Detection area forming method, crystal grain surface detection method and crystal grain surface hot spot judgment method

By forming a detection area on the integrated circuit layout and using AFM for non-destructive testing, the problems of insufficient detection resolution and long time in the prior art are solved, and efficient and fast nano-level hot spot detection on the production line is achieved.

CN119024006AActive Publication Date: 2024-11-26XIAMEN UNIV

Patent Information

Application Number
CN202411124113.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-11-26
Estimated Expiration
2044-08-15

AI Technical Summary

Technical Problem

When detecting hot spots on the grain surface after CMP processing, the resolution cannot reach the nanoscale, and the detection time is long, making it difficult to popularize and apply on the production line.

Method used

By selecting the characterization area on the integrated circuit layout, a detection area is formed, and the local area size is within the range of 0.5 microns to 50 microns. AFM is used to perform non-destructive testing, shortening the detection time and improving the detection compression ratio.

Benefits of technology

It realizes efficient and lossless nanoscale detection on the production line, shortens the detection time, improves the detection compression ratio, and can quickly determine whether there are hot spots on the grain surface.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119024006A_ABST
    Figure CN119024006A_ABST
Patent Text Reader

Abstract

The invention discloses a detection area forming method, a crystal grain surface detection method and a crystal grain surface hot spot judgment method. The detection area forming method comprises the following steps: selecting characterization areas on an integrated circuit layout, and forming a detection area by each characterization area; wherein the characterization area is used for characterizing a local area with a similar structure, and the size of the local area is in a range of 0.5 micrometer to 50 micrometers. According to the grain surface detection method, the detection area formed by the detection area forming method is used for performing shape detection on the grain surface. The grain surface hot spot judgment method judges whether hot spots exist or not according to a shape detection result. Compared with the prior art, the method has a higher detection compression ratio, and is more beneficial to popularization and application of detection means with high resolution on a production line.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of surface detection of crystal grains after CMP processing, and specifically to a detection area forming method, a crystal grain surface detection method and a crystal grain surface hot spot determination method. Background Art

[0002] Chemical Mechanical Polishing (CMP), also known as chemical mechanical planarization, is a key process for achieving global uniform planarization of wafers in the semiconductor manufacturing process. It is also the most commonly used process in the manufacturing process (accounting for about 25% of all processes). Figure 1 As shown, the purpose of the CMP process is to achieve the flattening of the wafer surface. In the drawings of this patent, the dark parts are all semiconductor substrates, and the light parts are all functional parts made of metal materials. In this patent, the functional part refers to the other parts except the semiconductor substrate. The most typical functional part is made of metal material, and it can also include insulating parts. The metal part is often used to realize the functions of the circuit, such as connection and contact. The insulating part is used for insulation isolation, which is often formed by filling the grooves of the semiconductor material with insulating material.

[0003] Whether there are defects in the CMP process depends mainly on whether there are hot spots on the die on the wafer. Hot spots will have a lasting impact on the subsequent processes of semiconductor processing and will eventually lead to a decrease in yield. Therefore, it is very important to determine whether there are hot spots on the die after CMP processing to improve yield. The generation of hot spots may come from the design of the integrated circuit layout, or it may come from the materials used in the CMP process or the relevant parameters of the CMP process. Therefore, determining whether there are hot spots on the die after CMP processing is also very important for improving the layout design and adjusting the CMP process. In reality, there are two main manifestations of hot spots after the CMP process, one of which is Figure 2 As shown in the figure, it is manifested as erosion of the functional part. Erosion occurs in the functional part of the metal material and generally occurs along the width direction of the metal line. Therefore, the detection of erosion requires nanometer-level detection of the shape (especially height) within the size range of nanometers or tens of nanometers. Another method is Figure 3 As shown, it manifests as a dip in the surface area. The dip occurs in the surface area at the micrometer level or tens of micrometers, and is strongly related to the line width and density of the functional parts in the surface area. Nanoscale detection of the shape is also required.

[0004] In order to determine whether there are hot spots on the die on the wafer, it is necessary to perform a topographic inspection (also known as roughness inspection) on the die surface after each CMP process. The resolution of this inspection should be at the nanometer level, and in order to achieve production line efficiency, this inspection should be completed within 4 hours.

[0005] In the prior art, white light interferometer is the default option for grain surface shape detection. Blunt et al. used white light interferometer to measure the surface roughness of semiconductor polished substrates and epitaxial layers, and Zhang et al. also applied it to measure the step height of sputtered functional layers. Studies have shown that although white light interferometry basically meets the requirements in terms of longitudinal resolution (0.1 nanometers) and measurement time (about 8 hours per grain), its lateral resolution is micron-level or slightly less than 1 micron, so the detection accuracy cannot meet the requirements. Moreover, when the thickness of the grain to be measured is too small (less than 50 nanometers), the grain itself will appear translucent. At this time, light scattering will affect the measurement results. For this reason, a layer of material needs to be applied to the surface of the grain to achieve white light interferometry, but it will damage the grain at the same time.

[0006] Atomic force microscopy (AFM) can achieve a spatial resolution of 1nm in both the X-axis and Y-axis directions, and has a higher resolution in the Z-axis direction (height direction), which can provide true three-dimensional atomic-level surface images. At the same time, AFM does not require any special treatment of the sample, and the probability of sample damage during the preparation process is small, so non-destructive testing can be performed. As one of the powerful surface morphology detection tools with nanoscale resolution, AFM fully meets the resolution (accuracy) requirements of semiconductor grain surface morphology detection, and will not damage the grains during the measurement process. Xu et al. used AFM to non-destructively measure the density of etch pits (Etch Pits) produced after CMP of GaN wafers. Shi et al. also realized the AFM in-situ observation of local CMP behavior on sapphire for the first time. However, the imaging speed of AFM is slow. Taking the lateral resolution of 10nm as an example, it takes about 5 minutes to image an area of ​​100 square microns. At this time, measuring a typical 33×27mm 2 Therefore, it is difficult to popularize and apply the high-resolution detection methods such as AFM to detect the shape of the grain surface on the production line. Summary of the invention

[0007] The purpose of this application is to overcome the above-mentioned defects or problems existing in the background technology, and to provide a detection area formation method, a grain surface detection method, a grain surface hotspot judgment method, a computer program, a computing device and a detection equipment, which have a higher detection compression ratio than the existing technology, and are more conducive to the popularization and application of high-resolution detection means on the production line.

[0008] In order to achieve the above purpose, the following technical solutions are adopted:

[0009] The first technical solution relates to a detection area forming method, which is used to determine the detection area when performing surface detection on a grain after CMP processing. It selects a characterization area on an integrated circuit layout and forms a detection area from each characterization area; the characterization area is used to characterize a local area with a similar structure, and the dimensions of the local area on the X-axis and Y-axis are in the range of 0.5 microns to 50 microns.

[0010] The second technical solution is based on the first technical solution, wherein the dimensions of the local area in the X-axis and the Y-axis are in the range of 10 microns to 30 microns.

[0011] The third technical solution is based on the second technical solution, wherein the dimensions of the local area in the X-axis and the Y-axis are 20 microns.

[0012] The fourth technical solution is based on the first technical solution, and includes the following steps: dividing the part to be measured on the integrated circuit layout into several local areas, and all local areas cover the part to be measured; the part to be measured covers all processing area groups, and each processing area group includes processing areas with similar structures; all local areas are classified based on structural similarity to obtain local area groups; each local area group selects at least one local area as a characterization area, and the collection of all characterization areas forms a detection area.

[0013] The fifth technical solution is based on the fourth technical solution, wherein adjacent local areas overlap with each other.

[0014] The sixth technical solution is based on the fifth technical solution, wherein the local area is obtained by sliding a virtual window on the simulated part.

[0015] The seventh technical solution is based on the fourth technical solution, wherein the simulated measurement part is determined by the following steps: edge diffusion is performed on each processing area in the integrated circuit layout to form a corresponding diffusion area; all diffusion areas are classified based on structural similarity to obtain diffusion area groups; at least one diffusion area is selected from each diffusion area group as the simulated measurement area, and the set of all simulated measurement areas forms the simulated measurement part.

[0016] The eighth technical solution is based on the seventh technical solution, wherein the structural similarity is rotational symmetry and / or mirror symmetry.

[0017] The ninth technical solution is based on the seventh technical solution, wherein the first path of the proposed survey part is the shortest among all the alternative proposed survey parts; the alternative proposed survey parts are a collection of alternative proposed survey areas, the alternative proposed survey parts cover all diffusion area groups and each diffusion area group has and only has a preset number of alternative proposed survey areas; the first path is the shortest one-way path in the alternative proposed survey part passing through the center points of all the alternative proposed survey areas.

[0018] The tenth technical solution is based on any one of the fourth to ninth technical solutions, wherein the second path of the detection area is the shortest among all the alternative detection areas; the alternative detection areas are a set of alternative representation areas, the alternative detection areas cover all local area groups and each local area group has and only has a preset number of alternative representation areas; the second path is the shortest one-way path in the alternative detection area passing through the center points of all the alternative representation areas.

[0019] The eleventh technical solution relates to a method for detecting a grain surface, which is used to detect a grain surface after CMP processing. The method performs shape detection on the grain surface based on a detection area formed by a detection area forming method as described in any one of the first to tenth technical solutions.

[0020] The twelfth technical solution is based on the eleventh technical solution, wherein AFM is used to perform shape detection on the surface of the grain.

[0021] The thirteenth technical solution is based on the twelfth technical solution, wherein the dimensions of the local area in the X-axis and Y-axis are less than or equal to the dimensions of the X-axis and Y-axis of the maximum imaging range of the AFM.

[0022] The fourteenth technical solution relates to a method for determining a hotspot on the surface of a grain, which is used to determine whether there is a hotspot on the surface of a grain after CMP processing. The method uses the grain surface detection method described in any one of the tenth to thirteenth technical solutions to perform a shape detection on the surface of the grain. If there is erosion of the functional part and / or depression of the characterization area in the detection results, it is determined that there is a hotspot on the surface of the grain.

[0023] Compared with the prior art, the above solution has the following beneficial effects:

[0024] Norma Rodriguez et al. pointed out in their paper "Hotspot Prevention Using CMP Model in Design Implementation Flow" that CMP planarization depends on the density and width of the metal portion in the local area on the grain. At different densities and widths, different degrees of dishing and erosion may occur.

[0025] Colin Hui et al. pointed out in their paper "Hotspot detection and design recommendation using silicon calibrated CMP model" that depression and erosion occur during the CMP process, forming depressions or erosions in the lower layer, resulting in copper residues or copper accumulation in the upper layer. Depression and erosion strongly depend on the density and line width of the design pattern. Wide features are more prone to depression than narrow features.

[0026] From the above papers, we can know that the functional part structure of the local area on the surface of the grain (reflected as a pattern structure on the integrated circuit layout) is a key factor affecting the hotspot distribution of the local area after the CMP process is completed. The structure of the functional part covers two important indicators: the density and line width of the design pattern. Of course, the local area without a functional part should also be regarded as a functional part structure. Local areas with different structures have different possibilities of hotspot formation during the CMP process.

[0027] The first technical solution is based on the rules revealed in the above paper, and uses characterization areas to characterize local areas with similar structures. This allows only a small number of characterization areas to be detected among a large number of local areas with similar structures to characterize the possibility of hot spots in these local areas, thereby greatly improving the detection compression ratio, providing a basis for shortening the detection time of grains, and being more conducive to the popularization and application of high-resolution detection methods on the production line.

[0028] In the first technical solution, the size of the local area is comparable to the size of the surface area where the depression may occur, and thus the detection of both depression and erosion hot spots can be covered.

[0029] In the first technical solution, the integrated circuit layout refers to the design layout of the integrated circuit on the die. In some processes, multiple CMP processes need to be performed. If each CMP process has a different integrated circuit topology, the integrated circuit layout refers to the design layout corresponding to the CMP process. If it is a planar semiconductor integrated circuit, the integrated circuit layout is the final design layout of the integrated circuit. The first technical solution uses the integrated circuit layout to select the detection area, which has the advantages of being more convenient and having a shorter detection time than forming a macroscopic image with low resolution during the detection process and then dividing the local area based on the macroscopic image.

[0030] The second and third technical solutions are optimizations for the size of the local area.

[0031] The fourth technical solution provides a specific method for determining the characterization area and then forming the detection area. In the fourth technical solution, the intended detection part covers all processing area groups, and all local areas cover the intended detection part, which can ensure that the set of local areas used for similarity classification can cover processing areas of all structural types.

[0032] In the fifth technical solution, adjacent local areas overlap with each other, and can traverse all functional part structures within the scope of the intended test part, further reducing the impact of factors related to the local area division (such as the starting position) on the detection results, which is conducive to the characterization area to better characterize the possibility of hot spots appearing.

[0033] The sixth technical solution is a specific implementation of the fifth technical solution.

[0034] In the seventh technical solution, the simulated test part is based on the diffusion area, which not only enables the simulated test part to include all types of processing areas, but also includes areas adjacent to the processing areas, and can more comprehensively express the structures of various local areas. Diffusion areas are classified based on structural similarity, and at least one diffusion area is selected from each diffusion area group as the simulated test area to jointly form the simulated test part, which can reduce the area of ​​the simulated test part, and use the repeatability of the macroscopic structure of the grain surface to reduce the number of local areas used for classification, while ensuring that the functional parts with different structural similarities are included in the simulated test part to avoid missed detection during the detection process.

[0035] In the eighth technical solution, the structural similarity is rotational symmetry and / or mirror symmetry, and therefore, can cover all different functional part structures.

[0036] In the ninth technical solution, the first path of the proposed test part is the shortest among all the alternative proposed test parts, which can provide a basis for shortening the final detection path as much as possible under the premise of traversing all diffusion area groups, thereby further shortening the detection time.

[0037] In the tenth technical solution, the detection area has the shortest second path among all candidate detection areas, which can shorten the final detection path as much as possible under the premise of traversing all local area groups, thereby further shortening the detection time.

[0038] The eleventh technical solution performs shape detection on the surface of the grain based on the detection area formed by the above-mentioned detection area forming method. It can utilize a high compression ratio to shorten the detection time of the grain, and is more conducive to the popularization and application of high-resolution detection methods on the production line.

[0039] The twelfth technical solution is a preferred implementation of the tenth technical solution. AFM is used to detect the surface shape of the grains. Compared with the use of white light interferometer, it can not only ensure the resolution, but also perform non-destructive detection of the grains. It is no longer necessary to apply material on the surface of the grains to avoid damaging the inspected grains.

[0040] In the thirteenth technical solution, the size of the local area is less than or equal to the size of the maximum imaging range of the AFM, so that when the AFM detects the characterization area, it only needs to move the probe without moving other parts, and the error of the detection field of view will not be caused by moving other parts. The correspondence between the detected characterization area in the actual detection process and the characterization area on the integrated circuit layout can be guaranteed.

[0041] In the fourteenth technical solution, the above-mentioned grain surface detection method is used to perform shape detection on the grain surface, and based on the detection results, it is determined whether there is a hot spot on the grain surface. Compared with the existing technology, it can more quickly determine whether a hot spot exists, which is more conducive to the popularization of high-resolution detection methods on the production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solution of the embodiment, the following briefly introduces the drawings required for use:

[0043] Figure 1 It is a schematic diagram of a partial longitudinal section of a wafer before and after the CMP process;

[0044] Figure 2 A schematic diagram of a local longitudinal section of a wafer containing an erosion hotspot;

[0045] Figure 3 A schematic diagram of a partial longitudinal section of a wafer containing a recessed hot spot;

[0046] Figure 4 The integrated circuit layout in the first embodiment is shown;

[0047] Figure 5 The diffusion region in the first embodiment is shown;

[0048] Figure 6 The diffusion area group in Example 1 is shown;

[0049] Figure 7 The part to be measured in the first embodiment is shown;

[0050] Figure 8 The local area in the first embodiment is shown;

[0051] Fig. 9 The local area clusters in the first embodiment are shown;

[0052] Fig.10 The detection path in the first embodiment is shown;

[0053] Fig.11 The detection area and detection path in the first embodiment are shown;

[0054] Fig.12The integrated circuit layout in the second embodiment is shown;

[0055] Fig.13 It shows the typical local areas of the local area group in the second embodiment;

[0056] Fig.14 The detection path in the second embodiment is shown. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments will be described clearly and completely below in conjunction with the accompanying drawings.

[0058] Embodiment 1

[0059] See also Figure 4 , Figure 4 The integrated circuit layout in the first embodiment is shown. In the embodiment, the integrated circuit layout refers to the design layout of the integrated circuit corresponding to the grain. In some processes, multiple CMP processes need to be performed. If each CMP process has a different integrated circuit topology, the integrated circuit layout refers to the design layout corresponding to the CMP process. If it is a planar semiconductor integrated circuit, the integrated circuit layout is the final design layout of the integrated circuit. Using the integrated circuit layout to select the detection area has the advantages of being more convenient and having a shorter detection time than forming a macroscopic image with low resolution during the detection process and then dividing the local area based on the macroscopic image.

[0060] like Figure 4 As shown, in this embodiment, the integrated circuit board Figure X The size in the X-axis direction is 12.5 microns, and the size in the Y-axis direction is 8.5 microns. In this embodiment, there are 6 processing areas on the integrated circuit layout, namely A1 to A6. Each processing area only includes functional parts. That is, in this embodiment, the processing area does not include non-functional parts. The functional part in this embodiment is a metal part, specifically a copper layer. Among them, A1, A3, A4 and A6 have the same size, the size in the X-axis direction is 2 microns, and the size in the Y-axis direction is 3 microns; A2 and A5 have the same size, the size in the X-axis direction is 6 microns, and the size in the Y-axis direction is 3 microns. The distance between two adjacent processing areas along the X-axis direction is 0.25 microns, and the distance between two adjacent processing areas along the Y-axis direction is 0.5 microns. The distance from the processing area to the adjacent edge of the integrated circuit layout is 1 micron along the X-axis direction and 1 micron along the Y-axis direction.

[0061] In this embodiment, based on the integrated circuit layout, the detection area E and the detection path F are formed through the following steps.

[0062] Step 1: Perform edge diffusion on each processing area in the integrated circuit layout to form a diffusion area. The simulated measurement part is based on the diffusion area, which not only enables the simulated measurement part to include all types of processing areas, but also includes areas adjacent to the processing areas, and can more comprehensively express the structures of various local areas. Figure 4 and Figure 5 . Figure 4 and Figure 5 FIG. 2 shows the diffusion region in this embodiment. Figure 4 As shown, Figure 4 The red frame in the middle is the diffusion area formed after the edge of the processing area A1 diffuses. In this embodiment, the distance of each edge of the processing area diffusing outward is 1 micron. The specific method of setting the diffusion distance will be described in detail later. Figure 5 As shown, the processing areas A1 to A6 form corresponding diffusion areas B1 to B6 after edge diffusion;

[0063] Step 2: Classify all diffusion regions based on rotational symmetry and / or mirror symmetry to obtain diffusion region groups; in this embodiment, the diffusion region groups are classified by structural similarity, specifically rotational symmetry and mirror symmetry. In other embodiments, it may also be rotational symmetry or mirror symmetry. In some cases where the accuracy of the results is not high, classification may also be based on structural similarity indicators such as Hamming distance. Classifying the diffusion regions based on rotational symmetry and / or mirror symmetry, and selecting at least one diffusion region in each diffusion region group as the proposed test region to jointly form a proposed test part can reduce the area of ​​the proposed test part, and use the repeatability of the macroscopic structure of the grain surface to reduce the number of local areas used for classification, while ensuring that all functional part structures are included in the proposed test part C to avoid missed detection during the detection process. See for details. Figure 6 . Figure 6 The diffusion area groups in this embodiment are shown. Figure 6 As shown, in this embodiment, the number of diffusion region groups is two, namely BC1 and BC2. Among them, the diffusion region group BC1 is a set of diffusion regions B1, B3, B4 and B6, wherein B1 and B3 are mirror-symmetric along the X-axis direction, B1 and B4 are mirror-symmetric along the Y-axis direction, and B1 and B6 are 180 degrees rotationally symmetric to each other. The diffusion region group BC2 is a set of diffusion regions B2 and B5, wherein B2 and B5 are mirror-symmetric along the Y-axis direction. Any diffusion region in the diffusion region group BC1 is not mirror-symmetric or rotationally symmetric or mirror-symmetric plus rotationally symmetric with any diffusion region in the diffusion region group BC2;

[0064] Step 3: At least one diffusion area is selected from each diffusion area group to form a simulated test area, and the set of all simulated test areas forms a simulated test part C; in this embodiment, one diffusion area is selected from each diffusion area group to form a simulated test area. In this embodiment, the simulated test part C has the shortest first path among all the candidate simulated test parts. The alternative simulated test part is a set of alternative simulated test areas, and the alternative simulated test part covers all diffusion area groups and each diffusion area group has and only has a preset number of alternative simulated test areas; the first path is the shortest one-way path through the center point of all the alternative simulated test areas in the alternative simulated test part. The simulated test part C has the shortest first path among all the alternative simulated test parts, which can provide a basis for shortening the final detection path F as much as possible under the premise of traversing all diffusion area groups, thereby shortening the detection time. In this embodiment, the preset number is 1. Specifically, this embodiment includes 8 alternative simulated test parts, each of which is a set of two alternative simulated test areas, and the two alternative simulated test areas are from two diffusion area groups respectively. The 8 alternative proposed measurement parts are respectively the set of B1 and B2, the set of B1 and B5, the set of B3 and B2, the set of B3 and B5, the set of B4 and B2, the set of B4 and B5, the set of B6 and B2, and the set of B6 and B5. In this embodiment, the first path is the path between the center points of two alternative proposed measurement areas in the alternative proposed measurement part. When the number of diffusion area groups exceeds 3, the first path is the shortest one-way path passing through the center points of all alternative proposed measurement areas in the alternative proposed measurement part. According to the above rules, the first path of the alternative proposed measurement part including B1 and B2 is the shortest. Of course, the first path of the alternative proposed measurement part including B3 and B2, the alternative proposed measurement part including B4 and B5, and the alternative proposed measurement part including B5 and B6 is also the shortest. Any one of the above four alternative proposed measurement parts can be selected as the proposed measurement part C. See. Figure 7 , Figure 7 FIG. 4 shows the simulated detection portion C in this embodiment. Figure 7 As shown, in this embodiment, the intended test portion C is a set of the intended test areas B1 and B2. It should be noted that in other embodiments, the intended test portion C may also be a portion of the integrated circuit layout covering all processing area groups. The intended test portion may also be the entire integrated circuit layout. Of course, if the entire integrated circuit design layout is selected as the intended test portion C, the calculation time for forming the detection area E and the detection path F will be longer, so it is inefficient;

[0065] Step 4: Divide the area of ​​the intended measurement part C into several local areas, and all local areas cover the intended measurement part C; all local areas cover the intended measurement part C, which can ensure that the set of local areas used for similarity classification can cover the processing areas of all structural types. In this embodiment, adjacent local areas overlap with each other. In other embodiments, adjacent local areas may also be adjacent to each other. Adjacent local areas overlap with each other, and all functional part structures within the scope of the intended measurement part can be traversed, which further reduces the influence of factors related to the division of local areas (such as the starting position) on the detection results, which is conducive to the characterization area to better characterize the possibility of hot spots.

[0066] See also Figure 7 and Figure 8 , Figure 7 and Figure 8 FIG. 2 shows a local area in this embodiment. Figure 7 As shown in FIG. 1 , in this embodiment, the size of the local area along the X-axis direction is 2 microns, and the size along the Y-axis direction is also 2 microns. Specifically, in this embodiment, the local area is moved along the measured portion C as shown in FIG. Figure 7 The virtual window of the red frame shown in is obtained, and the size of the virtual window is the same as the size of the local area. When sliding the virtual window, the step value is 1 micron each time. In order to have a higher detection efficiency, the size of the local area can be set to between 0.5-50 microns, more preferably between 10 microns and 30 microns, and more preferably to 20 microns. It should be noted that the size of the local area may or may not be correlated with the setting of the aforementioned diffusion distance. In order to improve efficiency, generally the diffusion distance can be set to be less than or equal to the size of the local area. More preferably, the diffusion distance can be set to the step value when sliding the window, that is, the size of the overlapping part of adjacent local areas. As Figure 8 As shown, according to the above method of dividing local areas, the number of local areas in this embodiment is 40, from D1 to D40;

[0067] Step 5: Classify all local regions based on the similarity of the functional part structures in the local regions to obtain local region groups. The similarity of the functional part structures should at least include the similarity of the proportion of the functional part to the area of ​​the local region and the similarity of the line width of the functional part. The so-called structure of the functional part refers to the shape of the functional subdomain if the functional part only contains one functional subdomain, the shape of each functional subdomain and the positional relationship between the functional subdomains if the functional part contains more than two functional subdomains, and if the local region does not contain a functional part, it should be divided into a separate local region group. In the classification based on the similarity of the functional part structure, one classification method that may lead to a large number of local region groups is classification based on rotational symmetry and / or mirror symmetry, that is, only local regions with rotational symmetry and / or mirror symmetry can be classified into a local region group. The present embodiment adopts a classification method based on rotational symmetry and / or mirror symmetry. On this basis, local regions with rotational symmetry and / or mirror symmetry after a slight translation can also be classified into the same local region group. Specifically, see Fig. 9 , Fig. 9 The eight local area groups formed based on the above classification method in this embodiment are shown. Among them, the local area group DC1 includes the local area D1; the local area group DC2 includes the local areas D2, D5 to D9, D11 and D12; the local area group DC3 includes the local areas D3, D4 and D10; the local area group DC4 includes the local areas D12, D15 to D19, D22, D25 to D29; the local area group DC5 includes the local areas D13, D14, D20, D23, D24 and D30; the local area group DC6 includes the local areas D33, D34 and D40; the local area group DC7 includes the local areas D32 and D35 to D39; the local area group DC8 includes the local area D31;

[0068] Step 6: Each local area group selects at least one local area to form a representation area, and the set of all representation areas forms a detection area E; in this embodiment, the detection area E has the shortest second path among all the candidate detection areas; the alternative detection area is a set of candidate representation areas, and the alternative detection areas cover all local area groups and each local area group has and only has a preset number of candidate representation areas; the second path is the shortest one-way path in the alternative detection area passing through the center point of all the alternative representation areas. The detection area has the shortest second path among all the alternative detection areas, which can shorten the final detection path as much as possible while traversing all the local area groups, thereby further shortening the detection time. In this embodiment, the preset number is 1. See. Fig.10 and Fig.11 , Fig.10 and Fig.11 The detection area E and the detection path F are shown. Fig.10 and Fig.11 As shown, in this embodiment, the detection area E includes 8 representation areas, namely D1, D2, D3, D12, D23, D33, D32 and D31. The 8 representation areas cover all 8 local area groups and each local area group has only one representation area. In this embodiment, the number of candidate representation areas is very large, and can be selected from Fig.10 It can be seen intuitively that the second path of the detection area E ( Fig.10 The red path in ( ) is the shortest. Fig.11 As shown, the detection path F is obtained based on the second path of the detection area E. Specifically, in this embodiment, optionally, the detection path F can be D1 through D2, D3, D12, D23, D33, D32 to D31 in sequence.

[0069] In this embodiment, the above-mentioned detection area formation method can be implemented by executing a computer program. In the computer program, the algorithm for generating diffusion area groups from each diffusion area and the algorithm for generating local area groups from each local area can be a combination of one or more of a support vector machine (SVM), a K nearest neighbor algorithm (KNN), a random forest, a convolutional neural network (CNN), a transfer learning algorithm, and a reinforcement learning algorithm. In this embodiment, a support vector machine (SVM) is selected. The algorithm for generating the simulated measurement part based on each diffusion area group and the algorithm for generating the detection area E based on each local area group can be a combination of one or more of a genetic algorithm, a Dijkstra algorithm, a grid algorithm, a particle swarm optimization (PSO) algorithm, and an ant colony optimization (ACO) algorithm. Specifically, this embodiment selects a genetic algorithm.

[0070] This embodiment also discloses a detection device, which includes a detection device, a computing device and a control device. The detection device is an atomic force microscope (AFM), which is used to perform shape detection on the grains after CMP processing and output the shape detection results. Compared with the use of white light interferometer, the use of AFM to perform surface shape detection on the grains can not only ensure the resolution, but also perform non-destructive detection on the grains, and no longer need to apply materials on the surface of the grains to avoid damaging the inspected grains. In this embodiment, the X-axis direction size of the maximum imaging range of the AFM is 2 microns, and the Y-axis direction size is 2 microns. Therefore, in this embodiment, the size of the local area is equal to the size of the maximum imaging range of the AFM, and in other embodiments, the size of the local area can be set to be smaller than the size of the maximum imaging range of the AFM. The size of the local area is less than or equal to the size of the maximum imaging range of the AFM, so that when the AFM detects the characterization area, it only needs to move the probe without moving other parts, and the error of the detection field of view will not be caused by moving other parts. It can ensure the correspondence between the characterization area detected in the actual detection process and the characterization area on the integrated circuit layout. The computing device includes an input unit, an output unit, a storage unit and a computing unit. The input unit is used to obtain the integrated circuit layout, the output unit is used to output the detection area and the detection path, the storage unit is used to store the above-mentioned computer program, and the computing unit is used to call and execute the computer program to generate the detection area and the detection path based on the integrated circuit layout. The control device is used to obtain the detection area and the detection path from the computing device to control the detection device to perform shape detection on the grains according to the detection path within the detection area. In this embodiment, the AFM realizes the shape detection in the selected local area by moving the probe, and the movement of the probe can adopt a raster scanning mode or a line scanning mode. The present embodiment adopts a raster scanning mode. In this embodiment, the movement of the AFM between the characterization areas is mainly realized by the displacement of other parts except the probe, such as relying on the movement of the cantilever or the support arm or other mechanism equipped with the probe along the X-axis and the Y-axis.

[0071] This embodiment also discloses a method for determining a hot spot on a grain surface, which is used to determine whether a hot spot exists on the surface of a grain after CMP processing. Specifically, the method for determining a hot spot on a grain surface determines whether there is erosion of a functional part and / or depression of a characterization area based on the detection results output by the above-mentioned detection device. Optionally, in each characterization area, if the height of the functional part is lower than the height of the semiconductor substrate and the height difference exceeds a first threshold, it should be determined that there is erosion of the functional part. The first threshold can be set to 10 nanometers, or it can be specifically set according to process requirements. If the overall height of the characterization area with a functional part is lower than the overall height of the characterization area without a functional part or with less functional parts, and the height difference exceeds a second threshold, it should be determined that there is a depression in the characterization area. The second threshold can be set to 20-50 nanometers, or it can be specifically set according to process requirements. As long as there is any one of erosion or depression, it should be determined that there is a hot spot on the grain surface. The above-mentioned grain surface detection method is used to perform shape detection on the grain surface, and based on the detection results, it is determined whether there is a hot spot on the grain surface. Compared with the prior art, it can more quickly determine whether a hot spot exists, which is more conducive to the popularization and application of high-resolution detection methods on the production line.

[0072] In this embodiment, the total area of ​​the grain is 106.25 square microns, and the detection area E includes 8 characterization areas, each of which has an area of ​​4 square microns, so the area of ​​the detection area is 32 square microns, and the detection compression ratio is 3.32. Of course, this embodiment is only for illustrating the method of this patent. In actual applications, due to the high repeatability of the processing area and the high similarity of the local area, the actual compression ratio is much greater than the compression ratio of this embodiment.

[0073] Embodiment 2

[0074] See also Fig.12 , Fig.12 FIG. 1 is a schematic diagram of the integrated circuit layout of the crystal grain in the second embodiment. Fig.12As shown, the integrated circuit layout in the second embodiment includes two types of processing areas, which are arranged along the X-axis and repeated as a repeating unit. The entire integrated circuit layout has 48 repeating units along the X-axis direction, and the entire integrated circuit layout has 96 repeating units along the Y-axis direction. In a repeating unit, the size of the processing area on the left is 10 microns in the X-axis direction, and the size in the Y-axis direction is 10 microns. There are both functional parts and non-functional parts in the processing area. The functional part is lighter in color and is the copper layer, and the non-functional part is darker in color and is the semiconductor substrate. The functional part includes 8 square rings nested in each other, each ring has a line width of 0.3 microns, and the rings are spaced 0.3 microns apart from each other; the size of the processing area on the right is 9.9 microns in the X-axis direction, and the size in the Y-axis direction is 10 microns. There are also both functional parts and non-functional parts in the processing area. The functional part is lighter in color and is the copper layer, and the non-functional part is darker in color and is the semiconductor substrate. There is only one functional part, and several square non-semiconductor substrate sub-domains are sandwiched therein. The semiconductor substrate sub-domains are square, and the sizes in the X-axis direction and the Y-axis direction are both 0.3 microns, and the spacing between them along the X-axis direction and the spacing along the Y-axis direction are also 0.3 microns. In each repeating unit, the interval between the left and right processing areas is 0.3 microns; along the X-axis direction, the interval between the two repeating units is 0.3 microns; along the Y-axis direction, the interval between the two repeating units is 0.3 microns. Based on the above dimensions, in Example 2, the area of ​​each repeating unit is 20120 square microns, there are 4608 repeating units in total, and the area of ​​the entire integrated circuit layout is approximately 92712960 square microns.

[0075] The detection area and detection path of Example 2 can be obtained in the same manner as Example 1. The difference from Example 1 is that in Example 2, the diffusion distance is 5 microns, the size of the local area in the X-axis direction is 10 microns, the size in the Y-axis direction is 10 microns, and the step value is 5 microns. In this embodiment, the method of obtaining local area groups by using structural similarity is to classify local areas with rotational symmetry and / or mirror symmetry after a small translation into the same local area group. See Fig.13 , Fig.13 The representative areas of each local area group in this embodiment are shown. Fig.13 As shown, in this embodiment, there are 11 local region groups, namely DC01 to DC11, and the characterization area in each local region group is as follows Fig.13 shown. Fig.14 FIG. 2 shows the detection path in this embodiment. Fig.14 As shown, the detection path in this embodiment is indicated by a red line. The corresponding detection area is also determined by the set of characterization areas.

[0076] The computing device for outputting the detection area and detection path in the second embodiment is basically the same as that in the first embodiment, and the detection equipment for performing shape detection on the grains after CMP processing is also basically the same as that in the first embodiment, except that the maximum imaging range of the AFM is 10 microns × 10 microns.

[0077] The method for detecting the grain surface and the method for determining the hot spots on the grain surface in the second embodiment are the same as those in the first embodiment.

[0078] In Example 2, the detection area includes 11 characterization areas, each characterization area has an area of ​​100 microns, the detection area has an area of ​​1100 square microns, and the detection compression ratio is 84284.5. Therefore, the detection area formation method of this patent can achieve a large detection compression ratio, which is conducive to greatly shortening the detection time and is conducive to the popularization and application of high-resolution detection methods on the production line.

[0079] It can be seen from the above embodiments that characterization areas are used to characterize local areas with similar structures, so that among a large number of local areas with similar structures, only a small number of characterization areas need to be detected to characterize the possibility of hot spots in these local areas, thereby greatly improving the detection compression ratio, shortening the detection time for grains, and being more conducive to the popularization and application of high-resolution detection methods on the production line.

[0080] The above embodiment uses the integrated circuit layout to select the detection area, which has the advantages of being more convenient and having a shorter detection time compared to forming a macro image with a low resolution and then dividing the local area based on the macro image during the detection process.

Claims

1. A method for forming a detection area, which is used to determine a detection area when performing surface detection on a crystal grain after CMP processing, characterized in that: Selecting characterization areas on the integrated circuit layout and forming a detection area from each characterization area; the characterization area is used to characterize a local area with a similar structure, and the dimensions of the local area on the X-axis and the Y-axis are in the range of 0.5 microns to 50 microns.

2. The detection area forming method according to claim 1, characterized in that: The dimensions of the local area in the X-axis and the Y-axis are in the range of 10 micrometers to 30 micrometers.

3. The detection area forming method according to claim 2, characterized in that: The dimensions of the local area in the X-axis and the Y-axis are 20 micrometers.

4. The detection area forming method according to claim 1, characterized in that: It includes the following steps: Dividing the portion to be tested on the integrated circuit layout into a plurality of local regions, wherein the portion to be tested covers all processing region groups, each processing region group includes processing regions with similar structures, and all local regions cover the portion to be tested; Classify all local regions based on structural similarity to obtain local region groups; Each local region cluster selects at least one local region as a representation region, and the set of all representation regions forms a detection region.

5. The detection area forming method according to claim 4, characterized in that: Adjacent local areas overlap each other.

6. The detection area forming method according to claim 5, characterized in that: The local area is obtained by sliding a virtual window on the part to be measured.

7. The detection area forming method according to claim 4, characterized in that: Determine the part to be measured by the following steps: Perform edge diffusion on each processing area in the integrated circuit layout to form a corresponding diffusion area; All diffusion areas were classified based on structural similarity to obtain diffusion area groups; At least one diffusion area is selected from each diffusion area group as a simulated measurement area, and the collection of all simulated measurement areas forms a simulated measurement part.

8. The detection area forming method according to claim 7, characterized in that: The structural similarity is rotational symmetry and / or mirror symmetry.

9. The detection area forming method according to claim 7, characterized in that: The first path of the proposed measured part is the shortest among all the alternative proposed measured parts; the alternative proposed measured parts are a collection of alternative proposed measured areas, the alternative proposed measured parts cover all diffusion area groups and each diffusion area group has and only has a preset number of alternative proposed measured areas; the first path is the shortest one-way path in the alternative proposed measured parts passing through the center points of all the alternative proposed measured areas.

10. The detection area forming method according to any one of claims 4 to 9, characterized in that: The detection area has the shortest second path among all candidate detection areas; the candidate detection area is a set of candidate representation areas, the candidate detection area covers all local area groups and each local area group has and only has a preset number of candidate representation areas; The second path is the shortest one-way path in the candidate detection area passing through the center points of all the candidate representation areas.

11. A method for detecting a grain surface, which is used to detect a grain surface after CMP processing, characterized in that: Based on the inspection region formed by the inspection region forming method according to any one of claims 1 to 10, the surface of the crystal grain is inspected for shape.

12. The method for detecting the grain surface according to claim 11, characterized in that: AFM is used to detect the shape of the surface of the grain.

13. The method for detecting the grain surface according to claim 12, wherein: The dimensions of the local area in the X-axis and the Y-axis are smaller than or equal to the dimensions of the maximum imaging range of the AFM in the X-axis and the Y-axis.

14. A method for determining hot spots on a grain surface, which is used to determine whether there are hot spots on a grain surface after CMP processing, and is characterized in that: The grain surface detection method according to any one of claims 10 to 13 is used to perform shape detection on the grain surface. If erosion of the functional part and / or depression of the characterization area are found in the detection results, it is determined that there is a hot spot on the grain surface.

Citation Information

Patent Citations

  • Voltage contrast based fault and defect inference in logic chips

    CN107533103A

  • CMP defect detection method and device, storage medium and processor

    CN110851946A

  • Wafer back surface solid particle detection method

    CN111128780A

  • Polycrystal layer photoetching process hot spot detection method based on image processing

    CN117173176A

  • Photoetching hot spot detection method based on 2D overlapping judgment

    CN117891143A

Cited By

  • Detection area formation method, grain surface detection method, and grain surface hotspot determination method

    WO2026036631A1