranges the measurement region for the structure in the image, and performs measurement related to the structure using a portion of the image within the measurement region.

[0021] Effects of the Invention

[0022] According to the representative embodiment of the present disclosure, the measurement technology of the overlap offset etc. can stably, in other words, more accurately measure the overlap offset etc. Other problems, structures, and effects are shown in the embodiments of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a diagram showing the configuration of the measurement system according to the first embodiment.

[0024] Figure 2 This is a diagram showing the configuration of a computer system serving as a computer in the first embodiment.

[0025] Figure 3 This is a functional block diagram of the measurement system according to the first embodiment.

[0026] Figure 4A In the first embodiment, an example of a design structure to be measured is an XY plan view.

[0027] Figure 4B In the first embodiment, an example of a design structure to be measured is an XZ cross-sectional view.

[0028] Figure 5 In implementation mode 1, relative to Figure 4AXY plan view of the structure when there is an offset in the designed structure.

[0029] Figure 6 In implementation mode 1, relative to Figure 5 XY plan view of the structure when there is deviation due to process variation in the structure.

[0030] Figure 7A In embodiment 1, the shooting Figure 6 Schematic diagram of the SE image obtained by construction.

[0031] Figure 7B In embodiment 1, the shooting Figure 6 Schematic diagram of the BSE image obtained by construction.

[0032] Figure 8 This is a flowchart of the processing of the learning unit in the first embodiment.

[0033] Figure 9A This is a diagram showing Example 1 of a region segmentation image of a sample image in the first embodiment.

[0034] Figure 9B This is a diagram showing a second example of a region segmentation image of a sample image in the first embodiment.

[0035] Figure 10 This is a flowchart of the process related to the setting of the measurement area creation rule in the first embodiment.

[0036] Figure 11A In implementation mode 1, Figure 9A Figure 1 shows an example of a regular object region in a region segmentation image.

[0037] Figure 11B In implementation mode 1, Figure 9B Figure 1 shows an example of a regular object region in a region segmentation image.

[0038] Figure 12 This is a diagram showing a setting example of measurement area creation rules in the first embodiment.

[0039] Figure 13A This is a diagram showing the first part of an example GUI screen in the first embodiment.

[0040] Figure 13B This is a diagram showing the second part of the GUI screen example in the first embodiment.

[0041] Figure 14 This is a flowchart of the processing in the measurement execution phase in the first embodiment.

[0042] Figure 15This is an explanatory diagram showing the generation of a region segmentation image, the generation of a measurement region, and the like in the first embodiment.

[0043] Figure 16A This is an explanatory diagram regarding edge detection and centroid calculation of an image arranged from a measurement region in the first embodiment.

[0044] Figure 16B This is a diagram showing an example of superimposed measurement results in the first embodiment.

[0045] Figure 16C This is a diagram showing an example of measuring the amount of overlap deviation in the Y direction in the measurement region arrangement image in the first embodiment.

[0046] Figure 16D This is a diagram showing an example of measuring the amount of overlap deviation in the X direction in the measurement region arrangement image in the first embodiment.

[0047] Figure 16E This is a diagram showing another example of center-of-gravity calculation in a modification of the first embodiment.

[0048] Figure 17 It is a diagram showing an example of arrangement of measurement regions in images in comparative examples (Comparative Example 1, Comparative Example 2) with respect to the first embodiment.

[0049] Figure 18 This is a functional block configuration diagram of the measurement system according to the second embodiment.

[0050] Figure 19 This is a diagram showing an example of generating measurement regions in a region segmentation image in the second embodiment.

[0051] Figure 20 This is a diagram showing details of the generation of the measurement area in the second embodiment.

[0052] Figure 21 This is a diagram showing an example of edge detection and center of gravity calculation in the second embodiment.

[0053] Figure 22 This is a diagram showing the concept and example of setting the measurement area in the first embodiment.

[0054] Figure 23 This is a diagram showing the concept and example of setting the measurement area in the first embodiment.

[0055] Figure 24 This is a diagram showing an example of generation of a measurement area based on a measurement area generation rule when there is a pattern shift in the first embodiment.

[0056] Figure 25This is a diagram showing an example of measuring the size of a pattern in the first embodiment.

[0057] Figure 26 This is an enlarged view showing the measurement area boundary setting column in the first embodiment. DETAILED DESCRIPTION

[0058] The following describes the embodiments of the present disclosure in detail with reference to the accompanying drawings. In principle, identical components are denoted by the same reference numerals in the accompanying drawings, and duplicate descriptions are omitted. In the accompanying drawings, components are sometimes depicted without indicating their actual position, size, shape, or range to facilitate understanding of the invention, but this is not intended to be limiting.

[0059] In the description, when describing program-based processing, the program, function, processing unit, etc. are sometimes used as the main body for description, but the main body of the hardware related to them is the processor, or the controller, device, computer, system, etc. composed of the processor, etc. The computer uses the processor to appropriately use resources such as memory and communication interfaces while executing the processing according to the program read from the memory. In this way, the specified function, processing unit, etc. are realized. The processor is composed of semiconductor devices such as CPU / MPU and GPU, etc., for example. Processing is not limited to software program processing and can also be implemented through dedicated circuits. Dedicated circuits can apply FPGA, ASIC, CPLD, etc.

[0060] The program can be pre-installed on the target computer as data, or it can be distributed to the target computer as data from a program source. The program source can be a program distribution server on a communication network, or it can be a non-transitory computer-readable storage medium, such as a memory card or disk. The program can also be composed of multiple modules. The computer system can also be composed of multiple devices. The computer system can also be composed of a client, a server system, a cloud computing system, an IoT system, etc. Various data and information are composed of structures such as tables and lists, but are not limited to these. Identification information, identifiers, IDs, names, numbers, etc. can be interchangeable.

[0061] [Overview of Embodiments]

[0062] The measurement system of the embodiment includes a microscope and a processor. In other words, the microscope is a charged particle beam device, an imaging device, or the like. In other words, the processor is a computer or computer system equipped with a processor. The measurement system of the embodiment measures predetermined parameter values, such as pattern dimensions and overlay offset, for samples such as semiconductor devices. In other words, it measures object values.

[0063] In the measurement system of the embodiment, for example, during the measurement process production phase, a processor generates a segmented image of the sample image based on a sample image obtained by photographing a sample pattern. The processor then obtains and sets measurement region generation rules related to the measurement target pattern based on a user's confirmation of the segmented image. The measurement region generation rules are rules for generating a measurement region for the measurement target pattern structure and can be set by the user on a screen.

[0064] For example, during the measurement execution phase of the measurement system, the processor acquires a measured image, which is an image of the sample pattern captured by a microscope. The measurement system acquires a measured image (in other words, a SEM image), which is an image of a sample, such as a semiconductor wafer, having a predetermined structure, such as a three-dimensional pattern, captured by a microscope (e.g., an SEM).

[0065] The processor of the measurement system generates a region segmentation image based on the measured image. The region segmentation image is an image segmented according to the region of each pattern structure. In one embodiment, the measurement system generates the region segmentation image based on the measured image through unsupervised machine learning.

[0066] The measurement system processor generates a measurement area based on the measured image and a measurement area generation rule, and places the measurement area within the pattern structure of the measured image. The measurement system processor obtains and references the measurement area generation rule applied to the regional elements of the pattern structure, applies the measurement area generation rule to the region segmentation image of the measured image (particularly, the designated rule target region), and generates the measurement area for the regional elements of the pattern structure. The measurement system processor places the measurement area within the measured image.

[0067] The processor of the measurement system then measures predetermined parameter values, in other words, measurement target values, such as pattern structure dimensions and overlap offsets, using the portion within the measurement area of ​​the measured image, and stores and outputs the measurement results.

[0068] The measurement region generation rule is a rule for specifying and determining the boundary of the measurement region using, for example, correction values, in other words, relative relationships or differences, based on coordinate information of region elements in a pattern structure in the region segmentation image.

[0069] Furthermore, in the measurement system of the embodiment, a portion of the measurement area generation rule includes a measurement area permission determination rule. In other words, the measurement area permission determination rule can be attached to and set within the measurement area generation rule. The measurement area permission determination rule is a rule used to determine whether or not measurement areas can be generated and placed for regional elements of a pattern structure. When the processor of the measurement system applies the measurement area generation rule to the region segmentation image to generate measurement areas, it determines whether or not to generate and place measurement areas according to the measurement area permission determination rule. If the processor determines that the measurement area is not generated or placed, it does not generate or place measurement areas.

[0070] <Implementation Method 1>

[0071] use Figure 1 The following figures illustrate the measurement system and other aspects of the first embodiment of the present disclosure. The measurement system of the first embodiment uses images of semiconductor devices captured using a microscope as the measured image and uses a computer to measure the amount of overlay offset. The measurement method of the first embodiment is executed by the computer of the measurement system of the first embodiment.

[0072] [Measurement system]

[0073] Figure 1 The structure of a measuring system 100 according to Embodiment 1 is shown. The measuring system 100 includes a scanning electron microscope (SEM) 101 , which is a type of charged particle beam device, a host computer 104 , an input / output device 105 , a first sub-computer 107 , and a second sub-computer. Figure 1 The various structural elements in the system are connected to each other via a network 110, such as a bus, LAN, WAN, cable, signal line or other communication means, and can appropriately exchange signals / data / information.

[0074] SEM 101 includes a main body 101A and a controller 102. SEM 101 captures a pattern (e.g., a three-dimensional pattern structure) of a semiconductor 201 (e.g., a wafer) as an inspection object, and generates and supplies a captured image. Main body 101A irradiates sample 201 with a charged particle beam b1, generating and outputting detection signals a1 and a2. Controller 102 is a control and image generation device that controls the entire SEM 101, drives and controls main body 101A, and generates an image to be measured based on detection signals a1 and a2 from main body 101A. Controller 102 supplies signals / data a3, such as the captured image, to a host computer 104, etc.

[0075] The host computer 104 is connected to the controller 102 and includes a main processor 103 as at least one processor. The host computer 104 uses the image captured by the SEM 101 as the measured image and processes the image to determine the values ​​of predetermined parameters such as pattern size and overlap offset. The host computer 104 stores and outputs measurement result data.

[0076] The input / output device 105 is operated by a user U1 to input instructions, settings, various data / information, etc. to the host computer 104, etc., and output measurement results, etc. The input / output device 105 includes input devices such as a mouse, keyboard, and microphone, and output devices such as a display, printer, and speakers. The input / output device 105 may also be a client terminal device such as a PC connected via the network 110. The user U1 is the person who uses this measurement system to perform measurement operations or manage the measurement system.

[0077] In other words, the host computer 104, or the portion consisting of the host computer 104 and the input / output device 105, constitutes a computer system. The computer system may also be a client-server system in which the host computer 104 serves as a server and the input / output device 105 serves as a client. Alternatively, the input / output device 105 may be integrally implemented in the host computer 104.

[0078] The first sub-computer 107 and the second sub-computer 109 are sub-computers for the host computer 104. The first sub-computer 107 is connected to the controller 102 and other devices and includes a first sub-processor 106 as at least one processor. The second sub-computer 109 is connected to the controller 102 and other devices and includes a second sub-processor 108 as at least one processor. Input and output can be performed from the input / output device 105 to each sub-computer, or the input / output device can be provided for each sub-computer. Alternatively, the input / output device can be integrated into the sub-computer.

[0079] exist Figure 1 In the configuration example, as the main process in the measurement system 100, the main computer 104 performs processing related to overlay measurement (the measurement process creation process and measurement process described later). As sub-processes assisting the main computer 104, the first sub-computer 107, etc., as a sub-computer, performs processing related to machine learning. More than one sub-computer is provided. This is not limiting, and in a modified example, the measurement process may be performed by the first sub-computer 107 or the second sub-computer 109. Furthermore, multiple computers (e.g., the main computer 104 and a sub-computer, or multiple sub-computers) may each perform measurement processing in parallel and in a distributed manner.

[0080] exist Figure 1The configuration example shown includes two sub-computers in addition to the main computer 104, but the present invention is not limited thereto. In other configurations, only the main computer 104 may be provided, with the main computer 104 performing all calculations and processing, including measurement and learning. The measurement system 100 may be any system that includes at least one computer system such as the main computer 104 and that acquires and processes measured images.

[0081] In addition, Figure 1 Although a single SEM 101 is shown, multiple microscopes may be present, or a server computer, etc., may be used in place of a microscope to store images captured by the SEM. If the SEM 101 is a server computer, the server computer stores images of semiconductor patterns captured by the SEM in a storage device, such as a hard disk drive (HDD). The server computer provides data such as images in response to requests from a host computer 104 or the like.

[0082] Furthermore, the operator responsible for the host computer 104, which performs the measurement processing as the primary process, and the operator responsible for the sub-computers performing machine learning, etc., may be different. For example, the operator responsible for the host computer 104 may collaborate with an operator providing machine learning services to outsource learning to the sub-computers performing machine learning or receive learning results. The sub-computers performing machine learning, etc., may also be constructed as a cloud computing system on the Internet.

[0083] The host computer 104 executes the measurement recipe creation process, described below. A measurement recipe is a series of control and setting information related to overlay measurement. In this embodiment, part of the measurement recipe also includes information such as the measurement area generation rules for overlay measurement. Furthermore, the host computer 104 executes the measurement process, described below. This measurement process measures the overlay offset, etc., according to the measurement recipe.

[0084] In addition, in the case of a client-server system, for example, the following actions are performed. User U1 accesses the server serving as the host computer 104 from the client PC serving as the input / output device 105. The server provides the client PC with a screen accompanied by a graphical user interface (GUI). The server sends the GUI screen data (for example, a web page) for this purpose to the client PC. The client PC displays the GUI screen on the display based on the received screen data. User U1 observes the GUI screen and inputs instructions, settings, etc. The client PC sends the input information to the server. The server performs processing corresponding to the received input information. For example, the server performs measurement process settings and overlapping measurement processing, stores the processing results, and sends the GUI screen data (which may also be only update information) for displaying the processing results to the client PC. The client PC updates the display of the GUI screen based on the received screen data. User U1 can observe the GUI screen to confirm the processing results, such as the measurement process, measurement results, etc.

[0085] [SEM]

[0086] exist Figure 1 In the figure, SEM 101 includes a main body 101A containing a sample chamber, which includes a sample stage, or movable stage 202, on which a semiconductor 201 serving as a sample 201 is placed. The movable stage 202 is, for example, movable in the illustrated X and Y directions, which are radial and horizontal directions, but is not limited thereto. Alternatively, it may be movable in the vertical Z direction, or may be rotated or tilted in any of the axial directions. Although not shown, the main body 101A and the controller 102 also include a drive circuit, etc., for driving and controlling the movable stage 202.

[0087] The main body 101A includes an electron gun 203 , a detector 204 , a detector 205 , a condenser lens 206 , an objective lens 207 , an aligner 208 , an ExB filter 209 , a deflector 210 , and the like.

[0088] The electron gun 203 generates a charged particle beam b1 that irradiates the sample 201. The condenser lens 206 and the objective lens 207 converge the charged particle beam b1 on the surface of the sample 201. The aligner 208 is configured to generate an electric field for aligning the charged particle beam b1 with the objective lens 207. The ExB filter 209 (ExB: Electromagnetic Field Orthogonality) is a filter for capturing secondary electrons emitted from the sample 201 and directed to the detector 204. The deflector 210 is a device for scanning the charged particle beam b1 across the surface of the sample 201.

[0089] Detector 204 is a secondary electron detector (in other words, a first detector) that primarily detects secondary electrons (SE) as particles generated from sample 201 and outputs a detection signal a1. Detector 205 is a backscattered electron detector (BSE, also called reflected electrons) that primarily detects backscattered electrons (BSE, also called reflected electrons) as particles generated from sample 201 (in other words, a second detector) and outputs a detection signal a2.

[0090] The controller 102 receives and inputs detection signals a1 from the detector 204 and a2 from the detector 205, performs analog and digital conversion on these signals, and generates digital images. The generated images become the sample image and the measured image. Specifically, in this embodiment, the controller 102 is configured to generate an SE image based on the signal a1 from the detector 204, and a BSE image based on the signal a2 from the detector 205. The SE image is an image primarily obtained from secondary electrons, while the BSE image is an image primarily obtained from backscattered electrons. These SE and BSE images are stored in association.

[0091] In this embodiment, the SEM 101 is configured to have two detection systems, in other words, two channels, of the detector 204 and the detector 205 in the main body 101A, and can generate two images. However, the present invention is not limited thereto, and any microscope having one or more channels may be used.

[0092] The controller 102 can be a computer system equipped with a processor, memory, a communication interface, etc., or a system or device implemented using dedicated circuitry. The controller 102 temporarily stores image and other data a3 generated based on the detection signal in a memory resource. The controller 102 transmits the image and other data a3 to, for example, a host computer 104 via a communication interface. The host computer 104 receives, inputs, and acquires the image and other data a3 from the controller 102 and stores it in its own memory resources. Furthermore, the memory resources used by a computer such as the host computer 104 can also exist as external storage resources (e.g., a database server) on the network 110.

[0093] Charged particle beam devices / shooting devices such as SEM101 sometimes have multiple channels or systems for detection and shooting. In this example, SEM101 has at least two detection channels for SE detection and BSE detection. That is, as described above, it is possible to generate two images, SE images and BSE images, based on the two detection signals a1 and a2. In this embodiment, the controller 102 generates an SE image and a BSE image based on the detection signals a1 and a2, and further, synthesizes the SE image and the BSE image into one image (set as a composite image) by accumulating them. Furthermore, the SE image, the BSE image, or the composite image, in particular, the composite image in this embodiment, can be used to generate the region segmentation image described later. In addition, as the measured image, an SE image, a BSE image, or a composite image can be used, and selection can be made.

[0094] [Computer System]

[0095] Figure 2 Express as Figure 1 A configuration example of a computer system including a host computer 104 and other computers. Figure 2 The computer system is primarily composed of a computer 1000. Computer 1000 includes a processor 1001, a memory 1002, a communication interface device 1003, an input / output interface device 1004, and the like, which are interconnected via a bus or other architecture. Input device 1005 and output device 1006 may also be externally connected to input / output interface device 1004. Examples of input device 1005 include a keyboard, a mouse, and a microphone. Examples of output device 1006 include a display, a printer, and a speaker. Figure 2 The input device 1005 and the output device 1006 are Figure 1 The input and output device 105 corresponds to the input and output device 105.

[0096] Memory 1002 stores data and information such as a control program 1002A, setting information 1002B, image data D1, measurement process data D2, measurement result data D3, and screen data D4. Control program 1002A is a computer program that causes processor 1001 to execute processes. Setting information 1002B contains setting information for control program 1002A and user setting information. Image data D1 is data representing captured images acquired from SEM 101. Measurement process data D2 is data representing the measurement process configured for overlay measurement. Measurement process data D2 includes setting information such as measurement area generation rules D5. Furthermore, the measurement process may include information such as the imaging conditions used by SEM 101 for imaging, or these may be configured as separate process / setting information. Measurement result data D3 is data representing the results of overlay measurement and includes information such as the overlay offset D6.

[0097] The screen data D4 is data for providing a GUI screen (for example, a Web page) or the like to the user U1 .

[0098] The processor 1001 is configured to include, for example, a CPU, ROM, and RAM. The processor 1001 executes processing according to a control program 1002A stored in the memory 1002. Thus, the predetermined functions and processing units of the measurement system 100 are implemented as execution modules. These execution modules are implemented when the computer system is activated.

[0099] The communication interface device 1003 is connected via Figure 1 The network 110 performs communication processing with the controller 102 of the SEM 101, other computers, or external devices such as the input and output device 105 (client terminal).

[0100] Computer systems are not limited to Figure 2 The structural example may be any system as long as it is composed of one or more processors and one or more memories.

[0101] The functions and processing of measurement system 100 are described in detail below. In Embodiment 1, the overlay offset is described as a predetermined parameter value used as the measurement target value. The overlay offset measurement process, described later, includes measuring the shape, size, and center coordinates (or center of gravity) of the target pattern using methods such as edge detection. The characteristic concepts and functions of this disclosure are not limited to overlay offset measurement and can be similarly applied to measuring the shape, size, and center coordinates of such patterns.

[0102] [Function block structure]

[0103] Figure 3 This is a functional block diagram of the process performed in the measurement system 100 in the first embodiment. This process generally includes a measurement process preparation stage 301 and a measurement execution stage 302. Figure 3 The functional block structure diagram is understood as a process flow chart.

[0104] The measurement process preparation stage 301 is a stage for preparing and setting the measurement process related to the target sample 201. For example, Figure 1 The host computer 104 in the measurement system 100 performs the measurement process creation phase 301. The information of the created measurement process is stored in a storage resource within the measurement system 100, such as a memory of the host computer 104.

[0105] The measurement execution phase 302 is a phase in which the measurement of the overlap offset and the like related to the target sample 201 is performed according to the measurement process. For example, Figure 1The host computer 104 in the measurement execution phase 302 performs the measurement processing. As a result of this measurement processing, measurement result data including the measured overlay offset etc. is obtained. The measurement result data is stored in a storage resource within the measurement system 100, such as the memory of the host computer 104.

[0106] Furthermore, various data and information such as measurement processes, measurement results, and system setting information are stored in any storage resource in the measurement system 100. For example, these data and information are not limited to the memory of the host computer 104, but may also be stored in a database server (not shown) or an external storage medium (e.g., a memory card).

[0107] The measurement process creation stage 301 includes a learning unit 304 and a measurement region generation rule creation unit 307 as main functional blocks. The measurement execution stage 302 includes a region division unit 310, a measurement region generation unit 312, and an overlap measurement unit 314 as main functional blocks. Each of these functional blocks can be implemented by processing on any computer, such as by program processing on a processor, but is not limited to this and can also be implemented by dedicated circuits.

[0108] For example, the learning unit 304 and the measurement region generation rule creation unit 306 are implemented by the main processor 103 of the host computer 104 reading corresponding programs from a memory (not shown) and executing processing in accordance with the programs. Furthermore, the region segmentation unit 306, the measurement region generation unit 312, and the overlap measurement unit 314 are implemented by the main processor 103 of the host computer 104 reading corresponding programs from a memory (not shown) and executing processing in accordance with the programs. Alternatively, if processing is performed by a sub-computer, the functional blocks may be implemented by the sub-processor 106 of the first sub-computer 107 or the sub-processor 108 of the second sub-computer 109 executing processing in accordance with the programs.

[0109] [Measurement process production stage]

[0110] right Figure 3 First, the overview of the measurement process production stage 301 will be described. Hereinafter, unless otherwise specified, the subject of each process is a computer or a processor.

[0111] The computer inputs a sample image 303. Sample image 303 is an image used as a sample for learning. A learning unit 304 inputs sample image 303, performs learning processing, and obtains a learning model 305 as the learning result. The computer obtains a segmented image 306 of sample image 303 as the output of learning model 305. In other words, learning model 305 performs machine learning on the correspondence between sample image 303 as input and segmented image 306 as output.

[0112] The measurement region generation rule generation unit 307 inputs the region segmentation image 306 of the sample image 303, processes it, and generates as output a measurement region generation rule 308. The measurement region generation rule 308 is a rule for generating a measurement region corresponding to the region segmentation image 306. In this embodiment, the user U1 sets the measurement region generation rule 308 by viewing the region segmentation image 306 on the screen.

[0113] Sample image 303 is a sample image collected in advance, obtained by capturing a pattern superimposed on the measurement object; in other words, it is a learning image or learning data. Learning model 305 is a machine learning model that calculates a segmented image based on an image (e.g., sample image 303) and is composed of parameters such as coefficients in the machine learning model. Learning unit 304 calculates and outputs the segmented image based on the structure and shading information of the pattern within sample image 303. In other words, the parameters of learning model 305 are adjusted and updated through learning and training. Segmented image 306 of sample image 303 is an image obtained by inputting sample image 303 used for calculation by learning unit 304 or another sample image 303 not used for calculation into learning model 305.

[0114] The learning unit 304 also provides the user U1 with a user interface for learning. Figure 1 The GUI screen is displayed on the display of the input / output device 105. Figure 3 In the figure, this user interface is shown as a "GUI" block. User U1 appropriately inputs necessary information through the GUI and confirms the output information. In this embodiment, the learning unit 304, the measurement area generation rule creation unit 307, and the overlap measurement unit 314 have corresponding GUIs (described later) that enable input and output based on user U1.

[0115] The measurement area generation rule creation unit 307 creates and sets a measurement area generation rule 308 for generating a measurement area to be placed on the measurement target pattern based on a pair of the sample image 303 and the region segmentation image 306 of the sample image 303. Furthermore, the measurement area generation rule creation unit 307 provides the user U1 with a GUI for creating and setting the measurement area generation rule 308.

[0116] [Measurement execution phase]

[0117] Next, the outline of the measurement execution phase 302 will be described. A computer, such as the host computer 104, inputs the measured image 309 obtained from the SEM 101. The measured image 309 is obtained from the SEM 101 during the overlay measurement. Figure 1 The SEM 101 and particularly the controller 102 are supplied with an image of the object for measuring the overlap offset and the like.

[0118] The segmentation unit 310 refers to the learning model 305 and infers a segmented image 311 from the measured image 309. The segmentation unit 310 inputs the measured image 309 to the learning model 305 that has been sufficiently trained, and obtains the segmented image 311 of the measured image 309 as an inference result of the learning model 305.

[0119] The measurement region generating unit 313 receives the region segmentation image 311 of the measured image 309 as input and generates measurement regions from the region segmentation image 311 based on the reference to the measurement region generating rule 308. The measurement region generating unit 313 then generates a measurement region arrangement image 313, which is an image in which the generated measurement regions are arranged on the measurement target pattern of the measured image 309.

[0120] Overlap measurement unit 314 receives measurement area configuration image 313 as input, measures the overlap offset and other information based on the information within the measurement area in measurement area configuration image 313, and generates measurement result data 315 including the overlap offset and other information. Overlap measurement unit 314 stores measurement result data 315 in a memory resource and outputs it to a GUI screen.

[0121] The processes of the region segmentation unit 310, the measurement region generation unit 313, and the overlap measurement unit 314 can be basically performed automatically. Furthermore, the user U1 can specify details of the overlap measurement method in the overlap measurement unit 314 through a GUI screen. The user U1 can also select and specify, for example, size, center point coordinates, overlap offset, and the like as measurement target parameter values ​​through the screen.

[0122] [Example of overlapping measurement objects]

[0123] Next, an example of overlapping the pattern structure of the sample 201 to be measured will be described. The amount of overlap deviation, one of the parameters to be measured, refers to the amount of overlap deviation between the upper and lower patterns in the three-dimensional pattern structure of the sample 201.

[0124] Figure 4A This is an XY plan view of the upper surface, in other words, the surface, of the semiconductor wafer, which is the sample 201 to be measured and overlapped. Figure 4B Is to express Figure 4A An XZ cross-sectional view of a corresponding cross-sectional structure design example. Here, the X-axis and Y-axis are two orthogonal axes that define the top surface of the semiconductor wafer, and the Z-axis is an axis in the height and depth directions that are orthogonal to the X-axis and Y-axis. The X-axis is sometimes referred to as the horizontal axis, and the Y-axis is sometimes referred to as the vertical axis. Figure 4A and Figure 4B Indicates the design structure.

[0125] Figure 4APlanar region 401 is a portion of the wafer's upper surface and, in this example, schematically includes eight patterns as shown. The patterns here represent semiconductor structures and, in this example, are represented by circles in the XY plane. Specific examples of these circular patterns include Hall elements. Figure 4A The cross-sectional view along the X-axis along the AB line is Figure 4B The cross-sectional structure 402 in FIG. 4 , in other words, the cross-sectional area 402 .

[0126] exist Figure 4A In the planar region 401, the upper pattern 403a, the upper pattern 403b, the upper pattern 403c and the upper pattern 403d are formed on the surface of the wafer. Figure 4B The upper layer 411, in other words, the pattern of the first layer. Regarding these upper layer patterns, the areas that appear to be circular in the XY plane. These upper layer patterns have the same prescribed size, such as a prescribed diameter.

[0127] The lower layer patterns 404a, 404b, 404c and 404d are lower layers formed on the surface of the wafer. Figure 4B The lower layer 412, in other words, the second layer, is located in the lower layer 412. In the XY plane, these lower layer patterns overlap with the upper layer patterns and are partially obscured, resulting in a crescent-shaped area (a shape where a portion of the arc of a circle is missing). These lower layer patterns have the same predetermined size, such as a predetermined diameter, which in this example is smaller than the diameter of the upper layer patterns.

[0128] exist Figure 4B In the embodiment, upper layer 411 has upper layer patterns 403a and 403b formed on the boundary line 423 with lower layer 412, in other words, on the upper surface of lower layer 412. These upper layer patterns are covered, for example, by insulating film region 421. Upper surface 431 is the upper surface (XY plane) of region 421 of upper layer 411. Lower layer patterns 404a and 404b are formed in lower layer 412 below boundary line 423. These lower layer patterns are covered, for example, by insulating film region 422.

[0129] exist Figure 4AIn the figure, the dotted rectangle region 405 represents unit cell structure 405 and is an example of a pattern repeatedly formed in the X and Y directions. The wafer structure includes areas not shown, and regions 405 forming such unit cell structures 405 are repeatedly arranged in a limited number in each of the X and Y directions. For example, within unit cell structure 405a, an upper pattern 403a and a lower pattern 404a are arranged at a certain position in the Y-axis direction, on the AB line, and an upper pattern 403c and a lower pattern 404c are arranged at another position in the Y-axis direction, on the CD line. Similarly, within unit cell structure 405b, an upper pattern 403b and a lower pattern 404b are arranged at a certain position in the Y-axis direction, and an upper pattern 403d and a lower pattern 404d are arranged at another position in the Y-axis direction.

[0130] As shown in the figure, a pair of upper and lower patterns arranged adjacent to each other when viewed in the XY plane is sometimes referred to as a pattern pair or set. This pattern pair is a pattern structure that overlaps in the Z-axis direction, and is a pattern structure in which the upper pattern overlaps on the upper side of the lower pattern. Figure 4A In the example of , there are four pattern pairs in the region 401, and these pattern pairs are arranged so as to be separated from each other in the XY directions.

[0131] exist Figure 4A In FIG. 1 , lower pattern 404a and lower pattern 404c are designed so that their centers of gravity in the X direction coincide, as indicated by a vertical dashed line. Lower pattern 404b and lower pattern 404d are designed so that their centers of gravity in the X direction coincide, as indicated by a vertical dashed line. Furthermore, upper pattern 403a and lower pattern 404a are designed so that their centers of gravity in the Y direction coincide, as indicated by line AB. Upper pattern 403c and lower pattern 404c are designed so that their centers of gravity in the Y direction coincide, as indicated by line CD. The center of gravity is, for example, the center point of a circle.

[0132] Such unit cell structure 405 is repeated in the X direction. Therefore, in the planar region 401, for example, the Y coordinates of the centers of gravity of upper pattern 403a, lower pattern 404a, upper pattern 403b, and lower pattern 404b are aligned. The Y coordinates of the centers of gravity of upper pattern 404c, lower pattern 404d, upper pattern 404d, and lower pattern 403d are aligned.

[0133] Furthermore, in the unit cell structure 405, for example, the unit cell structure 405a, the difference dxa between the center X coordinates of the upper pattern 403a and the center X coordinates of the lower pattern 404a, and the difference dxc between the center X coordinates of the upper pattern 403c and the center X coordinates of the lower pattern 404c are designed to have opposite signs and the same absolute value. The difference dxc and the difference dxd within the unit cell structure 405a are also similar, and are designed to satisfy dxa = dxc and dxc = dxd.

[0134] For example, in the unit cell structure 405a, the upper pattern 403a overlapping the lower pattern 404a on the upper side is offset to the left (-X) in the X direction, and the upper pattern 403c overlapping the lower pattern 404c on the upper side is offset to the right (+X) in the X direction. These offsets are correct offsets and displacements in design. Figure 4A In the XY top view, only a portion of the circular top surface of the lower pattern 404a is visible (the moon shape with the arc portion missing on the left), and only a portion of the circular top surface of the lower pattern 404c is visible (the moon shape with the arc portion missing on the right). The same is true for the unit cell structure 405b.

[0135] Figure 5 is relative to Figure 4 ( Figure 4A 、 Figure 4B ) is an XY top view of an example in which an overlap offset exists in a design example. The overlap offset is caused by, for example, some factors in the semiconductor manufacturing process, such as process variations that are greater than a certain level, and is caused by the design value ( Figure 4A ) of undesirable differences and fluctuations.

[0136] exist Figure 5 In the middle, the plane area 501 and Figure 4A The plane area 401 corresponds to the Figure 5 As an example of overlapping offset, the situation where the positions of the patterns as a whole are offset in the XY direction is shown. The upper pattern 503a, the upper pattern 503b, the upper pattern 503c and the upper pattern 503d are formed on the surface of the wafer, the pattern of the upper layer 411, and Figure 4A The lower pattern 504a, the lower pattern 504b, the lower pattern 504c and the lower pattern 504d are the lower layers formed on the surface of the wafer, the patterns of the lower layer 412, and the lower patterns 504a, 504b, 504c and 504d are the lower layers formed on the surface of the wafer, and the patterns of the lower layer 412 are the same as those of the lower pattern 403a to 403d. Figure 4A Corresponding to the lower layer patterns 404a~404d in.

[0137] For convenience of explanation, regions 511 to 514 represent approximate regions of each pattern pair of an upper layer pattern and a lower layer pattern that are adjacent and overlapped vertically on the XY plane.

[0138] exist Figure 5 In the example of FIG. 5 , the lower layer patterns 504a to 504d are formed relative to the upper layer patterns 503a to 503d. Figure 4A Such a design position is offset upward (+Y) in the Y direction. Figure 5 In the planar area 501, there is an overlapping offset in the Y direction. The overlapping offset in the Y direction is measured between adjacent upper and lower patterns. In this example, as shown in the figure, the overlapping offset in the Y direction is set to the difference between the Y coordinate of the center of gravity of the upper pattern (represented as the center point of the circle) and the Y coordinate of the center of gravity of the lower pattern (represented as the center point of the circle). For example, in a certain area 511, the overlapping offset 505a in the Y direction is the value obtained by subtracting the center Y coordinate of the upper pattern 503a from the center Y coordinate of the lower pattern 504a. Similarly, the overlapping offsets 505b, 505c, and 505d in the Y direction in each area are the values ​​obtained by subtracting the center Y coordinate of the upper patterns 503b, 504c, and 504d from the center Y coordinate of the lower patterns 504b, 504c, and 504d, respectively. In Figure 5 In FIG, the center point (X coordinate, Y coordinate) of the circle corresponding to the center of gravity is represented by a black dot.

[0139] In addition, Figure 5 The center of gravity position shown in the figure is conceptual; it is not necessarily accurate or easy to calculate from the image during measurement. In conventional art examples, it is sometimes impossible to accurately determine the center of gravity based on such an upper and lower pattern pair, particularly the moon shape of the lower pattern. This is because pattern boundaries may not be cleaned or process variations are large. Therefore, using the measurement area setting methods used in conventional art examples, it is sometimes impossible to accurately detect the area edges and calculate the center of gravity with high precision.

[0140] In this example, in addition to this, the lower layer patterns 504a to 504d are formed from the upper layer patterns 503a to 503d. Figure 4A The design position of is offset to the right (+X) in the X direction. That is, there is also an overlap offset in the X direction. However, in this embodiment, the overlap offset in the X direction is not the object of measurement. The following is the measurement of the overlap offset in the Y direction, that is, Figure 5 The measurement system 100 has a function of being able to perform high-precision and easy measurement even when targeting such overlapping offsets.

[0141] and, Figure 6 Is relative to Figure 5 The XY top view shows the situation where the size and position of each pattern are offset due to process variation. Figure 6 A second example is shown in which there is an overlap offset. Process variations may include not only variations not intended by the manufacturer but also changes in parameters of a desired manufacturing process.

[0142] exist Figure 6 In the planar region 601 (including regions 611 to 614), upper pattern 603a, upper pattern 603b, upper pattern 603c, and upper pattern 603d are patterns formed on the surface of the chip, and lower pattern 604a, lower pattern 604b, lower pattern 604c, and lower pattern 604d are patterns formed at the lower layer of the chip.

[0143] Lower layer pattern 604a has a smaller size (in this example, the diameter of the circle) than lower layer pattern 504a. Lower layer pattern 604b is formed at a position offset to the left (-X) in the X direction relative to lower layer pattern 504b. Upper layer pattern 603c has a larger size (in this example, the diameter of the circle) than upper layer pattern 503c. Upper layer pattern 604d is formed at a position offset to the right (+X) in the X direction relative to upper layer pattern 504d.

[0144] As in this example, the process variation of individual patterns that deviate from the measured image, such as the patterns in regions 611 to 614, may be relatively large relative to the pattern size. For example, any deviation or variation generated during the wafer manufacturing process is reflected in the pattern structure of the actual wafer, such as Figure 6 As in the example, the size and position of each pattern fluctuate. This results in particularly overlapping offsets 605a to 605d. Such fluctuations in the actual object also manifest as fluctuations in the pattern in the measured image. In this case, the accuracy of measuring overlapping offsets, etc., may also decrease.

[0145] [Sample images and SEM images]

[0146] Figure 3 Sample image 303 is an image captured before overlay offset measurement is applied. It is an image of wafer 201, the target of overlay measurement, or a wafer image close to the captured image of wafer 201. Sample image 303 may be captured by SEM 101 performing overlay measurement, or may be captured and collected using another microscope, such as an SEM, that has image quality similar to that of SEM 101.

[0147] Figure 7 Figure 7A 、 Figure 7B ) was taken with SEM101 Figure 6 The structure of the wafer 201 as in the example, in other words, an example of an SEM image of the actual object. Figure 7A The image 701 is the SE image 701 obtained based on the signal a1 of the detector 204. Figure 7B Image 702 is a BSE image 702 obtained based on signal a2 from detector 205. Sample image 303 is composed of one or more pairs of SE image 701 and BSE image 702. Each image may be a collection of multiple images obtained by repeatedly capturing the same pattern area multiple times, or may be an image obtained by accumulating multiple images.

[0148] In addition, the actual image is multi-grayscale, but Figure 7A etc. as a schematic diagram of several smear pattern areas. Figure 7A In SE image 701, for example, region 711 corresponds to the upper pattern. The circular boundary (e.g., edge region 714) is clearly and brightly displayed, and therefore, is depicted as a white ring. For example, region 712 corresponds to the lower pattern. It is located in the lower layer in the Z direction and, therefore, appears darker. Furthermore, background region 713 appears the darkest.

[0149] Furthermore, when generating the region segmentation image described later, edge region 714 may be segmented into regions different from the region of the upper pattern based on brightness differences, for example. For example, the upper pattern may be segmented into two regions: a circular region and an annular edge region 714 located on the outer periphery of the circle. Even in such a case, the functions described in the embodiment can be applied by selecting an appropriate region type when setting measurement region generation rule 308. Furthermore, by assigning identifiers to each of these two or more regions as region types representing the upper pattern, the functions described in the embodiment can be applied in the same manner.

[0150] exist Figure 7B Similarly, in the BSE image 702, for example, region 721 corresponds to the upper pattern and is relatively clearly and brightly displayed. Figure 7A The brightness of area 711 is high. For example, area 722 is the area corresponding to the lower pattern and is located in the lower layer in the Z direction. Therefore, it appears darker. The brightness of area 722 is higher than Figure 7A The brightness of area 712 is high. This is because BSE is more likely to capture the underlying structure than SE. Furthermore, background area 723 appears darkest. Generally, SE contains more information about the sample surface, while BSE contains more information about the interior of the sample surface.

[0151] exist Figure 7A In the SE image 701, for example, when focusing on the boundary region 715 between the region 711 and the region 712, a relatively bright edge region 714 exists, and therefore the boundary between the region 711 and the region 712 is easily understood. Figure 7B In the BSE image 702, for example, when focusing on the boundary region 725 between the region 721 and the region 722, the brightness difference between the region 721 and the region 722 is relatively small, so the boundary between the region 721 and the region 722 is relatively bright. Figure 7A In other words, the boundary between area 721 and area 722 is unclear. This can sometimes cause the boundary between the upper and lower patterns to become unclear in the image. In this case, for example, it's difficult to clearly detect the area of ​​the lower pattern, making it difficult to accurately calculate the center of gravity, and thus making it more difficult to measure the overlap offset.

[0152] In addition, Figure 7A For example, the brightness difference between the region 712 of the lower pattern and the region 713 of the background is relatively small. Therefore, the brightness difference between the region 712 and the region 713 is relatively small. Figure 7B In this case, it is more difficult to measure the overlap offset.

[0153] As in the example above, in the measured image of wafer 201, pattern boundaries may become unclear. In particular, the boundary between overlapping lower and upper patterns may become unclear. In this case, due to the influence of process variations, the measurement accuracy of overlap offsets and other parameters may decrease. In contrast, in the embodiments, this issue can be addressed by setting appropriate measurement area generation rules.

[0154] [Learning Department]

[0155] Figure 8 It is used for Figure 3 The flowchart will be used to explain the process of the learning unit 304 generating the learning model 305. Here, the host computer 104 performs the learning process, but the learning process may be performed by a sub-computer as described above.

[0156] In step S801, the learning unit 304 obtains a sample image 303, for example Figure 7A Such SE image 701 and Figure 7BIn step S802, the learning unit 304 performs known processing such as contrast adjustment of the SE image 701 and the BSE image 702, and integration in the (R, G, B) direction, to obtain a preferred pair of the SE image 701 and the BSE image 702. In the first embodiment, in the learning for generating the region segmentation image 306, a composite image of the SE image and the BSE image is specifically used as the designated image type in the input sample image 303.

[0157] In step S803, after initially setting the initial values ​​of the parameters, i.e., the coefficients, of the learning model 305, the learning unit 304 calculates the learning model 305 in such a way that the regional segmentation image can be inferred based on the structural and brightness features within the image when the sample image 303 is input, thereby generating the learning model 305.

[0158] Thus, a calculation method of a learning model for generating a region segmentation image for an image by a user through teaching an image with no labels in each pixel, in other words, a method for generating a region segmentation image based on unsupervised learning, can be realized, for example, by the technology of the following known literature.

[0159] Ji, Xu, Joao F. Henriques, and Andrea Vedaldi. "Invariant information clustering for unsupervised image classification and segmentation." Proceedings of the IEEE / CVF International Conference on Computer Vision. 2019.

[0160] In this embodiment, the technology of the above-mentioned known literature is applied to generate the region segmentation image 306. In this case, for example, the input data to the learning model 305 based on the convolutional neural network (CNN) is a synthetic image based on the SEM 101, and the output data has a region type for each pixel.

[0161] Then, in step S804, it is determined whether the learning model 305 is obtained, which generates a desired segmented image capable of segmenting the regions corresponding to the patterns of the overlapping measurement objects in the sample image 303. In other words, it is determined whether the learning model 305 is sufficiently learned.

[0162] If the region corresponding to the measurement target pattern in the sample image 303 is segmented in the generated region segmentation image (Yes), in step S805, the learning unit 305 stores the generated learning model 305 in a storage resource (not shown), for example Figure 1 in the memory of the host computer 104.

[0163] Finally, in step S806, the learning unit 304 saves the region segmentation image 306 of the sample image 303 obtained in the above calculation process in a storage resource (not shown), for example Figure 1 in the memory of the host computer 104.

[0164] In addition, as described above, in this embodiment, unsupervised learning is applied to the learning unit 304. However, the present invention is not limited thereto, and supervised learning may also be applied to the learning unit 304. The learning unit 304 may also perform learning using supervised data, which is an image in which labels of pattern structures within semiconductor devices in sample images (not shown) are assigned to each pixel of the image.

[0165] In addition, the learning unit 304 applies a method of separating the patterns of each layer by setting a threshold value for separating each distribution from the histogram of the gradation value of the sample image, rather than a machine learning model.

[0166] [Region segmentation image]

[0167] Figure 9( Figure 9A 、 Figure 9B ) means about Figure 3 The region segmentation image 306 of the sample image 303 output by the learning unit 305 is Figure 8 Two examples of the region segmentation image 306 of the sample image 303 output in the learning model generation of step S803.

[0168] Figure 9A Indicates that the image was taken using SEM101 Figure 4A The region segmented image 306 of the sample image 303 obtained by the structure of is the region segmented image 306A. Figure 9B Represents Figure 6The region segmentation image 306 of the sample image (SE image 701 and BSE image 702) of FIG7 corresponding to the example is the region segmentation image 306B. In this example, the region segmentation image 306 has three region types as the types of regions contained in the image. In the legend 906, three region types are shown. The first region type (value = 0) is the region corresponding to the upper pattern, and is represented by a stripe region. The second region type (value = 1) is the region corresponding to the lower pattern, and is illustrated by a gray region. The third region type (value = 2) is the region corresponding to the background, and is illustrated by a white region. Here, the contour positions of these region types may also be inconsistent with the contour positions of the corresponding patterns.

[0169] Figure 9A The region segmentation image 306A includes region elements 903a, 903b, 903c, 903d, region elements 904a, 904b, 904c, 904d, and region element 905a as region elements formed by region segmentation. Figure 9B The region segmentation image 306B includes region elements 906a, 906b, 906c, 906d, region elements 907a, 907b, 907c, 907d, and region element 905b as region elements formed by region segmentation.

[0170] In these two examples, for example, region elements 903a, 903b, 903c, and 903d in region segmentation image 306A are region types with the same identifier, e.g., belonging to the first region type, corresponding to the upper pattern region. Furthermore, region elements 904a, 904b, 904c, and 904d are region types with the same identifier, e.g., belonging to the second region type, corresponding to the lower pattern region. Furthermore, region element 905a is a region type different from the first and second region types, e.g., belonging to the third region type, corresponding to the background.

[0171] [Measurement Area Generation Rule Creation Department]

[0172] Figure 10 It is used for Figure 3 The following is a flowchart illustrating the process by which the measurement area generation rule creation unit 307 creates the measurement area generation rule 308. While the details of this process will be described later, an overview of this process will be provided here. In this embodiment, user U1 performs the configuration operation. Therefore, in the following description, the main subject of the actions in each step, or in other words, the trigger, is primarily user U1, but the main subject of the corresponding processing is a computer or processor, such as the host computer 104. The host computer 104 executes the corresponding processing (e.g., the configuration processing) based on the operation input by user U1.

[0173] In step S1001, user U1 Figure 3 The input / output terminal 105 connected to the host computer 104 selects a region segmentation image 306 (described later) of the sample image 303 stored in a storage resource (not shown), such as a memory of the host computer 104. Figure 13A At this time, the selected area segmentation image 306 and the sample image 303 corresponding to the area segmentation image 306 are read. In addition, at this time, the measurement target pattern can also be specified.

[0174] Next, in step S1002, the user U1 selects the type of image to be placed in the measurement area on the screen (described later). Figure 13A In this embodiment, the image types that can be selected here are Figure 7A Such SE images and Figure 7B For example, when considering the configuration of the measurement area for the lower pattern, Figure 7A SE images 701 and Figure 7B In the BSE images 702 , the influence on the measurement is different, and the user U1 can select the BSE image 702 , for example.

[0175] Next, in step S1003, the user U1 designates, on the region segmentation image, a region to be the target of setting a measurement region generation rule, namely, a rule target region (hereinafter referred to as a rule target region). Figure 13A At this time, if the measurement area generation rule 308 is common within the sample image, the rule target area as the designated range may be one. If it is a partial area within the sample image, the rule target area is set for each partial area.

[0176] In addition, in this embodiment, in step S1003, the user U1 freely sets the rule subject area containing the area element. The method of setting the rule subject area can be listed as follows. Figure 4AWhen viewed as a rectangle (e.g., ), for example, unit cell 405a includes area elements of four patterns: area elements 403a, 404a, 403c, and 404c. Adjacent patterns include a first pair of area elements 403a and 404a and a second pair of adjacent area elements 403c and 404c. Furthermore, the upper and lower patterns overlap differently in the first and second pairs. For the first pair, the lower pattern overlaps with the upper pattern, offset to the right in the X direction. In contrast, for the second pair, the lower pattern overlaps with the upper pattern, offset to the left in the X direction. Considering the application of measurement area generation rule 308, it is preferable to apply different measurement area generation rules to pairs with different overlapping patterns. Therefore, user U1 sets a first rule target area that includes the first pair and a second rule target area that includes the second pair. In this specific example, the area of ​​unit cell 405a can be divided into two upper and lower areas.

[0177] This embodiment describes a scenario where user U1 sets the rule-applicable area via a screen. However, the present invention is not limited to this. A computer can also automatically set the rule-applicable area. For example, the computer can use image analysis or learning to detect regional elements other than the background within the image that deviate from the background, such as the pair of upper and lower patterns described above. The computer then sets the rule-applicable area for each deviating regional element.

[0178] In step S1004, the user U1 selects the boundary of the measurement area to be set as the setting of the measurement area creation rule set in the rule target area (described later). Figure 13B 1305).

[0179] Figure 22 This is an explanatory diagram of the measurement area, and is a schematic diagram of an XY top view when the measurement area is generated and arranged in the area element corresponding to the lower layer pattern in the area segmentation image. Figure 22 , as a set of pattern pairs, shows an area element E1 of the first area type corresponding to the upper pattern, an area element E2 of the second area type corresponding to the lower pattern, and a measurement area 2201 for area element E2 of the lower pattern. In a screen described later, user U1 specifies the boundaries of measurement area 2201 for measuring the Y-direction overlap offset for area element E2 of the lower pattern. In this example, measurement area 2201 is a rectangle, with the four sides of the rectangle designated as boundaries: top, bottom, left, and right. The method for specifying these boundaries is arbitrary, and various known GUIs can be applied. Furthermore, measurement area 2201 is not limited to a rectangle; it can also be an ellipse, for example.

[0180] When the measurement area is set to a rectangle, for example, the positions of the top, bottom, left, and right sides of the rectangle are set sequentially. For example, user U1 can operate cursor 2210 (e.g., a mouse pointer) on the screen to specify the positions of the top, bottom, left, and right sides of the rectangle. Alternatively, the top left and bottom right points of the rectangle can be specified by clicking, etc. Alternatively, these positions can be input using coordinate values, etc. Alternatively, the left and right centers and top and bottom centers of the rectangle can be specified, and the difference from these centers to the left and right, and top and bottom can be specified. The width of the rectangle in the X direction and the width in the Y direction can also be specified.

[0181] In step S1005, user U1 selects the region type (e.g., lower layer pattern) to be used in setting the measurement region generation rule 308 (1305B, described later). Furthermore, user U1 sets the coordinate information (in other words, the reference position) of the region element corresponding to the region type (1305C, described later). Furthermore, user U1 sets a correction value for the measurement region boundary based on the coordinate information (in other words, the reference position) of the region element (1305C, described later). In other words, this correction value is used to determine the measurement region boundary based on a relative relationship or difference with respect to the reference position.

[0182] Examples of the correction value include a correction value for determining the boundary of the maximum or minimum value in the X direction of the measurement area using the maximum or minimum value in the X direction of the selected area element (eg, lower layer pattern) as the reference position coordinate.

[0183] In addition, the reference position and correction value may be set as follows: The reference position may be set to the centroid coordinates of the selected area element (e.g., the lower pattern), that is, the approximate centroid of the area segmentation image, and the center coordinates of the measurement area may be determined based on the centroid coordinates using the desired correction value.

[0184] exist Figure 22 In the example shown in FIG2 , the measurement area 2201 is set to have an X-coordinate X2 as the right boundary, an X-coordinate X4 as the left boundary, a Y-coordinate Y1 as the top boundary, and a Y-coordinate Y2 as the bottom boundary. First, the Y-coordinates Y1 and Y2 are set to the same size as the Y-direction width of the area element E2 containing the lower pattern, as shown in the figure. The X-coordinates X1 and X2 are specified using the reference position and correction value for the area element E2 of the lower pattern.

[0185] For the X coordinate of the right boundary, for example, the X-direction maximum value (in other words, the right end) of area element E2 of the lower pattern is designated as the reference position. The right end point corresponding to the X-direction maximum value is point PX1 in the illustration, and has an X-coordinate of X1. Alternatively, for example, -2 pixels (two pixels to the left in the X direction) is designated as the correction value 2202 from the reference position. In this case, the X coordinate of the right boundary becomes X2, which is a position shifted two pixels to the left from the X-coordinate X1 of point PX1, which is the X-direction maximum value, by the correction value 2202.

[0186] Furthermore, for the X coordinate of the left boundary, for example, the maximum value in the X direction of area element E1 of the upper pattern (point PX2) is specified as the reference position. Point PX2 has an X coordinate of X3. Furthermore, for example, "+2 pixels" (two pixels to the right in the X direction) is specified as the correction value 2203 from the reference position. In this case, the X coordinate of the left boundary becomes X4, which is the position shifted two pixels to the right from point PX2, the reference position, by the correction value 2203.

[0187] In this way, the measurement area can be determined based on the relative positional relationship with respect to the regional elements in the segmented image. User U1 can set the relative positional relationship while viewing the segmented image on the screen. In other words, the measurement area generation rule 308 is a rule for generating the measurement area based on the relative positional relationship with respect to the regional elements in the segmented image. The regional element serving as the basis for the relative relationship can be the measurement target pattern itself (e.g., the lower layer pattern) or another adjacent pattern (e.g., the upper layer pattern).

[0188] about Figure 22 In the example setting of measurement area 2201, the reference position is the right end of area element E2 of the lower pattern for the right boundary of measurement area 2201, and the right end of area element E1 of the upper pattern for the left boundary. Furthermore, within measurement area 2201, at each position in the X direction, the area of ​​the lower pattern is included, while the area of ​​the upper pattern and its boundary with the upper pattern are excluded. Furthermore, the Y-direction contour also excludes areas that serve only as background. Thus, using measurement area 2201 facilitates detection of the lower pattern's edges and addresses variations in pattern size and position due to process variations.

[0189] Not limited to Figure 22 As another setting example, the following is also possible: The area element corresponding to the pattern taking the reference position may be only the lower layer pattern or only the upper layer pattern.

[0190] In another setting example, when the area element of the pattern taking the reference position is only the lower layer pattern, the following is done. The X coordinate of the right boundary of the measurement area is Figure 22 Similarly, the right end of the lower-layer pattern's area element E2 is set as the reference position. The X coordinate of the left boundary of the measurement area is used as the reference position, and the X-direction maximum value of the lower-layer pattern's area element E2 (point PX1) is set. For example, a correction value of "-13 pixels" is specified. This results in the left boundary being, for example, at X coordinate X4. In this case, both the left and right sides of the measurement area are determined relative to the lower-layer pattern.

[0191] In another setting example, when the area element of the pattern taking the reference position is only the upper layer pattern, the following is done. The X coordinate of the left edge of the measurement area is Figure 22 Similarly, the right end of the upper-layer pattern's area element E1 is set as the reference position. The X coordinate of the right boundary of the measurement area is used as the reference position, and the maximum value (point PX2) in the X direction of the upper-layer pattern's area element E1 is set. A correction value, such as "+13 pixels," is specified. Thus, the right boundary becomes, for example, X coordinate X2. In this case, both the left and right sides of the measurement area are determined relative to the upper-layer pattern.

[0192] Each of the above-described setting examples can also be selected based on whether the boundary between patterns in the image is clear or unclear. For example, if the boundary between the upper and lower patterns in the image is unclear, and the boundary between the lower pattern and the background area is clear, the right and left sides of the measurement area can be determined using the right end of area element E2 of the lower pattern as a reference position, without using the boundary with the upper pattern as a reference. Conversely, if the boundary between the lower pattern and the background area in the image is unclear, the right and left sides of the measurement area can be determined using the right end of area element E1 of the upper pattern as a reference position, without using this boundary as a reference.

[0193] As in the above-described example, the measurement region creation rule 308 can be set so as to avoid portions that may be unclear on the image, thereby enabling more appropriate measurement.

[0194] Return to Figure 10 In step S1006, it is determined whether the settings of all measurement areas related to the rule target area are completed. If the settings of all measurement areas are completed (yes), in step S1007, the user U1 sets the measurement area determination rule (described later). Figure 13BColumn 1306). Here, the measurement area determination rule refers to a rule that does not set a measurement area if it is determined that measurement of the measurement target pattern cannot be performed based on the coordinate information of the area elements of the area segmentation image. In other words, the measurement area determination rule does not create a measurement area if a specified condition, such as measurement failure, is met when attempting to generate a measurement area according to the measurement area generation rule 308. The specified condition may be, for example, a condition that the width of the measurement area is less than a specified value.

[0195] In step S1008, it is determined whether all measurement area creation rules 308 related to the measurement target patterns of all rule target areas have been set. If the setting is complete (yes), in step S1009, the computer saves the set measurement area creation rules 308 as part of the measurement process.

[0196] [Rule subject area]

[0197] Figure 11A and Figure 11B Indicates Figure 9A and Figure 9B The region segmentation image 306 ( 306A, 306B) of the sample image 303 shown is used as an example of setting the rule target region 1101 and the rule target region 1102 . Figure 11A and Figure 11B Indicates that the region segmentation image 306 is obtained by the method ( Figure 10 In step S1003), the rule object area is set. Figure 11A represents the rule target areas 1101 (1101a, 1101b, 1101c, 1101d) set in the area segmentation image 306A, Figure 11B The rule target areas 1102 ( 1102 a , 1102 b , 1102 c , and 1102 d ) set in the area segmentation image 306B are shown.

[0198] In this case, in this example, the regions indicated by the dashed boxes containing the pattern pairs in the two upper sets (Set1 and Set2) of segmented image 306 are the first type of rule-bound regions. Furthermore, the regions indicated by the dashed boxes containing the pattern pairs in the two lower sets (Set3 and Set4) of segmented image 306 are the second type of rule-bound regions. As described above, the overlapping positional relationships between adjacent upper and lower patterns in the upper and lower pattern pairs differ in the left-right X-direction, resulting in different types of rule-bound regions. To identify the types of rule-bound regions, the symbols RA and RB are also assigned.

[0199] That is, in Figure 11A In the region segmentation image 306A, the rule subject area 1101 includes the rule subject areas 1101a and 1101b as the first type of rule subject areas RA, and includes the rule subject areas 1101c and 1101d as the second type of rule subject areas RB. Figure 11B In the region segmentation image 306B, the rule subject area 1102 includes rule subject areas 1102 a and 1102 b as the first type of rule subject areas RA, and includes rule subject areas 1102 c and 1102 d as the second type of rule subject areas RB.

[0200] For example, the rule target area 1101a is a rectangular area containing Set1 as a pattern pair, and includes the area element 903a of the upper pattern (first area type) and the area element 904a of the lower pattern (second area type). The measurement area generation rule 308 is set for each type of rule target area. The rule target area RA is associated with the first measurement area generation rule, and the rule target area RB is associated with the second measurement area generation rule. For example, in the first type of rule target area RA, the measurement area generation rule 308 is set. Figure 22 The measurement region generation rule 308 for generating the measurement region is similarly set in the second type of rule target region RB for generating the measurement region for the lower pattern which is arranged on the left side of the upper pattern in the X direction. Figure 22 The X direction in the figure can be considered in the opposite way.

[0201] [Example of measurement area generation rule]

[0202] exist Figure 12 In FIG. 1 , an example of measurement area creation rule 308 that has been set and saved is shown in table format. The upper table shows measurement area creation rule 1201 as example 1, and the lower table shows measurement area creation rule 1202 as example 2. The measurement area creation rule 1201 is for Figure 11A 、 Figure 11B The measurement area generation rule 308 is set for the first type of rule target area RA and especially for the application to the lower layer pattern. Figure 11A 、 Figure 11B The measurement region generation rule 308 is set based on the second type of rule target region RB and, in particular, the application to the lower layer pattern.

[0203] The table of the measurement area creation rule 308 includes, for example, “measurement area boundary”, “area type”, “coordinate information of area element”, and “correction value” as items.

[0204] In addition, Figure 12 In the setting of such measurement area creation rules, although not shown in the figure, the correspondence between each rule target area and the measurement area creation rule to be applied is set. An ID is assigned to each rule target area and each measurement area creation rule, and their correspondence is maintained in data management.

[0205] The measurement area creation rule 1201 in Example 1 consists of five rule elements shown in rows #1 to #5. When the measurement area is rectangular, these rule elements define the following as the boundaries (or center) of the measurement area: 1. Minimum X coordinate (in other words, the left edge), 2. Maximum X coordinate (in other words, the right edge), 3. Center Y coordinate, 4. Minimum Y coordinate (in other words, the bottom edge), and 5. Maximum Y coordinate (in other words, the top edge).

[0206] Figure 22 The concept of generating the measurement area 2301 based on the rule elements #1 to #5 in the measurement area generating rule 1201 of Example 1 is shown. Figure 23 The concept of generating the measurement area 2302 based on the rule elements #1 to #5 in the measurement area generating rule 1202 of Example 2 is shown.

[0207] In the rule element #1 of Example 1, the minimum X coordinate (in other words, the position on the left) is determined as the boundary of the measurement area. Regarding the area element that serves as the basis for determining this boundary, the area type is 0 (first area type). As the coordinate information of this area element, the reference position is the maximum X coordinate (in other words, the position at the right end), and the correction value (in other words, the relative relationship) is +3 pixels in the X direction from this reference position. Based on this reference position and correction value, the minimum X coordinate is determined as the boundary of the measurement area. In the rule element #2, the maximum X coordinate (in other words, the position on the right) is determined as the boundary of the measurement area. The area type is 1 (second area type). The reference position is the maximum X coordinate of the area element. The correction value is -5 pixels in the X direction.

[0208] In the rule element #3, the center Y coordinate is determined as the measurement area boundary. The area type is 1 (second area type). The reference position is the Y coordinate of the center of gravity of the area element. The correction value is 0 pixels in the Y direction. In the rule element #4, the minimum Y coordinate (bottom edge) is determined as the measurement area boundary. The area type is None, the reference position is None, and the correction value is -20 pixels from the center Y coordinate. In the rule element #5, the maximum Y coordinate (top edge) is determined as the measurement area boundary. The area type is None, the reference position is None, and the correction value is +20 pixels from the center Y coordinate.

[0209] exist Figure 22In this example, the centroid Y coordinate Y0 of the area element E2 of the second area type is calculated. This centroid Y coordinate Y0 is the center Y coordinate of the measurement area 2201. The maximum Y coordinate Y1 of the measurement area 2201 is determined at a position +20 pixels above the centroid Y coordinate Y0 (upper edge B1), and the minimum Y coordinate Y2 of the measurement area 2201 is determined at a position -20 pixels below the centroid Y coordinate Y0 (lower edge B2). Furthermore, the maximum X coordinate X3 (right end position) of the area element E1 of the first area type is calculated. The minimum X coordinate X4 (left side B4) of the measurement area 2201 is determined at a position +3 pixels (correction value 2203) in the X direction from this reference position. Furthermore, the maximum X coordinate X1 (right end position) of the area element E2 of the second area type is calculated. The maximum X coordinate X2 (right side B3) of the measurement area 2201 is determined at a position -5 pixels (correction value 2202) in the X direction from this reference position.

[0210] The measurement area generation rule 1202 of Example 2 is similarly composed of the five rule elements shown in rows #1 to #5. The differences from the rule in Example 1 are as follows. For rule element #1, the area type is set to 1 (second area type), the reference position is the minimum X coordinate (the left position), and the correction value is +5 pixels in the X direction. For rule element #2, the area type is set to 0 (first area type), the reference position is the minimum X coordinate (the left position), and the correction value is -3 pixels in the X direction.

[0211] exist Figure 23 The center Y coordinate, maximum Y coordinate, and minimum Y coordinate of the measurement area 2301 are Figure 22 The minimum X coordinate X5 (left end position) of the area element E2 of the second area type is calculated, and the minimum X coordinate X6 (left side) of the measurement area 2301 is determined at a position +5 pixels (correction value 2302) in the X direction from its reference position. Furthermore, the minimum X coordinate X7 (left end position) of the area element E1 of the first area type is calculated, and the maximum X coordinate X8 (right side) of the measurement area 2301 is determined at a position -3 pixels (correction value 2303) in the X direction from its reference position.

[0212] Supplement the coordinates of the center of gravity. Figure 12In the example, as shown in #3, the Y coordinate of the center of gravity of the regional element in the regional segmentation image is also used as the reference position. It should be noted that this is different from the Y coordinate of the center of gravity of the measurement object pattern in the measured image to be measured using the measurement area. The Y coordinate of the center of gravity of the regional element is an approximate center of gravity coordinate that is different from the accurate center of gravity coordinate of the lower pattern of the measured image. Although the Y coordinate of the center of gravity of the regional element is not accurate, it indicates the tendency of the lower pattern to shift upward or downward in the Y direction. Therefore, if the Y coordinate of the center of gravity of the regional element is used as a reference, correction values ​​are taken at the top and bottom, and the boundaries (top and bottom) of the measurement area are set, the upper and lower edges of the lower pattern can be included in the measurement area with a high probability.

[0213] In other words, in Embodiment 1, by creating measurement area generation rules based on relative relationships based on information about regional elements in the segmented image, it is possible to generate an appropriate measurement area even if the location of the measurement target pattern in the measured image is uncertain. Conventional technology cannot accurately configure a measurement area without knowing the location of the measurement target pattern in the measured image. In contrast, in Embodiment 1, the coordinates of the ends and center of gravity of regional elements such as the underlying pattern can be determined from the segmented image. Using these approximate coordinates as a reference, an appropriate measurement area is generated using calibration values ​​and placed in the measured image. Based on this measurement area, the edges and center of gravity of the pattern can be calculated with higher accuracy than those in the segmented image. Consequently, measurement accuracy can be improved.

[0214] In the above Figure 22 as well as Figure 23 In the example setting of measurement area creation rule 308, the reference position is selected based on ease of image detection. As described above, the image type used to determine the reference position can be selected from, for example, SE images and BSE images. This does not require a single image type; multiple images can be used. In other words, measurement area creation rule 308 can be set so that the reference position is detected based on an image type that is easily detectable.

[0215] Figure 24 The above Figure 12 The example of the measurement area generation rule 308 is applied to the existence of Figure 11B This is an explanatory diagram of a case where one set of region segmented images with such positional shifts is used. In this case, a measurement region 2401 as shown in the figure is obtained. Figure 24 The lower pattern area element E2 and Figure 22Although there are changes in size and position compared to the area element E2 of the lower pattern, measurement area 2401, like measurement area 2201, can capture the area that constitutes the appropriate outline of the lower pattern (excluding the boundary with the upper pattern and the portion of the background area). Therefore, based on measurement area 2401, the edge of the area of ​​the lower pattern can be detected and the Y coordinate of the center of gravity can be calculated.

[0216] [GUI for creating measurement area generation rules]

[0217] Figure 13A as well as Figure 13B Indicates based on Figure 3 An example of a GUI screen for creating and setting the measurement area creation rule 308 of the measurement area creation rule creation unit 307 is shown. Figure 13A The first part of the "Measurement area creation rule setting" screen 1301 is shown. Figure 13B Indicates the second part of the screen 1301. Figure 13A In the screen 1301, there are a "Selection of measurement target pattern" column 1302, a "Selection of area segmentation image and image type for arranging measurement area" column 1303, and a "Selection of set rule target area" column 1304. Figure 13B In the example, screen 1301 includes a "Measurement area creation rule setting" column 1305, a "Measurement area permission determination rule setting" column 1306, an "Apply measurement area creation rule" column 1307, and a "Rule name setting" column 1308. The white arrow image is an example of an operating cursor 1309, which the user U1 can manipulate with a mouse or the like.

[0218] Column 1302 is in Figure 10 Column 1303 is a GUI for selecting and setting the measurement target pattern to which the set measurement area creation rule 308 is applied in step S1001. Column 1303 is a GUI for selecting a representative area segmentation image 306 from among the multiple area segmentation images 306 for setting in step S1002, and for selecting the type of image to which the measurement area is to be configured. Column 1304 is a GUI for specifying the area to which the measurement area creation rule 308 is applied, i.e., the rule target area, in step S1003. Column 1305 is a GUI for setting the measurement area creation rule 308 in steps S1004 and S1005. In this case, column 1305 selects an area element and area type within the area segmentation image 306, and sets the coordinates of the upper, lower, left, and right boundaries of the measurement area using correction values ​​representing the relative positional relationship of the coordinates (reference positions) of the area elements.

[0219] Column 1306 is a GUI for setting the measurement region permission determination rule in step S1007. Column 1307 is a GUI for displaying the measurement region allocation results when the set measurement region creation rule 308 is applied to the measured image of a different sample image. Column 1308 is a GUI for naming the set measurement region creation rule 308 and storing it in the measurement recipe storage unit.

[0220] exist Figure 13A In the "Selection of Measurement Target Pattern" field 1302, for example, a measurement target pattern can be selected from a list box. In other words, a pattern to be used for generating a measurement area based on a measurement area generation rule can be selected. In this example, the options include an upper layer pattern and a lower layer pattern.

[0221] In the "Selection of Region Segmentation Image and Image Type for Measurement Region Configuration" column 1303, in the "Region Segmentation Image" column 1303A on the left, you can select the region segmentation image 306 and display and confirm the contents of that region segmentation image. In the "Image Type" column 1303B on the right, you can select the image type for measurement region configuration and display and confirm the image corresponding to that image type. In this example, the options are SE, BSE, and Mix. Mix is ​​a composite image of the SE and BSE images.

[0222] In the "Selection of Set Rule Target Area" column 1304, in the left column 1304A, you can select "Coordinate Specification" and "Manual Specification" as the method for setting the rule target area. You can also specify the area size, starting coordinates, spacing, and number of repetitions. In the right column 1304B, the set rule target area in the segmented image 306 is displayed and confirmed as a result. For example, the set rule target areas r1 and r2 are displayed within dotted lines.

[0223] In addition, Figure 13A The above mentioned Figure 4A And so on, the area of ​​the unit cell is selected.

[0224] exist Figure 13BIn the "Measurement Area Generation Rule Settings" column 1305, a rule target area selected in column 1304 is displayed in column 1305A on the left, and an enlarged display area can be specified within the rule target area. For example, enlarged display area 1305a can be specified. Furthermore, in the lower portion of column 1305A on the left (column 1305D), you can select from the X and Y directions, and you can select from "left and right settings" and "center settings." Selecting the X and Y directions means selecting a rule related to the X direction or the Y direction. "Left and right settings" and "center settings" mean selecting a rule related to the left or right, up or down, or center position in the selected direction (X or Y). For example, if "left and right settings" is selected, you can further select from "right" and "left."

[0225] In the center column 1305B, the enlarged display area specified in column 1305A is displayed in an enlarged manner. Within this enlarged display area, the area type can be specified. In this example, the values ​​can be selected from 0 (first area type), 1 (second area type), and 2 (third area type). In this example, the second area type, which is the lower layer pattern, is selected.

[0226] Similarly, in the right column 1305C, in the enlarged display of the image of the image type selected in the column 1303B, the user U1 can set the position of the boundary of the measurement area for the area element of the area type specified in the column 1305B. Figure 22 Such a setting. Figure 26 An enlarged view of column 1305C is shown in the figure. Below column 1305C (column 1305D), the coordinate information (reference position) for the area element specified in column 1305B can be selected from "maximum," "minimum," and "center of gravity." Furthermore, a correction value corresponding to the reference position can be specified using, for example, the number of pixels. Alternatively, as shown in the figure, user U1 can operate a cursor, for example, by moving line 1305c (X-direction position) left and right, thereby confirming and specifying the correction value and the position of the corresponding boundary. Line 1305c is a GUI corresponding to the specification in column 1305D below. In this example, to determine the right boundary of the measurement area in the X direction, the right boundary is specified to be located at a position -10 pixels (px) from the maximum X coordinate (right end) of the area element of area type 1 (second area type) selected in column 1305B.

[0227] When the Apply button 1305D is pressed, the setting of the measurement area creation rule 308 in the column 1305 is saved and applied. Figure 12 In this way, the setting data of the measurement area creation rule 308 is saved.

[0228] The "Measurement Area Permission Determination Rule Settings" field 1306 allows you to set measurement area permission determination rules. For example, you can set a condition such as "the horizontal width of the measurement area is less than 2 pixels (px)" by making the number of pixels of width and parameters such as "less than" and "less than or equal to" variable.

[0229] In the "Apply Measurement Region Generation Rule" field 1307, any segmented region image can be specified. Image 1307B shows the application of the measurement region generation rule 308 specified in field 1305 to the sample image / measured image corresponding to the specified segmented region image. Aa, Ab, and Ac are examples of generated and configured measurement regions. By observing this image 1307B, the user U1 can confirm the appropriateness of the measurement region generation rule 308. For example, the user U1 can confirm the appropriateness of the rule by specifying another image with a similar pattern structure to the measurement target pattern, attempting to apply the measurement region generation rule, and observing the results.

[0230] In the “setting of rule name” column 1308 , the measurement area creation rule 308 set on the above-mentioned screen 1301 can be named and saved.

[0231] As described above, according to the present embodiment, the user U1 can designate and input items required for user input in order to set the measurement area creation rule 308 through the GUI screen, and can set an appropriate measurement area creation rule 308 .

[0232] exist Figure 26 shows an enlarged view of the measurement region boundary setting field 1305C. In this example, a line 1305c for specifying the correction value in the X direction is displayed superimposed on the enlarged display area 2601 of the designated BSE image. In this field 1305C, a region element 2610 corresponding to the underlying pattern in the corresponding region segmentation image is displayed superimposed on the enlarged display area 2601 of the designated BSE image, for example, by a dotted line. For example, user U1 observes the BSE image and, using the right end position 2602 of the region element 2610 of the underlying pattern as a reference position, moves line 1305c left and right in the X direction, thereby specifying an X coordinate 2603 based on the correction value 2604 as the measurement region boundary (e.g., the right side). Furthermore, in this example, as shown in the figure, there is an offset between the outline of region element 2610 and the edge of the underlying pattern region in the actual BSE image. The user can also confirm this offset through the GUI.

[0233] As in the above example, the edge of the pattern structure in the actual image is not necessarily consistent with the outline of the region element in the region segmentation image. In such a case, in the first embodiment, the GUI screen, such as the one described above, is used. Figure 22 In this way, the measurement area generation rule 308 can be set to address this issue, ensuring that the measurement area is appropriate. Specifically, when observing the Y-axis contour at various X-axis positions, the measurement area boundaries can be specified by adjusting the correction value based on the region element's contour as a reference position, ensuring that only the background area and the lower pattern area are included in the Y-axis contour. Alternatively, the measurement area boundaries can be specified to exclude X-axis positions where the upper pattern, lower pattern, and background are difficult to distinguish due to unclear boundaries.

[0234] [Overlap Measurement]

[0235] Figure 14 yes Figure 3 Flowchart of the measurement of the overlap offset in the measurement execution phase 302. First, in step S1401, a computer such as the host computer 104 obtains the measured image 309. In this embodiment, Figure 7A Such SE images and Figure 7B Such a BSE image. In this case, the measured image 309 is assigned accompanying information such as a position ID. After acquiring the measured image 309, in step S1402, the region segmentation unit 310 generates a region segmentation image 311 based on the measured image 309 and the learning model 305. Then, first, in step S1403, the computer arranges the rule-targeted region in the region segmentation image 311. If the measurement region generation rule 308 is common within the measured image 309, step S1403 can be skipped.

[0236] Next, in step S1404, the measurement region generation unit 312 of the computer determines the size and position of the measurement region for measuring each measurement target pattern within the measured image 309 based on the information in the region segmentation image 311 and the measurement region generation rule 308, and places the generated measurement region within the measured image 309. In step S1405, the measurement region generation unit 312 determines whether measurement regions have been determined for all measurement target patterns within the measured image 309. If not, the process returns to step S1404 and repeats the same process.

[0237] After determining the measurement area, in step S1406, the computer's overlay measurement unit 314 detects the edge of the pattern to be measured using the portion of the measurement area within the measured image 309. In step S1407, the overlay measurement unit 314 uses the edge coordinates detected in step S1406 to calculate the centroid coordinates of the pattern to be measured. In step S1407, the overlay measurement unit 314 uses these centroid coordinates to calculate the overlay offset associated with the pattern to be measured. This is not limited to the overlay offset; measurements of pattern dimensions, etc., are also possible. Finally, in step S1408, it is determined whether measurement of all the patterns to be measured in the measured image 309 has been completed. If not, the process returns to step S1402 and repeats the same process.

[0238] [Region segmentation image, rule target area, and measurement area]

[0239] Figure 15 As Figure 14 The generation process in the process of FIG is an explanatory diagram showing a specific example of the generation of the region segmentation image 311 in step S1402, the configuration of the rule target area in step S1403, and the generation and configuration of the measurement area in step S1404. Figure 15 , as an example, a case where a measurement area for measuring a lower layer pattern is determined is shown.

[0240] First, the computer obtains the SE image 309A and the BSE image 309B of the measured image 309 and performs the same image processing as that used to generate the learning model 305. The segmentation unit 310 generates a segmented image 311 of the measured image 309 by referring to the learning model 305 stored in a storage unit (not shown).

[0241] about Figure 15 In the segmented region image 311 of the measured image 309 in the example, the regions corresponding to the upper pattern are region elements 1503a, 1503b, 1503c, and 1503d. These region elements belong to the same first region type, the same region type as the region elements corresponding to the upper pattern in the segmented region image 306 of the sample image, and have a common identifier. Meanwhile, the regions corresponding to the lower pattern are region elements 1504a, 1504b, 1504c, and 1504d. These region elements belong to the same second region type, the same region type as the region elements corresponding to the lower pattern in the segmented region image 306 of the sample image, and have a common identifier. Furthermore, the background region element 1500 is also of the same third region type as the region element corresponding to the background in the segmented region image 306 of the sample image.

[0242] After generating the segmented region image 311, the measurement region generating unit 312 arranges the rule target region (dashed-line frame in the figure) in the segmented region image 311. The measurement region generating unit 312 refers to the measurement region generating rule 308 stored in a storage unit (not shown) for each region element of the rule target region in the segmented region image 311, thereby generating measurement regions 1505a, 1505b, and 1505c for each underlying pattern as the measurement region 1505.

[0243] Here, in Figure 15 In the example shown in FIG. 1 , measurement area generation unit 312 further refers to the measurement area availability determination rule within measurement area generation rule 308 and does not generate a measurement area for lower layer pattern 1504d with respect to area elements 1503d and 1504d in the lower right set. In other words, measurement area generation unit 312 determines that the upper and lower edges of the lower layer pattern are unmeasurable based on the coordinate information of area elements 1503d and 1504d, as per the measurement area availability determination rule. Therefore, measurement area generation unit 312 does not generate this measurement area. In portions of the pattern structure where this measurement area is not allocated, overlapping measurements of patterns where accurate edge detection is not possible can be eliminated.

[0244] exist Figure 15 In the example, the BSE image 309B is selected as the image type for measuring the lower pattern. The measurement area generating unit 312 generates a measurement area configuration image 313 by applying and configuring these measurement areas 1505 (1505a, 1505b, 1505c) to the BSE image 309B corresponding to the image type for measuring the lower pattern. In addition, the function described in the embodiment performs processing such as generating the region segmentation image 311 and the measurement area based on the input source image, that is, the same measured image 309. Therefore, when Figure 15 When such a measurement region creation result 1509 is arranged on the measured image 309 to generate the measurement region arrangement image 313 , no unnecessary offset occurs.

[0245] In addition, as in the example of the set on the lower right, when the measurement area is not generated and arranged according to the measurement area determination rule, the measurement system 100 can also output to the user U1 through the screen that the measurement area is not generated and arranged according to the measurement area determination rule for this part.

[0246] [Edge detection and calculation of center of gravity coordinates within the measurement area]

[0247] Figure 16A As Figure 14The generation process in the process of FIG is an explanatory diagram showing a specific example of edge detection within the measurement area in step S1406 and calculation of the center of gravity coordinates in step S1407 after the above measurement area is configured. Figure 16B This is a table showing an example of the processing result of the calculation of the overlap offset amount in step S1408.

[0248] exist Figure 16A The image 1601B on the left is equivalent to Figure 15 The image 1601A on the right shows measurement areas arranged in BSE image 309B, namely measurement area arrangement image 313. Image 1601A on the right shows measurement areas arranged in BSE image 309A, namely measurement area arrangement image 313. In measurement area arrangement image 1601A, measurement areas are arranged for the upper layer pattern, while in measurement area arrangement image 1601B, measurement areas are arranged for the lower layer pattern. Measurement area arrangement image 1601A includes measurement areas 1605a, 1605b, 1605c, and 1605d as measurement areas 1605.

[0249] The same configuration of the measurement area relative to the upper pattern can also be used Figure 15 Alternatively, if the configuration of the measurement area has no effect on the measurement accuracy of the overlay offset, the configuration of the measurement area relative to the upper pattern may also be used using the conventional method. In the example of the lower pattern, since the configuration of the measurement area has an effect on the measurement accuracy of the overlay offset, the method of this embodiment is used instead of the conventional method. Figure 15 The method shown.

[0250] Edge detection result 1602B shows an example of edge detection of a lower-layer pattern from its contour within measurement area 1505 in measurement area configuration image 1601B. Edge detection result 1602A shows an example of edge detection of an upper-layer pattern from its contour within measurement area 1605 in measurement area configuration image 1601A. These results correspond to the edge detection results of step S1406. In this example, the measurement area is for measurement in the Y direction, so the contour here is the brightness contour in the Y direction. Edges a1 and a2 represent the Y-direction edge positions of a lower-layer pattern. Edges b1 and b2 represent the Y-direction edge positions of an upper-layer pattern.

[0251] For example, within measurement area 1505a of a region element of a lower-layer pattern 1611, the contour at a certain X position (indicated by a dashed line) is referenced. The contour at this X position includes the upper and lower edges of the lower-layer pattern in the Y direction between the background areas. This X position can also be set as the center of the X-direction width of measurement area 1505a. Alternatively, within measurement area 1505a, the edge positions can be statistically calculated based on the contours at each X position. The same applies to measurement area 1605a of a region element of an upper-layer pattern 1612.

[0252] The barycentric coordinate display image 1603B is an example of the result of the barycentric Y coordinate of the lower pattern calculated based on the edge detection result 1602B of the lower pattern. The barycentric coordinate display image 1603A is an example of the result of the barycentric Y coordinate of the upper pattern calculated based on the edge detection result 1602A of the upper pattern. These results are equivalent to the results of the barycentric coordinate calculation in step S1407. The barycentric Y coordinate and the representative X coordinate of each X coordinate used for measurement are shown with an × mark, in other words, the barycentric position based on the X coordinate of the center. For example, based on the Y coordinate Ya1 of edge a1 and the Y coordinate Ya2 of edge a2, the barycentric Y coordinate of the regional element of a certain lower pattern is obtained as BY1=(Ya1+Ya2) / 2 (referred to as BY1). Similarly, based on the Y coordinate Yb1 of edge b1 and the Y coordinate Yb2 of edge b2, the barycentric Y coordinate of the regional element of a certain upper pattern is obtained as BY2=(Yb1+Yb2) / 2 (referred to as BY2).

[0253] As shown in the barycenter coordinate display image 1603B, the coordinates (barycenter Y coordinate BY1 and barycenter X coordinate BX1) of the barycenter (× mark) of a certain lower-layer pattern are calculated. Similarly, the barycenters (× marks) of other lower-layer patterns are calculated for each measurement area. Similarly, as shown in the barycenter coordinate display image 1603A, the coordinates (barycenter Y coordinate BY2 and barycenter X coordinate BX2) of the barycenter (× mark) of a certain upper-layer pattern are calculated. Similarly, the barycenters (× marks) of other upper-layer patterns are calculated for each measurement area.

[0254] [Overlap measurement results]

[0255] Figure 16B Table 1600 is an example of overlay measurement result data, showing the data of the center Y coordinate of the lower pattern (set as YL) and the center Y coordinate of the upper pattern (set as YU) for each set of adjacent patterns, and the calculated value of the overlay offset (set as OD) calculated based on each YL and YU using the following formula 1. This is equivalent to the result of step S1408. Here, the data is shown using the ( Figure 16A ) as the center coordinates.

[0256] OD=YL-YU…Equation 1

[0257] A set of adjacent patterns is a pair of overlapping upper and lower patterns, such as Figure 11A The measurement object pattern of each regular object area corresponds to each other, and an ID is added for data management. In this example, the set ID is set to Set1 to Set4. For example, for Set1, the center Y coordinate of the lower pattern is Y coordinate 1 (YLa), the center Y coordinate of the upper pattern is Y coordinate 2 (YUa), and the overlap offset calculation value ODa = YLa-YUa. Y coordinate 1 (YLa) can use the above-mentioned BY1, and Y coordinate 2 (YUa) can use the above-mentioned BY2. For Set4, since the center Y coordinate of the lower pattern is not measured according to the configuration without measurement area, OD is also not measured.

[0258] Alternatively, the overlap offset value (OD) for each set can be calculated, for example, by using a statistical method such as arithmetic averaging to calculate the overlap offset for the entire measured image. This method is not limited to arithmetic averaging; alternative methods include multiplicative averaging and median value calculation. Furthermore, deviations such as standard deviation can be calculated and output.

[0259] [Overlap offset]

[0260] Figure 16C As an illustration of the overlap offset, a schematic enlarged view of a measurement area configuration image 1601B (313) is shown, showing an image in which measurement areas are configured within a BSE image containing lower-layer patterns. The background area is shown in white. Furthermore, the center of gravity positions (GLYa, GLYb, GLYc) of the center of gravity Y coordinates (YLa, YLb, YLc) of the measurement area 1505 containing each lower-layer pattern are superimposed and illustrated with an x-mark. Furthermore, the center of gravity positions (GUa, GUb, GUc, GUd) of the center of gravity Y coordinates (YUa, YUb, YUc, YUd) of the upper-layer pattern side are superimposed and illustrated with an x-mark.

[0261] Regarding a set Set1, Figure 22 Similarly, the measurement area MYa for the lower pattern PLa is set by taking the correction value for the reference position. The measurement area MYa includes the lower pattern PLa at each position in the X direction and does not include the portion that is only the background area. In addition, the measurement area MYa does not include the boundary with the upper pattern. Therefore, it is possible to use the measurement area MYa as described above ( Figure 16A ) The edge and center Y coordinate of the lower pattern are appropriately calculated. In this way, by setting an appropriate measurement area to remove unnecessary areas that may reduce accuracy, the center Y coordinate can be calculated with high accuracy.

[0262] For the upper and lower patterns in Set 1, the calculated Y-direction overlap offset (ODa) is ODa = (YLa - YUa). Similarly, for Set 2, the calculated Y-direction overlap offset (ODb) is ODb = (YLb - YUb). For Set 3, the calculated Y-direction overlap offset (ODc) is ODc = (YLc - YUc). The overlap offset is represented by, for example, pixel distance within the image, coordinate differential values, and the like.

[0263] When the overlap shift amount of the entire measured image is calculated from the overlap shift amount calculation values ​​(ODa, ODb, ODc) of each set, for example, by arithmetic averaging, the overlap shift amount is obtained as (ODa+ODb+ODc) / 3.

[0264] Figure 16D This figure shows an example of an image in which a measurement area is configured for measuring the X-direction overlap offset and calculating the X-direction center of gravity of a lower pattern. Furthermore, a measurement area generation rule 308 suitable for calculating the X-direction center of gravity is set, and a measurement area is generated based on this rule, such as shown in the figure. For example, regarding Set 1, a measurement area MXa is configured for the lower pattern PLa. The center of gravity GLXa (including the X-coordinate XLa of the center of gravity) of the lower pattern PLa is calculated based on the measurement area MXa. Similarly, the center of gravity GUXa (including the X-coordinate XUa of the center of gravity) of the upper pattern PUa is calculated. The calculated value ODXa of the X-direction overlap offset between the lower pattern PLa and the upper pattern PUa is ODXa = (XLa - XUa).

[0265] As in the above example, for the same measurement target pattern, different measurement areas are set in the X and Y directions according to the measurement area creation rules 308, and different centroid coordinates (e.g., centroid GLYa and centroid GLXa) are calculated for each area. The centroid coordinates of these measurement areas are generated and calculated so that the overlap offset can be calculated with high accuracy.

[0266] [Variation: Center of gravity calculation]

[0267] As a modified example, as described above, the center of gravity of a certain measurement target pattern may be calculated by combining the X coordinate of the center of gravity calculated according to the measurement area generation rule in the X direction and the Y coordinate of the center of gravity calculated according to the measurement area generation rule in the Y direction.

[0268] Figure 16EThis is an explanatory diagram of the calculation of the center of gravity in this modified example. In the target image, there is a certain set of lower pattern PL and upper pattern PU. In this example, the upper pattern PU overlaps with the lower pattern PL to a large extent, and most of the lower pattern PL is obscured. The boundary of the lower pattern PL that is hidden and cannot be seen is represented by a dotted circle, and the center of the dotted circle is represented by a center point. With respect to the upper pattern PU, it is assumed that the center of gravity GPU is obtained. With respect to the lower pattern PL, the measurement area MY is set according to the measurement area generation rule in the Y direction. The center of gravity Y coordinate GLY is calculated based on the measurement area MY. In addition, with respect to the lower pattern PL, the measurement area MX is set according to the measurement area generation rule in the X direction. The center of gravity X coordinate GLX is calculated based on the measurement area MX. The overlapping offset can be calculated by the described method based on these two center of gravity coordinates (GLX, GLY).

[0269] Here, in a modified example, the center of gravity of the lower pattern PL is calculated by combining these two types of center-of-gravity coordinates (GLX, GLY). As one method, as shown in the figure, the center-of-gravity X coordinate GLX and the center-of-gravity Y coordinate GLY can also be set as the position coordinates (GLX, GLY) of the new center of gravity GPL1. As another method, the center-of-gravity X coordinate GLX and the center-of-gravity Y coordinate GLY can be connected by a straight line, and the midpoint of this line can be used as the new center of gravity GPL2. Then, the overlap offset can be calculated using the center of gravity GPL1 or GPL2 of the lower pattern PL and the center of gravity GPU of the upper pattern PU. For example, when using the center of gravity GPL1, the offsets dx and dy in each direction are obtained as shown by the arrows in the figure.

[0270] [Effects of Embodiment 1, etc.]

[0271] As described above, according to the first embodiment, overlay offsets and other measurements can be made stably, or more accurately, even when pattern boundaries are unclear or when process variations are large, resulting in significant variations in the size or position of the pattern within the measured image. According to the first embodiment, by generating a segmented image of the measured image, the positional relationship between individual measurement target patterns that deviate within the measured image, such as a collection of adjacent individual upper and lower layer patterns, can be determined. According to the first embodiment, by applying predetermined measurement region generation rules to the segmented image, appropriate measurement regions can be generated and positioned within the measured image based on the effects of process variations on the pattern. Furthermore, according to the first embodiment, the overlay offsets and other measurements of the measured pattern can be calculated with high accuracy using these measurement regions.

[0272] In the technology described in Patent Document 2, a standard measurement area is set for a standard image to be measured. Figure 17 It is an explanatory diagram of a comparative example. Figure 17An image 1701 of Comparative Example 1 is an example in which a standard measurement region 1705 (1705a, 1705b, 1705c, 1705d) is set for a standard measured image. The measurement region 1705b of the upper right set is enlarged and schematically illustrated.

[0273] Therefore, in the technology such as patent document 2, when the Figure 6 When the setting of the measurement region is applied to the measured image obtained by the structure in which the size and position of such a pattern vary greatly, the measurement region cannot appropriately capture only the lower layer pattern. Figure 17 The image 1702 of the comparative example 2 is for Figure 6 The measured image with such positional and dimensional deviations is shown in the case where the standard measurement area settings are applied, as in Comparative Example 1. Image 1702 includes measurement areas 1706 (1706a, 1706b, 1706c, and 1706d), which are identical in position and shape to measurement areas 1705 (1705a, 1705b, 1705c, and 1705d). Measurement area 1706b in the upper right corner is enlarged and schematically illustrated.

[0274] For example, in image 1702, measurement area 1706b includes a background portion (e.g., the position of the dashed line) in the X direction that lacks the underlying pattern. In this measurement area 1706b, the edges of the background portion lacking the underlying pattern are detected, making it impossible to perform proper edge detection. Measurement area 1706c is set to include both the background portion and the upper-layer pattern, and the edges of both are detected, making it impossible to perform proper edge detection. Furthermore, since measurement area 1706d lacks a portion in the X direction that contains only the underlying pattern, proper edge detection is impossible.

[0275] Therefore, in measurement area 1706b and the like, erroneous center-of-gravity coordinates, or in other words, low-precision center-of-gravity coordinates, are calculated based on erroneous edge detection results of the underlying pattern. Furthermore, in the measurement based on these center-of-gravity coordinates, erroneous overlay offsets, or in other words, low-precision overlay offsets, are calculated and output.

[0276] Furthermore, because the number of measurement areas in image 1702 is fixed, measurement area 1706d is configured even when edge detection, which is necessary for calculating the centroid coordinates of the lower-layer pattern, is not possible. Consequently, edge detection cannot be performed in measurement area 1706d, and the centroid coordinates of the lower-layer pattern cannot be calculated. Alternatively, edge detection of the upper-layer pattern may be erroneously performed in measurement area 1706d.

[0277] On the other hand, according to embodiment 1, Figure 15As shown in FIGURE 3, even when the size and position of the pattern within the measured image vary significantly, an appropriate measurement area for capturing the pattern to be measured can be configured according to measurement area generation rule 308. Furthermore, in Embodiment 1, by also using measurement area availability determination rules, it is possible to omit the configuration of a measurement area in situations where edge detection, necessary for calculating the centroid coordinates, is not possible. Therefore, in Embodiment 1, the use of an appropriate measurement area improves the measurement accuracy of, for example, the overlap offset.

[0278] Furthermore, even in the case of measured images with unclear edges, the pattern is binarized based on the inference results, and the centroid coordinates are calculated from the binarized image. Therefore, in the case of the technique described in Patent Document 1, if the entire boundary between the upper and lower patterns is not correctly inferred, it may be impossible to calculate the correct centroid coordinates.

[0279] On the other hand, in the first embodiment, the user U1 can set the measurement region generation rule 308 ( Figure 13A Therefore, it is possible to exclude unclear boundary areas such as the boundary between the upper pattern and the lower pattern, and the boundary between the lower pattern and the background from the measurement area, and configure appropriate measurement areas only in areas where the lower pattern exists (e.g. Figure 22 、 Figure 26 ). In addition, the edge of the pattern to be measured can be automatically measured based on the continuous light and dark outlines within the measurement area of ​​the measured image 309 ( Figure 16A Thus, according to the first embodiment, unlike the technique of Patent Document 1, even when the boundary of a pattern is not clear, the measurement accuracy of the overlap offset and the like can be improved by using an appropriate measurement area.

[0280] Furthermore, when the lower pattern is always obscured by the upper pattern, or when a large portion of the lower pattern is obscured by the upper pattern, it is generally difficult to accurately calculate the centroid coordinates of the lower pattern, including the obscured area, based on the binary image of the unobscured area of ​​the lower pattern. In contrast, in Embodiment 1, for example, by configuring an appropriate measurement area based on the measurement area generation rule 308 for accurately calculating the centroid coordinates in the Y direction, it is possible to accurately calculate the centroid coordinates in the Y direction of the lower pattern, including the obscured area, even in such situations.

[0281] As semiconductor devices continue to become increasingly miniaturized, process variations have a relatively greater impact on pattern dimensions, among other factors. Furthermore, with respect to three-dimensional structures, boundaries between upper and lower patterns are becoming less clear. Conventional methods of configuring standard measurement areas are unable to track actual pattern variations. In this method, the width, number, and positional relationships between measurement areas are fixed. In contrast, according to Embodiment 1, measurement area generation rules are used. This allows the measurement area serving as the measurement range to be optimized based on the actual position and size of the regional elements corresponding to the pattern structure of the measured image, allowing for the configuration of appropriate measurement areas according to individual pattern structures. This makes it easier to detect and measure pattern edges and centers of gravity, improving the accuracy of measurements of overlap offsets and other factors.

[0282] In addition, the edges of the region elements in the region segmentation image may not match the edges of the actual pattern structure due to unclearness or the like (for example, Figure 26 ), but their deviations tend to be somewhat biased. In contrast, in the first embodiment, by using correction values, measurement region generation rules are applied to arrange measurement regions in areas that are reliably estimated to be the pattern to be measured—in other words, to exclude unreliable areas. In the first embodiment, the measurement region generation rules are applied to the regional elements of the region segmentation image to determine the relationship between the contour of the pattern structure and the correction of the measurement region. This improves measurement accuracy.

[0283] [Dimensional Measurement]

[0284] Figure 25 As an example of measuring parameter values ​​other than the overlap offset, the following example illustrates measuring the dimensions of the underlying pattern. For a certain set, the dimensions of the underlying pattern PL are measured in the X and Y directions. The center of gravity 2501 (center of gravity X coordinate GX1, center of gravity Y coordinate GY1) is obtained from the measurement area MY for the Y direction, and the center of gravity 2502 (center of gravity X coordinate GX2, center of gravity Y coordinate GY2) is obtained from the measurement area MX for the X direction.

[0285] As described above, regarding the shape of the lower layer pattern PL, since the edge can be detected from each measurement area, the shape of the boundary with the background area (for example, an arc) can be measured.

[0286] Using edges detectable within the measurement area MY, for example, the maximum and minimum edge points p1 and p2, the Y-direction width 2503 of the portion of the underlying pattern PL that is not obscured and visible can be calculated. Similarly, using edges detectable within the measurement area MX, for example, the maximum and minimum edge points p3 and p4, the X-direction width 2504 of the portion of the underlying pattern PL that is not obscured and visible can be calculated.

[0287] The width 2504 in the X direction and the width 2503 in the Y direction are approximate widths of a portion of the lower layer pattern PL that is not shielded and is visible.

[0288] When it is desired to measure the coordinates of the center point of the lower pattern PL, for example, the Figure 16E Alternatively, you can use the shape and width measured above to calculate the center point coordinates.

[0289] As in the above-described example, by using the measurement region creation rule 308 and the measurement region, the size of the pattern and the like can also be measured.

[0290] [Variation of Embodiment 1]

[0291] In the first embodiment, the case of unsupervised learning has been described as described above, but the structural elements can also be changed as follows.

[0292] As a variation, supervised learning can also be applied. For example, Figure 3 The learning unit 304 uses, in addition to the sample images 303, supervised data in which the user U1 has annotated the sample images 303 with respect to regional elements to learn the learning model 305. This clarifies the correspondence between the region types and pattern types at the time of generating the region segmented image 306, making it easier to set the measurement region generation rule 308.

[0293] As a modified example, the generation of the region segmentation image 306 may be performed by applying a rule-based method instead of machine learning. Figure 3 The learning model 305 generates a segmented image 306 from the sample image 303 by processing the basic rules. User U1 sets the generated rules (segmented image generation rules) on the screen. During measurement, the segmentation unit 310 uses these rules to generate a segmented image 311 from the measured image 309. This modified example also achieves similar effects to those of the first embodiment.

[0294] <Implementation Method 2>

[0295] The second embodiment is described. The basic structure of the second embodiment is the same as that of the first embodiment. Hereinafter, the structural parts of the second embodiment that are different from those of the first embodiment will be described. The main structural differences in the second embodiment are as follows: Figure 3 Function block structure.

[0296] In the first embodiment, the user U1 confirms the region division image ( Figure 13AIn the second embodiment, a method for creating and setting measurement area generation rules 308 is described, in which a measurement system generates measurement areas by obtaining measurement area generation rules from a storage unit (not shown) without requiring user U1 to confirm the area segmentation image on the screen. In the second embodiment, the measurement area generation rules are pre-designed and prepared as a basic rule program.

[0297] Figure 18 The functional block structure in the second embodiment is shown. Figure 18 and Figure 3 The difference lies in the measurement area generation rule creation unit 1807 and the measurement area generation rule 1808. Figure 18 , there is no region segmentation image 306. In this configuration, unlike the generation of the learning model 305, the user U1 does not check the region segmentation image, but preliminarily creates and sets a measurement region generation rule 1808 that does not depend on the measurement object of the overlap offset amount.

[0298] Regarding the measurement area generation rule 1808, taking the case of generating a measurement area for measuring a lower layer pattern as an example, the following rules can be cited. This rule, for example, is a rule for identifying the area type corresponding to the upper layer pattern and the area type corresponding to the lower layer pattern, and arranging multiple measurement areas in the direction of the boundary of the area element of the lower layer pattern. ( Figure 19 Here, as a method for identifying the region type corresponding to the upper-layer pattern and the region type corresponding to the lower-layer pattern, for example, there is a method for performing identification by obtaining brightness information of the SE image or BSE image at the location of the region element of each region type and comparing them. This identification step can be skipped in the case where the region segmentation image is obtained through supervised learning and application of basic rules, which is a variation of Example 1. The measurement region generation rule 1808 includes, for example, a rule for excluding the region element of the upper-layer pattern from the measurement region. In addition, the width, height, etc. of each of the multiple measurement regions are set in the measurement region generation rule 1808.

[0299] Figure 19This example shows the generation of a measurement region configuration image 313 from measured image 309. Measured image 1900, for example, represents a portion of a certain set in BSE image 309B. Based on measured image 1900, region segmentation image 1901 is generated as an example of region segmentation image 311. Based on measurement region generation rule 1808, measurement region 1903 is generated for region segmentation image 1901. Then, measurement region 1903 is configured in measured image 1900, resulting in measurement region configuration image 1902 as an example of measurement region configuration image 313. Region segmentation image 1901 includes region elements 1901U representing an upper pattern (first region type) and region elements 1901L representing a lower pattern (second region type) in a certain set. This example shows the generation of measurement region 1903 for measuring the lower pattern.

[0300] Figure 20 As Figure 19 A partially enlarged view of a region segmentation image 1901 shows how multiple measurement regions 1903 are generated for region elements 1901L of the lower pattern according to measurement region generation rule 1808. Measurement region generation rule 1808 generates a desired number (set as m) of measurement regions 1903 {A1, A2, ..., Am} along a boundary 1905 (e.g., an arc) between region elements 1901L of the lower pattern and the background, excluding a boundary 1904 with the upper pattern 1901U. Furthermore, measurement region generation rule 1808 arranges, for each measurement region 1903, a rectangular measurement region 1903 extending in a normal direction 1906 relative to a tangent line of the boundary 1905. A computer processor executes processing in accordance with the program for measurement region generation rule 1808.

[0301] In this example, the width W1 of the rectangle in the tangential direction and the height W2 (the width in the longitudinal direction) of each measurement region 1903 in the normal direction 1906 can be set and pre-set. Furthermore, the number and spacing of measurement regions 1903 arranged along the boundary 1905 can be set.

[0302] In this example, multiple measurement areas 1803 are arranged at intervals on the boundary 1905, with some covering the boundary 1905 and some not. However, the present invention is not limited to this. Multiple measurement areas 1803 may be arranged on the boundary 1905 so as to cover the entire boundary 1905. For example, in another measurement area generation rule, a partially missing annular area may be generated as one measurement area to coincide with the arc of the boundary 1905 of the area element 1901L of the lower layer pattern.

[0303] exist Figure 18 During measurement execution phase 1802, when measurement is performed, measurement region generation unit 312 generates measurement region 1903 according to measurement region generation rule 1808 based on region segmentation image 311 (1901) automatically generated from measured image 309 (1900), thereby obtaining measurement region configuration image 313 (1902). Overlap measurement unit 314 then uses measurement region 1903 to measure the amount of overlap offset, etc., based on measurement region configuration image 313 (1902).

[0304] Figure 21 express Figure 20 The measurement area configuration image 1902 corresponding to the example of FIG1 shows an example of edge calculation and centroid calculation of the lower pattern 2101L using the measurement area 103. In this example, during measurement, the overlay measurement unit 314 can detect the edge corresponding to the boundary 2105 of the lower pattern 2101L (the portion that falls within the rectangular measurement area 1903) based on the brightness profile of the portion within each measurement area 1903 of the lower pattern 2101L. For example, in measurement area A1, edge point E1 can be detected based on the profile at a position such as line 2106 along the normal direction 1906. That is, in this example, multiple edge portions (edge ​​points E1, E2, ..., Em) can be detected corresponding to multiple (m) measurement areas 1903 (A1 to Am).

[0305] After edge detection, the overlay measurement unit 314 calculates, for example, the Y coordinate of the center of gravity of the lower layer pattern 1901L based on the multiple edge portions. For example, the overlay measurement unit 314 calculates the Y coordinate of the center of gravity 1921 based on the coordinates of the multiple (m) edge points {E1, E2, ..., Em}, for example, by statistical calculation. Alternatively, the overlay measurement unit 314 may calculate the X coordinate of the center of gravity 1922 based on the coordinates of the multiple (m) edge points. Furthermore, the overlay measurement unit 314 also calculates, for example, the Y coordinate and the X coordinate of the center of gravity for the upper layer pattern 1901U using the same method as for the lower layer pattern or a conventional method.

[0306] Then, the overlay measurement unit 314 can calculate the overlay offset in the Y direction based on the difference between the Y coordinates of the center of gravity of the lower layer pattern 1901L and the Y coordinates of the center of gravity of the upper layer pattern 1901U. Similarly, the overlay offset in the X direction can be calculated.

[0307] To supplement the above-mentioned edge detection, the edge (boundary) of the region element 1901L corresponding to the lower-layer pattern in the region segmentation image 1901 may not coincide with the edge of the lower-layer pattern 2101L in the measured image 1900. However, there is a certain inclination and sense of distance relative to the edge of the region element 1901L, and the edge of the lower-layer pattern 2101L exists nearby. Therefore, in the above-mentioned measurement region generation rule 1808, the region segmentation image 1901 is used to configure a measurement region 1903 having a size in the normal direction 1906 of the edge (boundary 1905) of the region element 1901L corresponding to the lower-layer pattern. This ensures that the edge of the lower-layer pattern 2101L in the measured image 1900 falls within this measurement region 1903 with a high probability. Therefore, the edge of the lower-layer pattern 2101 can be detected with high precision within this measurement region 1903, and the center of gravity can be calculated with high precision based on this edge.

[0308] [Effects of Embodiment 2, etc.]

[0309] As described above, according to Embodiment 2, user U1 does not need to observe and confirm the segmented region image. Instead, user U1 can set a common measurement region generation rule 1808 for measurement target patterns with different pattern shapes and overlap offsets. This reduces the amount of setup work required and achieves similar effects as Embodiment 1. By creating and setting a program for measurement region generation rules designed based on basic rules once, only those rules can be selected for application thereafter. This facilitates the setup of measurement region generation rules by user U1. Alternatively, programs for multiple measurement region generation rules can be prepared.

[0310] While the embodiments of the present disclosure have been described above in detail, they are not limited to the embodiments described above and can be modified in various ways without departing from the main purpose. In addition to the essential structural elements, each embodiment can include the addition, deletion, or replacement of structural elements. Unless otherwise specified, each structural element can be single or multiple. A combination of the various embodiments and variations is also possible.

[0311] As described in the embodiments and modifications, there may be multiple types of measurement area creation rules. Alternatively, a combination of multiple types of measurement area creation rules may be employed, and a rule may be selected and applied from these rules.

[0312] The disclosed technology improves the measurement accuracy of semiconductor devices. Therefore, it contributes to achieving high levels of economic productivity through technological advancement and innovation, which is crucial for achieving the Sustainable Development Goals (SDGs), particularly Project 8, "Jobs and Economic Growth."

[0313] The technology disclosed in the present invention is not limited to the above-mentioned embodiments and includes various modifications. For example, in the structural elements of the measurement system 100, the input and output device 105 may also be a touch panel. Processors such as the main processor 104 may include an MPU, a CPU, a GPU, an FPGA, a quantum processor or other computable semiconductor devices. The computer constituting the measurement system 100 may be, for example, a PC (personal computer), a tablet terminal, a smart phone, a server computer, a blade server, a cloud server, etc., or a collection of computers. The controller 102, the main computer 104, the first sub-computer 107 and the second sub-computer 109 may also share part or all of the hardware. In addition, the program related to the overlap measurement may also be stored in a computer-readable non-volatile storage medium, etc. In this case, the program may also be read from an external recording medium input and output device not shown and executed by the processor.

[0314] (Note)

[0315] The following may also be an embodiment. A measurement system according to an embodiment is a semiconductor device measurement system including a microscope and a processor. The processor obtains an image of the structure of the semiconductor device captured by the microscope, obtains a measurement region generation rule related to the structure, generates a measurement region for arranging the structure based on the image and the measurement region generation rule, arranges the measurement region for the structure in the image, and performs measurement related to the structure using a portion within the measurement region in the image.

[0316] The program of the embodiment is a program that causes a computer having a processor to execute processes, and the processes that the processor is caused to execute include: obtaining an image of the structure of a semiconductor device obtained by photographing it with a microscope; obtaining a measurement area generation rule related to the structure; generating a measurement area for configuring the structure based on the image and the measurement area generation rule; configuring the measurement area for the structure in the image; and performing measurements related to the structure using a portion within the measurement area of ​​the image.

[0317] The storage medium in the embodiment is a non-transitory computer-readable storage medium storing the above-mentioned program, such as a memory card or a disk.

[0318] In the measurement system of the embodiment, the processor sets a rule target area, i.e., an area in the region segmentation image containing a structure to which the measurement region generation rule is applied, in response to a user's operation of confirming the region segmentation image on a screen, and applies the measurement region generation rule to the rule target area to generate a measurement region.

[0319] Furthermore, as a configuration, the processor arranges the measurement region in the SE image when measuring the upper layer pattern, and arranges the measurement region in the BSE image when measuring the lower layer pattern.

[0320] Furthermore, the processor detects an edge of the structure from a portion within the measurement area, calculates a center of gravity of the structure from the edge, and performs measurement based on the center of gravity.

[0321] Description of Reference Numerals

[0322] 100…Measurement system, 101…SEM, 101A…Main body, 102…Controller, 104…Main computer, 105…Input / output device, 107…First sub-computer, 109…Second sub-computer, 201…Specimen, 303…Sample image, 304…Learning unit, 305…Learning model, 306…Region segmentation image, 307…Measurement region generation rule preparation unit, 308…Measurement region generation rule, 309…Measured image, 310…Region segmentation unit, 311…Region segmentation image, 312…Measurement region generation unit, 313…Measurement region configuration image, 314…Overlap measurement unit, 315…Measurement result data

Claims

1. A semiconductor device measurement system having a microscope and a processor, characterized in that: The processor obtains an image of the structure of the semiconductor device taken by the microscope, obtains a measurement area generation rule related to the structure, generates a measurement area for configuring the structure based on the image and the measurement area generation rule, configures the measurement area for the structure in the image, and uses a portion within the measurement area of ​​the image to perform measurements related to the structure.

2. The measurement system according to claim 1, characterized in that The processor generates a region segmentation image based on the image, and applies the measurement region generation rule to region elements included in the region segmentation image based on the region segmentation image and the measurement region generation rule, thereby generating the measurement region for arranging the structure.

3. The measurement system according to claim 2, characterized in that The processor obtains a sample image as the image, generates the region segmentation image based on the sample image, obtains the measurement region generation rule set based on the region segmentation image of the sample image, obtains a measured image of the structure of the semiconductor device photographed by the microscope during measurement, generates the region segmentation image based on the measured image, generates the measurement region for configuring the structure based on the region segmentation image of the measured image and the measurement region generation rule, configures the measurement region for the structure of the measured image, and uses a portion within the measurement region of the measured image to perform measurement related to the structure.

4. The measurement system according to claim 1, wherein: The structure is a three-dimensional structure having at least a lower layer pattern and an upper layer pattern overlapping the lower layer pattern. The processor measures an overlap offset of a set of the lower layer pattern and the upper layer pattern of a measurement object.

5. The measurement system according to claim 2, characterized in that The processor provides a screen displaying the region segmentation image and the measurement region creation rule to a user, and sets the measurement region creation rule in response to the user's confirmation operation on the region segmentation image on the screen.

6. The measurement system according to claim 2, characterized in that The structure is a three-dimensional structure having at least a lower layer pattern and an upper layer pattern overlapping the lower layer pattern. The region segmentation image has region types according to region elements. The region types include at least the lower layer pattern, the upper layer pattern, and a background region. The measurement region generation rule includes the region type for generating the region element to be applied to the structure of the measurement object.

7. The measurement system according to claim 6, characterized in that The measurement area creation rule includes coordinate information of the area element of the area type serving as a reference for generating the measurement area, and a correction value indicating a relative relationship with the coordinate information serving as the reference.

8. The measurement system according to claim 7, characterized in that The measurement area generation rule has the maximum value or minimum value in the specified direction among the specified area elements as the coordinate information serving as the benchmark, and has the correction value in the specified direction as the correction value, and is capable of specifying coordinate information selected from the coordinate information of the upper pattern and the coordinate information of the lower pattern as the coordinate information serving as the benchmark for generating the measurement area to be applied to the lower pattern.

9. The measurement system according to claim 1, wherein: The measurement area generation rule includes a measurement area availability determination rule for determining whether measurement can be performed by generating the measurement area for the structure. When a negative determination is made according to the measurement area permission determination rule, the processor does not generate the measurement area for the structure.

10. The measuring system according to claim 9, characterized in that The measurement region permission determination rule includes a condition that, when the measurement region is to be generated for the structure according to the measurement region generation rule, the measurement region is not generated if the width of the measurement region is smaller than or equal to a specified width.

11. The measurement system according to claim 3, characterized in that The processor learns a learning model for generating the region segmentation image based on the sample image, sets the measurement region generation rule based on the region segmentation image generated by the learning model, and generates the region segmentation image based on the measured image using the learning model during the measurement. The learning is unsupervised machine learning that does not use supervised data in which labels indicating region types are assigned to pixels of the image as input, or supervised machine learning that uses the supervised data as input.

12. The measurement system according to claim 3, characterized in that The measurement area generation rule is set to generate the measurement area by program processing of a basic rule. The processor generates the measurement area by program processing of the basic rule based on the area segmentation image and the measurement area generation rule.

13. The measurement system according to claim 2, characterized in that The image types include an SE image captured by detecting secondary electrons (SE) of the structure of the semiconductor device through the microscope, a BSE image captured by detecting backscattered electrons (BSE), and a composite image obtained by combining the SE image and the BSE image. The processor generates the region segmentation image as the image of the designated image type based on the synthesized image, and arranges the measurement region in accordance with the structure of the image of the designated image type.

14. A computer in a semiconductor device measurement system, the measurement system comprising a microscope and a processor, the computer comprising the processor, wherein: The processor obtains an image of the structure of the semiconductor device taken by the microscope, obtains a measurement area generation rule related to the structure, generates a measurement area for configuring the structure based on the image and the measurement area generation rule, configures the measurement area for the structure in the image, and uses a portion within the measurement area of ​​the image to perform measurements related to the structure.

15. A measurement method in a measurement system for a semiconductor device, the measurement method being executed by a computer, the measurement system comprising a microscope and the computer, the computer comprising a processor, wherein: The steps executed by the computer include: a step of obtaining an image of the structure of the semiconductor device photographed by the microscope; a step of obtaining a measurement region generation rule related to the structure; generating a measurement area for arranging the structure based on the image and the measurement area generation rule; a step of arranging the measurement area on the structure of the image; as well as The step of performing measurement related to the structure using a portion of the image within the measurement area.

Images