Semiconductor observation system and overlay measurement method

Abstract

To improve a success rate of overlay measurements.SOLUTION: An image to be measured of a sample that is captured by a microscope is acquired, a first degree that indicates a degree in which a second layer (upper layer) of the sample transmits a first layer (lower layer) is acquired, a first layer template image and a second layer template image indicating pattern shapes of the first layer and the second layer are acquired, pattern matching processing of the second layer is performed based on the second layer template image and the image to be measured to acquire a second position deviation amount related to the second layer and an area recognized as the second layer on the image to be measured, a consideration range of the image to be measured in pattern matching processing of the first layer is determined based on the first degree and the area recognized as the second layer, pattern matching processing of the first layer is performed based on the consideration range, the first layer template image and the image to be measured. A first position deviation amount related to the first layer and an area recognized as the first layer on the image to be measured are acquired, and an overlay is measured based on the second position deviation amount and the first position deviation amount.SELECTED DRAWING: Figure 6

Description

Technical Field

[0001] The present invention relates to a measuring apparatus for semiconductor patterns and a computer program, and particularly to an overlay measuring apparatus that measures the dimensions of a pattern and the overlay error of patterns across a plurality of layers based on an image obtained by a charged particle beam apparatus or the like.

Background Art

[0002] Patterns manufactured by recent semiconductor processes have advanced in miniaturization and multi-layer structuring, and reduction of the overlay error of patterns across a plurality of layers of an exposure apparatus (hereinafter simply referred to as overlay) is required. Therefore, it is considered that the importance of measuring the overlay with high accuracy and feeding back to the exposure apparatus becomes increasingly high.

[0003] As such means for overlay measurement, a scanning electron microscope (SEM), which is a type of charged particle beam apparatus, is widely used. The SEM outputs an imaging image (hereinafter referred to as a measured image) by detecting reflected electrons and the like obtained when an electron beam irradiates a semiconductor sample. By performing appropriate image processing on this measured image and calculating the positions of the patterns of each layer to be measured for overlay measurement, the overlay can be measured.

[0004] There are mainly two methods for calculating the position of patterns in each layer using image processing. One method involves performing pattern matching on each layer between a template image and the image under measurement, and calculating the position where the matching score is maximized as the position of the pattern in each layer. The other method involves focusing on the change in brightness at the edges of the patterns on the image under measurement, detecting the edges of the patterns in each layer, and calculating the position of the center of the edges as the position of the patterns in each layer. Which method is better depends on the image under measurement, but generally, the latter method is effective when the edges of the pattern being measured are clear, while the former method is effective when they are unclear or when part of the edge of the pattern in the layer of interest is hidden by other layers. This invention targets the former method, and the following description is based on the assumption of the former pattern matching process.

[0005] Generally, overlay images reflect multiple layers of patterns. This can result in some lower-layer patterns being obscured by upper layers, leading to a low matching score and failure of the matching process, or incorrect matching with patterns from unintended layers. To reduce such pattern matching failures, it is effective to exclude information from layers unrelated to the target layer during the pattern matching calculation.

[0006] Patent documents 1 and 2 serve as background technologies in this field. Patent document 1 describes a method for dividing the template image and the image under measurement into upper and lower layer patterns based on luminance information, and then performing pattern matching separately between the upper layer pattern regions and between the lower layer pattern regions of the template image and the image under measurement. Patent document 2 describes a method for generating mask processing regions that are not considered during pattern matching using the edge information of the design data of the sample under measurement. Both methods can reduce pattern matching failures by avoiding the use of information from layers unrelated to the layer being measured. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] Japanese Patent Publication No. 2013-168595 [Patent Document 2] Japanese Patent Publication No. 2011-90470 [Overview of the project] [Problems that the invention aims to solve]

[0008] With the increasing multilayer structure of semiconductor patterns, conventional low-energy electron beams have been insufficient to reveal patterns in the lower layers. Therefore, in recent years, high-acceleration SEMs (SEMs) irradiating with high-energy electron beams (e.g., 15 keV or higher) have begun to be used for overlay measurements. In images obtained with high-acceleration SEMs, due to the high energy of the electron beam, the lower layers may appear to show through to the upper layers.

[0009] If, for an image under measurement where the overlapping portion of the upper and lower layers is transparent, region division or masking as shown in Patent Documents 1 and 2 is carelessly performed and information from the upper layer region is removed during pattern matching of the lower layer, the information of the lower layer that is transparent to the upper layer will be erased, which may actually worsen the success rate of pattern matching.

[0010] In view of the above problems, the present invention aims to improve the success rate of pattern matching for each layer of complex multilayer structures, and the success rate of overlay measurement based thereon. [Means for solving the problem]

[0011] The present invention, to give one example, is a semiconductor observation system having a microscope and a processor for measuring the overlay between a first layer, which is a predetermined layer of a sample having two or more layers formed on it, and a second layer, which is one or more layers above the first layer. The aforementioned processor, The image of the sample to be measured, captured by the microscope, A first degree is obtained that indicates the extent to which the second layer transmits through the first layer. A first layer template image showing the pattern shape of the first layer and a second layer template image showing the pattern shape of the second layer are obtained. By performing a pattern matching process for the second layer based on the second layer template image and the image under measurement, a second positional displacement amount for the second layer and a region on the image under measurement recognized as the second layer are obtained, where the second positional displacement amount is the positional displacement amount between the second layer template image and the image under measurement with respect to the second layer. Based on the first degree and the region recognized as the second layer, the first layer consideration range, which is the range of the image under measurement to be considered in the pattern matching process of the first layer, is determined. By performing a pattern matching process for the first layer based on the first layer consideration range, the first layer template image, and the image under measurement, a first positional displacement amount for the first layer and a region on the image under measurement recognized as the first layer are obtained, where the first positional displacement amount is the positional displacement amount between the first layer template image and the image under measurement with respect to the first layer. The overlay is measured based on the second and first positional displacement amounts. [Effects of the Invention]

[0012] According to the present invention, it is possible to improve the success rate of pattern matching for each layer of complex multilayer structures, and the overlay measurement based thereon. [Brief explanation of the drawing]

[0013] [Figure 1A] This diagram illustrates the problem of upper and lower layers being transparent in the SEM image being measured. [Figure 1B] This diagram illustrates the problem of upper and lower layers being transparent in the SEM image being measured. [Figure 2A] This figure illustrates the challenges of brightness changes due to etching in SEM images. [Figure 2B] This figure illustrates the challenges of brightness changes due to etching in SEM images. [Figure 2C] This is a diagram for explaining the problem of luminance change due to etching in the SEM measurement target image. [Figure 3] This is a diagram showing the outline of SEM. [Figure 4] This is a block diagram showing the outline of the overlay measurement apparatus in Example 1. [Figure 5A] This is a diagram showing the measurement target image for explaining the overlay measurement process in Example 1. [Figure 5B] This is a diagram showing the measurement target image for explaining the overlay measurement process in Example 1. [Figure 6] This is a flowchart of the overlay measurement process in Example 1. [Figure 7] This is a diagram showing an example in which design data and the like are aligned with the measurement target image of FIG. 5B. [Figure 8] This is a diagram showing the degree of transmission between layers for the measurement target image of FIG. 5B. [Figure 9] This is a diagram showing the flow of the overlay measurement process and intermediate processed images for the measurement target image of FIG. 5B. [Figure 10] This is a diagram showing an example of a GUI screen for setting the automatic adjustment function of imaging conditions and image processing conditions in Example 2.

Best Mode for Carrying Out the Invention

[0014] Hereinafter, examples of the present invention will be described with reference to the drawings.

Examples

[0015] First, details of the problems to be solved by the present invention will be described. As described above, in the measurement target image of a high-acceleration SEM, due to the high energy of the electron beam, there are cases where the lower layer can be seen through the upper layer. Examples of cases where the upper layer transmits through the lower layer and cases where it does not transmit through the lower layer on the measurement target image will be described using FIGS. 1A and 1B.

[0016] In Figures 1A and 1B, the middle and lower sections show the image under measurement, while the upper section shows a cross-sectional view of the overlay sample under measurement at the position indicated by XX' on the image under measurement. Furthermore, the sample exemplified in Figures 1A and 1B consists of an upper layer consisting of a vertical line pattern 401 and a lower layer consisting of a base pattern 402 which is the background for the line pattern 401 and a hole pattern 403 located in the background between the lines, as shown in the upper cross-sectional view.

[0017] Figure 1A shows an ideal layout with no overlay between the upper and lower layers. In Figure 1A, in the measured images 404 and 405, although there are differences in pattern contrast, the entire pattern 401-403 is displayed. In contrast, Figure 1B shows an example where there is an overlay between the upper line pattern 401 and the lower hole pattern 403. As shown in the upper cross-sectional view of Figure 1B, the hole pattern 403 is embedded beneath the upper line pattern 401. In this case, in the measured image 406, information about the portion of the lower hole pattern 403 covered by the upper layer is lost. On the other hand, in the measured image 407, the portion of the lower hole pattern 403 covered by the upper layer is also visible through the upper line pattern 401.

[0018] For images under measurement where the degree to which the upper layer transmits through the lower layer, that is, the brightness difference between the upper layer and the overlapping portion of the image under measurement, differs, if region division or masking processing as shown in Patent Documents 1 and 2 is carelessly performed and information from the upper layer region is removed during pattern matching of the lower layer, the information of the lower layer that is transmitted through the upper layer will be erased, which may actually worsen the success rate of pattern matching.

[0019] To address this, this embodiment applies a method that adaptively changes the range considered during pattern matching calculations based on the degree to which the upper layer transmits through the lower layer, that is, the brightness difference of each layer in the overlapping portion of the image being measured. However, the degree to which the upper layer transmits through the lower layer varies depending on the structure and material of the sample being measured, as well as imaging conditions such as the acceleration voltage of the high-acceleration SEM, making it difficult to predict before image acquisition. Therefore, the degree to which the upper layer transmits through the lower layer is made available to the user as an input and adjustment of the processing conditions for pattern matching after they have confirmed the image being measured.

[0020] In addition to measuring the overlapping error between the upper and lower patterns as described above, overlay measurement also includes measuring the positional displacement between the original pattern and the pattern formed by etching. In this case, instead of the degree of transmission, a change in brightness occurs due to etching, which can lead to the same problems as in the case where the upper layer transmits light through the lower layer.

[0021] Figures 2A, 2B, and 2C illustrate the challenges of brightness changes due to etching in SEM images. In Figures 2A, 2B, and 2C, the lower panel shows the image under measurement, and the upper panel shows a cross-sectional view of the overlay sample at the position indicated by XX' on the image under measurement. The sample exemplified in Figures 2A, 2B, and 2C is composed of two materials: insulating material 1001 and non-insulating material 1002.

[0022] Figure 2A shows a cross-sectional view of the sample before etching and an example of the image under measurement. In the image under measurement, the insulating material 1001 appears as a dark background like pattern 1003, and the non-insulating material 1002 appears as a bright foreground like pattern 1004. When an equally spaced groove pattern 1005 is etched onto this sample, an overlay sample under measurement is obtained as shown in Figures 2B and 2C. Figure 2B shows an example of an ideal layout, and Figure 2C shows an example where the groove pattern 1005 is formed with a misalignment. As shown in the cross-sectional view and example of the image under measurement in Figure 2B, the region 1006 where the groove pattern is etched onto the insulating material 1001 and the region 1007 where the groove pattern is etched onto the non-insulating material 1002 are both darker than the brightness of the original patterns 1003 and 1004. Even in the case where the groove pattern 1005 is formed with a misalignment, as shown in Figure 2C, the brightness on the image under measurement changes depending on the overlapping relationship between the insulating material 1001, the non-insulating material 1002 and the groove pattern 1005.

[0023] To address cases where etching causes changes in brightness, a method that adaptively changes the range of consideration during pattern matching calculations between the etched pattern area and other areas is considered effective, similar to the aforementioned measured image where the upper layer transmits through the lower layer. Hereafter, when referring to the degree of transmission in this embodiment, it will include the degree of brightness change caused by this etching.

[0024] Figure 3 shows an overview of the SEM that forms the basis of this embodiment. In Figure 3, the SEM 1100 extracts an electron beam 1103 from an electron source 1101 by an extraction electrode 1102 and accelerates it by an accelerating electrode (not shown). This beam is then focused by a condenser lens 1104, which is a type of focusing lens, and then scanned one-dimensionally or two-dimensionally over the sample 1109 by a scanning deflector 1105. The electron beam 1103 is decelerated by a negative voltage applied to an electrode built into the sample stage 1108 and focused by the lens action of the objective lens 1106 before being irradiated onto the sample 1109.

[0025] When the electron beam 1103 irradiates the sample 1109, electrons 1110, such as secondary electrons and backscattered electrons, are emitted from the irradiated area. The emitted electrons 1110 are accelerated towards the electron source by an acceleration effect based on the negative voltage applied to the sample, collide with the conversion electrode 1112, and generate secondary electrons 1111. The secondary electrons 1111 emitted from the conversion electrode 1112 are captured by the detector 1113, and the output of the detector 1113 changes depending on the amount of secondary electrons captured. The brightness of a display device (not shown) changes according to this output. For example, when forming a secondary electron image, the image of the scanning region is formed by synchronizing the deflection signal to the scanning deflector 1105 with the output of the detector 1113.

[0026] The SEM illustrated in Figure 3 is a device capable of applying a high voltage (e.g., 15kV or more) to an accelerating electrode (not shown), and by irradiating the sample with a high-acceleration electron beam, the electron beam can reach embedded patterns and other elements not exposed on the sample surface. Furthermore, while Figure 3 illustrates an example where electrons emitted from the sample are first converted by a conversion electrode and then detected, the configuration is not limited to this. For example, it is possible to arrange the detection surfaces of the electron doubling tube and detectors along the trajectory of the accelerated electrons. In addition, the conversion electrode 1112 and detector 1113 do not need to be a single unit; a configuration with multiple detection surfaces divided in the azimuthal and elevation directions relative to the optical axis, and a detector corresponding to each detection surface, is also possible. In this configuration, it is possible to simultaneously acquire the number of measured images corresponding to the number of detectors in a single imaging scan.

[0027] The control device 1120 controls each component of the SEM and also has functions to form an image based on detected electrons and to measure the pattern width of a pattern formed on a sample based on the intensity distribution of detected electrons, called a line profile. The control device 1120 also includes an SEM control device that mainly controls the optical conditions of the SEM and a signal processing device that processes the detection signal obtained by the detector 1113. The SEM control device includes a scanning control device for controlling the beam scanning conditions (direction, velocity, etc.). The control device 1120 also has a storage medium (not shown) that stores programs that cause a computer (CPU) to perform image processing and calculations as described later.

[0028] Figure 4 shows an overview of the overlay measurement device in this embodiment. The overlay measurement device acquires the image to be measured from an image generation device such as the SEM exemplified in Figure 3 and performs overlay measurement. The image to be measured by the overlay measurement device may be received via communication from the image generation device such as the SEM, acquired via portable memory such as a USB (Universal Serial Bus, registered trademark) memory, or acquired by reading the image to be measured that has already been stored in the built-in memory. Furthermore, the device consisting of the SEM shown in Figure 3 and the overlay measurement device shown in Figure 4 is referred to as a semiconductor observation system.

[0029] In Figure 4, the overlay measurement device includes an input / output unit 205 that inputs the image to be measured and outputs measurement results, etc., an operation unit 204 that inputs necessary information such as the degree to which the upper layer is transparent over the lower layer from a GUI (Graphical User Interface) screen, a storage unit 206 that stores the image to be measured, a template image, and overlay measurement conditions such as the degree to which the upper layer is transparent over the lower layer used for overlay measurement, a control unit 201 equipped with a processor 202 that performs calculation processing required for overlay measurement by software processing realized by executing a program (hereinafter sometimes referred to as the overlay measurement program) based on image information input from the input / output unit 205 and information such as measurement conditions input from the operation unit 204 and the storage unit 206, and a display unit 203 that displays the input image to be measured, the overlay measurement results obtained based on the image to be measured, and GUI images, etc.

[0030] Next, the overlay measurement process in this embodiment will be described. To simplify the explanation, the measurement images of a three-layered sample shown in Figures 5A and 5B will be used as examples of the measurement images. The sample shown in Figures 5A and 5B consists of, in order from the top layer, a vertical line pattern 501 of the third layer, a horizontal line pattern 502 of the second layer, a hole pattern 503 of the first layer, and a base layer 504.

[0031] Figure 5A shows an example of a measured image in an ideal layout without an overlay. Figure 5B shows an example of a measured image in a layout where the horizontal line pattern 502 of the second layer is shifted upwards relative to the vertical line pattern 501 of the third layer, and the hole pattern 503 of the first layer is shifted downwards to the right. In this example of a measured image, the area of ​​the vertical line pattern 501 of the third layer is transparent to the horizontal line pattern 502 of the second layer, but not to the hole pattern 503 of the first layer. Also, the area of ​​the horizontal line pattern 502 of the second layer is transparent to the hole pattern 503 of the first layer.

[0032] Figure 6 is a flowchart of the overlay measurement process in this embodiment. In Figure 6, first, the input / output unit 205 reads the image to be measured from the SEM, which is an image generation device. Alternatively, the storage unit 206 stores the image to be measured and its associated ID in advance, and when the user inputs the ID using the operation unit 204, it is compared with the ID in the storage unit 206 and the image to be measured associated with it is read (step S101).

[0033] Next, the processor 202 reads reference data associated with the image under measurement from the memory unit 206 and obtains template images for the upper and lower layers (step S102). Here, the reference data is, for example, design data that shows the layout of the patterns of each layer. The template images for the upper and lower layers are line drawings that show the edges of the patterns of each layer based on this design data. Alternatively, they are images created from the design data to simulate the appearance of the image under measurement. A simple method of simulation is to fill in the pattern areas of each layer in the design data with the average brightness values ​​of the pattern areas of each layer in the corresponding image under measurement. It is also effective to apply a Gaussian filter to account for image blurring during imaging, or to add fluctuations to the pattern edges that account for roughness caused by the manufacturing process. Furthermore, it is also effective to pre-learn the conversion relationship between the design data and the image under measurement using deep learning, and then use this learned model to generate simulated images from the design data. Alternatively, the reference data is an averaged image of multiple images under measurement that have been pre-stored in the memory unit 206. While the image being measured generally contains an unknown overlay, under the assumption that this overlay follows a probability distribution with a mean of zero, an image with an overlay close to zero can be obtained by averaging multiple images. Using this as reference data, one possible method for creating template images for the upper and lower layers is to apply a region division process, such as the one described in Patent Document 1, to the averaged image, set the brightness of the averaged image in the resulting upper and lower regions, and use images with zero brightness or the brightness of the background region in regions other than the upper and lower layers. Next, the processor 202 obtains the degree to which the upper layer penetrates the lower layer for all layers of the object to be measured (step S103). The degree to which the upper layer penetrates the lower layer is a value obtained by comparing the region where the upper and lower layers overlap with the region where the upper and lower layers do not overlap. For example, in the region where the upper and lower layers overlap, if the upper layer penetrates the lower layer and the lower layer is sufficiently visible compared to the region where the upper and lower layers do not overlap, the value is 1; if it is almost invisible and not penetrating, the value is 0. In other words, the degree to which the upper layer penetrates the lower layer is a value that can take an upper or lower limit with respect to the region where the upper layer covers the lower layer. The upper limit means that the upper layer covering the lower layer is not visible in the measured image, and the lower limit means that the lower layer covered by the upper layer is not visible in the measured image. Alternatively, the degree to which the upper layer penetrates the lower layer may be selected to be between 0 and 1. In other words, the degree to which the upper layer transmits through the lower layer may be a value that is less than the upper limit and greater than the lower limit, with respect to the area where the upper layer covers the lower layer. For example, if the user views the image to be measured displayed on the display unit 203 and compares the area where the upper and lower layers overlap with the area where they do not overlap, and it appears that the upper layer transmits through about half of the lower layer, then the value may be set to 0.5. By inputting these values ​​using the operation unit 204, the processor 202 obtains the degree to which the upper layer transmits through the lower layer. Alternatively, the degree to which the upper layer transmits through the lower layer may be automatically determined using the brightness of the image to be measured. For example, if the brightness of the lower layer pattern is x1, the brightness of the upper layer pattern is x2, and the brightness of the area where the upper and lower layer patterns overlap is x, then α is the real value when the brightness x is expressed in linear form as a linear sum such as equation x = α(x1 + x2) or equation x = αx1 + x2. This real value α may be automatically calculated by the processor 202 using the luminances x1 and x2 of each layer, based on the user inputting the luminances x1 and x2 of each layer via the operation unit 204. In this case, the luminances x1 and x2 of each layer may be determined by the user viewing the image under measurement and the design data displayed on the display unit 203, and adjusting the position of the design data on the operation unit 204 to match the pattern of each layer in the image under measurement.

[0034] Figure 7 shows an example of aligning design data with the image under measurement shown in Figure 5B. In Figure 7, the design data for the upper layer 603 is aligned with the upper layer pattern, the design data for the middle layer 602 with the middle layer pattern, and the design data for the lower layer 601 with the lower layer pattern, relative to the example image 506 shown in Figure 5B. The average value and mode of luminance within each pattern area of ​​the design data aligned to each layer's pattern are then determined as the luminance x1, x2 for each layer and the luminance x of the overlapping area between the upper and lower layers. Using the luminances x1, x2, and x thus obtained, the degree α to which the upper layer transmits through the lower layer may be calculated according to the linear form described above. Here, design data refers to data that shows the pattern layout of each layer of the object under measurement. This is, for example, electronic data that shows the edges of the circuit pattern as line segments when designing the circuit pattern under measurement using CAD (Computer Aided Design). Such design data may be stored in the storage unit 206 in advance, linked to the image to be measured, and read together with the image to be measured when the ID is matched. In this case, instead of design data, a measuring cursor used when measuring dimensions such as pattern width may be used. Measuring cursors are generally recorded in the pattern's dimensional measurement condition data. Therefore, the storage unit 206 may store the dimensional measurement condition data, and the user may read and use it by entering the ID associated with it in the operation unit 204. Furthermore, instead of design data, the user may place rectangular cursors or the like on the layer-specific patterns in the operation unit 204 while looking at the image to be measured displayed in the display unit 203. Alternatively, the degree to which the upper layer transmits through the lower layer may be a value calculated physically. For example, it may be a transmittance calculated using a desk calculation or an electron beam simulator from the structural information of the sample, including the material and thickness of each layer, and the energy of the electron beam. In other words, the degree to which the upper layer transmits through the lower layer may be entered by the user via the GUI as described above, or it may be calculated by a processor.

[0035] Through the above process (step S103), the processor 202 obtains the degree to which the upper layer permeates the lower layer for all layers. Hereafter, the degree to which the upper layer permeates the lower layer for all layers will be referred to as the degree of inter-layer permeability. For the three-layer sample shown in Figures 5A and 5B, the degree of inter-layer permeability 701 is shown in Figure 8. In the case of a three-layer sample, the degree of inter-layer permeability is expressed by three variables: the degree to which the third layer permeates the second layer, the degree to which the third layer permeates the first layer, and the degree to which the second layer permeates the first layer. For a sample with all N layers, the degree of inter-layer permeability is expressed by the number of variables shown in the following equation (1).

[0036]

number

[0037] The degree of transparency between the aforementioned layers is displayed on the GUI screen and can be edited by the user. This allows the user to check whether the automatically calculated degree of transparency between the layers is a reasonable value and to correct the value as needed.

[0038] Next, the processor 202 performs overlay measurement processing by pattern matching based on the image to be measured read in steps S101 to S103, the template images for each layer, and the degree of transparency between layers (steps S104 to S109).

[0039] Figure 9 shows the flow of the overlay measurement process and the intermediate processed image for the image under measurement in Figure 5B. The procedure for the overlay measurement process in this embodiment will be explained below using Figures 6 and 9.

[0040] In Figure 9, first, the processor 202 sets the topmost layer, the third layer, as the target layer for pattern matching, and performs pattern matching of the third layer using the template image 803 of the third layer and the image under measurement 804 (step S104 in Figure 6). This yields the matching result 805 of the third layer and the pattern position of the third layer. Furthermore, from the matching result 805 of the third layer, the region image 806 of the third layer, which is the region recognized as the third layer, is obtained. The region image 806 of the third layer is an image in which the brightness value is 255 within the region recognized as the third layer and 0 everywhere else. The method for obtaining the region image 806 of the third layer is, for example, if the template image of the third layer is an image generated from design data, then the region within the third layer region in the original design data, or the region obtained by applying expansion / contraction processing to it, is recognized as the third layer region. Alternatively, it may be obtained by performing region extraction processing using brightness thresholding on the template image of the third layer. Alternatively, the region in which the pattern matching score is above a certain threshold may be recognized as the third layer region.

[0041] Next, the processor sets the second layer, which is one layer below, as the target layer for pattern matching (step S105), and uses the region image 806 of the third layer, which is above the second layer, and the degree 901 to which the upper layer is transparent to the second layer, obtained from the degree of transparency between layers 701 input in step S102, to derive the consideration range 807 of the image under measurement to be used for pattern matching of the second layer (step S106). Hereafter, the consideration range of the image under measurement used for pattern matching of the nth layer will be referred to as the nth layer consideration range. The second layer consideration range 807 is represented by data with the same number of pixels as the image under measurement, and within the region of the third layer, it is an image with a value of 0.5 to which the third layer is transparent to the second layer, and a value of 1 in other regions. Note that the brightness of the image may actually take the form of an integer between 0 and the largest integer (e.g., 255). In this case, the aforementioned 1 may be considered to correspond to the largest integer. Furthermore, the scope of consideration can also be seen as indicating the degree to which each pixel is considered when performing pattern matching, and is therefore sometimes called a weight map.

[0042] Next, pattern matching for the second layer is performed using the second layer consideration range 807, the second layer template image 802, and the image under measurement 804 to obtain the second layer matching result 808. Then, similar to the third layer, a region image 809 of the second layer, which is the region recognized as the second layer, is obtained (step S107). As a specific method for calculating pattern matching, for example, zero-mean normalized cross-correlation (ZNCC) is used. Equation (2) below shows an example of the formula for calculating the matching score in this embodiment using the second layer consideration range 807 and ZNCC in step S107.

[0043]

number

[0044] Here, f(x,y) is the template image, g(x,y) is the image under measurement, and w(x,y) is the range to consider. f(x,y), g(x,y), and w(x,y) represent images, and x and y represent the coordinates of pixels that make up the image. dx and dy indicate the number of pixels to shift in the x and y directions when calculating the score. f￣ (overline) and g￣ (overline) are the average values ​​of images f(x,y) and g(x,y), respectively. The matching score using the above equation (2) is calculated within the given range of dx and dy, i.e., the search range for pattern matching, and the dx and dy that maximize this score are calculated as the position of the target pattern. In addition to the above ZNCC, the score calculation formula may also be the sum of the absolute differences in pixel values ​​between the template image and the image under measurement, or the sum of the squares of the differences in pixel values ​​between the template image and the image under measurement.

[0045] Next, if there are still layers to be measured, as shown in Figure 9 (Step S108), the processor sets the layer below, the first layer, as the target layer for pattern matching (Step S105). Then, using the region image 806 of the third layer and the region image 809 of the second layer, which are above the first layer, and the degree 902 of the upper layer's transparency through the first layer, obtained from the degree of transparency between layers 701, the first layer consideration range 810 is derived (Step S106). The first layer consideration range 810 is an image that, similar to the second layer consideration range 807, has a value of 0.0 in this example for the degree of the third layer's transparency through the first layer within the third layer's region, and a value of 0.5 in this example for the degree of the second layer's transparency through the first layer within the second layer's region. Note that in the overlapping region of the third and second layers, it is necessary to separately derive the degree of the overlapping region of the third and second layers' transparency through the first layer. One possible derivation method is to consider the minimum value or product of the degree to which the third layer transmits through the first layer and the degree to which the second layer transmits through the first layer. This is because, generally, as the layers overlap, the number of backscattered electrons obtained from the lower layers decreases, and the signal of the lower layer pattern in the image becomes smaller. Another possible derivation method is to consider the minimum value or product of the degree to which the third layer transmits through the second layer and the degree to which the second layer transmits through the first layer.

[0046] Then, using the first layer consideration range 810, the first layer template image 801, and the image under measurement 804, pattern matching for the first layer is performed to obtain the first layer matching result 811 and the first layer pattern position (step S107). The score calculation shown in equation (2) above is also performed to calculate the first layer pattern position.

[0047] If processing is completed for all layers to be measured (step S108), finally, the amount of overlay between layers is calculated using the pattern position of each layer (step S109).

[0048] Through the above process, by adaptively performing pattern matching calculations as shown in equation (2) using the consideration range of each layer based on the region of each layer and the degree of transparency between layers, it is possible to improve the success rate of pattern matching for each layer of complex multilayer structures and the overlay measurement based on this.

[0049] In this embodiment, as shown in Figures 6 and 9, we have demonstrated an example where pattern matching is performed sequentially from the top layer. Generally, the number of backscattered electrons detected when irradiated with an electron beam increases with the upper layers, and the brightness and signal-to-noise ratio (S / N ratio) tend to be higher in the measured image. Therefore, performing pattern matching sequentially from the top layer is considered an effective method. However, the brightness of the measured image changes not only depending on the brightness of each layer, but also on the sample structure such as holes and trenches, and the material of the sample. In particular, when metal wiring is formed in the lower layer, the brightness and S / N ratio of the lower metal wiring pattern may be higher than that of the upper layer pattern. In such cases, it is not always necessary to perform pattern matching sequentially from the top layer, but rather to perform the pattern matching process from the lower layer, which has higher brightness and S / N ratio.

[0050] Furthermore, although this embodiment describes a sample with a three-layer structure, the layer structure is not limited to three layers and can be applied to any number of layers. Also, although this embodiment describes overlay between consecutive layers, the overlay error between discontinuous layers may also be measured. [Examples]

[0051] This embodiment describes a GUI for performing the overlay measurement shown in Example 1.

[0052] Figure 10 shows an example of a GUI screen for setting the automatic adjustment function for imaging conditions and image processing conditions in this embodiment. In Figure 10, the overlay measurement box 300 displays the ID input box and preview box 301 (step S101) of the image to be measured as described in Embodiment 1, a box 302 (step S102) for acquiring template images for each layer from reference data, a box 303 (step S103) for inputting and displaying the degree of transparency between layers, and buttons 304 (steps S104-S109) for executing overlay measurement using the input image to be measured, the template images of each layer, and the degree of transparency.

[0053] According to this embodiment, it is possible to specify items that require user input in order to perform the overlay measurement shown in Example 1.

[0054] As illustrated above, the present invention has the effect of improving the success rate of pattern matching for each layer of complex multilayer structures and the overlay measurement based thereon. Therefore, the present invention contributes to achieving a high level of economic productivity through technological advancement and innovation, particularly in relation to SDG 8, "Decent Work and Economic Growth," for realizing the SDGs (Sustainable Development Goals).

[0055] Furthermore, the present invention is not limited to the embodiments described above, and various modifications are included. For example, in the components of the overlay measurement device, the input / output unit 205 is an input / output interface, the operation unit 204 may be a keyboard or a touch panel, and the storage unit 206 may be a storage medium such as semiconductor memory or a hard disk. Also, the processor 202 includes a microprocessor, CPU (Central Processing Unit), GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), quantum processor, or other computationally capable semiconductor device. Furthermore, the overlay measurement device including the control unit 201 equipped with the processor 202 may be a computer such as a personal computer, tablet terminal (computer), smartphone, server computer, blade server, cloud server, or a collection of computers. Also, the overlay measurement device may have multiple components as described above. A collection of multiple computers is an example. The overlay measurement device may share some or all of the hardware with the control device 1120. Also, the overlay measurement program may be stored in a computer-readable non-volatile memory medium. In this case, the overlay measurement program is read from the medium and executed by the processor.

[0056] Furthermore, the embodiments described above are explained in detail for the purpose of clearly illustrating the present invention, and are not necessarily limited to those having all the configurations described. [Explanation of symbols]

[0057] 201: Control unit, 202: Processor, 203: Display unit, 204: Operation unit, 205: Input / Output unit, 206: Storage unit, 701: Degree of transparency between layers, 801: Template image of the first layer, 802: Template image of the second layer, 803: Template image of the third layer, 804: Image under measurement, 805: Matching result of the third layer, 806: Region image of the third layer, 807: Consideration range of the image under measurement used for pattern matching of the second layer (consideration range of the second layer), 808: Matching result of the second layer, 809: Region image of the second layer, 810: Consideration range of the image under measurement used for pattern matching of the first layer (consideration range of the first layer), 811: Matching result of the first layer

Claims

1. A semiconductor observation system comprising a microscope and a processor for measuring the overlay between a first layer, which is a predetermined layer of a sample having two or more layers formed on it, and a second layer, which is one or more layers above the first layer, The aforementioned processor, The image of the sample to be measured, captured by the microscope, A first degree is obtained that indicates the extent to which the second layer transmits through the first layer. A first layer template image showing the pattern shape of the first layer and a second layer template image showing the pattern shape of the second layer are obtained. By performing a pattern matching process for the second layer based on the second layer template image and the image under measurement, a second positional displacement amount for the second layer and a region on the image under measurement recognized as the second layer are obtained, where the second positional displacement amount is the positional displacement amount between the second layer template image and the image under measurement with respect to the second layer. Based on the region recognized as the first layer and the second layer, the first layer consideration range, which is the range of the image under measurement in the pattern matching process of the first layer, is determined. By performing a pattern matching process for the first layer based on the first layer consideration range, the first layer template image, and the image under measurement, a first positional displacement amount related to the first layer and a region on the image under measurement recognized as the first layer are obtained, where the first positional displacement amount is the positional displacement amount between the first layer template image and the image under measurement related to the first layer. A semiconductor observation system characterized by measuring an overlay based on the second positional displacement and the first positional displacement.

2. In the semiconductor observation system according to claim 1, The sample has a third layer which is one or more layers below the first layer. The aforementioned processor, A second degree is obtained that indicates the extent to which each of the second and first layers transmits through the third layer. A third layer template image showing the pattern shape of the third layer is obtained, Based on the second degree, the region recognized as the first layer, and the region recognized as the second layer, the third layer consideration range, which is the range of the image under measurement to be considered in the pattern matching process of the third layer, is determined. By performing a pattern matching process for the third layer based on the third layer consideration range, the third layer template image, and the image under measurement, the third positional displacement amount related to the third layer and the region on the image under measurement recognized as the third layer are obtained, where the third positional displacement amount is the positional displacement amount between the third layer template image and the image under measurement with respect to the third layer. A semiconductor observation system characterized by measuring an overlay based on the third positional displacement, the first positional displacement, and the second positional displacement.

3. In the semiconductor observation system according to claim 1, The first degree is a value that can take on a value at least below the upper limit and above the lower limit, With respect to the region in which the second layer covers the first layer, The aforementioned upper limit means that the second layer covering the first layer is not visible in the image being measured. The aforementioned lower limit means that the first layer covered by the second layer is not visible in the image being measured. A semiconductor observation system characterized by the following features.

4. In the semiconductor observation system according to claim 1, The aforementioned first degree is a value that can take an upper or lower limit, With respect to the region in which the second layer covers the first layer, The aforementioned upper limit means that the second layer covering the first layer is not visible in the image being measured. The aforementioned lower limit means that the first layer covered by the second layer is not visible in the image being measured. A semiconductor observation system characterized by the following features.

5. In the semiconductor observation system according to claim 1, The first degree is the brightness x of the region in the measured image where the second layer and the first layer overlap, and the brightness x of the first layer. １ and the brightness x of the second layer ２ A semiconductor observation system characterized by being a real value α when expressed as a linear sum using and coefficient α.

6. In the semiconductor observation system according to claim 1, A semiconductor observation system characterized in that the second layer template image and the first layer template image are design data having layout information of the patterns of each layer of the sample.

7. In the semiconductor observation system according to claim 1, The second layer template image and the first layer template image are at least one of the following: A line drawing image generated based on design data having region information of the pattern of each layer of the sample, an image created to simulate the appearance of the image under measurement based on the design data, A semiconductor observation system characterized by the following:

8. In the semiconductor observation system according to claim 1, A semiconductor observation system characterized in that the second layer template image and the first layer template image are aggregated average images of a plurality of the images to be measured.

9. In the semiconductor observation system according to claim 1, The semiconductor observation system is characterized in that the second layer template image and the first layer template image are images obtained by performing layer recognition processing using brightness information or the like on an averaged sum of multiple images under measurement, thereby separating them into a second layer and a first layer.

10. In the semiconductor observation system according to claim 1, The semiconductor observation system is characterized in that the first layer consideration range is represented by data having the same number of pixels as the image under measurement, and each pixel has a value corresponding to the first degree.

11. In the semiconductor observation system according to claim 1, The semiconductor observation system is characterized in that the pattern matching process calculates a matching score based on normalized cross-correlation values ​​between the image under measurement and a template image, weighting it based on a consideration range, and calculating the position where the matching score is maximized as the pattern position.

12. In the semiconductor observation system according to claim 1, The semiconductor observation system is characterized in that the first degree is a change in brightness due to etching of the sample.

13. In the semiconductor observation system according to claim 1, A semiconductor observation system characterized in that the arrangement of the second layer and the first layer is from bottom to top, rather than from top to bottom, and the pattern matching process is performed from bottom to top.

14. In the semiconductor observation system according to claim 1, The semiconductor observation system is characterized in that the first degree is displayed on a GUI screen and can be viewed and edited by the user.

15. An overlay measurement method performed by a processor that measures the overlay between a first layer, which is a predetermined layer of a sample having two or more layers formed on it, and a second layer, which is one or more layers above the first layer, The image of the sample being measured is acquired, A first degree is obtained that indicates the extent to which the second layer transmits through the first layer. A first layer template image showing the pattern shape of the first layer and a second layer template image showing the pattern shape of the second layer are obtained. By performing a pattern matching process for the second layer based on the second layer template image and the image under measurement, a second positional displacement amount for the second layer and a region on the image under measurement recognized as the second layer are obtained, where the second positional displacement amount is the positional displacement amount between the second layer template image and the image under measurement with respect to the second layer. Based on the region recognized as the first layer and the second layer, the first layer consideration range, which is the range of the image under measurement in the pattern matching process of the first layer, is determined. By performing a pattern matching process for the first layer based on the first layer consideration range, the first layer template image, and the image under measurement, a first positional displacement amount related to the first layer and a region on the image under measurement recognized as the first layer are obtained, where the first positional displacement amount is the positional displacement amount between the first layer template image and the image under measurement related to the first layer. An overlay measurement method characterized by measuring the overlay based on the second positional displacement and the first positional displacement.

16. In the overlay measurement method described in claim 15, The sample has a third layer which is one or more layers below the first layer. A second degree is obtained that indicates the extent to which each of the second and first layers transmits through the third layer. A third layer template image showing the pattern shape of the third layer is obtained, Based on the second degree, the region recognized as the first layer, and the region recognized as the second layer, the third layer consideration range, which is the range of the image under measurement to be considered in the pattern matching process of the third layer, is determined. By performing a pattern matching process for the third layer based on the third layer consideration range, the third layer template image, and the image under measurement, the third positional displacement amount related to the third layer and the region on the image under measurement recognized as the third layer are obtained, where the third positional displacement amount is the positional displacement amount between the third layer template image and the image under measurement with respect to the third layer. An overlay measurement method characterized by measuring the overlay based on the third positional displacement, the first positional displacement, and the second positional displacement.

17. In the overlay measurement method described in claim 15, The first degree is a value that can take on a value at least below the upper limit and above the lower limit, With respect to the region in which the second layer covers the first layer, The aforementioned upper limit means that the second layer covering the first layer is not visible in the image being measured. The aforementioned lower limit means that the first layer covered by the second layer is not visible in the image being measured. An overlay measurement method characterized by this feature.

18. In the overlay measurement method described in claim 15, The aforementioned first degree is a value that can take an upper or lower limit, With respect to the region in which the second layer covers the first layer, The aforementioned upper limit means that the second layer covering the first layer is not visible in the image being measured. The aforementioned lower limit means that the first layer covered by the second layer is not visible in the image being measured. An overlay measurement method characterized by the following features.

19. In the overlay measurement method described in claim 15, The first degree is the brightness x of the region in the measured image where the second layer and the first layer overlap, and the brightness x of the first layer. １ and the brightness x of the second layer ２ An overlay measurement method characterized by being a real value α when expressed as a linear sum using and coefficient α.

20. In the overlay measurement method described in claim 15, The overlay measurement method is characterized in that the second layer template image and the first layer template image are design data having layout information of the patterns of each layer of the sample.

Images