Method, device and equipment for generating training sample for training photoresist model

Abstract

The application discloses a training sample generation method, device and equipment for training a photoresist model. The method comprises the following steps: obtaining a first training sample, wherein the first training sample comprises a first mask pattern and a corresponding first photoresist label pattern; inputting the first training sample into the photoresist model to obtain a defect area in the first mask pattern, wherein the defect area is an area in which a difference between a photoresist pattern predicted according to the photoresist model and the first photoresist label pattern is greater than a preset difference threshold; determining a target layout area corresponding to the defect area in the first mask layout according to the defect area; and generating a second training sample for training the photoresist model according to the target layout area and the first photoresist label pattern. In the embodiment of the application, a mask layout in which the model training effect is poor is obtained through the photoresist pattern, and more training samples are obtained through the mask layout, so that the training effect of the model is improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of semiconductor manufacturing, and particularly relates to a training sample generation method, device and equipment for training a photoresist model. BACKGROUND

[0002] With the feature size of a design layout in large-scale integrated circuit manufacturing being continuously reduced, process equipment can only complete manufacturing under certain pattern conditions at an increasingly extreme process size. However, the exposure cost in the chip manufacturing process is high, and the manufacturing cycle is long. Therefore, in order to more efficiently and high-quality detect the manufacturability of the design layout, it is necessary to determine the photoresist pattern corresponding to the design layout.

[0003] In the related art, a deep learning model can be used to simulate the process from a mask layout to a photoresist pattern. However, under the background of high cost and long cycle of the chip exposure process, there is less data for training the deep learning model, which leads to a large error between the photoresist image determined by the deep learning model and the actual production photoresist image. Therefore, how to increase the training samples of the deep learning model is a technical problem that the related technical personnel urgently need to solve. SUMMARY

[0004] Embodiments of the present application provide a training sample generation method, device and equipment for training a photoresist model, which obtains a mask layout with poor model training effect of a photoresist pattern acquisition model, and obtains more training samples through the mask layout, thereby improving the training effect of the model.

[0005] In a first aspect, embodiments of the present application provide a training sample generation method for training a photoresist model, comprising: obtaining a first training sample, the first training sample comprising a first mask pattern and a first photoresist label pattern corresponding to the first mask pattern, the first mask pattern being a pattern obtained by sampling a first mask layout, and the first photoresist label pattern being a photoresist pattern obtained by correcting the first mask layout through an optical proximity effect correction model; inputting the first training sample into a photoresist model to obtain a defect region in the first mask pattern, the defect region being a region where the difference between the photoresist pattern predicted according to the photoresist model and the first photoresist label pattern is greater than a preset difference threshold; determining a target layout region corresponding to the defect region in the first mask layout according to the defect region; and generating a second training sample for training the photoresist model according to the target layout region and the first photoresist label pattern.

[0006] In an embodiment, the first training sample is input into the photoresist model to obtain a defect region in the first mask pattern, including: predicting a photoresist pattern of the first mask pattern according to the photoresist model to obtain a first predicted photoresist pattern; determining deviation information between the first predicted photoresist pattern and the first photoresist label pattern; and determining the defect region according to the deviation information in a case where the deviation information is greater than a preset deviation threshold.

[0007] In an embodiment, the deviation information between the predicted photoresist pattern and the photoresist label pattern is determined, including: obtaining a first coordinate value of a first feature point of the predicted photoresist pattern and a second coordinate value of a second feature point corresponding to the first feature point on the photoresist label pattern; and calculating a distance between the first coordinate value and the first coordinate value, and taking the distance value as the deviation information.

[0008] In an embodiment, the defect region is determined according to the deviation information, including: counting a number of the first feature points corresponding to the preset deviation threshold; and determining a region including the first feature points as the defect region in a case where the number of the first feature points is greater than a number threshold.

[0009] In an embodiment, a second training sample for training the photoresist model is generated according to the target layout region and the first photoresist label pattern, including: sampling the target layout region to obtain a second mask pattern; simulating the second mask pattern through an optical proximity correction model to obtain a photoresist pattern of the second mask pattern, and taking the photoresist pattern as a second photoresist label pattern of the second mask pattern; and generating the second training sample for training the photoresist model according to the second mask pattern and the corresponding second photoresist label pattern.

[0010] In an embodiment, the photoresist pattern of the second mask pattern is simulated through the optical proximity correction model, including: obtaining optical information and photoresist parameter information; determining a light intensity distribution image corresponding to the second mask pattern through the optical information and the photoresist parameter information; and obtaining the photoresist pattern of the second mask pattern by intercepting the light intensity distribution image according to a preset light intensity threshold.

[0011] In an implementation, before the first training sample is input into the photoresist model, the method further comprises: obtaining a training sample set, the training sample set comprising a plurality of third mask patterns and corresponding third photoresist label patterns, the third mask patterns being patterns obtained by sampling a third mask plate pattern, and the third photoresist label patterns being photoresist patterns obtained by simulating the third mask plate pattern by an optical proximity effect correction model; predicting photoresist patterns of the third mask patterns according to the preset photoresist model to obtain third predicted photoresist patterns; determining a learning error of the photoresist model according to errors between the third predicted photoresist patterns and the third photoresist label patterns; in a case where the learning error is greater than a preset error threshold, adjusting parameters in the preset photoresist pattern obtaining model, and returning to predicting photoresist patterns of the third mask patterns according to the preset photoresist model to obtain third predicted photoresist patterns until the learning error is less than or equal to the preset error threshold, and obtaining a trained photoresist pattern obtaining model.

[0012] In an implementation, after the trained photoresist pattern obtaining model is obtained, the method further comprises: updating the trained photoresist model by using the training sample set and the second training sample.

[0013] In a second aspect, an embodiment of the present application provides a training sample generation device for training a photoresist model, the device comprising: a first obtaining module configured to obtain a first training sample, the first training sample comprising a first mask pattern and a corresponding first photoresist label pattern, the first mask pattern being a pattern obtained by sampling a first mask plate pattern, and the first photoresist label pattern being a photoresist pattern obtained by correcting the first mask plate pattern by an optical proximity effect correction model; a second obtaining module configured to input the first training sample into the photoresist model to obtain a defect region in the first mask pattern, the defect region being a region where a difference between a photoresist pattern predicted according to the photoresist model and the first photoresist label pattern is greater than a preset difference threshold; a determining module configured to determine a target layout region corresponding to the defect region in the first mask plate pattern according to the defect region; and a generating module configured to generate a second training sample for training the photoresist model according to the target layout region and the first photoresist label pattern.

[0014] In a third aspect, an embodiment of the present application provides a training sample generation device for training a photoresist model, the device comprising: a processor and a memory storing computer program instructions; and the processor implements the training sample generation method for training the photoresist model in the first aspect or any implementation of the first aspect when executing the computer program instructions.

[0015] In a fourth aspect, a computer-readable storage medium is provided. The computer-readable storage medium stores computer program instructions. When the computer program instructions are executed by a processor, the computer program instructions implement the training sample generation method for training a photoresist model according to any one of the first aspect or the implementation forms of the first aspect.

[0016] In a fifth aspect, a computer program product is provided. When instructions in the computer program product are executed by a processor of an electronic device, the instructions cause the electronic device to perform the training sample generation method for training a photoresist model according to the first aspect or any one of the implementation forms of the first aspect.

[0017] The training sample generation method, the device, and the apparatus for training a photoresist model according to the embodiments of the present application determine a defect area in which the training effect of the photoresist model is poor in the first training sample, and determine a target layout area corresponding to the defect area. Further, the second training sample for training the photoresist model is generated by using the target layout area and the first photoresist label pattern in the first training sample, so as to increase the training sample of the photoresist model and improve the accuracy of the model. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the embodiments of the present application will be briefly introduced. For those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0019] Figure 1 A flowchart of the training sample generation method for training a photoresist model provided by an embodiment of the present application is shown;

[0020] Figure 2 A flowchart of the training sample generation method for training a photoresist model provided by another embodiment of the present application is shown;

[0021] Figure 3 A schematic diagram of a C2C detector provided by an embodiment of the present application is shown;

[0022] Figure 4 A flowchart of the training sample generation method for training a photoresist model provided by another embodiment of the present application is shown;

[0023] Figure 5 A flowchart of the photoresist pattern generation method of a second mask pattern provided by an embodiment of the present application is shown;

[0024] Figure 6A flowchart of a method for generating training samples for training a photoresist model is shown.

[0025] Figure 7 A structural diagram of an apparatus for generating training samples for training a photoresist model is shown.

[0026] Figure 8 A structural diagram of an apparatus for generating training samples for training a photoresist model is shown. DETAILED DESCRIPTION

[0027] The features and exemplary embodiments of the various aspects of the present application will be described in detail below with reference to the drawings. The following detailed description is merely intended to explain the present application, and is not intended to limit the present application. The present application can be implemented without some of the specific details, which are well known to those skilled in the art. The following description of the embodiments is merely provided to give a better understanding of the present application by showing examples of the present application.

[0028] It should be noted that the terms such as first and second, etc., are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or device. Without more limitations, the elements defined by the statement "include" do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0029] In the field of semiconductor manufacturing technology, the manufacturability of a design layout can be determined more efficiently and with higher quality by determining a photoresist pattern corresponding to a mask layout.

[0030] In the related art, a photoresist pattern can be determined by an optical proximity correction (OPC) model. However, as chips move towards smaller processes, the physical and chemical effects in the photolithography process become more complex and difficult to predict, so the process of determining the photoresist pattern by the OPC model is more cumbersome and takes longer. It can be understood that the photoresist pattern is the pattern formed on the wafer surface after the photoresist in the related photolithography process (for example, exposure, development, etc.). And the photoresist pattern can usually accurately reproduce the design pattern on the mask.

[0031] In addition, in the related art, the process from the mask pattern to the photoresist pattern can also be simulated by a deep learning model. However, due to the high cost and long cycle of the exposure process in chip manufacturing, there is less training data for training the deep learning model, which leads to unsatisfactory training effect of the deep learning model. Therefore, how to increase the training samples of the deep learning model is a technical problem that the related technical personnel urgently need to solve.

[0032] To solve the technical problems in the prior art, the embodiments of the present application provide a training sample generation method, device and equipment for training a photoresist model. First, the training sample generation method for training the photoresist model provided by the embodiments of the present application is introduced.

[0033] Figure 1 The flowchart of the training sample generation method for training the photoresist model provided by an embodiment of the present application is shown. As shown in Figure 1 The training sample generation method for training the photoresist model in the embodiments of the present application includes the following steps:

[0034] S110, obtaining a first training sample, the first training sample including a first mask pattern and a corresponding first photoresist label pattern.

[0035] The first mask pattern is a pattern obtained by sampling the first mask pattern. The first photoresist label pattern is a photoresist pattern obtained by correcting the first mask pattern by an optical proximity correction model.

[0036] In the embodiments of the present application, the first mask pattern can be a pattern obtained by sampling the first mask pattern. In an example, the first mask pattern can be obtained by randomly sampling the first mask pattern. In another example, the first mask pattern can also be obtained by feature matching to determine the first mask pattern in the first mask pattern. The first mask pattern obtained by feature matching can be a pattern that needs to be paid attention to by the technician in advance, or it can also be some experience pattern.

[0037] In the embodiments of the present application, the first photoresist tag pattern corresponding to the first mask pattern can be a photoresist pattern obtained by correcting the first mask plate pattern through an optical proximity correction (OPC) model. It can be understood that, due to the high cost and long cycle of the exposure process in chip manufacturing, obtaining the actual generated photoresist pattern corresponding to the first mask pattern will cause high cost and take a long time, which does not meet the needs of the chip manufacturing industry for efficient production and cost control. Moreover, the OPC model can well simulate the process from the mask plate pattern to the generated photoresist pattern, and the photoresist pattern obtained through the OPC model is also closer to the pattern in the design layout. Therefore, the photoresist pattern obtained by correcting the first mask plate pattern through the OPC model can be used as the photoresist tag pattern corresponding to the first mask plate pattern.

[0038] S120, input the first training sample into the photoresist model to obtain a defect area in the first mask pattern.

[0039] The defect area is an area in which the difference between the photoresist pattern predicted by the photoresist model and the first photoresist tag pattern is greater than a preset difference threshold. The photoresist model can be a pre-trained photoresist model.

[0040] In the embodiments of the present application, the first training sample can be input into the photoresist training model, so that the photoresist model simulates the first training sample, thereby determining the defect area. It can be understood that the defect area can be an area in which the photoresist model has poor training effect on the first sample. That is, the defect area can be an area in which the difference between the photoresist pattern predicted by the photoresist model and the first photoresist tag pattern is greater than a preset difference threshold.

[0041] S130, determining a target layout area corresponding to the defect area in the first mask plate pattern according to the defect area.

[0042] In the embodiments of the present application, the area in the first mask plate pattern corresponding to the defect area can be used as the target layout area.

[0043] S140, generating a second training sample for training the photoresist model according to the target layout area and the first photoresist tag pattern.

[0044] In the embodiments of the present application, the second training sample can be generated according to the target layout area and the first photoresist tag pattern corresponding to the target layout area. The second training sample can be used to train the photoresist model.

[0045] In an example, the target mask pattern can be obtained in the target layout area by sampling the target layout area.

[0046] In the embodiment of the present application, the first training sample including the first mask pattern and the corresponding first photoresist label pattern is obtained, and the first training sample is input into the photoresist model, so as to determine the defect area in the first mask pattern. Further, the target layout area corresponding to the defect area in the first mask pattern can be determined according to the determined defect area, and the second training sample for training the photoresist model is generated by using the target layout area and the first photoresist label pattern.

[0047] In the embodiment of the present application, the first training sample is input into the photoresist model, so as to determine the area with poor training effect in the first training sample, i.e. the defect area, and determine the target layout area corresponding to the defect area. Further, the second training sample for training the photoresist model is generated by using the target layout area and the first photoresist label pattern in the first training sample, so as to increase the training sample of the photoresist model and improve the accuracy of the model.

[0048] In order to determine the defect area, as another implementation manner of the present application, another implementation manner of the training sample generation method for training the photoresist model is also provided, which is specifically described in the following embodiment.

[0049] Figure 2 The flowchart of the training sample generation method for training the photoresist model provided by another embodiment of the present application is shown. As shown in the figure, the training sample generation method for training the photoresist model in the embodiment of the present application includes the following steps: Figure 2

[0050] S210, a first training sample is obtained, and the first training sample includes a first mask pattern and a corresponding first photoresist label pattern.

[0051] The step S210 is consistent with the step S110, and the steps S250-S260 are consistent with the steps S130-S140, which will not be described in detail here.

[0052] S220, the photoresist pattern of the first mask pattern is predicted according to the photoresist model, and a first predicted photoresist pattern is obtained.

[0053] In the embodiment of the present application, the first mask pattern can be input into the photoresist model, and the first predicted photoresist pattern is determined by predicting the first mask pattern by the photoresist model. It can be understood that the first predicted photoresist pattern can be predicted by the photoresist model, and the photoresist can leave a pattern on the wafer surface after passing through the first mask pattern and the related photoetch process.

[0054] ​S230, determine deviation information between the first predicted photoresist pattern and the first photoresist label pattern.

[0055] In the embodiments of the present application, the deviation information between the first predicted photoresist pattern and the first photoresist label pattern is determined by matching the first predicted photoresist pattern and the first photoresist label pattern.

[0056] In some optional embodiments, the first coordinate value of the first feature point of the first predicted photoresist pattern and the second coordinate value of the second feature point corresponding to the first feature point on the first photoresist label pattern can be acquired; the distance value between the first coordinate value and the first coordinate value is calculated, and the distance value is taken as the deviation information.

[0057] For example, the first feature point and the second feature point can be the same positions on the first predicted photoresist pattern and the first photoresist label pattern respectively. Further, the difference between the coordinate values corresponding to the same positions on the first predicted photoresist pattern and the first photoresist label pattern is determined, and the difference between the coordinate values is taken as the deviation information.

[0058] In an example, the calculation of the deviation information can be realized by a contour comparison detector, for example, the C2C (contour to contour) detector. Figure 3 A schematic diagram of the C2C detector provided by an embodiment of the present application is shown. As shown in the figure, Figure 3 The contour 10 can be the first predicted photoresist pattern, the contour 20 can be the first photoresist label pattern, and the feature points 1, 2 and 3 can be set, and the corresponding deviation information, i.e. the distance between the contour 10 and the contour 20, can be calculated.

[0059] In another example, the calculation of the deviation information can be realized by an EPE (Edge Placement Error) detector. It can be understood that the above-mentioned method for calculating the deviation information is used for exemplary description, and the calculation of the deviation information can also be realized by the END (line-end / space-end) detection and the CDV (Critical Dimension Variation) detection, which is not limited in the embodiments of the present application.

[0060] S240, in the case that the deviation information is greater than the preset deviation threshold, determining the defect region according to the deviation information.

[0061] In the embodiments of the present application, in a case where it is determined that the deviation information is greater than the preset deviation threshold, the defect region can be determined according to the deviation information. The preset deviation threshold can be set by a technician in advance, and the size of the preset deviation threshold can be related to the photoresist model accuracy. For example, if a higher photoresist model accuracy is required, the preset deviation threshold can be set to a smaller value; if a higher photoresist model accuracy is required, the preset deviation threshold can be set to a larger value.

[0062] In some optional embodiments, the number of first feature points corresponding to the deviation information greater than the preset deviation threshold can be counted; in a case where it is determined that the number of first feature points is greater than a number threshold, the region including the first feature points is determined as the defect region.

[0063] In the embodiments of the present application, the region with more defect positions in the first mask pattern can be taken as the defect region. The defect positions in the first mask pattern can be positions with deviation information greater than the preset deviation threshold.

[0064] It can be understood that, in the embodiments of the present application, the number of first feature points corresponding to the deviation information greater than the preset deviation threshold is counted, and in a case where the number of first feature points is greater than a number threshold, the region including the first feature points is determined as the defect region. That is, in the embodiments of the present application, the region with more defects in the first mask pattern is taken as the defect region, so that the detection of the region with poor photoresist model training effect is realized.

[0065] S250, determining a target layout region corresponding to the defect region in the first mask layout according to the defect region.

[0066] S260, generating a second training sample for training the photoresist model according to the target layout region and the first photoresist label pattern.

[0067] In the embodiments of the present application, the first training sample including the first mask pattern and the corresponding first photoresist label pattern is obtained, and the first training sample is input into the photoresist model, so that the first predicted photoresist pattern is obtained, and the deviation information between the first predicted photoresist pattern and the first photoresist label pattern is calculated. In a case where it is determined that the deviation information is greater than the preset deviation threshold, the defect region in the first mask pattern is determined according to the deviation information. Further, the target layout region corresponding to the defect region in the first mask layout can be determined according to the determined defect region; and then the second training sample is generated by using the target layout region and the first photoresist label pattern.

[0068] In the embodiment of the present application, the area with more defects in the first mask pattern is taken as the defect area, so as to detect the area with poor training effect of the photoresist model and determine the area with poor training effect of the photoresist model.

[0069] In order to determine the defect area, as another implementation manner of the present application, the present application further provides another implementation manner of the training sample generation method for training the photoresist model, which is specifically described in the following embodiment.

[0070] Figure 4 A flowchart of the training sample generation method for training the photoresist model provided by still another embodiment of the present application is shown. As shown in the figure, the training sample generation method for training the photoresist model in the embodiment of the present application includes the following steps: Figure 4

[0071] S410, obtaining a first training sample, the first training sample including a first mask pattern and a first photoresist label pattern corresponding to the first mask pattern.

[0072] S420, inputting the first training sample into the photoresist model to obtain a defect area in the first mask pattern.

[0073] S430, determining a target layout area corresponding to the defect area in the first mask pattern according to the defect area.

[0074] The steps S410-S430 are consistent with the steps S110-S130, and thus will not be described in detail.

[0075] S440, sampling the target layout area to obtain a second mask pattern.

[0076] In the embodiment of the present application, the second mask pattern can be a pattern obtained by sampling the target layout area. The sampling operation on the target layout area can be random sampling.

[0077] For example, sampling the target layout area can obtain a plurality of second mask patterns.

[0078] S450, simulating the second mask pattern by using an optical proximity correction model to obtain a photoresist pattern of the second mask pattern, and taking the photoresist pattern as a second photoresist label pattern of the second mask pattern.

[0079] ​In the embodiment of the present application, the second mask pattern can be input into the OPC model, and the photoresist pattern corresponding to the second mask pattern is predicted by using the OPC model, so as to obtain the photoresist pattern corresponding to the second mask pattern. It can be understood that the photoresist pattern predicted by the OPC model has high consistency with the pattern in the design layout. Therefore, the photoresist pattern predicted by the OPC model for the second mask pattern can be used as the second photoresist tag pattern.

[0080] For example, the second mask pattern and the second photoresist tag pattern have a corresponding relationship.

[0081] S460, according to the second mask pattern and the corresponding second photoresist tag pattern, a second training sample for training the photoresist model is generated.

[0082] In the embodiment of the present application, the second mask pattern and the corresponding second photoresist tag pattern can be used as the second training sample to train the photoresist model.

[0083] In the embodiment of the present application, the target layout area can be sampled to obtain the second mask pattern, and the second photoresist tag pattern corresponding to the second mask pattern is determined. It can be understood that the target layout area is an area with poor training effect for the photoresist model. At this time, the second training sample obtained from the target layout area can make the photoresist model compensate for the poor training effect on the target layout area when training according to the second training sample, and improve the accuracy of the model.

[0084] In order to obtain the second photoresist tag pattern, as another implementation manner of the present application, another implementation manner of the training sample generation method for training the photoresist model is provided, which is described in the following embodiment.

[0085] Figure 5 A flowchart of a photoresist pattern method of a second mask pattern provided by an embodiment of the present application is shown. As shown in Figure 5 The photoresist pattern method of the second mask pattern in the embodiment of the present application includes the following steps:

[0086] S510, obtaining optical information and photoresist parameter information.

[0087] In the embodiment of the present application, the optical information can be optical information in the photoetching process. For example, the optical information can include information related to the light source, including the wavelength and power of the light.

[0088] In the embodiment of the present application, the photoresist parameter information can include related parameter information of the photoresist used in the photoetching process. For example, the photoresist parameter information can include sensitivity, viscosity and other parameter information.

[0089] S520, determine the light intensity distribution image corresponding to the second mask pattern through the optical information and the photoresist parameter information.

[0090] In the embodiment of the present application, the light intensity distribution image of the light passing through the second mask pattern in the photoresist can be simulated through the optical information and the photoresist parameter information.

[0091] S530, intercept the light intensity distribution image according to the preset light intensity threshold to obtain the photoresist pattern of the second mask pattern.

[0092] In the embodiment of the present application, the preset light intensity threshold can be the light intensity value determined in the training process of the OPC model. It can be understood that the preset light intensity threshold can reflect whether the photoresist will be exposed and chemically reacted. For example, the light intensity region higher than the threshold in the light intensity distribution image can be regarded as an exposed region (i.e., a region in which the photoresist needs to be removed), and the light intensity region lower than the threshold in the light intensity distribution image can be regarded as an unexposed region (i.e., a region in which the photoresist is retained). Through the above-mentioned manner, the photoresist pattern of the second mask pattern can be intercepted from the light intensity distribution image, that is, the pattern shape on the photoresist after lithography can be predicted.

[0093] In the embodiment of the present application, the optical information and the photoresist parameter information are obtained, and the light intensity distribution image corresponding to the second mask pattern is determined through the optical information and the photoresist parameter information. The photoresist pattern of the second mask pattern is obtained by intercepting the light intensity distribution image according to the preset light intensity threshold, so as to realize the acquisition of the photoresist pattern. It can be understood that the photoresist pattern determined through the OPC model has a high matching degree with the pattern in the design layout, and is used as the true value in the training process of the photoresist model, so as to ensure the accuracy of the photoresist model.

[0094] In order to realize the training of the photoresist model, as another implementation manner of the present application, the present application further provides another implementation manner of the training sample generation method for training the photoresist model, which is specifically described in the following embodiment.

[0095] Figure 6 The flowchart of the training sample generation method for training the photoresist model provided by another embodiment of the present application is shown. As shown in the figure, Figure 6 After the first training sample is input into the photoresist model, the training sample generation method for training the photoresist model in the embodiment of the present application further includes the following steps:

[0096] S610, obtain a training sample set.

[0097] The training sample set includes a plurality of third mask patterns and corresponding third photoresist label patterns. The third mask pattern is a pattern obtained by sampling a third mask plate, and the third photoresist label pattern is a photoresist pattern obtained by simulating the third mask plate through an optical proximity effect correction model.

[0098] In the embodiments of the present application, the third mask plate in the training sample set can be distinguished from the first mask plate. Therefore, the third mask pattern obtained by sampling the third mask plate can be different from the first mask pattern.

[0099] In the embodiments of the present application, the training sample set can be used to train the photoresist model to obtain a trained photoresist model.

[0100] S620, predicting a photoresist pattern of the third mask pattern according to a preset photoresist model to obtain a third predicted photoresist pattern.

[0101] In the embodiments of the present application, the third mask pattern can be input into the preset photoresist model, so that the predicted photoresist pattern predicts the photoresist pattern corresponding to the third mask pattern to obtain the third predicted photoresist pattern.

[0102] In the embodiments of the present application, the preset photoresist model can be a function expression combined by related technical personnel to simulate the process of forming an image in the photoresist by light passing through the third mask pattern in the third mask plate.

[0103] S630, determining a learning error of the preset photoresist model according to the error between the third predicted photoresist pattern and the third photoresist label pattern.

[0104] In the embodiments of the present application, the learning error of the predicted photoresist model can be determined through the error between the third predicted photoresist pattern and the third photoresist label pattern. In an example, the error between the third predicted photoresist pattern and the third photoresist label pattern can be calculated by using a C2C detector. In another example, the error between the third predicted photoresist pattern and the third photoresist label pattern can also be calculated by using an EPE detector. It can be understood that the error between the third predicted photoresist pattern and the third photoresist label pattern can also be calculated by other ways, which are not limited in the embodiments of the present application.

[0105] S640, determining whether the learning error is greater than a preset error threshold, if yes, performing step S650; if no, ending the training to obtain a trained photoresist model.

[0106] S650, adjusting parameters in the preset photoresist pattern acquisition model.

[0107] In the embodiments of the present application, whether the calculated learning error is greater than the preset error threshold is determined to determine the subsequent operation. Wherein, in the case of determining that the learning error is greater than the preset error threshold, it can be determined that the photoresist model does not meet the requirements, at this time, the parameters in the preset photoresist model can be adjusted, and the third photoresist pattern is re-input into the photoresist model after the parameter adjustment, the third predicted photoresist pattern is re-predicted, and the difference between the re-predicted third predicted photoresist pattern and the third photoresist label pattern is calculated, and the learning error is re-calculated. Further, the relationship between the re-calculated learning error and the preset error threshold is determined, if the re-calculated learning error is still greater than the preset error threshold, the photoresist model parameters are continuously adjusted until the learning error is less than or equal to the preset error threshold, it is determined that the model training is completed, and the trained photoresist model is obtained.

[0108] Wherein, in the case of determining that the learning error is less than or the preset error threshold, the photoresist model can be determined, and it is determined that the model training is completed, at this time the training can be ended, and the trained photoresist model is obtained.

[0109] In the embodiments of the present application, the third mask pattern in the training sample is input into the preset photoresist model to obtain the third predicted photoresist pattern. Further, the third predicted photoresist pattern and the third photoresist label pattern are used to determine the deviation calculation learning error. And through the learning error, it is judged whether the photoresist model is completed. It can be understood that the photoresist model trained through the above process has high accuracy for its training sample, and the mask pattern similar to the third mask pattern in the training sample can be quickly and accurately predicted to the photoresist pattern.

[0110] In some optional embodiments, the trained photoresist model can be updated by using the training sample set and the second training sample. It can be understood that the second training sample is a training sample obtained by the area where the model training effect is not good. Therefore, retraining the trained photoresist model by using the training sample set and the second training sample can make up for the defects of the photoresist model in the initial training process, thereby improving the accuracy of the model for predicting the photoresist pattern.

[0111] Based on the training sample generation method for training the photoresist model provided in the above embodiments, correspondingly, the present application also provides a specific implementation manner of a training sample generation device for training the photoresist model. Please see the following embodiments.

[0112] First see Figure 7 The training sample generation device 700 for training the photoresist model provided in the embodiments of the present application includes the following units:

[0113] The first obtaining module 701 is configured to obtain a first training sample, the first training sample comprising a first mask pattern and a corresponding first photoresist label pattern, the first mask pattern being a pattern obtained by sampling a first mask plate pattern, and the first photoresist label pattern being a photoresist pattern obtained by correcting the first mask plate pattern by using an optical proximity effect correction model;

[0114] The second obtaining module 702 is configured to input the first training sample into a photoresist model to obtain a defect region in the first mask pattern, the defect region being a region in which a difference between a photoresist pattern predicted according to the photoresist model and the first photoresist label pattern is greater than a preset difference threshold.

[0115] The determining module 703 is configured to determine, according to the defect region, a target layout region corresponding to the defect region in the first mask plate pattern.

[0116] The generating module 704 is configured to generate, according to the target layout region and the first photoresist label pattern, a second training sample for training the photoresist model.

[0117] As an implementation manner of the present application, the second obtaining module 702 inputs the first training sample into the photoresist model to obtain the defect region in the first mask pattern in the following manner: predicting a photoresist pattern of the first mask pattern according to the photoresist model to obtain a first predicted photoresist pattern; determining deviation information between the first predicted photoresist pattern and the first photoresist label pattern; and determining the defect region according to the deviation information in a case where it is determined that the deviation information is greater than a preset deviation threshold.

[0118] As an implementation manner of the present application, the second obtaining module 702 determines the deviation information between the predicted photoresist pattern and the photoresist label pattern in the following manner: obtaining a first coordinate value of a first feature point of the first predicted photoresist pattern and a second coordinate value of a second feature point corresponding to the first feature point on the first photoresist label pattern; and calculating a distance between the first coordinate value and the first coordinate value, and taking the distance value as the deviation information.

[0119] As an implementation manner of the present application, the determining module 703 determines the defect region according to the deviation information in the following manner: counting a number of the first feature points corresponding to the deviation information greater than the preset deviation threshold; and determining a region comprising the first feature points as the defect region in a case where it is determined that the number of the first feature points is greater than a number threshold.

[0120] As an implementation form of the present application, the generating module 704 generates the second training sample for training the photoresist model according to the target layout region and the first photoresist tag pattern in the following manner: sampling the target layout region to obtain a second mask pattern; simulating the second mask pattern by the optical proximity correction model to obtain a photoresist pattern of the second mask pattern, and taking the photoresist pattern of the second mask pattern as a second photoresist tag pattern of the second mask pattern; and generating the second training sample for training the photoresist model according to the second mask pattern and the corresponding second photoresist tag pattern.

[0121] As an implementation form of the present application, the generating module 704 simulates the second mask pattern by the optical proximity correction model to obtain the photoresist pattern of the second mask pattern in the following manner: obtaining optical information and photoresist parameter information; determining the light intensity distribution image corresponding to the second mask pattern by the optical information and the photoresist parameter information; and obtaining the photoresist pattern of the second mask pattern by intercepting the light intensity distribution image according to a preset light intensity threshold.

[0122] As an implementation form of the present application, the device further comprises a training module 705, which is configured to: obtain a training sample set before the first training sample is input into the photoresist model, the training sample set comprising a plurality of third mask patterns and corresponding third photoresist tag patterns, the third mask pattern being a pattern obtained by sampling a third mask plate, and the third photoresist tag pattern being a photoresist pattern obtained by simulating the third mask plate by the optical proximity correction model; and predict the photoresist pattern of the third mask pattern by the preset photoresist model to obtain a third predicted photoresist pattern.

[0123] determining a learning error of the predicted photoresist model according to the error between the third predicted photoresist pattern and the third photoresist tag pattern; adjusting the parameters in the preset photoresist pattern acquisition model in a case where the learning error is greater than a preset error threshold, and returning to predict the photoresist pattern of the third mask pattern by the preset photoresist model to obtain the third predicted photoresist pattern until the learning error is less than or equal to the preset error threshold, thereby obtaining the trained photoresist model.

[0124] As an implementation form of the present application, after obtaining the trained photoresist model, the training module 705 is further configured to update the trained photoresist model by using the training sample set and the second training sample.

[0125] Figure 8 A hardware structure schematic diagram of the training sample generation method for training the photoresist model provided by the embodiment of the present application is shown.

[0126] The training sample generation method for training a photoresist model can include a processor 301 and a memory 302 having stored computer program instructions.

[0127] In particular, the processor 801 can include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement one or more embodiments of the present application.

[0128] The memory 802 can include a mass storage for data or instructions. By way of example, and not limitation, the memory 802 can include a hard disk drive (HDD), a floppy disk drive, flash memory, an optical disc (e.g., a CD or DVD), a tape drive, a USB drive, or a combination of two or more of these. The memory 802 can include removable or non-removable (or fixed) media, where appropriate. The memory 802 can be internal or external to the integrated gateway disaster recovery device, as appropriate. In certain embodiments, the memory 802 is nonvolatile, solid-state memory.

[0129] The memory can include read-only memory (ROM), random-access memory (RAM), magnetic disk storage mediums, optical storage mediums, flash memory devices, electrical, optical, or other physically tangible / moφhological memory storage devices. Thus, in general, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software that, when executed (by one or more processors), is operable to

[0130] The processor 801 implements any of the above-described embodiments of the training sample generation method for training a photoresist model by reading and executing computer program instructions stored in the memory 802.

[0131] In one example, the training sample generation device for training a photoresist model can further include a communication interface 803 and a bus 810. As shown, the processor 801, the memory 802, and the communication interface 803 are connected by the bus 810 and complete communication among each other. Figure 8

[0132] The communication interface 803 is mainly used to realize the communication between the modules, devices, units and / or equipment in the embodiments of the present application.

[0133] ​Bus 810 includes a hardware, software, or both that couples components of the online data traffic metering device to each other. As an example and not by way of limitation, the bus can include an accelerated graphics port (AGP) or other graphics bus, an enhanced industry standard architecture (EISA) bus, a front-side bus (FSB), a HyperTransport (HT) interconnect, an industry standard architecture (ISA) bus, an InfiniBand (IB) interconnect, a low pin count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a peripheral component interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a serial advanced technology attachment (SATA) bus, a video electronics standards board (VLB) bus, or another suitable bus or a combination of two or more of these. Where appropriate, bus 810 can include one or more buses. Although this application describes and shows a particular bus, this application contemplates any suitable bus or interconnect.

[0134] The training sample generation device for training a photoresist model can execute the training sample generation method for training a photoresist model in the embodiments of the application to achieve the training sample generation method for training a photoresist model and the device for training a photoresist model described in combination with Figure 1 and Figure 7 the above embodiments.

[0135] In addition, in combination with the training sample generation method for training a photoresist model in the above embodiments, the embodiments of the application can provide a computer storage medium to implement. The computer storage medium has computer program instructions stored thereon; the computer program instructions are executed by a processor to implement any one of the training sample generation methods for training a photoresist model in the above embodiments.

[0136] The embodiments of the application also provide a computer program product, including a computer program, the computer program is executed by a processor to implement the method of the training sample generation for training a photoresist model in the above embodiments.

[0137] It needs to be clear that the application is not limited to the specific configurations and processes described above and shown in the drawings. For the sake of brevity, detailed descriptions of well-known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between steps, after understanding the spirit of the application.

[0138] The functions noted in the description of the structural block diagrams above can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, and the like. When implemented in software, the elements of the present application are program or code segments that are used to perform the required tasks. The program or code segments can be stored in a machine-readable medium, or transmitted through a data signal carried in a carrier wave over a transmission medium or communication link. A "machine-readable medium" includes any medium that can store or transport information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, and the like. The code segments can be downloaded via computer networks such as the Internet, intranets, and the like.

[0139] It is also important to note that the examples mentioned in the present application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the steps mentioned above, that is, the steps can be performed in the order mentioned in the examples, or in an order different from the examples, or several steps can be performed simultaneously.

[0140] The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer program instructions can also be stored in a computer readable medium that can direct a computer, other programmable data processing apparatus, or other processing device to operate in a particular manner, such that the instructions stored in the computer readable medium produce an article of manufacture including instructions which implement the function / act specified in the flowchart and / or block diagram block or blocks. The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer program instructions can also be stored in a computer readable medium that can direct a computer, other programmable data processing apparatus, or other processing device to operate in a particular manner, such that the instructions stored in the computer readable medium produce an article of manufacture including instructions which implement the function / act specified in the flowchart and / or block diagram block or blocks.

[0141] The above merely describes a specific implementation of the present application. Those skilled in the art can clearly understand the specific working processes of the system, modules and units described above for the convenience and brevity of description, and can refer to the corresponding processes in the foregoing method embodiments, which will not be described herein again. It should be understood that the protection scope of the present application is not limited to this, and any person skilled in the art can easily think of various equivalent modifications or replacements within the technical range disclosed by the present application, and these modifications or replacements should be covered within the protection scope of the present application.

Claims

1. A method for generating training samples for training photoresist models, characterized in that, include: Obtain a first training sample, which includes a first mask pattern and its corresponding first photoresist label pattern. The first mask pattern is a pattern obtained by sampling the first mask pattern, and the first photoresist label pattern is a photoresist pattern obtained by correcting the first mask pattern through an optical proximity effect correction model. The first training sample is input into the photoresist model to obtain the defect region in the first mask pattern. The defect region is the region where the difference between the photoresist pattern predicted by the photoresist model and the first photoresist label pattern is greater than a preset difference threshold. Based on the defective region, a target layout region corresponding to the defective region is determined in the first mask layout. Based on the target layout area and the first photoresist label pattern, a second training sample is generated for training the photoresist model.

2. The method according to claim 1, characterized in that, The step of inputting the first training sample into the photoresist model to obtain the defect region in the first mask pattern includes: Based on the photoresist model, the photoresist pattern of the first mask pattern is predicted to obtain the first predicted photoresist pattern. Determine the deviation information between the first predicted photoresist pattern and the first photoresist label pattern; If the deviation information is determined to be greater than a preset deviation threshold, the defect area is determined based on the deviation information.

3. The method according to claim 2, characterized in that, The step of determining the deviation information between the predicted photoresist pattern and the photoresist label pattern includes: Obtain the first coordinate value of the first feature point of the first predicted photoresist pattern and the second coordinate value of the second feature point on the first photoresist label pattern corresponding to the first feature point; Calculate the distance between the first coordinate value and the first coordinate, and use the distance value as the deviation information.

4. The method according to claim 2 or 3, characterized in that, The step of determining the defect area based on the deviation information includes: Count the number of first feature points whose deviation information is greater than a preset deviation threshold; If the number of the first feature points is greater than the number threshold, the region including the first feature points is defined as the defect region.

5. The method according to claim 1, characterized in that, The step of generating a second training sample for training the photoresist model based on the target layout region and the first photoresist label pattern includes: The target area is sampled to obtain a second mask pattern; The second mask pattern is simulated using an optical proximity effect correction model to obtain the photoresist pattern of the second mask pattern, which is then used as the second photoresist label pattern of the second mask pattern. Based on the second mask pattern and the corresponding second photoresist label pattern, a second training sample is generated for training the photoresist model.

6. The method according to claim 5, characterized in that, The step of simulating the second mask pattern using an optical proximity effect correction model to obtain the photoresist pattern of the second mask pattern includes: Acquire optical information and photoresist parameter information; The light intensity distribution image corresponding to the second mask pattern is determined using the optical information and photoresist parameter information. The light intensity distribution image is cropped according to a preset light intensity threshold to obtain the photoresist pattern of the second mask pattern.

7. The method according to claim 1, characterized in that, Before inputting the first training sample into the photoresist model, the method further includes: A training sample set is obtained, which includes multiple third mask patterns and their corresponding third photoresist label patterns. The third mask pattern is a pattern obtained by sampling the third mask pattern, and the third photoresist label pattern is a photoresist pattern obtained by simulating the third mask pattern through an optical proximity effect correction model. The photoresist pattern of the third mask is predicted according to the preset photoresist model to obtain the third predicted photoresist pattern. The learning error of the preset photoresist model is determined based on the error between the third predicted photoresist pattern and the third photoresist label pattern. If the learning error is determined to be greater than a preset error threshold, the parameters in the preset photoresist model are adjusted, and the photoresist pattern of the third mask pattern is predicted according to the preset photoresist model to obtain the third predicted photoresist pattern, until the learning error is less than or equal to the preset error threshold, and the trained photoresist model is obtained.

8. The method according to claim 7, characterized in that, After obtaining the trained photoresist model, the method further includes: The trained photoresist model is updated using the training sample set and the second training sample.

9. A training sample generation device for training photoresist models, characterized in that, The device includes: The first acquisition module is used to acquire the first training sample, which includes a first mask pattern and its corresponding first photoresist label pattern. The first mask pattern is a pattern obtained by sampling the first mask pattern, and the first photoresist label pattern is a photoresist pattern obtained by correcting the first mask pattern through an optical proximity effect correction model. The second acquisition module is used to input the first training sample into the photoresist model to obtain the defect region in the first mask pattern. The defect region is the region where the difference between the photoresist pattern predicted by the photoresist model and the first photoresist label pattern is greater than a preset difference threshold. The determining module is used to determine the target layout area corresponding to the defect area in the first mask layout based on the defect area; The generation module is used to generate a second training sample for training the photoresist model based on the target layout area and the first photoresist label pattern.

10. A training sample generation device for training photoresist models, characterized in that, The device includes: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, it implements the training sample generation method for training a photoresist model as described in any one of claims 1-8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions, which, when executed by a processor, implement the training sample generation method for training a photoresist model as described in any one of claims 1-8.

12. A computer program product, characterized in that, When the instructions in the computer program product are executed by the processor of the electronic device, the electronic device performs the training sample generation method for training a photoresist model as described in any one of claims 1-8.

Images