Hot spot avoidance method for manufacturing integrated circuits

Abstract

The present disclosure relates to a hotspot avoidance method for manufacturing integrated circuits. A method includes: cutting multiple images from a layout of an integrated circuit; generating a first plurality of hash values, each of which is from one of the multiple images; loading a second plurality of hash values ​​stored in a hotspot library; and comparing each of the first plurality of hash values ​​with each of the second plurality of hash values. The comparison step includes calculating a similarity value between each of the first plurality of hash values ​​and each of the second plurality of hash values. The method also includes comparing the similarity value with a predetermined threshold similarity value, and in response to a result that the similarity value is greater than the predetermined threshold similarity value, recording a position of a corresponding image having the result. The position is the position of the corresponding image in the layout.

Description

Technical Field

[0001] The present disclosure relates to hot spot avoidance methods for fabricating integrated circuits. Background Art

[0002] In the manufacture of integrated circuits, process-related defects such as topological hotspots may be found after the corresponding process is completed by taking physical measurements from the manufactured wafers. For example, to find defects related to chemical mechanical polishing (CMP), several stages must be performed (including the circuit design stage, the circuit layout stage, manufacturing and performing CMP on the physical wafer, and measuring the physical wafer) to find topological defects. This process usually takes three months. Summary of the invention

[0003] According to an embodiment of the present disclosure, a method for manufacturing a semiconductor device is provided, comprising: cutting out a plurality of images from a layout of an integrated circuit; generating a first plurality of hash values, each of which is from one of the plurality of images; loading a second plurality of hash values ​​stored in a hotspot library; comparing each of the first plurality of hash values ​​with each of the second plurality of hash values, wherein the comparison comprises calculating a similarity value between each of the first plurality of hash values ​​and each of the second plurality of hash values; comparing the similarity value with a predetermined threshold similarity value; and in response to a result that the similarity value is greater than the predetermined threshold similarity value, recording a position of a corresponding image having the result, wherein the position is a position of the corresponding image in the layout.

[0004] According to another embodiment of the present disclosure, a method for manufacturing a semiconductor device is provided, comprising: cutting out a plurality of images from a layout of an integrated circuit; generating a plurality of hash values, each hash value being from one of the plurality of images; searching from a hotspot library to find similar hash values ​​similar to the plurality of hash values, wherein the hotspot library stores hash values ​​indexed to images having hotspots; and marking positions of some of the plurality of images associated with the similar hash values ​​on the layout of the integrated circuit.

[0005] According to another embodiment of the present disclosure, a system for manufacturing a semiconductor device is provided, comprising: a library stored in a tangible medium, the library comprising a plurality of entries, each entry comprising: a hash value; an image associated with the hash value, wherein the image comprises a hotspot; a recipe configured to reduce the hotspot; and topological information of the hotspot. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Various aspects of the present disclosure may be best understood from the following detailed description when read in conjunction with the accompanying drawings. Note that, in accordance with standard industry practice, various features are not drawn to scale. In fact, the size of various features may be arbitrarily increased or reduced for clarity of discussion.

[0007] Figures 1 to 4 A cross-sectional view of a structure on which a chemical mechanical polishing process is performed and the results are shown in accordance with some embodiments.

[0008] Figure 5 A schematic flow in the design and fabrication of an integrated circuit according to some embodiments is shown.

[0009] Figure 6 A process flow for constructing a hotspot library according to some embodiments is shown.

[0010] Figure 7 An example image and generated hash value are shown in accordance with some embodiments.

[0011] Figure 8 A schematic diagram of an example wafer with hot spots is shown in accordance with some embodiments.

[0012] Fig. 9 An example cropped image is shown in accordance with some embodiments.

[0013] Fig.10 Grouping of hash values ​​is shown in accordance with some embodiments.

[0014] Fig.11 and Fig.12 The represented areas in the cropped image are shown with different pattern densities and line widths according to some embodiments.

[0015] Fig.13 A process flow for determining possible hotspots using a hotspot library is shown according to some embodiments.

[0016] Fig.14 Cropping of a layout into a plurality of cropped images is shown in accordance with some embodiments.

[0017] Fig.15 A process flow for finding a recipe corresponding to a hotspot is shown according to some embodiments.

[0018] Fig.16 A graphical representation of a search recipe is shown in accordance with some embodiments.

[0019] Fig.17 An example recipe according to some embodiments is shown.

[0020] Fig.18 A process of improving a hotspot prevention model and improving a recipe according to some embodiments is shown.

[0021] Fig.19 A system for performing a task according to some embodiments is shown. DETAILED DESCRIPTION

[0022] The following disclosure provides many different embodiments or examples for realizing different features of the present invention. Specific examples of components and arrangements are described below to simplify the present disclosure. Of course, these are merely examples and are not intended to be limiting. For example, in the following description, forming a first feature above or on a second feature may include an embodiment in which the first feature and the second feature are formed in direct contact, and may also include an embodiment in which additional features may be formed between the first feature and the second feature so that the first feature and the second feature may not be in direct contact. In addition, the present disclosure may repeat reference numerals and / or letters in various examples. This repetition is for the purpose of simplicity and clarity, and does not itself indicate the relationship between the various embodiments and / or configurations discussed.

[0023] Additionally, spatially relative terms (e.g., "below," "beneath," "below," "above," "upper," etc.) may be used herein to facilitate describing the relationship of one element or feature illustrated in the figures relative to another element(s) or feature(s). These spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. The device may be otherwise oriented (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein may likewise be interpreted accordingly.

[0024] According to some embodiments, a hotspot avoidance method for manufacturing integrated circuits is provided. According to some embodiments, a system and process for predicting hotspots and using the predicted hotspots to find the best recipe are shown. Some variations of some embodiments are discussed. The embodiments discussed herein will provide examples to enable the subject matter of the present disclosure to be performed or used, and those of ordinary skill in the art will readily understand the modifications that can be made while remaining within the intended scope of the different embodiments. Throughout the various views and illustrative embodiments, similar reference numbers are used to indicate similar elements. Although method embodiments may be discussed as being performed in a particular order, other method embodiments may be performed in any logical order.

[0025] Throughout this specification, the term "hot spot" refers to defects generated during the integrated circuit manufacturing process, rather than design-related defects. In other words, the term "hot spot" refers to process-related defects. An example of a hot spot is a defect generated during a chemical mechanical polishing (CMP) process, which will be referred to in detail in the following sections. Figures 1 to 4A detailed discussion will be given, and hot spots may also refer to other types of defects, including but not limited to defects related to the etching process (for example, portions that are intended to be removed but are not removed during the etching process), defects related to the deposition process, etc.

[0026] Figures 1 to 4 The deposition of some features, the CMP process, and several possible results due to CMP are shown. It should be understood that Figures 1 to 4 An example structure of a CMP process for forming a metal line is shown, but an actual CMP process may be applied to a different structure, a different material, etc. Figure 1 , a wafer 10 is provided. The wafer 10 includes a base layer 20, which may include, for example, a silicon substrate and overlying structures and layers, and the details are not shown. A plurality of trenches may be formed in a dielectric layer extending into the base layer 20. A deposition process is then performed to deposit a glue layer 22, and a metal material 24 is deposited over the glue layer 22. According to some embodiments of the present disclosure, the glue layer 22 may be formed of, or include, titanium, titanium nitride, tantalum, tantalum nitride, etc. The filling material 24 may include copper, aluminum copper, etc. Due to the topology of the trenches in the base layer 20, the deposited metal material 24 has a non-planar top surface that can reflect the topology of the base layer 20. A CMP process is performed to remove excess portions of the metal material 24, producing a plurality of conductive features 26 (including 26A and 26B), which may include metal lines, vias, contact plugs, etc., such as Figure 2 , Figure 3 or Figure 4 shown.

[0027] Due to various factors, such as the topology of the substrate layer 20 (eg, the density and width of the trenches), and the recipe of the CMP process, different results may be achieved, such as Figure 2 , Figure 3 or Figure 4 As shown. Throughout the specification, the term "recipe" refers to a set of process conditions, such as the number of (sub)steps in a CMP process, the type of slurry, the flow rate of the slurry, the down force of the wafer relative to the polishing pad, trimming, rotation speed, etc. Therefore, a recipe includes a fixed set of process conditions. When one or more of the process conditions in a recipe are changed, it can be considered that another recipe is generated.

[0028] Figure 2 shows the ideal situation to be achieved. Figure 2 In the embodiment of the present invention, the top surfaces of all resulting conductive features 26 (including 26A and 26B) are coplanar regardless of the width and pattern density of conductive features 26 .

[0029] Figure 3The actual situation is shown, which is not ideal, but still acceptable. Due to the pattern loading effect, the portion of the metal material 24 with a higher density and / or a larger width is polished more than the portion of the metal material 24 with a lower density and / or a smaller width, thereby producing a dishing effect, generating grooves. The depth D1 of the grooves is less than the design specification, and therefore no hot spots are generated. For example, the design specification may require that the dishing depth is less than about 10nm. Because the depth D1 of all grooves is less than the specification, the result is acceptable and the grooves are not hot spots.

[0030] Figure 4 The example of generating hot spots when the grooves are wide and / or the pattern density of the grooves is high is shown. For example, the depression depth D2 of the wide grooves is greater than the design specification (e.g., 10 nm). These irregular grooves may cause problems for subsequent processes, which may include circuit shorts, circuit disconnections, etc., depending on the specific circuit design. Throughout the specification, irregular grooves are used as example hot spots to explain the concepts of the present disclosure. In addition, it should be understood that Figure 3 and Figure 4 Over-polishing in the CMP process is shown, while under-polishing where some portions are less polished (thus these portions are above the top surface of the base layer 20) may also occur, and hot spots may also be generated when the resulting hump is not standardized. Hot spots may cause production yield to decrease and need to be eliminated or at least reduced to within specifications.

[0031] Figure 5 A schematic flow in the design and manufacture of an integrated circuit according to some embodiments of the present disclosure is shown. First, a circuit design is provided (process 30), and the design may include a schematic diagram of the circuit. Next, a layout of the circuit is prepared (process 32). Based on the layout, a model provided according to an embodiment of the present disclosure is used to predict hot spots and the location of the hot spots in the layout (process 34). Fig.13 The process 300 in FIG. 1 shows the details of predicting hot spots. The generation, use, and improvement of the model are discussed in detail in the subsequent paragraphs. Throughout the specification, the model is referred to as the hot spot prevention model.

[0032] After the hot spots are predicted, a recipe that can generate the least number of hot spots (hereinafter referred to as the selected recipe) is selected based on the predicted hot spots (process 36), so that by performing CMP using the selected recipe, the number of hot spots is minimized and the severity (e.g., depths D1 and D2, such as Figure 4 shown) is minimized. Fig.15The details of selecting a recipe are shown in process 400 in FIG. The recipe is then used to fabricate the circuit on the wafer and is used to perform a CMP process on the physical wafer (process 38). It should be understood that until the point in time when the selected recipe is selected, no CMP process may be performed on any wafer that implements this particular layout. After the CMP process, the resulting polished wafer may be tested (process 40) to verify the presence and location of hot spots. The test results may also be used to improve the hot spot prevention model, which includes the improvement process in Fig.18 In process 500 shown.

[0033] In the following paragraphs, the process flow 200 for constructing and improving the hotspot library is discussed in detail. Figure 6 ), process flow 300 of using a hotspot library to predict hotspots on a circuit layout ( Fig.13 ), and a process flow 400 for suggesting a selected recipe ( Fig.15 ). Together, these processes provide a solution for predicting and eliminating (or at least minimizing) hot spots without actually performing a process (e.g., a CMP process) on a physical wafer.

[0034] refer to Figure 6 , a process flow 200 for generating and improving a hotspot library is provided. Referring to process 202, a training layout of a chip implementing a circuit is provided. The chip layout may be in the form of a Graphic Data System (GDS) format, or any other suitable format. Throughout the specification, the layout may alternatively be referred to as a GDS file. It should be understood that the training layout may be specifically used to generate a hotspot library, but not for mass production of a product, or the training layout may be a production layout that will be implemented on a production wafer.

[0035] Then, experimental wafers are fabricated to implement the training GDS. Figure 8 4 shows a schematic diagram of a corresponding wafer 42, which includes a plurality of chips 44, wherein a layout (training GDS) is implemented in each chip 44. After performing the CMP process, a test is performed to measure the surface topography (surface topography) of the wafer 42 ( Figure 6 The process 204 in FIG. 204 is performed to identify the hot spot 46 in the wafer 42. The location of the hot spot 46 in the wafer 42 is recorded, such as Figure 6 As shown in process 208 in . Because multiple hot spots 46 can be found in process 204, multiple locations in wafer 42 are recorded.

[0036] Next, if Figure 6 As shown in process 210 in FIG. 1 , for each hotspot 46 found, an image is cropped from the layout, which image may be in the form of a line GDS file. For example, Fig. 9 An example cropped image is shown. Assuming that hotspot 46 is found at location 48, an image surrounding hotspot 46 is cropped from the row layout data. The image may be rectangular, and may be square. The length L1 and width W1 of the cropped image are selected so that the surrounding environment around hotspot 46 is large enough to include surrounding features whose pattern and density may produce hotspot 46, but not so large as to include features that are not conducive to the generation of hotspot 46. For example, length L1 and width W1 may be in a range between about 64 μm and about 512 μm. Because single or multiple hotspots 46 may be found in process 204, single or multiple images may be cropped from the layout of wafer 42.

[0037] refer to Figure 6 In process 212, a hash value is generated from the cropped image. Figure 7 An example for generating a hash value from an image is shown. Because images cannot be indexed and cannot be searched, images are represented by a hash value, which is a unique digital representation of the image. The hash value and the image are in a one-to-one correspondence, so that the same image will generate the same hash value, and different images will generate different hash values. In addition, images that are similar to each other will generate similar hash values, and the similarity of the hash values ​​can be calculated. The similarity of the hash values ​​also represents the similarity of the images. For example, the similarity of the hash values ​​can be between 0 and 1, where a value of 0 indicates that the images are completely different from each other, and a value of 1 indicates that the images are the same. Existing algorithms and tools can be used to perform the generation of hash values ​​from images and the calculation of the similarity of hash values. For example, the discrete cosine transform (DCT) algorithm used by perceptual hashing (pHash) is a known available algorithm.

[0038] The hash value can be obtained by an intermediate value represented by a two-dimensional matrix, which is then converted into a hash value represented by a series of numbers and letters. For example, Figure 7 Three example images are shown, Image A, Image B, and Image C. The details of the images are not shown. Image A shows a person wearing thick clothes sitting on the snow with a tree in the snow. Image B is similar to Image A, except that Image B has been equalized from Image A, adjusting the color and contrast. Image C shows a person's face with goggles on the forehead and flames surrounding the face. On the right side of each of Image A, Image B, and Image C, an 8×8 two-dimensional matrix is ​​provided, which is generated from the corresponding image and / or two-dimensional matrix. Hash values ​​(which include numbers and letters) are shown on the right side of each matrix. Return to reference Figure 6 When a single or multiple hot spots 46 are found in process 204 , the single or multiple images are cropped, and a single or multiple hash values ​​are generated in process 212 .

[0039] refer to Figure 6In process 214, the plurality of hash values ​​are grouped into one or more groups by a grouping algorithm, wherein the grouping is performed based on the similarity of the hash values, and similar hash values ​​are grouped in the same hash group. Fig.10 Explain the example grouping algorithm. Fig.10 A plurality of circles shown in a two-dimensional space are shown for visually showing hash values, each circle representing a hash value generated from a cropped image. In the grouping algorithm, a plurality of hash values ​​(which include hash values ​​H1 to H13 shown as examples) are processed one by one. It is assumed that the order of processing is the sequence number of the hash value (e.g., 1 to 13). When processing the hash value H1, because there are no other hash values ​​and there is no previously generated hash group, a first hash group G1 is generated, and the hash value H1 is placed in the first group G1. The first placed hash value H1 is considered to be the center of the first group G1.

[0040] Next, the second hash value H2 is processed. A similarity value between the hash value H2 and the center of the group G1 (the center is H1) is calculated. Assuming that the similarity value is greater than a predetermined threshold similarity value, the hash values ​​H1 and H2 are considered to be similar to each other, and the hash value H2 belongs to the hash group G1. Throughout the specification, two hash values ​​whose similarity values ​​are greater than a predetermined threshold similarity value are referred to as similar hash values. Their corresponding images are also referred to as similar images. The hash value H2 is added to the hash group G1. According to some embodiments, the threshold similarity value is 0.9, and other values ​​may be used.

[0041] Assuming that the next processed hash value is H3, the similarity value between hash value H3 and the center H1 of hash group G1 is calculated. Further assuming that the similarity value is equal to or less than a predetermined threshold similarity value, hash values ​​H1 and H3 are considered to be dissimilar, and hash value H3 does not belong to hash group G1. Therefore, a second hash group G2 is generated, and hash value H3 is placed in hash group G2. Hash value H3 is the center of hash group G2.

[0042] In the subsequent process, each of the remaining hash values ​​H4 to H13 is processed one by one to calculate their similarities with the centers of existing hash groups (e.g., G1 and G2) so that it can be determined to which hash group the newly processed hash value belongs, or whether a new hash group should be generated. Fig.10 An example is shown in which hash value H12 is not similar to any of the centers (eg, H1 and H3), so that an additional hash group G3 is generated and hash value H12 is placed in hash group G3. Other hash values ​​H4-H11 and H13 are in hash groups G1 or G2.

[0043] refer to Figure 6In process 216, the center of each hash group is obtained, which can be the hash values ​​that are first placed in each hash group. After obtaining the center of the hash group, the non-center hash values ​​are discarded because each center is similar to and can represent other hash values ​​in its group. In other words, the cropped image represented by the discarded hash values ​​in the same hash group is similar to the cropped image represented by the center hash value of the hash group. The hash values ​​of the center hash values ​​in different hash groups are not similar to each other. Otherwise, if two center hash values ​​are similar to each other, the two center hash values ​​will have been placed in the same hash group, and as a result, only one of the hash values ​​will become the center and the other hash value will be discarded.

[0044] refer to Figure 6 In process 218, a hotspot library (which comprises a database) entry is formed for each hash value that is not discarded, and the hash value that is not discarded is the center of the hash group. According to some embodiments, multiple recipes are generated, such as Figure 6 As shown in process 206. Fig.18 The generation and improvement of multiple recipes are discussed. The multiple recipes may also include empirical recipes known to remove hot spots from certain types of images. Each recipe in the multiple recipes corresponds to Fig.18 The test GDS shown has its hash value. The hash value of the non-discarded center of the hash group is compared with the hash value of the GDS file corresponding to the recipe (by calculating the similarity value), and the recipe (the corresponding test GDS of the recipe is closest to the center hash value) is associated with each center hash value. Each center hash value will be associated with a recipe.

[0045] In addition to the recipe, the cropped image (from which the corresponding center hash value is generated) is associated with the center hash value. Similarly, as will be discussed in subsequent paragraphs, the expected topological information (e.g., whether the hot spot is under-polished or over-polished, and the depth of the depression or the height of the protuberance) is also associated with the recipe (as shown in reference Fig.18 will be discussed). Fig.18 The expected topological information can also be obtained in the process shown. Therefore, each hotspot library entry includes a hash value, a corresponding cropped image, a corresponding recipe, and a corresponding topological information. A plurality of hotspot library entries are generated. The index of the hotspot library entry can be a hash value. These hotspot library entries are stored in a database in the hotspot library 222, such as Figure 6 As shown in process 220.

[0046] Likewise Figure 6As shown, through the process in process flow 200, a hotspot prevention model 223 can be constructed and updated. Hotspot prevention model 223 incorporates the relationship between GDS files and hotspots discussed previously, and uses GDS files or cropped images (or their corresponding hash values) as input parameters, and outputs hotspots as output parameters.

[0047] Fig.11 and Fig.12 Schematically shows an example cropped image stored in the hotspot library together with the hash value. It should be understood that Fig.11 and Fig.12 The diagram is schematic, in which the outlines of some larger areas 52 are shown, and the outlines of some smaller areas are not shown. In addition, inside each area, there are multiple patterns such as parallel stripes, and the patterns in the illustrated areas 52 are not shown. The size, shape, and density of the patterns in the illustrated areas 52 may be different from each other. The different patterns, pattern densities, etc. that form the environment around the hotspot 46 are the cause of the hotspot. For example, in Fig.11 In the example, region 50 has a smaller line width that is much smaller than the line width of its surrounding region 52. Region 50 may also have a higher pattern density that is much higher than the pattern density of its surrounding region 52. This creates a hotspot 46. When similar images with similar environments are found in other GDS files, it is expected that hotspots may appear.

[0048] Fig.12 The cropped image saved in the hotspot library is shown. Similarly, the area 52 showing the outlines of some larger areas is shown, and the outlines of some smaller areas are not shown. In addition, inside each area 52, there are multiple patterns such as parallel strips that are not shown. The size, shape and density of the patterns in the shown area 52 can be different from each other, thereby generating corresponding hotspots 46.

[0049] In the previous discussion, it is assumed that at the beginning of the process flow 200, the hotspot library 222 has not yet been generated, and the hash groups and central hash values ​​have not been previously generated. Therefore, a new hash group will be generated, and the hotspot library will be generated from scratch. Once the hotspot library 222 is generated, the hotspot library 222 can be continuously improved using new training GDS files, which can be mass production GDS files, or GDS files that are specifically used for training purposes rather than production. The new GDS files will be repeated. Figure 6202, 204, 208, 210, 212 and 214 in the process. Accordingly, multiple new hot spots are found from the newly manufactured wafer implementing the new GDS file, and therefore, multiple new images are cropped. Then, multiple new hash values ​​are generated from the newly cropped images. Then, the newly generated hash values ​​are processed one by one, and the similarity of the newly generated hash values ​​to the existing central hash values ​​(stored in the hot spot library 222) is calculated. It can be understood that at this time, the stored central hash value is still the center of the hash group in theory, except that each hash group is a single-member group (non-central members have been discarded) with only one hash value (which is the center of the previously processed hash group). Calculate the similarity with each central hash value stored in the hot spot library 222 to determine whether the newly processed hash value belongs to an existing hash group. If the newly processed hash value belongs to one of the existing groups, because similar hot spots already exist in the hot spot library 222, the newly processed hash value and its corresponding cropped image, recipe, topology information, etc. will be discarded. If the newly processed hash value is not similar to any stored central hash value, the newly processed hash value and its corresponding cropped image, recipe, and topology information will be saved as a new entry in the hotspot library 222. Through this process, the hotspot library 222 can be improved.

[0050] Fig.13 A process flow 300 is shown, which is a process flow for determining possible hot spots in a new GDS file (new layout) using the hot spot library 222. Referring to process 302, a new GDS file (layout) is provided. Next, in process 304, the new GDS file is cropped into a plurality of cropped images, each of which has a size of L2×W2, such as Fig.14 For example, Fig.14 5 shows a design of a layout 55 that is divided into an array of images 56 having a length L2 and a width W2. According to some embodiments, the length L2 is equal to the length L1 ( Figure 8 ), and the width W2 is equal to the width W1. According to other embodiments, the length L2 may be greater than or less than the length L1 ( Figure 8 ), and width W2 can be greater than or less than width W1.

[0051] Referring to process 306, each cropped image 56 is processed to generate a hash value. The method of generating the hash value is similar to that of referring to Figure 6 The method discussed in process 212 in , and therefore will not be repeated.

[0052] refer to Fig.13In process 308, the (central) hash value stored in the hotspot library 222 is loaded into the computer and corresponding software. Each newly generated hash value is compared with each hash value loaded from the hotspot library 222 to compare their similarities, as shown in process 310. For example, when the newly generated hash value is similar to a hash value loaded from the hotspot library, it is determined that the newly generated hash value and the corresponding cropped image have been represented by similar hash values ​​stored in the hotspot library 222 (process 312). It is also determined that (one or more) hotspots may be generated in the corresponding cropped image. Therefore, the location of the corresponding image in the corresponding GDS file is marked (process 314). For example, when the cropped image is located from Fig.14 When the layout 55 in the array is partitioned in row 2 and column 3, the position (2, 3) will be marked. By comparing all newly generated (cropped image) hash values ​​with all hash values ​​in the library, the GDS 55 ( Fig.14 ) and marks the corresponding position of each hotspot in the corresponding GDS. The marked position can be used in the future. For example, after the CMP process, the part of the chip at the marked position is checked to determine whether the hotspot has been successfully eliminated or at least reduced.

[0053] On the other hand, if none of the newly generated hash values ​​are similar to any hash values ​​loaded from the hotspot library, then it is determined that the newly generated hash value and the corresponding cropped image are not represented by any hotspot library entry in the hotspot library 222 (process 312). In other words, no hash value is found from the GDS file, and the process can end (process 316). Therefore, the layout can be manufactured without paying attention to hotspots.

[0054] Fig.15 A process flow 400 is shown for determining and suggesting recipes for performing a CMP process to eliminate or at least reduce the amount of Fig.13 Referring to process 402, obtain the hot spots found in process flow 300 ( Fig.13) in process 314. If no hotspots are found, the process flow ends. If one or more hotspots are found, a comparison is performed to find a hotspot in the hotspot library 222 that is similar to the hotspot found. In order to perform the comparison, the hash values ​​stored in the hotspot library 222 are first loaded into the corresponding tools and computers, and the similarity value between the hash value of the hotspot found and the corresponding hash value stored in the hotspot library is calculated (by calculating the similarity value), as shown in process 406. Depending on the total number of hotspots found in the GDS, there may be one or more similarity values, each corresponding to a hotspot found. The hotspots (their hash values) are arranged according to the corresponding similarity values ​​(process 406), wherein hotspots with higher similarity values ​​have higher arrangements and higher priorities than hotspots with lower similarity values.

[0055] Referring to process 408 in process flow 400, for example, by indexing to the corresponding stored central hash value in hotspot library 222, a GDS file, topology information, and corresponding recipe corresponding to the arranged hotspot are found from hotspot library 222. The topology information is analyzed, and one of the found recipes is selected (process 410). According to some embodiments, the selected recipe is the recipe corresponding to the highest arranged hash value. According to some embodiments, considering other factors, the selected recipe is the recipe corresponding to a hash value that is not the highest arranged. The selected recipe can therefore be used to perform a CMP process on the corresponding physical wafer.

[0056] Fig.16 Shows Fig.15 A graphical representation of processes 406 and 410 in process flow 400 in FIG. Fig.16 As shown, multiple hash values ​​of the found hot spots are generated (represented by their two-dimensional matrices), which corresponds to Fig.15 Next, further according to Fig.15 In process 406, similarity values ​​between the discovered hot spots and their corresponding representative hot spots in the hot spot library 222 are calculated, and the discovered hot spots (and their hash values) are arranged. Fig.16 As shown, the order of the hash values ​​shown is rearranged to show the arrangement. Fig.16 Also shown are multiple recipes and GDS files corresponding to the ranked hash values. Next, one of the recipes is selected (process 410), and in this example the selected recipe is recipe B. In other embodiments, the recipe with the highest ranking (recipe A) may be selected.

[0057] Fig.17An example recipe for a CMP process is shown. Each recipe may include (sub) steps to be performed in the CMP process, and parameters to be used in each step. In the example shown, there are four steps, step 1, step 2, step 3, and step 4, each of which is performed with multiple parameters, and the parameters change between steps. For example, there may be head rotation A, head rotation B, slurry flow A (flow rate of the first slurry), slurry flow B (flow rate of the second slurry), down pressure (wafer head on the polishing pad), trimming open / close (open or close the pad regulator), etc. The X-axis represents the time of the CMP process, and the Y-axis represents the parameters and their corresponding values. For example, for each value, at any time, when the corresponding strip exists, the corresponding parameter is turned on, and if the strip is wider (in the Y direction), the corresponding parameter has a higher value. For example, slurry B has a higher flow rate in the initial stage of step 1, and is then closed for the remaining time of step 1. Slurry B has a relatively small flow rate throughout step 2, and is closed throughout step 3 and throughout step 4. The pad conditioner (indicated by trim open / closed) is opened with relatively less downforce during steps 1 and 2, and is opened with relatively higher downforce during steps 3 and 4. Throughout this specification, when a recipe is referred to as being adjusted, this refers to a combination of adjustment steps, parameters, and parameter values, meaning that when any parameter is adjusted, the recipe is considered to be adjusted.

[0058] Fig.18 The method for formulating and training the hotspot prevention model 223 ( Figure 6 ) process. Throughout the specification, the hotspot prevention model 223 may be alternatively referred to as a machine learning (ML) model because the model can be improved by learning during the improvement process of the process flow 500. The improved recipes (recipes A, B, C, and D) can be used as Figure 6 The stored recipe in process 206 in.

[0059] refer to Fig.18 , multiple GDS (layouts) A, B, C, and D are provided. GDS A, B, C, and D can be the same as each other, can be slightly different from each other, or can be completely different from each other. Each recipe improvement process will improve the recipe through iteration. For example, GDS A is provided (process 502), and GDS A is fed to the hotspot prevention model 223 ( Figure 6 ), so that the hot spots that may occur are generated and output by the hot spot prevention model 223. Then, recipe A is suggested, and in process 400 ( Fig.15) shows the generation of the suggested (selected) recipe A. The selected recipe A is then used to perform a CMP process on a physical wafer implementing GDS A. Measurements are then performed on the wafer to determine hot spots and topological information on the wafer, and wafer results are generated (process 506). During the measurement, locations on the wafer (which were measured in process 314 ( Fig.13 ) is marked in order to determine in process 312 ( Fig.13 ) has been eliminated or at least reduced. If the hot spots are eliminated or at least reduced, it is determined that the recipe is beneficial and may or may not be further improved. If the hot spots are not reduced or even worsened, other recipes are needed.

[0060] Depending on the wafer results, the training data 508 (which may include the measurement results) is fed back to the hotspot prevention model 223, and the hotspot prevention model 223 is updated (process 510). For example, when the measurement results indicate that some new hotspots are found, or some expected hotspots do not exist, the hotspot prevention model will be updated so that the hotspot prevention model will output the newly found hotspots instead of the non-existent models.

[0061] Furthermore, based on the measurement results, recipe A can be modified, for example, to eliminate the remaining found hot spots. Then, another wafer can be manufactured and CMP performed using the modified recipe A, and measurements can be performed to determine the hot spots and topological information. This constitutes an iteration, and the iteration can continue until the results are satisfactory.

[0062] GDS A, B, C, D can be identical to each other, and the initial recipes A, B, C, D can be selected to be different, so that the recipes can be improved in different directions, and the best recipe can be finally selected among the modified recipes A, B, C, and D (each recipe is modified in its own iteration). It is also possible to use GDS A, B, C, D that are different from each other, so that the resulting model can cover different layouts, and more recipes can be generated for different layouts.

[0063] Fig.19 A tool 600 is schematically shown for performing the above-described tasks, including but not limited to calculating, determining, and storing the hotspot library 222. For example, the processes shown in the process flows 200, 300, 400, and 500 can all be performed using a computer (processor) 602 including hardware and software (computer program code). The program code of the tool 600 can be embodied on a non-transitory storage medium, such as a hard drive, a disk, etc. The hotspot library 222, which can be implemented in a storage device such as a hard drive, is electrically connected and signal-connected to the computer 602 for storage and retrieval.

[0064] Embodiments of the present disclosure have some advantageous features. By predicting hot spots and selecting recipes to reduce / eliminate the predicted hot spots, hot spots can be found without manufacturing a physical wafer and measuring the physical wafer. A first physical wafer can be manufactured using a recipe that is expected to eliminate potential hot spots. Manufacturing cycle time can be significantly reduced, for example, by one-third.

[0065] According to some embodiments of the present disclosure, a method includes: cutting a plurality of images from a layout of an integrated circuit; generating a first plurality of hash values, each of which is from one of the plurality of images; loading a second plurality of hash values ​​stored in a hotspot library; comparing each of the first plurality of hash values ​​with each of the second plurality of hash values, wherein the comparison includes calculating a similarity value between each of the first plurality of hash values ​​and each of the second plurality of hash values; comparing the similarity value with a predetermined threshold similarity value; in response to a result that the similarity value is greater than the predetermined threshold similarity value, recording a position of a corresponding image having the result, wherein the position is a position of the corresponding image in the layout. In an embodiment, the plurality of images form an array, and the position includes a row number and a column number of the corresponding image in the array. In an embodiment, the method further includes: manufacturing the integrated circuit on a wafer, wherein the manufacturing includes performing a chemical mechanical polishing process on the wafer; and finding a hotspot from the position, wherein the hotspot is a defect in the wafer caused by the chemical mechanical polishing process. In an embodiment, cutting the plurality of images includes: dividing the layout into an array of images, and cutting each of the plurality of images in the array. In an embodiment, the predetermined threshold similarity value is 0.9. In an embodiment, the method further comprises: cropping an additional plurality of images from an additional layout of an additional integrated circuit; generating a third plurality of hash values, each of which is from one of the additional plurality of images; comparing each of the third plurality of hash values ​​to all hash values ​​stored in the hotspot library to find a set of hash values ​​similar to the third plurality of hash values; and ranking the similarity values ​​of the set of hash values. In an embodiment, the method further comprises selecting a recipe associated with one of the set of hash values.

[0066] According to some embodiments of the present disclosure, a method includes: cutting a plurality of images from a layout of an integrated circuit; generating a plurality of hash values, each hash value being from one of the plurality of images; searching from a hotspot library to find similar hash values ​​similar to the plurality of hash values, wherein the hotspot library stores hash values ​​indexed to images having hotspots; and marking the positions of some of the plurality of images associated with the similar hash values ​​on the layout of the integrated circuit. In an embodiment, the method also includes: implementing the layout of the integrated circuit on a wafer, wherein the implementation includes performing a chemical mechanical polishing process on the wafer using a recipe; and inspecting a position on the wafer to determine the hotspot at the position. In an embodiment, the method also includes: determining a recipe based on the similar hash values ​​that have been found. In an embodiment, the similar hash values ​​are associated with a plurality of recipes in the hotspot library, and wherein the recipe is selected from the plurality of recipes. In an embodiment, the recipe includes: a first duration and flow rate of a slurry used in a chemical mechanical polishing process, and respective second durations and amplitudes of a trim and a down force used in the chemical mechanical polishing process. In an embodiment, each of the plurality of images has a square shape, and the length and width of the square are in a range between about 64 μm and about 256 μm. In an embodiment, the hotspot library includes a plurality of entries, each entry including a hash value, an image, a recipe, and topology information. In an embodiment, the hotspot library is indexed by a hash value.

[0067] According to some embodiments of the present disclosure, a system includes: a library stored in a tangible medium, the library including a plurality of entries, each entry including: a hash value; an image associated with the hash value, wherein the image includes a hotspot; a recipe configured to reduce the hotspot; and topological information of the hotspot. In an embodiment, the system also includes: a tool including software, wherein the software is configured to generate a hash value from the image. In an embodiment, the similarity value of any pair of hash values ​​in the plurality of entries stored in the library is less than about 0.9. In an embodiment, the hotspot includes a groove or a protrusion that appears at the center of the image. In an embodiment, the recipe includes process conditions configured to reduce the hotspot.

[0068] The features of several embodiments are summarized above so that those skilled in the art can better understand the various aspects of the present disclosure. Those skilled in the art should understand that they can easily use the present disclosure as a basis for designing or modifying other processes and structures to achieve the same purpose of the embodiments introduced herein and / or achieve the same advantages of the embodiments introduced herein. Those skilled in the art should also recognize that such equivalent constructions do not depart from the spirit and scope of the present disclosure, and they can make various changes, substitutions and modifications herein without departing from the spirit and scope of the present disclosure.

[0069] Example 1. A method for manufacturing a semiconductor device, comprising: cutting out multiple images from a layout of an integrated circuit; generating a first plurality of hash values, each of which is from one of the multiple images; loading a second plurality of hash values ​​stored in a hotspot library; comparing each of the first plurality of hash values ​​with each of the second plurality of hash values, wherein the comparison includes calculating a similarity value between each of the first plurality of hash values ​​and each of the second plurality of hash values; comparing the similarity value with a predetermined threshold similarity value; and in response to a result that the similarity value is greater than the predetermined threshold similarity value, recording a position of a corresponding image having the result, wherein the position is a position of the corresponding image in the layout.

[0070] Example 2. A method according to Example 1, wherein the plurality of images form an array, and the position includes a row number and a column number of the corresponding image in the array.

[0071] Example 3. The method according to Example 1 further includes: manufacturing the integrated circuit on a wafer, wherein the manufacturing includes performing a chemical mechanical polishing process on the wafer; and finding a hot spot from the location, wherein the hot spot is a defect in the wafer caused by the chemical mechanical polishing process.

[0072] Example 4. The method of Example 1, wherein cropping the plurality of images comprises dividing the layout into an array of images and cropping each of the plurality of images in the array.

[0073] Example 5. The method according to Example 1, wherein the predetermined threshold similarity value is 0.9.

[0074] Example 6. The method according to Example 1 also includes: cropping additional multiple images from additional layouts of additional integrated circuits; generating a third plurality of hash values, each of which is from one of the additional multiple images; comparing each of the third plurality of hash values ​​with all hash values ​​stored in the hotspot library to find a group of hash values ​​similar to the third plurality of hash values; and arranging the similarity values ​​of the group of hash values.

[0075] Example 7. The method of Example 6, further comprising: selecting a recipe associated with a hash value in the set of hash values.

[0076] Example 8. A method for manufacturing a semiconductor device, comprising: cutting out multiple images from a layout of an integrated circuit; generating multiple hash values, each hash value coming from one of the multiple images; searching from a hotspot library to find similar hash values ​​that are similar to the multiple hash values, wherein the hotspot library stores hash values ​​indexed to images having hotspots; and marking the locations of some of the multiple images associated with the similar hash values ​​on the layout of the integrated circuit.

[0077] Example 9. The method according to Example 8 further includes: implementing a layout of the integrated circuit on a wafer, wherein the implementation includes performing a chemical mechanical polishing process on the wafer using a recipe; and inspecting the location on the wafer to determine a hot spot at the location.

[0078] Example 10. The method of Example 9, further comprising: determining the recipe based on the similar hash values ​​that have been found.

[0079] Example 11. The method of Example 10, wherein the similar hash values ​​are associated with a plurality of recipes in the hotspot library, and wherein the recipe is selected from the plurality of recipes.

[0080] Example 12. The method of Example 9, wherein the recipe includes: a first duration and flow rate of a slurry used in the chemical mechanical polishing process, and second respective durations and amplitudes of trim and downforce used in the chemical mechanical polishing process.

[0081] Example 13. The method of Example 8, wherein each of the plurality of images has a square shape, the length and width of the square being in a range between 64 μm and 256 μm.

[0082] Example 14. The method of Example 8, wherein the hotspot library comprises a plurality of entries, each entry comprising a hash value, an image, a recipe, and topology information.

[0083] Example 15. The method of Example 14, wherein the hotspot library is indexed by hash values.

[0084] Example 16. A system for manufacturing a semiconductor device, comprising: a library stored in a tangible medium, the library comprising a plurality of entries, each entry comprising: a hash value; an image associated with the hash value, wherein the image comprises a hotspot; a recipe configured to reduce the hotspot; and topological information of the hotspot.

[0085] Example 17. The system of Example 16 further comprises: a tool comprising software, wherein the software is configured to generate the hash value from the image.

[0086] Example 18. The system of Example 16, wherein a similarity value of any pair of hash values ​​in the plurality of entries stored in the repository is less than 0.9.

[0087] Example 19. A system according to Example 16, wherein the hotspot comprises a groove or a protrusion appearing at the center of the image.

[0088] Example 20. The system of Example 16, wherein the recipe includes process conditions configured to reduce the hot spots.

Claims

1. A method for manufacturing a semiconductor device, comprising: cropping a plurality of images from a layout of an integrated circuit; generating a first plurality of hash values, each of which is from one of the plurality of images, and each of the first plurality of hash values ​​comprises numbers and letters; loading a second plurality of hash values ​​stored in the hotspot library; comparing each of the first plurality of hash values ​​to each of the second plurality of hash values, wherein the comparing comprises calculating a similarity value between each of the first plurality of hash values ​​and each of the second plurality of hash values; comparing the similarity value to a predetermined threshold similarity value; and In response to a result in which the similarity value is greater than the predetermined threshold similarity value, a position of a corresponding image having the result is recorded, wherein the position is a position of the corresponding image in the layout.

2. The method according to claim 1, wherein: The plurality of images form an array, and the position includes a row number and a column number of the corresponding image in the array.

3. The method according to claim 1, further comprising: fabricating the integrated circuit on a wafer, wherein the fabricating comprises performing a chemical mechanical polishing process on the wafer; and A hot spot is found from the location, wherein the hot spot is a defect in the wafer generated by the chemical mechanical polishing process.

4. The method according to claim 1, wherein: Cropping the plurality of images includes dividing the layout into an array of images and cropping each of the plurality of images in the array.

5. The method according to claim 1, wherein: The predetermined threshold similarity value is 0.

9.

6. The method according to claim 1, further comprising: cropping additional multiple images from additional layouts of additional integrated circuits; generating a third plurality of hash values, each of which is from one of the additional plurality of images; comparing each of the third plurality of hash values ​​to all hash values ​​stored in the hotspot library to find a set of hash values ​​similar to the third plurality of hash values; as well as The similarity values ​​of the group of hash values ​​are arranged.

7. The method according to claim 6, further comprising: A recipe associated with one of the set of hash values ​​is selected.

8. A method for manufacturing a semiconductor device, comprising: cropping a plurality of images from a layout of an integrated circuit; generating a plurality of hash values, each hash value being from one of the plurality of images, and each of the plurality of hash values ​​comprising numbers and letters; searching from a hotspot library to find similar hash values ​​similar to the plurality of hash values, wherein the hotspot library stores hash values ​​indexed to images having hotspots; as well as Positions of ones of the plurality of images associated with the similar hash values ​​on a layout of the integrated circuit are marked.

9. The method according to claim 8, further comprising: Implementing a layout of the integrated circuit on a wafer, wherein the implementing comprises performing a chemical mechanical polishing process on the wafer using a recipe; and The location on the wafer is inspected to determine a hot spot at the location.

10. The method according to claim 9, further comprising: The recipe is determined based on the similar hash values ​​that have been found.

11. The method according to claim 10, wherein: The similar hash values ​​are associated with a plurality of recipes in the hotspot library, and wherein the recipe is selected from the plurality of recipes.

12. The method according to claim 9, wherein: The recipe includes a first duration and flow rate of a slurry used in the chemical mechanical polishing process, and second respective durations and magnitudes of conditioning and downforce used in the chemical mechanical polishing process.

13. The method according to claim 8, wherein: Each of the plurality of images has a square shape, a length and a width of the square ranging between 64 μm and 256 μm.

14. The method according to claim 8, wherein: The hotspot library includes multiple entries, each entry including a hash value, an image, a recipe, and topology information.

15. The method according to claim 14, wherein: The hotspot library is indexed by hash values.

16. A system for manufacturing a semiconductor device, comprising: A library, stored in a tangible medium, the library comprising a plurality of entries, each entry comprising: Hash value, including numbers and letters; an image associated with the hash value, wherein the image includes a hotspot; a recipe configured to reduce the hot spots; and Topological information of the hotspot; Wherein, reducing the hot spots comprises: cropping a plurality of images from a layout of an integrated circuit; generating a plurality of hash values, each hash value being from one of the plurality of images, and each of the plurality of hash values ​​comprising numbers and letters; searching the library to find similar hash values ​​that are similar to the plurality of hash values; and Positions of ones of the plurality of images associated with the similar hash values ​​on a layout of the integrated circuit are marked.

17. The system of claim 16, further comprising: A tool comprising software, wherein the software is configured to generate the hash value from the image.

18. The system of claim 16, wherein: A similarity value of any pair of hash values ​​in the plurality of entries stored in the repository is less than 0.

9.

19. The system of claim 16, wherein: The hot spot includes a groove or a bump that appears at the center of the image.

20. The system of claim 16, wherein: The recipe includes process conditions configured to reduce the hot spots.

Images