Detection model after opc and method for establishing same, and method for optimizing opc verification

Abstract

This invention provides a detection model after OPC and its establishment method, as well as a method for optimizing OPC verification. The method for establishing the detection model after OPC includes: acquiring a first sample pattern after OPC, wherein the first sample pattern is prone to photoresist top morphology distortion after photolithography; acquiring an optical model of the first sample pattern and its corresponding distortion level; performing photolithography simulation on the optical model using different exposure parameters to obtain measurement data and measurement images of different distortion levels, wherein the measurement data and measurement images include linewidths and images reflecting the top morphology of the photoresist after exposure; and inputting the distortion level, measurement data, and measurement images corresponding to the optical model under different exposure parameters into a sample pattern library to establish a detection model, which is used to determine whether there are abnormal top morphology patterns of photoresist in the layout data after OPC. This invention is used to optimize the detection effect and efficiency of some hot spot defects.

Description

Technical Field

[0001] This invention relates to the field of integrated circuit manufacturing technology, and in particular to a detection model after OPC and its establishment method, as well as a method for optimizing OPC verification. Background Technology

[0002] In photolithography, the pattern on a photomask is projected onto photoresist through an exposure system, forming the corresponding pattern. However, due to optical factors (such as diffraction) or chemical reactions in the photoresist, the pattern formed in the photoresist may deviate from the pattern on the photomask. This deviation requires pre-modification of the pattern on the photomask using OPC (Optical Proximity Correction). When the photomask is exposed using OPC-corrected OPC, the pattern formed in the photoresist will match the designed pattern and meet the process requirements.

[0003] Currently, after OPC correction of the layout, OPC verification is performed on the entire layout data after OPC. Hotspot patterns are detected and corrected before OPC verification is performed again until they meet the requirements. However, in practice, some hotspot defects (such as abnormal morphology of the photoresist pattern after exposure) are not easily detected in the optical model or the exposed model (photoresist model) during the above OPC verification process, which seriously affects the effectiveness of OPC verification. On the other hand, if an additional etched model is added for simulation verification, the efficiency of OPC verification will be severely affected. Summary of the Invention

[0004] The purpose of this invention is to provide a detection model after OPC and its establishment method, as well as a method for optimizing OPC verification, to improve the detection effect and efficiency of some hot spot defects.

[0005] To solve the above-mentioned technical problems, the present invention provides a method for establishing a detection model after OPC, comprising:

[0006] The first sample pattern obtained after OPC has preset size parameters and exposure parameters. The first sample pattern is prone to photoresist top morphology distortion after photolithography.

[0007] Obtain the optical model of the first sample image and its corresponding distortion level;

[0008] Photolithography simulation was performed on the optical model using different exposure parameters to obtain measurement data and measurement images with different distortion levels. The measurement data and measurement images include line widths and images reflecting the top morphology of the photoresist after exposure.

[0009] The distortion levels, measurement data, and measurement images of the optical model under different exposure parameters are input into a sample image library to establish a detection model. This model is used to detect whether there are images in the layout data after OPC that match the optical model under the same or similar exposure parameters, and to determine whether there are images with abnormal top morphology of the photoresist in the layout data after OPC.

[0010] Optionally, the size parameters of the first sample image are adjusted to obtain multiple second sample images based on the first sample image. The optical models and corresponding distortion indices of the first sample image and each of the second sample images are obtained. The first sample image and each of the second sample images are divided into distortion levels from severe distortion to slight distortion according to the distortion indices. The size parameters of the multiple second sample images are set at equal intervals with the size parameters of the first sample image as the center.

[0011] Optionally, the severely distorted sample image is repaired to obtain a third sample image, and then the optical model and distortion index of the third sample image are obtained.

[0012] Optionally, the steps of obtaining the optical model and the distortion index include:

[0013] The optical model is derived based on the light source distribution of the exposure machine and the corresponding target wire diameter.

[0014] The maximum and minimum light intensities of the cross-sections at different positions of the photoresist in the optical model are obtained. The distortion index is the difference between the maximum and minimum light intensities on all cross-sections divided by the sum of the maximum and minimum light intensities.

[0015] Optionally, the step of performing photolithography simulation on each of the optical models using different exposure parameters includes:

[0016] Based on the optical models described, the exposure conditions that have the greatest impact on the photoresist within the normal process window are identified as the key exposure conditions.

[0017] Using the key exposure conditions as variables, multiple exposure parameter combinations are constructed within the normal process window range to perform photolithography simulation on each optical model to obtain corresponding measurement data and measurement images.

[0018] Optionally, when establishing the detection model, the patterns input into the sample pattern library include normal patterns and patterns with distorted top morphology of the photoresist, and the patterns with distorted top morphology of the photoresist are composed of cases with different distortion levels.

[0019] Optionally, when acquiring the measurement data after photolithography simulation,

[0020] Measure the depth of the photoresist and determine whether the depth is less than a preset height.

[0021] If not, the minimum linewidth of the photoresist in the cross-section is taken as the linewidth.

[0022] If so, the depth height shall be used as the measurement data.

[0023] Optionally, when it is difficult to accurately measure or determine the line width of a portion of the sample graphic, the line width of the sample graphic can be roughly estimated by the actual line width variation trend of a sample graphic with a size close to that of the sample graphic, and made to conform to the actual line width variation trend.

[0024] Based on another aspect of the present invention, a detection model after OPC is also provided, which is established using the method for establishing a detection model after OPC as described above.

[0025] Based on another aspect of the present invention, a method for optimizing OPC verification is also provided, comprising:

[0026] Perform OPC processing on the layout to obtain the OPC-processed layout data;

[0027] The layout data after OPC is processed using the detection model described above to determine whether there are distorted graphics. If so, the distorted graphics are corrected and the layout data after OPC is updated.

[0028] Perform OPC verification on the layout data after OPC.

[0029] In summary, this invention, after OPC correction and before OPC verification, sets up a post-OPC detection model to detect hot spot defects that are difficult to detect in related technologies using the post-OPC layout data. Unlike related technologies that use normal patterns to build models, the post-OPC detection model of this invention includes cases formed by the aforementioned patterns that are prone to hot spot defects under various exposure conditions, that is, it includes distortion cases with different distortion levels. Moreover, the measurement data and measurement images in the aforementioned distortion cases directly reflect the linewidth and image of the top morphology after photoresist exposure. Thus, the aforementioned difficult-to-detect hot spot defects can be specifically detected before the overall verification of the post-OPC layout data, thereby improving the effectiveness and efficiency of OPC correction and OPC verification. Attached Figure Description

[0030] Those skilled in the art will understand that the accompanying drawings are provided to better understand the invention and do not constitute any limitation on the scope of the invention.

[0031] Figure 1 This is a flowchart of a method for establishing a detection model after OPC according to an embodiment of the present invention;

[0032] Figure 2a A schematic diagram of a first sample graphic;

[0033] Figure 2b To and Figure 2a A diagram illustrating the corresponding quantitative distortion index;

[0034] Figure 3a This is a schematic diagram for measuring line width;

[0035] Figure 3b A schematic diagram for measuring the missing photoresist pattern at the top;

[0036] Figure 4 This is a flowchart of an optimized OPC verification method provided in an embodiment of the present invention.

[0037] In the attached diagram: 11 - Opaque area; 12 - Transparent area; 13 - Risk area; 14 - Center line; 15 - Photoresist cross section. Detailed Implementation

[0038] As described in the background section, the models (verifications) established in the related technologies are all based on normal patterns and normal cases (patterns that are not expected to have distortion). Therefore, when performing OPC correction and OPC verification, statistics, corrections and verifications are performed according to these undisputed data. However, for some patterns that are not friendly to light source settings (light source distribution), the normal pattern (ordinary model) cannot reflect the actual situation, such as the abnormal morphology of the photoresist pattern after exposure.

[0039] Therefore, the detection model and its establishment method after OPC, and the method for optimizing OPC verification provided by this invention, after OPC correction and before OPC verification, set up the detection model after OPC to detect the aforementioned difficult-to-detect hot spot defects in the OPC layout data. Unlike the related technologies that use normal patterns to establish models, the detection model after OPC provided by this invention includes cases formed by the aforementioned patterns that are prone to hot spot defects under various exposure conditions, that is, it includes distortion cases with different distortion levels. Moreover, the measurement data and measurement images in the aforementioned distortion cases directly reflect the linewidth and image of the top morphology after photoresist exposure. Thus, the aforementioned difficult-to-detect hot spot defects can be specifically detected before the overall verification of the OPC layout data, thereby improving the effect and efficiency of OPC correction and OPC verification.

[0040] To make the objectives, advantages, and features of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the drawings are all in a very simplified form and are not drawn to scale, and are only used to facilitate and clarify the explanation of the embodiments of this invention. Furthermore, the structures shown in the drawings are often part of the actual structures. In particular, different figures may emphasize different aspects and may sometimes use different scales.

[0041] As used in this invention, the singular forms “a,” “an,” and “the” include plural objects; the term “or” is generally used to mean “and / or”; the term “a number” is generally used to mean “at least one”; and the term “at least two” is generally used to mean “two or more”. Furthermore, the terms “first,” “second,” and “third” are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as “first,” “second,” or “third” may explicitly or implicitly include one or at least two of that feature, unless otherwise expressly indicated.

[0042] Figure 1 This is a flowchart of a method for establishing a detection model after OPC according to an embodiment of the present invention.

[0043] like Figure 1 As shown, the method for establishing the detection model after OPC provided in this embodiment includes the following steps:

[0044] S01: The first sample pattern after obtaining OPC has preset size parameters and exposure parameters. The first sample pattern is prone to photoresist top morphology distortion after photolithography.

[0045] S02: Obtain the optical model of the first sample image and the corresponding distortion level;

[0046] S03: Perform photolithography simulation on the optical model using different exposure parameters to obtain measurement data and measurement images with different distortion levels. The measurement data and measurement images include line widths and images reflecting the top morphology of the photoresist after exposure.

[0047] S04: Input the distortion level, measurement data and measurement images of the optical model under different exposure parameters into the sample graphic library to establish a detection model, which is used to detect whether there is a graphic in the layout data after OPC that matches the optical model under the same or similar exposure parameters, and thereby determine whether there is a graphic with abnormal top morphology of photoresist in the layout data after OPC.

[0048] First, step S01 is executed to obtain the first sample pattern after OPC, which has preset size parameters and exposure parameters. The first sample pattern is prone to distortion of the top morphology of the photoresist after photolithography.

[0049] The first sample pattern can be a hotspot pattern obtained from an actual product line. These hotspot patterns are prone to photoresist morphology abnormalities (photoresist top morphology distortion) after photolithography, i.e., abnormal photoresist linewidth and / or abnormal photoresist top morphology. Unlike other types of distortion (e.g., proximity effect, rounding of square corners, nonlinear changes, etc.), the distortions that may be involved in the first sample pattern are not obvious in the top view of the photoresist pattern (linewidth in the top view), but are very likely to form hotspot defects after etching using the above-mentioned photoresist pattern. Therefore, the first sample pattern can be selected from the hotspot database of the actual product line based on the above-mentioned hotspot defects. This embodiment does not limit the number of first sample patterns. Multiple first sample patterns with similar patterns are regarded as a class. This embodiment can include multiple sample patterns of different types that are prone to photoresist top morphology distortion, i.e., multiple first sample patterns. Each first sample pattern can be a typical pattern selected from the same class. The following uses one (class) of first sample patterns as an example to establish a detection model.

[0050] It should be noted that the first sample pattern mentioned above is the pattern after OPC, which has preset size parameters (after OPC) and exposure parameters. These parameters are the parameters used when the aforementioned hot spot defects are formed. The areas in the first sample pattern where the top morphology of the photoresist is prone to distortion can be considered risk areas, which are usually the areas with narrower linewidths.

[0051] Next, step S02 is executed to obtain the optical model of the first sample image and the corresponding distortion level.

[0052] First, the light source distribution and corresponding target line diameter size in the exposure conditions of the first sample pattern photolithography process (exposure machine) can be obtained. Then, combined with the first sample pattern, the optical model of the first sample pattern can be derived. Among them, the target line diameter size is related to the pattern size before OPC, and the light source distribution is related to the exposure machine. It can be understood that the detection model established in this embodiment is for the pattern after OPC under the same exposure machine and the same light source distribution.

[0053] Using the optical model of the first sample pattern, the distortion of the optical model in the risk area is estimated as a distortion index of the first sample pattern. The specific steps include: drawing a center line of the opaque area along the extension direction of the photoresist in the risk area; then drawing multiple tangents perpendicular to the center line at preset intervals, each tangent forming a photoresist cross-section; and obtaining the maximum light intensity I of all the photoresist cross-sections from the optical model. max With minimum value Imin And thus, the quantified distortion index on this optical model is obtained: the difference between the maximum and minimum light intensity divided by the sum of the maximum and minimum light intensities, i.e. (I max -I min ) / (I max + I min This serves as the distortion index for the optical model. It is understood that the aforementioned tangent, besides being tangent to the opaque area of ​​the photoresist to form a photoresist cross-section, is used to obtain the maximum light intensity I. max With minimum value I min In addition to being used to calculate distortion metrics, the aforementioned tangent can also be tangent to the light-transmitting area of ​​the photoresist to form a cross-section of the photoresist, thereby obtaining the maximum light intensity I. max With minimum value I min Used to calculate distortion metrics.

[0054] Here, it is preferable to extend the first sample pattern into a similar or related hotspot pattern to expand the coverage of the established detection model. Specifically, the size parameters of the first sample pattern can be sequentially increased and decreased at equal intervals, forming multiple sample patterns with different size parameters, which serve as second sample patterns. Of course, the first sample pattern may have more than one size parameter, such as the width (linewidth) of the opaque area and the width (linewidth) of the adjacent transparent area on the same photoresist cross-section. Therefore, the width of the opaque area and the width of the adjacent transparent area can be sequentially increased or decreased at equal intervals to construct multiple sample patterns (second sample patterns) with different size parameters. Then, similar to the first sample pattern, the optical model and corresponding distortion index of each second sample pattern are obtained. Next, the distortion indices of the first sample pattern and each second sample pattern are arranged in ascending or descending order, and at least three sample patterns and their optical models, ranging from severe distortion to slight distortion, are selected for subsequent photolithography simulation (exposure simulation). According to actual process requirements, the range of distortion indexes corresponding to severe distortion, general distortion and slight distortion can be defined, or more sample graphics and their optical models can be selected for subsequent simulation.

[0055] In addition, sample images with excessive distortion (e.g., severe distortion) can be selected from the first sample image and each of the second sample images. After OPC correction, these images can be used as the third sample image. Similar operations to those on the first sample image can be performed on the third sample image to obtain the optical model and corresponding distortion index of each third sample image. Furthermore, similar to obtaining the second sample image from the first sample image, the dimensional parameters of the third sample image can be increased or decreased to obtain the corresponding fourth sample image, thereby obtaining the optical model and distortion index of the first sample image.

[0056] Figure 2a This is a schematic diagram of a first sample graphic. Figure 2b To and Figure 2a A diagram illustrating the corresponding quantification distortion index. (For example...) Figure 2a As shown, in the risk area 13 of the first sample pattern, the line width (first line width) of the opaque area 11 is d, and the line width (second line width) of the translucent areas 12 on both sides of the opaque area is w. d and w are the minimum design rule line diameter dimensions. The first line width is extended to d-2, d-1, d, d+1, d+2, and the second line width is extended to w-2, w-1, w, w+1, w+2. The first and second line widths are combined and a portion is removed to obtain the dimensional parameters of the second sample pattern. The removed dimensional parameters include the parameters of the first sample pattern composed of d and w, and other dimensional parameter combinations that do not conform to the design rules. Figure 2b As shown, the center line is 14, and the cross-sections of the photoresist arranged at equal intervals are 15.

[0057] It should be noted that in practice, multiple second sample images can be obtained from the first sample image (by increasing or decreasing the size parameters of the first sample image), and then the optical model and distortion index (distortion level) can be obtained from the first sample image and each second sample image at the same time.

[0058] Next, step S03 is performed to perform photolithography simulation on the optical model using different exposure parameters to obtain measurement data and measurement images with different distortion levels. The measurement data and measurement images include linewidth and images reflecting the top morphology of the photoresist after exposure.

[0059] Before lithography simulation, based on each optical model, the exposure conditions that have the greatest impact on the photoresist morphology in the risk area within the normal process window (lithography process window) of each optical model are identified as critical exposure conditions, such as the exposure dose or defocus range. During lithography simulation of each optical model, in addition to performing simulation according to preset exposure parameters, simulation is also performed according to multiple combinations of exposure parameters within the normal process window. These multiple exposure parameter combinations can be obtained by sequentially increasing or decreasing the parameters corresponding to the critical exposure conditions within the normal process window. In other words, if some exposure parameters are close to the upper and / or lower limits of the critical exposure conditions within the normal process window, it is equivalent to changing (or even worsening) the exposure parameters corresponding to the critical exposure conditions.

[0060] After photolithography simulation, the photoresist in the risk area is measured to obtain corresponding measurement data and images. These measurement data and images must accurately reflect the morphology of the photoresist, especially its top surface. Specifically, when acquiring measurement data after photolithography simulation, the depth of the photoresist is first measured to determine if it is less than the preset height. If not, the depth meets expectations. The minimum linewidth of the photoresist in the cross-section is then measured as linewidth data, and the corresponding measurement image is obtained. If the depth does not meet expectations, it indicates an abnormal top surface morphology (missing top surface). In this case, the depth is used as measurement data to reflect the top surface morphology of the photoresist, and the corresponding measurement image is obtained. Next, based on the above measurement data and images, the optical model and corresponding exposure conditions after exposure simulation can be classified by distortion level to directly identify the top surface morphology of the photoresist and the corresponding linewidth data. Of course, the distortion level of the aforementioned optical model under different exposure conditions may be different from the distortion level before the original exposure simulation. However, the degree of distortion measured after the exposure simulation is used as the subsequent judgment standard. Moreover, there are also cases where the aforementioned optical model does not show distortion (basically a normal image) after exposure simulation under some exposure conditions.

[0061] It should be noted that if, during data measurement, the data (depth or linewidth) of some images is difficult to measure accurately or determine, a rough estimate can be made by observing the variation trend of sample patterns with similar dimensions (defining their measurement data) to ensure it conforms to the actual variation trend. In one example, the depth of the photoresist is less than the preset height. Besides using the depth of the photoresist to reflect its linewidth, a linewidth that matches (fits) the actual variation trend can also be estimated by observing the linewidth variation trends of several sample patterns with similar dimensional parameters but measurable linewidths. In another example, the depth of the photoresist matches the preset height, but both sides of the photoresist are sloping, with the smallest linewidth at the top. Besides using the smallest linewidth at the top as its linewidth, a linewidth W3 that matches (fits) the actual variation trend can also be estimated by observing the linewidth variation trends of several sample patterns with similar dimensional parameters but whose linewidths are easier to determine.

[0062] Figure 3a A schematic diagram for measuring line width, as shown below. Figure 3a As shown, the depth of the photoresist is as expected. In contrast to the related technologies that measure the bottom width of the photoresist as the line width, in this embodiment, the width W1 of the narrowest point between the two side walls is measured as the line width, instead of W2 as the line width in the related technologies. Figure 3b A schematic diagram for measuring the missing photoresist pattern at the top, as shown. Figure 3bAs shown, the depth H1 of the photoresist is lower than the preset height H2 (which does not meet expectations). Compared with the measurement of the bottom width of the photoresist as the line width in related technologies, in this embodiment, the depth W2 of the photoresist on both sides represents its line width. Of course, in practice, a line width W3 can also be estimated based on the line width change trend as above.

[0063] Next, step S04 is executed, in which the distortion level, measurement data and measurement images of the optical model under different exposure parameters are put into the sample graphic library to establish a detection model, which is used to detect whether there is a graphic in the layout data after OPC that matches the optical model under the same or similar exposure parameters, and thereby determine whether there is a graphic with abnormal top morphology of photoresist in the layout data after OPC.

[0064] A sample pattern with a single size parameter, its distortion index, measurement data, and measurement image under a single exposure parameter can be used as a case (example). Multiple cases, consisting of the distortion levels, measurement data, and measurement images of various sample patterns and their optical models under different exposure parameters, can be used as a sample pattern library to build the detection model. This ensures that a certain number of normal cases (without photoresist top distortion) and a certain number of distorted cases are simultaneously input into the detection model after OPC. In a preferred example, the ratio of normal cases to distorted cases can be approximately 1:1, and the distorted cases should include various types ranging from severe to slight distortion.

[0065] After establishing the detection model for the first sample image and similar sample images, the line width of the normal example can be used as a reference. Various other size parameters and cases under exposure conditions can be input to verify its distortion (distortion level) and define the line width specification that matches the distortion level.

[0066] Therefore, after inputting the layout data of the OPC sample and its exposure conditions into the detection model, the detection model can search for whether there are graphics similar to those in the model, and distinguish whether the graphics under the exposure conditions are normal or distorted patterns after exposure simulation, and the distortion level (distortion condition) when they are distorted patterns. It should be noted that the detection model is not used to accurately calculate the linewidth of the sample graphic after exposure simulation, but rather to detect whether there are distorted graphics similar to those in the detection model and their distortion levels.

[0067] This embodiment also provides a detection model after OPC, which is established using the method for establishing a detection model after OPC as described above.

[0068] This embodiment also provides a method for optimizing OPC verification.

[0069] Figure 4The flowchart shows the optimized OPC verification method provided in this embodiment.

[0070] like Figure 4 As shown, the method for optimizing OPC verification provided in this embodiment includes:

[0071] S11: Perform OPC processing on the layout to obtain the layout data after OPC;

[0072] S12: Perform detection processing on the layout data after OPC using the aforementioned detection model after OPC, determine whether there are distorted graphics, and if so, perform correction processing on the distorted graphics and update the layout data after OPC.

[0073] S13: Perform OPC verification on the layout data after OPC.

[0074] In step S11, OPC processing is performed on the OPC layout to obtain the OPC layout data. In addition, the preset exposure conditions and corresponding photolithography process window and other related process conditions also need to be obtained.

[0075] In step S12, the layout data of OPC and the corresponding process conditions are input into the detection model after OPC and the detection process is performed to distinguish whether there is distortion related to abnormal top morphology of photoresist and the corresponding distortion level. If so, the detected distorted pattern is repaired until it is verified to be qualified, and then the repaired pattern is replaced with the distorted pattern in the layout data after OPC, that is, the layout data after OPC is updated. If not, no additional processing is performed.

[0076] In step S13, the layout data after the OPC test is performed as a whole.

[0077] Post-OPC verification typically includes CD error, EPE, pinch-off error, and bridging error, but does not cover photoresist top morphology distortion. Therefore, post-OPC verification is insufficient to detect such distortion. In some related technologies, etching model simulation is superimposed on the exposure model simulation in the verification process to detect hotspot defects caused by photoresist top morphology distortion. This makes the OPC verification process lengthy and inefficient. Furthermore, etching model simulation is accompanied by etching load effects, leading to the capture of false hotspot defects (not caused by photoresist top morphology distortion) during simulation. In this embodiment, after OPC processing and before OPC verification, the aforementioned detection model is used to detect photoresist top morphology distortion separately. This detection model is derived from the physical parameters of actual photoresist top morphology distortion, enabling rapid and accurate detection of the distortion.

[0078] In summary, this invention, after OPC correction and before OPC verification, sets up a post-OPC detection model to detect hot spot defects that are difficult to detect in related technologies using the post-OPC layout data. Unlike related technologies that use normal patterns to build models, the post-OPC detection model of this invention includes cases formed by the aforementioned patterns that are prone to hot spot defects under various exposure conditions, that is, it includes distortion cases with different distortion levels. Moreover, the measurement data and measurement images in the aforementioned distortion cases directly reflect the linewidth and image of the top morphology after photoresist exposure. Thus, the aforementioned difficult-to-detect hot spot defects can be specifically detected before the overall verification of the post-OPC layout data, thereby improving the effectiveness and efficiency of OPC correction and OPC verification.

[0079] The above description is merely a description of preferred embodiments of the present invention and is not intended to limit the scope of the present invention in any way. Any changes or modifications made by those skilled in the art based on the above disclosure shall fall within the protection scope of the claims.

Claims

1. A method for establishing a detection model after OPC, characterized in that, The application relates to a method for detecting abnormal top profile of photoresist after OPC. The method comprises the following steps: obtaining a first sample pattern after OPC, which has preset size parameters and exposure parameters, and the first sample pattern is prone to top profile distortion of photoresist after photolithography; obtaining an optical model of the first sample pattern and a corresponding distortion level; performing photolithography simulation on the optical model by using different exposure parameters to obtain measurement data and measurement images of different distortion levels, wherein the measurement data and the measurement images include line width and images reflecting the top profile of photoresist after exposure; 2. The method of claim 1, wherein the method further comprises: putting the corresponding distortion levels, the measurement data and the measurement images of the optical model under different exposure parameters into a sample pattern library to establish a detection model, which is used for detecting whether there is a pattern matching the optical model under the same exposure parameter in the layout data after OPC, and determining whether there is a pattern with abnormal top profile of photoresist in the layout data after OPC.

3. The method of claim 2, wherein the method further comprises: adjusting the size parameters of the first sample pattern to obtain a plurality of second sample patterns based on the first sample pattern, obtaining optical models of the first sample pattern and each second sample pattern and corresponding distortion indexes, dividing the first sample pattern and each second sample pattern into distortion levels from severe distortion to slight distortion according to the distortion indexes, and setting the size parameters of the plurality of second sample patterns at equal intervals with the size parameters of the first sample pattern as the center.

4. The method of claim 2 or 3, wherein the method further comprises: performing repair processing on the sample pattern with severe distortion to obtain a third sample pattern, and then obtaining an optical model of the third sample pattern and a distortion index. The steps of obtaining the optical model and the distortion index comprise the following steps: deducing the optical model according to the light source distribution of an exposure machine and corresponding target line diameter size; 5. The method of claim 1, wherein the method further comprises: obtaining the maximum light intensity and the minimum light intensity of a cross section of photoresist at different positions in the optical model, and the distortion index is the difference between the maximum light intensity and the minimum light intensity divided by the sum of the maximum light intensity and the minimum light intensity. The steps of performing photolithography simulation on each optical model by using different exposure parameters comprise the following steps: finding out the exposure condition which has the greatest influence on photoresist in the normal process window of each optical model as a key exposure condition; 6. The method of claim 1, wherein the method further comprises: constructing a plurality of exposure parameter combinations in the normal process window range with the key exposure condition as a variable to perform photolithography simulation on each optical model to obtain corresponding measurement data and measurement images.

7. The method of claim 1, wherein the method further comprises: When the detection model is established, the patterns put into the sample pattern library include normal patterns and patterns with top profile distortion of photoresist, and the patterns with top profile distortion of photoresist are composed of different distortion levels. When the measurement data is obtained after photolithography simulation, measuring the depth height of photoresist to determine whether the depth height is less than a preset height, if not, measuring the minimum line width of photoresist in a cross section as the line width, 8. The method of claim 1, wherein the method further comprises: if yes, taking the depth height as the measurement data. When it is difficult to accurately measure or determine the line width of part of the sample patterns, the line width of the sample pattern is estimated according to the actual line width variation trend of the sample pattern with a size close to that of the sample pattern, and the line width is made to conform to the actual line width variation trend.

9. An OPC post detection model characterized by, The method is used to establish the detection model after OPC according to any one of claims 1 to 8.

10. A method of optimizing OPC verification, characterized by, It comprises: performing OPC processing on the layout to obtain layout data after OPC; performing detection processing on the layout data after OPC using the detection model after OPC according to claim 9 to determine whether there is a distorted pattern, if so, performing correction processing on the distorted pattern and updating the layout data after OPC; performing OPC verification on the layout data after OPC.

Images