non-transitory computer readable medium having instructions recorded thereon

Abstract

Described herein are non-transitory computer-readable media having instructions recorded thereon for a method of evaluating a selected set of patterns of a design layout. The method includes obtaining (i) a first set of patterns generated by a pattern selection process, (ii) first pattern data associated with the first set of patterns, (iii) property data associated with the first pattern data, and (iv) second pattern data associated with a second set of patterns. A machine learning model is trained based on the property data, where the machine learning model is configured to predict pattern data for an input pattern. The second set of patterns is input to the trained machine learning model to predict second pattern data for the second set of patterns. The first set of patterns is evaluated by comparing the second pattern data and the predicted second pattern data.

Description

Technical Field

[0001] This article generally describes improvements to metrology and lithography-related processes. More specifically, it describes devices, methods, and computer programs for evaluating pattern sets for metrological measurements or training models used in the patterning process. Background Technology

[0002] Photolithography projection equipment can be used, for example, in the manufacture of integrated circuits (ICs). In this case, a patterning apparatus (e.g., a mask) can contain or provide a pattern (“design layout”) corresponding to a single layer of the IC, and this pattern can be transferred onto a target portion (e.g., comprising one or more dies) on a substrate (e.g., a silicon wafer) by methods such as irradiating a target portion through the pattern on the patterning apparatus, which has been coated with a layer of radiation-sensitive material (“resist”). Typically, a single substrate contains multiple adjacent target portions, and the pattern is sequentially transferred to these target portions by the photolithography projection equipment, one target portion at a time. In one type of photolithography projection equipment, the entire pattern on the patterning apparatus is transferred onto a single target portion at a time; this type of equipment is generally referred to as a stepper. In an alternative equipment, generally referred to as a step-scanning apparatus, a projection beam scans across the patterning apparatus along a given reference direction (“scanning” direction) while the substrate moves parallel to or antiparallel to that reference direction. Different portions of the pattern on the patterning apparatus are progressively transferred onto a single target portion. Typically, because photolithography projection apparatuses will have a reduction ratio M (e.g., 4), the speed at which the substrate is moved, F, will be 1 / M times that of the projection beam scanning pattern forming apparatus. More information about photolithography apparatuses can be found, for example, in US 6,046,792, which is incorporated herein by reference.

[0003] Before a pattern is transferred from a patterning apparatus to a substrate, the substrate may undergo various processes such as coating, resist coating, and soft baking. After exposure, the substrate may undergo other processes (“post-exposure processes”) such as post-exposure baking (PEB), development, hard baking, and measurement / inspection of the transferred pattern. This array of processes is used as the basis for a single layer in the fabrication of a device (e.g., an IC). The substrate can then undergo various processes such as etching, ion implantation (doping), metallization, oxidation, chemical mechanical polishing, etc., all of which are designed to complete a single layer of the device. If multiple layers are required in the device, the entire process or its variations are repeated for each layer. Ultimately, the device will exist in each target portion on the substrate. These devices are then separated from each other using techniques such as dicing or sawing, so that individual devices can be mounted on a carrier, connected to pins, etc.

[0004] Therefore, manufacturing apparatuses such as semiconductor devices typically involve processing a substrate (e.g., a semiconductor wafer) using numerous manufacturing processes to form various features and multiple layers of the device. These layers and features are typically fabricated and processed using techniques such as deposition, photolithography, etching, chemical mechanical polishing, and ion implantation. Multiple devices can be fabricated on multiple dies on a substrate and then separated into individual devices. The device fabrication process can be considered a patterning process. A patterning process involves patterning steps, such as optical and / or nanoprinting lithography using patterning apparatus in a photolithography device to transfer a pattern from the patterning apparatus to the substrate, and typically, but optionally, involves one or more associated patterning processing steps, such as resist development by a developing apparatus, baking of the substrate using a baking tool, etching of the pattern using an etching apparatus, etc. Summary of the Invention

[0005] In an embodiment, a method is provided for evaluating a selected set of patterns from a design layout, such as for performing metrological measurements and / or generating a training dataset for a computational lithography machine learning model. Measuring a patterned substrate using metrological tools is a time-consuming process and impacts the throughput of semiconductor manufacturing processes (e.g., the number of chips manufactured per hour). Typically, there may be millions of patterns on a design layout that are expected to be printed on a chip. Measuring all of these patterns within the expected timeframe to meet the throughput specifications of the semiconductor manufacturing process is impractical. Therefore, a typically reduced set of patterns is selected from the design layout for measurement.

[0006] These measurements can be used for various purposes related to the patterning process. In one embodiment, the measurements can be used to monitor or adjust the patterning process in semiconductor manufacturing. In another embodiment, the measurements can be used for model calibration or training models related to the patterning process. Therefore, a reduced set of patterns selected for measurement can be evaluated, thus enabling accurate control and adjustment of the patterning process even with reduced measurements.

[0007] Additionally, accurate model calibration or training can be achieved even with reduced measurements. This also helps reduce the computation time and resources used during model calibration or training.

[0008] For example, if the selected pattern set represents millions of patterns in a design layout, or if the pattern set has sufficient pattern coverage, then when measurements of such a pattern set are used to train a model related to the patterning process, the trained model can accurately predict the characteristics of the patterned substrate. This disclosure provides a mechanism for evaluating one or more selected pattern sets without relying on their measurement data. The pattern sets can be generated separately from the pattern selection process. In this way, the evaluation can be advantageously performed before time-consuming measurement.

[0009] According to an embodiment, a method for evaluating a selected set of patterns is provided. The method includes: obtaining (i) a first set of patterns generated by a pattern selection process, (ii) first pattern data associated with the first set of patterns, (iii) feature data associated with the first pattern data, and (iv) second pattern data associated with a second set of patterns. A machine learning model is trained based on the feature data associated with the first patterns, wherein the machine learning model is configured to predict pattern data of patterns input into the machine learning model. A second set of patterns is input into the trained machine learning model to predict second pattern data for the second set of patterns. The first set of patterns is evaluated by comparing the second pattern data with the predicted second pattern data. The first set of patterns includes a first plurality of patterns of a design layout, and the second set of patterns includes a second plurality of patterns of a design layout. In an embodiment, the second set of patterns includes a full-chip layout.

[0010] In an embodiment, obtaining first pattern data involves generating a first contour or first image by executing a reference model configured to simulate a patterning process using a first set of patterns as input. Similarly, obtaining second pattern data involves generating a second contour or second image by executing a reference model configured to simulate a patterning process using a second set of patterns as input.

[0011] In an embodiment, the first pattern set may be a subset of the second pattern set.

[0012] In an embodiment, evaluating the first pattern set involves calculating the difference between the second pattern data and the predicted second pattern data. In an embodiment, absolute pattern coverage is determined as a function of absolute error associated with a trained machine learning model trained using the first pattern set. In an embodiment, relative pattern coverage can be determined as a function of relative error, which is a comparison between a first error range associated with the trained machine learning model trained using the first pattern set and a second error range associated with another pattern set.

[0013] In this embodiment, risk patterns within the design layout can be determined. These risk patterns are associated with model prediction errors that violate the expected error threshold. These risk patterns can be supplemented to a first pattern set to improve pattern coverage.

[0014] Once the selected pattern set is deemed to have sufficient coverage, it can be used for various purposes related to the patterning or measurement process. For example, the selected pattern set can be used to capture measurements of a patterned substrate via a measurement tool. Advantageously, the selected patterns improve measurement throughput. As another example, the selected pattern set can be used to train ML models related to the patterning process.

[0015] According to an embodiment, a computer system is provided, comprising a non-transitory computer-readable medium on which instructions are recorded. When executed by a computer, these instructions perform the method steps described above. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate certain aspects of the subject matter disclosed herein and, together with the description, help to explain some principles associated with the disclosed embodiments. In the drawings,

[0017] Figure 1 This is a block diagram of various subsystems of the photolithography projection apparatus according to an embodiment.

[0018] Figure 2 A flowchart illustrating an exemplary method for simulating lithography in a lithography projection apparatus according to an embodiment is shown.

[0019] Figure 3 It is a block diagram of the evaluation process for a given set of patterns in a design layout according to an embodiment.

[0020] Figure 4 This is an exemplary flowchart of a method for evaluating a given set of patterns (e.g., a subset of patterns) for design layout, according to an embodiment.

[0021] Figure 5 It is a box plot of contour-to-contour defect dimensions associated with different selected pattern sets according to an embodiment.

[0022] Figure 6 The illustration shows absolute and relative pattern coverage checks based on defect size according to an embodiment, where the defect size is associated with different sets of patterns obtained from different pattern selection processes.

[0023] Figure 7 The illustration shows a relative comparison of different pattern sets based on contour-to-contour error range according to an embodiment, wherein the different pattern sets are obtained from different pattern selection processes.

[0024] Figure 8 The illustration shows a risk pattern for error identification based on a trained machine learning model according to an embodiment.

[0025] Figure 9This is a block diagram of an example computer system according to an embodiment.

[0026] Figure 10 This is a schematic diagram of an exemplary extreme ultraviolet (EUV) lithography projection device according to an embodiment.

[0027] Figure 11 This is a schematic diagram of another photolithography projection device according to an embodiment.

[0028] Figure 12 According to the embodiments Figure 11 A more detailed view of the exemplary device in the image.

[0029] Figure 13 According to the embodiments Figure 11 and Figure 12 A more detailed view of the device's source collector module. Detailed Implementation

[0030] While specific references may be made to the manufacture of ICs herein, it should be clearly understood that the descriptions herein have many other possible applications. For example, it can be used to manufacture integrated optical systems, guide and detection patterns for magnetic domain memories, liquid crystal display panels, thin-film magnetic heads, and the like. Those skilled in the art will understand that, in the context of such alternative applications, any use of the terms “mask,” “wafer,” or “die” herein should be considered interchangeable with the more general terms “mask,” “substrate,” and “target portion,” respectively.

[0031] In this document, the terms “radiation” and “beam” can be used to cover all types of electromagnetic radiation, including ultraviolet radiation (e.g., wavelengths of 365, 248, 193, 157, or 126 nm) and EUV (extreme ultraviolet radiation, e.g., wavelengths in the range of about 5 to 100 nm).

[0032] Patterning apparatuses may include or can form one or more design layouts. Design layouts can be generated using CAD (Computer-Aided Design) programs, a process often referred to as EDA (Electronic Design Automation). Most CAD programs follow a predetermined set of design rules to create functional design layouts / patterning apparatuses. These rules are set by processing and design constraints. For example, design rules define spatial tolerances between devices (such as gates, capacitors, etc.) or interconnects to ensure that devices or lines do not interact with each other in undesirable ways. One or more design rule constraints may be referred to as “critical dimensions” (CDs). A critical dimension of an apparatus can be defined as the minimum width of a line or hole, or the minimum space between two lines or two holes. Therefore, CDs determine the overall size and density of the designed apparatus. Of course, one of the goals in apparatus fabrication is to faithfully reproduce the original design intent on the substrate (via the patterning apparatus).

[0033] The terms "mask" or "patterning apparatus" used herein can be broadly interpreted as general-purpose patterning apparatuses that can be used to impart a patterned cross-section to an incoming radiation beam, corresponding to a pattern to be created in a target portion of a substrate; in this context, the term "optical valve" may also be used. Besides classic masks (transmissive or reflective; binary, phase-shifting, hybrid, etc.), other examples of such patterning apparatuses include programmable mirror arrays and programmable LCD arrays.

[0034] An example of a programmable mirror array could be a matrix-addressable surface with a viscoelastic control layer and a reflective surface. The underlying principle behind such a device is that, for example, addressable regions of the reflective surface reflect incident radiation as diffracted radiation, while unaddressed regions reflect incident radiation as non-diffracted radiation. Using appropriate filters, the non-diffracted radiation can be filtered out from the reflected beam, leaving only the diffracted radiation; in this way, the beam is patterned according to the addressing pattern of the matrix-addressable surface. The desired matrix addressing can be performed using suitable electronic components.

[0035] An example of a programmable LCD array is given in U.S. Patent No. 5,229,872, which is incorporated herein by reference.

[0036] Figure 1 This is a block diagram of various subsystems of the photolithography projection apparatus 10A according to an embodiment. The main component is a radiation source 12A, which can be a deep ultraviolet excimer laser source or other types of sources, including: extreme ultraviolet (EUV) sources (as discussed above, the photolithography projection apparatus itself does not need to have a radiation source); irradiation optics, which, for example, define partial coherence (denoted as σ), and may include optics 14A, 16Aa, and 16Ab for shaping the radiation from source 12A; a pattern forming apparatus 18A; and a transmission optics 16Ac that projects an image of the pattern from the pattern forming apparatus onto a substrate plane 22A. An adjustable filter or aperture 20A at the pupil plane of the projection optics may limit the range of beam angles striking the substrate plane 22A, where the maximum possible angle defines the numerical aperture NA of the projection optics as NA = n sin(Θ). max ), where n is the refractive index of the medium between the substrate and the final element of the projection optics, and Θ max It is the maximum angle of the beam leaving the projection optics, which can still strike the substrate plane 22A.

[0037] In a photolithography projection apparatus, a source provides illumination (i.e., radiation) to a patterning apparatus, and projection optics guide and shape the illumination onto a substrate via the patterning apparatus. The projection optics may include at least some of components 14A, 16Aa, 16Ab, and 16Ac. The spatial image (AI) is the distribution of radiation intensity at the substrate level. A resist model can be used to calculate the resist image from the spatial image, examples of which can be found in U.S. Patent Application Publication No. US2009-0157360, the disclosure of which is incorporated herein by reference in its entirety. The resist model relates only to the properties of the resist layer (e.g., the effects of chemical processes occurring during exposure, post-exposure baking (PEB), and development). The optical properties of the photolithography projection apparatus (e.g., the properties of the illumination, the patterning apparatus, and the projection optics) specify the spatial image and can be defined in the optical model. Because the patterning apparatus used in a photolithography projection apparatus can be modified, it is desirable to separate the optical properties of the patterning apparatus from the optical properties of the rest of the photolithography projection apparatus, which includes at least the source and the projection optics. Details of the techniques and models used to transform design layouts into various lithographic images (e.g., spatial images, resist images, etc.), apply OPC using these techniques and models, and evaluate performance (e.g., according to process windows) are described in U.S. Patent Application Publications Nos. 2008-0301620, 2007-0050749, 2007-0031745, 2008-0309897, 2010-0162197, and 2010-0180251, the disclosures of each of which are incorporated herein by reference in their entirety.

[0038] According to embodiments of this disclosure, one or more images can be generated using various types of signals corresponding to pixel values ​​(e.g., intensity values) for each pixel. Depending on the relative values ​​of the pixels within the image, the signals can be referred to as, for example, weak signals or strong signals, as will be understood by those skilled in the art. The terms "strong" and "weak" are relative terms based on the intensity values ​​of pixels within the image, and the specific values ​​of intensity may not limit the scope of this disclosure. In embodiments, strong and weak signals can be identified based on selected thresholds. In embodiments, the threshold can be fixed (e.g., the midpoint between the highest and lowest intensities of pixels within the image). In embodiments, a strong signal can refer to a signal with a value greater than or equal to the average signal value of the entire image, while a weak signal can refer to a signal with a value less than the average signal value. In embodiments, relative intensity values ​​can be based on percentages. For example, a weak signal can be a signal with an intensity less than 50% of the highest intensity of pixels within the image (e.g., a pixel corresponding to a target pattern can be considered the pixel with the highest intensity). Furthermore, each pixel within the image can be considered a variable. According to this embodiment, the derivative or partial derivative can be determined relative to each pixel within the image, and the value of each pixel can be determined or modified based on a cost function-based evaluation and / or gradient-based calculation. For example, a CTM image may include pixels, where each pixel is a variable that can take any real value.

[0039] Figure 2 An exemplary flowchart for simulating lithography in a lithography projection apparatus according to an embodiment is illustrated. Source model 31 represents the optical characteristics of a source (including radiation intensity distribution and / or phase distribution). Projection optics model 32 represents the optical characteristics of a projection optics (including variations in radiation intensity distribution and / or phase distribution caused by the projection optics). Design layout model 35 represents the optical characteristics of a design layout (including variations in radiation intensity distribution and / or phase distribution caused by the design layout 33), which is a representation of a feature arrangement formed on or by a patterning apparatus. Spatial image 36 can be simulated using design layout model 35, projection optics model 32, and design layout model 35. Resist image 38 can be simulated from spatial image 36 using resist model 37. The lithography simulation can, for example, predict the contours and CDs in the resist image.

[0040] More specifically, note that source model 31 can represent the optical characteristics of a source, including but not limited to numerical aperture settings, illumination sigma (σ) settings, and any particular illumination shape (e.g., off-axis radiation sources such as ring, quadrupole, and dipole sources). Projection optics model 32 can represent the optical characteristics of a projection optics, including aberrations, distortion, one or more refractive indices, one or more physical dimensions, etc. Design layout model 35 can represent one or more physical properties of a physical pattern forming apparatus, such as those described in U.S. Patent No. 7,587,704, which is incorporated herein by reference in its entirety. The purpose of the simulation is to accurately predict, for example, edge placement, spatial image intensity slope, and / or CD, which can then be compared with the intended design. The intended design is typically defined as a pre-OPC design layout, which can be provided in a standardized digital file format such as GDSII or OASIS or other file formats.

[0041] Through this design layout, one or more portions can be identified, referred to as "clips". In embodiments, a set of clips is extracted, representing complex patterns in the design layout (typically approximately 50 to 1000 clips, although any number of clips can be used). These patterns or clips represent small parts of the design (e.g., circuits, cells, or patterns), and more specifically, clips typically represent small parts that require special attention and / or verification. In other words, a clip may be part of the design layout, or may be similar to or have similar behavior to a portion of the design layout, where one or more key features are identified through experience (including clips provided by the customer), through trial and error, or by running full-chip simulations. Clips may contain one or more test patterns or standard metric patterns.

[0042] The initial set of larger clips can be provided a priori by the client based on known key feature regions in the design layout requiring specific image optimization. Alternatively, in another embodiment, the initial set of larger clips can be extracted from the entire design layout using some automated (such as machine vision) or manual algorithm that identifies one or more key feature regions. These clips, or a reduced set of patterns therefrom, can also be used for various purposes related to the patterning process or measurement.

[0043] As used herein, the term "patterning process" generally refers to the process of creating an etched substrate by applying a specified light pattern as part of a photolithography process. However, "patterning process" can also include plasma etching, as many of the features described herein can provide benefits for forming printed patterns using plasma processing.

[0044] As used herein, the term "design pattern" or "target pattern" refers to an idealized pattern to be etched onto a substrate. The term "target layout" refers to a design layout that includes one or more target patterns.

[0045] As used herein, the terms “printed pattern” or “patterned substrate” refer to a physical pattern on a substrate based on target pattern imaging and / or etching. Printed patterns may include, for example, trenches, channels, recesses, edges, or other two-dimensional and three-dimensional features generated by photolithography processes.

[0046] As used herein, the term "process model" refers to a model that includes one or more models simulating the patterning process. For example, a process model may include an optical model (e.g., modeling a lens system / projection system used to transmit light during the photolithography process, and may include modeling the final optical image of the light entering the photoresist), a resist model (e.g., simulating the physical effects of the resist, such as chemical effects due to light), and an OPC model (e.g., which can be used to modify the target pattern to include sub-resolution resist features (SRAF), etc.).

[0047] Typically, in training patterning process models, controlling patterning processes, or other applications or semiconductor manufacturing-related processes, a subset of patterns for the design layout is selected and used to improve measurement throughput. As an example, a reduced set of patterns decreases measurement or modeling time and increases throughput. For instance, to improve patterning process and patterning accuracy, process models are trained using target patterns, mask patterns, substrate images, etc. OPC processes can include one or more trained models to generate better mask patterns, such as optical models, mask models, resist models, etching models, etc. For example, machine learning-assisted OPC significantly improves the accuracy of full-chip auxiliary feature (e.g., SRAF) placement while maintaining mask design consistency and runtime under control. Deep convolutional neural networks (DCNNs) can be trained using target layouts or target patterns and corresponding continuous transmission mask (CTM) images. These CTM images are optimized using an inverse mask optimization simulation process. The SRAF guide map generated by the CNN is then used to place the SRAF on the full-chip design layout.

[0048] When selecting a set of patterns for training, it is desirable to choose the set that provides the most information to the model. Currently, various methods are available for pattern selection. For example, pattern hashing techniques may be fast, but they perform best in exact matching rather than capturing pattern similarity. In another example, unsupervised image-based pattern imaging techniques (e.g., based on autoencoders) can capture pattern similarity in a higher-dimensional latent space. In model simulation-based pattern classification and selection techniques, spatial images or resist image parameter spaces that combine similarity from a model simulation perspective can be used.

[0049] This disclosure provides a mechanism for evaluating the performance achieved by a selected or reduced set of patterns for a design layout before collecting wafer metrology data. This evaluation can be used to verify pattern coverage, assess the pattern selection process, compare different selected pattern sets, or improve pattern coverage. Advantageously, evaluating the selected pattern set can provide insights into whether the selected pattern set provides sufficient pattern coverage for the design layout, whether more patterns should be added to improve pattern coverage, or other aspects of the patterning or metrology process. As an example, improving pattern coverage, in turn, makes metrology measurements more efficient. As another example, improving pattern coverage will improve training time and the quality of predictions made by the trained process model. For example, a trained process model can accurately predict the profile of the patterned substrate throughout the design layout.

[0050] In this embodiment, a mechanism is provided for evaluating a selected set of patterns for a design layout. For example, Figure 3 This is a block diagram illustrating the process of evaluating a given pattern of design layout by training and executing a machine learning model, according to an embodiment.

[0051] In embodiments, the pattern subset of the design layout can be obtained in any suitable manner. For example, the pattern subset of the design layout can be provided by a user or selected using a pattern selection process. The pattern subset of the design layout can be referred to as a given pattern set GP1 of the design layout, a first pattern set GP1, or a reduced pattern set GP1. A reference model can be applied to the given pattern set GP1 to generate or extract pattern data for training a machine learning (ML) model. For example, the pattern data can include simulated contours, standard metrics for measuring pattern characteristics, statistics associated with the physical properties of the pattern, individual positions of the pattern, the relative positions of the pattern to other patterns, or other pattern-related data that can be used to train the ML model. The trained ML model can be used for different applications related to the patterning process.

[0052] In one embodiment, a reference model can generate simulated contours of a given pattern set GP1 from which standard metric data is derived. The standard metric data is then used as training data for training an ML model. The trained ML model is then applied to a second pattern set to generate or extract second pattern data (e.g., contours or other predictions). The second pattern set differs from the given pattern set GP1 and can be a much larger pattern set than the given pattern set GP1, such as a full-chip dataset. The reference model is also applied to the second pattern set to generate corresponding first pattern data. The first pattern data generated by the two models are compared, and the differences are used to evaluate the given pattern set GP1 relative to the ML model training. In some embodiments, both models are configured to predict contours of the second pattern set. Contour-to-contour comparisons of the second dataset between the two models are used for pattern coverage evaluation.

[0053] More specifically, such as Figure 3 As shown, in operation 301, a reference model REFM1 associated with the patterning process can be applied to a given pattern GP1 to generate a simulated profile of the given pattern GP1. This disclosure is not limited to any specific form of input data for the model. Depending on the implementation and configuration of REFM1, the input data can be a polygon of the target pattern or a rendered image of the pattern set GP1. The reference model REFM1 can be one or more of, including but not limited to, optical models, mask models, resist models, etching models, etc. The reference model can be a trained machine learning model or a calibrated non-machine learning model (e.g., a physics-based model or an empirical model). In some other embodiments, in 301, the profile of the given pattern GP1 can be obtained from measurement data (e.g., images) captured during measurement of a previously patterned substrate, and the measurement data is stored in a database.

[0054] In operation 303, a standard metric is configured to extract characteristic data associated with the pattern. For example, a standard metric may be an edge placement (EP) standard metric or a point placed on the contour, such as characterizing the shape and / or size of the contour. As another example, a standard metric may be a critical dimension standard metric (CD standard metric), or a standard metric configured to measure spacing, space, curvature, or other physical properties of the pattern.

[0055] In operation 305, a standard metric can be used to train an ML model such that the trained ML model closely matches reference data, such as the output of a reference model REFM1. The machine learning model can be a convolutional neural network or any other suitable model configuration. This disclosure is not limited to a particular machine learning model. The machine learning model can be trained using ML algorithms, such as supervised or unsupervised learning. As an example, an ML model (e.g., a CNN or DCNN) is trained using simulated EP standard metric data. In other words, the trained ML is configured to mimic the output of the reference model REFM1.

[0056] In operation 308, a trained ML model can be applied to predict contours of a second set of patterns different from GP1. Without departing from the scope of this disclosure, the second dataset includes any number of patterns. In some embodiments, it can be significantly larger than GP1. In some embodiments, the second set of patterns can be a full-chip layout. Furthermore, a reference model REFM1 can be applied to generate a second set of contours for the second set of patterns. The ML-predicted contours can be compared with the contours generated by the reference model to evaluate the performance of the machine learning model, which is used as a performance evaluation for a given set of patterns GP1. For example, a full-chip contour-to-contour (C2C) comparison between the contours generated by the reference model and the contours generated by the trained ML model is used to characterize the pattern coverage of GP1.

[0057] As an example, if a given pattern set GP1 has good pattern coverage properties, then a trained ML model can provide good predictions for new patterns of a design layout not included in the given pattern set. For instance, if a design layout includes more than 1 million patterns, and the given pattern set GP1 includes 10,000 patterns representing those 1 million patterns, then a trained ML model can make good predictions for those 1 million patterns (e.g., contours that can be patterned on a substrate). In other words, the predictions of a trained ML model will be associated with low error values. For example, error values ​​can be characterized by the C2C difference between the ML-predicted contour and a reference contour (e.g., generated by a reference model REFM1). Error values ​​can be characterized by root mean square (RMS), standard deviation, range (e.g., between a minimum and a maximum value), distribution of error values, or other ways of characterizing the error associated with the model predictions.

[0058] On the other hand, if a given pattern set GP1 has poor pattern coverage, the trained ML model may provide poor predictions for new patterns not included in the first pattern set. In other words, the predictions of the trained ML model will be associated with high error values ​​when compared to the predictions of the reference model. For example, the C2C difference between the ML-predicted contour and the reference contour may be large.

[0059] Figure 5 The illustration shows exemplary metrics characterizing pattern coverage when different ML models are trained using different given patterns. These metrics could be the C2C difference across the entire chipset, the RMS of error values, standard deviation, etc. In this discussion, the C2C difference is used as an example and the scope of this disclosure is not limited to a specific metric. A first given pattern set 501 can be used to train the ML model. When the trained ML model is applied to a full-chip layout, the C2C difference (e.g., the difference between the ML predicted contour and the reference contour) can be large. The graph provides a visualization of outliers. In an embodiment, values ​​that deviate significantly from the mean may be outliers. For example, values ​​greater than 1 or less than -2.5. These outliers may correspond to patterns not included in the first given pattern set 501. Therefore, this indicates that the given pattern set 501 does not provide sufficient pattern coverage, and additional patterns should be included to improve pattern coverage. Similarly, different ML models can be trained using pattern sets 503, 505, and 507, respectively, and the corresponding C2C differences can be evaluated to determine the pattern coverage performance of the respective pattern sets. In some embodiments, the evaluation results of the ML models are compared to select the corresponding optimal pattern set.

[0060] In an embodiment, the C2C results can also identify outlier patterns associated with high model error. These identified C2C outlier patterns can be added to a given pattern GP1 as training data for retraining the ML model or training a new ML model. The evaluation of a given pattern set GP1 or a first pattern set relative to... Figures 4 to 8 Further detailed discussion.

[0061] Figure 4 This is an exemplary flowchart of a method 400 for evaluating the performance of a selected set of patterns for design layout relative to training a machine learning model, according to an embodiment. In this embodiment, training based on an ML model is evaluated to predict characteristics (e.g., contours) of any input pattern. Processes P401, P403, P405, and P407 of example implementations of method 400 are discussed.

[0062] Process P401 involves obtaining data associated with a first pattern set PS1 that will be used to train the ML model. The first pattern set PS1 is the pattern to be evaluated. In an embodiment, process P401 involves obtaining the first pattern set PS1 generated by a pattern selection process and first pattern data (e.g., a first contour CS1) associated with the first pattern set PS1. This discussion uses contours as exemplary pattern data to explain the concepts of this disclosure. However, the scope of this disclosure is not limited to contours, and other forms of pattern data (e.g., images) may be used. The pattern data generated or extracted for the set of patterns / pattern set may depend on the configuration of the input and output of the ML model. Pattern data (contours, standard metric data, or images) can be derived from mask images, spatial images, etched images, or developed images generated by corresponding simulations or modeling.

[0063] In an embodiment, process P401 may further involve obtaining feature data CHD1 associated with the first contour CS1. In an embodiment, feature data CHD1 may be used to train an ML model. In an embodiment, process P401 may further involve obtaining second pattern data (e.g., the second contour CS2) associated with a second pattern set PS2 (e.g., a full chip layout). In an embodiment, the first pattern set PS1 and the second pattern set PS2 are selected from the same design layout. However, in some other embodiments, pattern sets PS1 and PS2 may be included in different design layouts.

[0064] In one embodiment, the first pattern set PS1 may be a subset of a second pattern set PS2 selected from the design layout. In another embodiment, the first pattern set PS1 and the second pattern set PS2 may include several common patterns selected from the design layout. In yet another embodiment, the first pattern set PS1 may be different from the second pattern set PS2, but both sets are selected from the design layout. In one embodiment, the first pattern set PS1 includes a first plurality of patterns from the design layout, and the second pattern set PS2 includes a second plurality of patterns from the design layout. In another embodiment, the second pattern set PS2 includes more patterns than the first pattern set PS1. In yet another embodiment, the second pattern set PS2 includes a full-chip layout.

[0065] In one embodiment, obtaining a first contour CS1 involves generating the first contour CS1 by executing a reference model REFM1, which is configured to simulate a patterning process using a first pattern set PS1 as input. In another embodiment, obtaining a second contour CS2 involves generating the second contour CS2 by executing a reference model REFM1, which is configured to simulate a patterning process using a second pattern set PS2 as input.

[0066] In an embodiment, the reference model REFM1 includes one or more component models characterizing the patterning process. For example, the reference model REFM1 may include, but is not limited to, optical device models, mask models, resist models, etching models, ML resist models, ML etching models, etc. Accordingly, the first contour CS1 and the second contour CS2 may be, for example, a substrate-level resist contour or an etching contour. In an embodiment, the first contour CS1 and the second contour CS2 may be spatial image contours or mask image contours. As mentioned herein, the scope of this disclosure is not limited to contours, and other types of pattern data may be generated or extracted from pattern sets. In an embodiment, the reference model REFM1 may be a non-machine learning model, such as a physically and / or empirically based model.

[0067] In another embodiment, obtaining the first profile CS1 and the second profile CS2 involves obtaining the profiles from an image of a previously patterned substrate that has been captured by a measurement tool and stored in a database. In this embodiment, such a previously patterned substrate may include a first pattern set PS1 and a second pattern set PS2 with a design layout.

[0068] In an embodiment, the characteristic data CHD1 may be a standard metric generated from the first profile CS1. In an embodiment, the standard metric may be configured to quantify one or more physical properties of the pattern. For example, the standard metric may be a standard metric placed at multiple locations along the contour of the first profile CS1; a critical dimension (CD) standard metric configured to measure the CD value of the first profile CS1; a standard metric configured to measure a line; a standard metric configured to measure space; a standard metric configured to measure tip-to-tip structure; and / or a standard metric configured to measure contour differences.

[0069] Process P403 involves training an ML model based on feature data CHD1 associated with a first contour CS1. The ML model is configured to predict contours in response to pattern input. In embodiments, existing machine learning algorithms (e.g., supervised or unsupervised learning techniques) can be used to train the ML model. The ML model can be, but is not limited to, a convolutional neural network (CNN or DCNN). For example, training a CNN can use a cost function to determine the weights of different layers of the CNN. As an example, the cost function can be a function of the EP standard metric. During the training process, the EP standard metric or a given set of patterns can be input into the CNN, and the CNN generates predicted contours. Based on the predicted contours, the cost function, such as the edge placement error between the predicted contours and reference contours, can be evaluated, and the weights can be adjusted to minimize (or maximize) the cost function. After training the ML model, the ML model can be referred to as the trained ML model TML1.

[0070] Process P405 involves generating a predicted second profile PCS2 of the second pattern set PS2 by inputting the second pattern set PS2 into a trained ML model TML1.

[0071] Process P407 involves evaluating a first pattern set PS1 by comparing a second contour CS2 with a predicted second contour PCS2. In an embodiment, evaluating the first pattern set PS1 involves calculating the difference between the second contour CS2 and the predicted second contour PCS2. In an embodiment, the second contour CS2 may be associated with a full chip layout, and the predicted second contour PCS2 can be predicted by inputting the full chip layout into a trained ML model TML1. Based on these contours, the first pattern set PS1 can be evaluated, for example, based on statistics (e.g., outliers, variations, mean, median, distribution, etc.) associated with the C2C difference between contours CS2 and PCS2.

[0072] In an embodiment, the evaluation of the first pattern set PS1 involves: determining whether the difference violates a difference threshold (e.g., user-defined or statistically based); and classifying the first pattern set PS1 as having good pattern coverage in response to the difference not violating the difference threshold. Therefore, the first pattern PS1 can be considered acceptable for performing measurements to save measurement time while obtaining good measurement results to improve the patterning process.

[0073] In one embodiment, the evaluation of the first pattern set PS1 involves determining absolute pattern coverage as a function of the absolute error associated with a trained machine learning model directly trained using a second pattern set PS2. In another embodiment, the evaluation of the first pattern set PS1 involves determining relative pattern coverage as a function of relative error. For example, the relative error could be a comparison between a first error range associated with a trained machine learning model trained using the first pattern set PS1 and a second error range associated with another trained machine learning model using another pattern set.

[0074] Figure 6 The illustration depicts absolute and relative pattern coverage checks based on C2C error according to an embodiment, where the C2C error is associated with different sets of patterns obtained from different pattern selection processes. Figure 6In this example, bar TM1 corresponds to the baseline model error associated with a trained machine learning model (e.g., TML1). In this embodiment, the baseline model error refers to the error in the ML model predictions when the full-chip pattern set is used to train the ML model. For example, the full-chip pattern can be used as input to a reference model to generate simulated contours and corresponding standard metrics. These standard metrics can be used to train the ML model. The trained ML model can also be applied to the full-chip pattern to predict contours that can be printed onto a substrate. When the C2C difference between the predicted contour and the simulated contour (from the reference model) is calculated, ideally, the difference should be zero. However, small non-zero C2C differences may exist, which are referred to as the baseline model error. In this example, bar TM1 corresponds to this baseline model error.

[0075] exist Figure 6 In the illustrated example, when the first pattern set is used to train the ML model, the first bar SP1 corresponds to the C2C difference between the predicted contour and the reference contour of the trained ML model. For example, the first pattern set may include approximately 200 patterns selected from 1500 patterns in a DRAM design layout (using the first pattern selection method). Similarly, when the second pattern set is used to train the ML model, the second bar SP2 corresponds to the C2C difference between the predicted contour and the reference contour of the trained ML model. For example, the second pattern set may include approximately 200 patterns selected from 1500 patterns in a DRAM design layout (using the second pattern selection method). Similarly, each of the third bar SP3, the fourth bar SP4, the fifth bar SP5, and the sixth bar SP6 corresponds to the C2C difference between the predicted contour and the reference contour of the trained ML model when different pattern sets are used to train the ML model. Each pattern set in the pattern sets selected using different methods may include the same or different numbers of patterns.

[0076] In an embodiment, the absolute values ​​of the C2C differences V1, V2, V3, V4, V5, and V6 corresponding to pattern sets SP1, SP2, SP3, SP4, SP5, and SP6, respectively, can indicate the absolute error associated with the corresponding pattern set. A low absolute error (e.g., close to the baseline model error) is desirable because it indicates good pattern coverage provided by that particular pattern set. In an embodiment, a threshold error value can be set for comparison with the individual model errors to assess whether a particular pattern set provides good pattern coverage.

[0077] In an embodiment, each of the values ​​V1 through V6 can be compared to the baseline model error V7 to evaluate the relative error associated with a particular pattern relative to the baseline model error. In an embodiment, a set of patterns associated with error values ​​close to the baseline model error indicates that the set of patterns provides good pattern coverage. For example, a fourth set of patterns (corresponding to bar SP4) is associated with error value V4, which is substantially close to value V7. Therefore, the fourth set of patterns can be considered to provide good pattern coverage. In an embodiment, the relative error between values ​​V1 through V6 can be analyzed. Among different set of patterns, a set with a relatively lower error compared to others can be considered to have better pattern coverage.

[0078] Figure 7 This illustration demonstrates another way to evaluate the absolute and relative pattern coverage characteristics of a selected pattern set based on error range. Figure 6 Similar to the discussion in [the previous section], the reference error range (e.g., C2C difference range) associated with a trained ML model trained using full-chip patterning can be determined. For example, the reference error range can be represented as [example missing]. Figure 7 In this embodiment, the error range is the range between the maximum positive error value and the maximum negative error value. The error range associated with each different selected pattern set can be determined. For example, the error range can be calculated based on the C2C difference between the predicted contour and the reference contour by the trained model. For example, pattern set SP11 can be associated with error range ER11, another pattern set SP12 can be associated with error range ER12, and yet another pattern set SP13 can be associated with error range ER13. Figure 7 The illustration shows that the error range ER13 is smaller compared to ranges ER11 and ER12, indicating that pattern set SP13 provides better pattern coverage compared to other pattern sets. Error range ER13 is also substantially close to error range ER1, indicating that pattern set SP13 may provide sufficient pattern coverage for the entire chip.

[0079] Return to reference Figure 4 According to an embodiment, in process 407, the evaluation of the first pattern set PS1 involves identifying risk patterns within the design layout. In this embodiment, a risk pattern is a pattern whose prediction error is easily predicted by a reference model, a baseline ML model, or other models associated with the patterning process. In this embodiment, a risk pattern may be flagged as an outlier (e.g., associated with a difference value that violates a risk threshold). In this embodiment, such a risk pattern may be included in the first pattern set to improve pattern coverage.

[0080] Figure 8The illustration depicts a risk pattern for error identification based on an error generated by a trained machine learning model, according to an embodiment. In this embodiment, the risk pattern refers to a pattern of errors that are easily predicted by a reference model, a baseline ML model, or other models associated with the patterning process. In this embodiment, outlier error values ​​can be characterized by the C2C difference between the reference profile and the ML-predicted profile. In this embodiment, outlier error values ​​can be identified based on the standard deviation, comparing the value to the mean (example threshold), or other statistical methods to determine outliers in the dataset.

[0081] exist Figure 8 In this process, the C2C difference between the contour generated by the trained ML model and the reference contour is plotted for visual reference. As discussed in this paper, the C2C difference can be calculated using a trained ML model trained on selected pattern sets such as SP21, SP22, SP23, SP24, SP25, or SP26. Different ML models can be trained for each selected pattern set SP21 to SP26. The trained ML model is then used to predict the contour of the design layout and determine the C2C difference.

[0082] C2C difference curves provide a visual representation of identifiable outliers. However, those skilled in the art will understand that in some embodiments, outliers can be identified without a visual representation. The pattern corresponding to such outliers may be a risk pattern. Different risk patterns can be identified based on the selected set of patterns used to train the ML model. For example, when SP21 is used to train the ML model, risk pattern RP1 is identified from the design layout. Similarly, risk patterns RP2 and RP3 can be identified from the design layout.

[0083] In one embodiment, method 400 involves identifying a list of patterns (e.g., risk patterns, hotspots, etc.) to be examined by a measurement tool based on a first pattern set PS1. In another embodiment, measurement data associated with the pattern list can also be used to train or calibrate an ML model related to the patterning process.

[0084] In one embodiment, method 400 involves identifying locations of breaches corresponding to a difference threshold in a second pattern set PS2, supplementing a first pattern set PS1 with one or more patterns associated with the identified locations, and training another machine learning model using the supplemented first pattern set PS1. In one embodiment, the supplemented first pattern set PS1 corresponds to higher pattern coverage compared to the first pattern set PS1.

[0085] In embodiments, metrology tools determine metrological measurements, including but not limited to critical dimensions, overlap, and edge placement errors associated with a first set of patterns PS1 or a supplementary first set of patterns for a design layout patterned on a substrate. These measurements can also be used to control the lithography process, train machine learning models to generate mask patterns (e.g., defect detection models, etching models, resist models, OPC, SMO, etc.), train machine learning models to predict measurements of the design pattern, or for other lithography or metrology-related applications. Because the evaluation process described herein ensures sufficient design layout coverage for a given set of patterns, only a reduced number of measurements can be performed, resulting in significant time savings in metrology. Therefore, the throughput of the patterning process can be improved.

[0086] In embodiments, the methods discussed herein may be provided as one or more computer program products or as a non-transitory computer-readable medium having instructions recorded thereon, which, when executed by a computer, perform the operations of the method 400 discussed above. For example, Figure 9 The example computer system CS includes a non-transitory computer-readable medium (e.g., memory) containing instructions that, when executed by one or more processors (e.g., 104), implement operations for evaluating any given set of patterns for a design layout. For example, evaluating a user-provided set of patterns or a set of patterns selected using different pattern selection methods.

[0087] According to this disclosure, combinations and sub-combinations of the disclosed elements constitute separate embodiments. For example, a first combination includes evaluating a selected pattern set by training a machine learning model using the selected pattern set. A sub-combination may include evaluating selected patterns for pattern coverage. In another combination, risk patterns are determined based on C2C differences. A sub-combination may include employing a reference model to determine C2C differences.

[0088] Figure 9 This is a block diagram of an example computer system CS according to an embodiment. The computer system CS includes a bus BS or other communication mechanism for transmitting information and a processor PRO (or multiple processors) coupled to the bus BS for processing information. The computer system CS also includes main memory MM, such as random access memory (RAM) or other dynamic storage device, coupled to the bus BS for storing information and instructions to be executed by the processor PRO. The main memory MM can also be used to store temporary variables or other intermediate information during the execution of instructions to be executed by the processor PRO. The computer system CS also includes read-only memory (ROM) or other static storage device coupled to the bus BS for storing static information and instructions from the processor PRO. Storage devices SD, such as magnetic disks or optical disks, are provided and coupled to the bus BS for storing information and instructions.

[0089] A computer system (CS) can be coupled to a display (DS), such as a cathode ray tube (CRT) or flat panel or touchpad display, via a bus (BS) for displaying information to the computer user. Input devices (ID), including alphanumeric keys and other keys, are coupled to the bus (BS) to transmit information and command selections to the processor (PRO). Another type of user input device is a cursor control (CC), such as a mouse, trackball, or arrow keys, to transmit directional information and command selections to the processor (PRO) and control cursor movement on the display (DS). This input device typically has two degrees of freedom on two axes (a first axis, e.g., x) and a second axis, e.g., y), allowing the device to specify its position in a plane. Touchpad (screen) displays can also be used as input devices.

[0090] According to one embodiment, portions of one or more methods described herein can be executed by a computer system CS in response to a processor PRO executing one or more sequences of one or more instructions contained in main memory MM. Such instructions may be read into main memory MM from another computer-readable medium, such as a storage device SD. Execution of the sequence of instructions contained in main memory MM causes processor PRO to execute the process steps described herein. One or more processors in a multiprocessor arrangement may also be employed to execute the sequence of instructions contained in main memory MM. In alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions. Therefore, the description herein is not limited to any specific combination of hardware circuitry and software.

[0091] As used herein, the term "computer-readable medium" refers to any medium that participates in providing instructions to a processor (PRO) for execution. Such media can take many forms, including but not limited to non-volatile media, volatile media, and transmission media. Non-volatile media include, for example, optical discs or magnetic disks, such as storage devices (SDs). Volatile media include dynamic memory, such as main memory (MMs). Transmission media include coaxial cables, copper wires, and optical fibers, including wires forming a bus (BS). Transmission media can also take the form of sound waves or light waves, such as those generated during radio frequency (RF) and infrared (IR) data communications. Computer-readable media can be non-transient, such as floppy disks, retractable disks, hard disks, magnetic tape, any other magnetic media, CD-ROMs, DVDs, any other optical media, punched cards, paper tape, any other physical media with a perforated pattern, RAM, PROMs and EPROMs, flash memory EPROMs, any other memory chips, or memory cartridges. Non-transient computer-readable media may have instructions written on them. When executed by a computer, the instructions may implement any of the features described herein. Transient computer-readable media may include carrier waves or other propagating electromagnetic signals.

[0092] Various forms of computer-readable media can involve carrying one or more sequences of one or more instructions to a processor PRO for execution. For example, the instructions may initially be carried on a disk of a remote computer. The remote computer can load the instructions into its dynamic memory and transmit them over a telephone line using a modem. A modem local to the computer system CS can receive data over the telephone line and convert the data into an infrared signal using an infrared transmitter. An infrared detector coupled to a bus BS can receive the data carried in the infrared signal and place the data on the bus BS. The bus BS carries the data to main memory MM, from which the processor PRO retrieves and executes the instructions. The instructions received by the main memory MM may optionally be stored on a storage device SD before or after execution by the processor PRO.

[0093] The computer system CS may also include a communication interface CI coupled to the bus BS. The communication interface CI provides bidirectional data communication coupling with a network link NDL connected to the local network LAN. For example, the communication interface CI may be an Integrated Services Digital Network (ISDN) card or a modem to provide data communication connectivity with a corresponding type of telephone line. As another example, the communication interface CI may be a Local Area Network (LAN) card to provide data communication connectivity with a compatible LAN. A wireless link may also be implemented. In any such implementation, the communication interface CI transmits and receives electrical, electromagnetic, or optical signals carrying digital data streams representing various types of information.

[0094] A network link (NDL) typically provides data communication to other data devices over one or more networks. For example, a network link NDL can provide connectivity to a host computer (HC) via a local network (LAN). This can include data communication services provided via a global packet data communication network (now commonly referred to as the "Internet" INT). Local networks (LANs) (the Internet) all use electrical, electromagnetic, or optical signals that carry digital data streams. Signals over various networks, as well as signals on the network data link (NDL) and through the communication interface (CI) (carrying digital data to and from the computer system (CS)), are exemplary forms of carriers for transmitting information.

[0095] A computer system CS can send and receive messages, including program code, via networks (multiple), network data links (NDL), and communication interfaces (CI). In the Internet example, a host computer HC can transmit application request code via the Internet (INT), network data links (NDL), local network (LAN), and communication interface (CI). For example, such a download application can provide all or part of the methods described herein. The received code can be executed by the processor PRO upon receipt and / or stored in storage devices (SD) or other non-volatile storage devices for later execution. In this way, the computer system CS can obtain application code in carrier form.

[0096] Figure 10 This is a schematic diagram of a photolithography projection apparatus according to an embodiment.

[0097] A photolithography projection device may include an illumination system IL, a first stage MT, a second stage WT, and a projection system PS.

[0098] The irradiation system IL can adjust the radiation beam B. In this specific case, the irradiation system also includes a radiation source SO.

[0099] The first stage (e.g., pattern forming apparatus stage) MT may be provided with a pattern forming apparatus holder to hold the pattern forming apparatus MA (e.g., a mask), and is connected to the first locator to accurately position the pattern forming apparatus relative to the article PS.

[0100] The second stage (substrate stage) WT may be provided with a substrate holder to hold the substrate W (e.g., a silicon wafer coated with resist) and is connected to a second locator to accurately position the substrate relative to the article PS.

[0101] A projection system (“lens”) PS (e.g., a refractive, reflective, or reflective-refracting optical system) can image the radiating portion of the pattern forming apparatus MA onto a target portion C (e.g., comprising one or more dies) of the substrate W.

[0102] As described herein, the device can be transmissive (i.e., having transmissive pattern forming apparatus). However, it can also typically be reflective, for example (having reflective pattern forming apparatus). The device can employ a different kind of pattern forming apparatus than a classic mask; examples include programmable mirror arrays or LCD matrices.

[0103] A source SO (e.g., a mercury lamp or excimer laser, LPP (laser-generated plasma) EUV source) generates a radiation beam. This beam is fed into an irradiation system (irradiator) IL, either directly or after passing through adjustment components such as a beam expander Ex. The irradiator IL may include adjustment components AD for setting the outer and / or inner radial ranges of the intensity distribution in the beam (generally referred to as σ_outer and σ_inner, respectively). Additionally, it will typically include various other components such as an integrator IN and a concentrator CO. In this way, the beam B striking the patterning apparatus MA has the desired uniformity and intensity distribution in its cross-section.

[0104] In some embodiments, the source SO can be inside the housing of the photolithography projection device (e.g., this is often the case when the source SO is a mercury lamp), but it can also be located away from the photolithography projection device, with the radiation beam it generates being introduced into the device (e.g., by means of a suitable guide mirror); the latter scenario can be the case when the source SO is an excimer laser (e.g., based on KrF, ArF, or F2 laser irradiation).

[0105] The beam PB can then intercept the patterning apparatus MA, which is held on the patterning apparatus stage MT. After traversing the patterning apparatus MA, the beam PB can pass through the lens PL, which focuses the beam B onto the target portion C of the substrate W. With the aid of a second positioning component (and an interferometry component IF), the substrate stage WT can be moved precisely, for example, to position the different target portions C in the path of the beam PB. Similarly, for example, after the patterning apparatus MA is mechanically retrieved from the patterning apparatus library or during scanning, the first positioning component can be used to precisely position the patterning apparatus MA relative to the path of the beam B. Typically, movement of the stages MT and WT can be achieved using a long-stroke module (coarse positioning) and a short-stroke module (fine positioning). However, in the case of a stepper (as opposed to a stepping scanning tool), the patterning apparatus stage MT can be precisely connected to the short-stroke actuator, or it can be fixed.

[0106] The described tool can be used in two different modes: step mode and scan mode. In step mode, the patterning apparatus stage MT remains essentially stationary, and the entire patterning apparatus image is projected onto the target portion C in one pass (i.e., a single "flash"). The substrate stage WT can be offset in the x and / or y directions, allowing different target portions C to be irradiated by the beam PB.

[0107] In scanning mode, essentially the same scenario applies, except that the given target portion C is not exposed in a single "flash". Instead, the patterning stage MT is movable at a speed v in a given direction (the so-called "scanning direction," e.g., the y-direction), allowing the projection beam B to be used to scan the patterning image; concurrently, the substrate stage WT moves simultaneously in the same or opposite direction at a speed V = Mv, where M is the magnification of the lens PL (typically M = 1 / 4 or 1 / 5). In this way, a relatively large target portion C can be exposed without sacrificing resolution.

[0108] Figure 11 This is a schematic diagram of another photolithography projection apparatus (LPA) according to an embodiment.

[0109] The LPA may include a source collector module SO, an irradiation system (irradiator IL) configured to modulate the radiation beam B (e.g., EUV radiation), a support structure MT, a substrate stage WT, and a projection system PS.

[0110] The support structure (e.g., a pattern forming apparatus stage) MT can be configured to support the pattern forming apparatus (e.g., a mask or mask plate) MA and is connected to a first positioner PM configured to accurately position the pattern forming apparatus.

[0111] A substrate stage (e.g., a wafer stage) WT can be configured to hold a substrate (e.g., a wafer coated with resist) W and is connected to a second locator PW configured to accurately position the substrate.

[0112] The projection system (e.g., a reflective projection system) PS can be configured to project a pattern given by the pattern forming apparatus MA to the radiation beam B onto a target portion C (e.g., including one or more dies) of the substrate W.

[0113] As depicted here, the LPA can be reflective (e.g., using a reflective patterning apparatus). It is important to note that because most materials are absorbent in the EUV wavelength range, the patterning apparatus can have a multilayer reflector comprising, for example, multiple stacks of molybdenum and silicon. In one example, the multilayer reflector has 40 pairs of molybdenum and silicon layers, each layer being a quarter wavelength thick. Even smaller wavelengths can be produced using X-ray lithography. Since most materials are absorbent at both EUV and X-ray wavelengths, the thin sheet of patterned absorbing material on the morphology of the patterning apparatus (e.g., TaN absorber on top of a multilayer reflector) defines where features will be printed (positive resist) or not printed (negative resist).

[0114] The irradiator IL can receive an extreme ultraviolet (EUV) radiation beam from the source collector module SO. Methods for generating EUV radiation include, but are not limited to, converting a material into a plasma state having at least one element (e.g., xenon, lithium, or tin) and possessing one or more emission lines in the EUV range. In one such method, commonly referred to as laser-generated plasma (“LPP”), the plasma can be generated by irradiating a fuel (such as material droplets, streams, or clusters having spectral emission elements) with a laser beam. The source collector module SO can be part of an EUV radiation system that includes a laser (not shown in the original text). Figure 11 (As shown in the diagram) a laser beam is used to provide the excitation fuel. The resulting plasma emits output radiation, such as EUV radiation, which is collected using a radiation collector located in the source collector module. The laser and the source collector module can be separate entities, for example, when a CO2 laser is used to provide the laser beam for fuel excitation.

[0115] In this case, the laser may not be considered part of the lithography apparatus, and the radiation beam can be delivered from the laser to the source-collector module by means of a beam delivery system including, for example, suitable guide mirrors and / or beam expanders. In other cases, the source may be an integrated part of the source-collector module, such as when the source is a discharge-generated plasma EUV generator, often referred to as a DPP source.

[0116] An irradiator IL may include adjusters for adjusting the angular intensity distribution of the radiation beam. Typically, at least the outer and / or inner radial ranges of the intensity distribution in the pupil plane of the irradiator (generally referred to as σ_outer and σ_inner, respectively) can be adjusted. Additionally, the irradiator IL may include various other components, such as faceted fields and pupil reflector devices. The irradiator can be used to adjust the radiation beam to have a desired uniformity and intensity distribution in its cross-section.

[0117] A radiation beam B can be incident on a patterning apparatus (e.g., a mask) MA and patterned by the patterning apparatus, which is held on a support structure (e.g., a patterning apparatus stage) MT. After being reflected from the patterning apparatus (e.g., the mask) MA, the radiation beam B passes through a projection system PS, which focuses the beam onto a target portion C of a substrate W. The substrate stage WT can be precisely moved, for example, to position different target portions C within the path of the radiation beam B, by means of a second locator PW and a position sensor PS2 (e.g., an interferometer, a linear encoder, or a capacitive sensor). Similarly, a first locator PM and another position sensor PS1 can be used to precisely position the patterning apparatus (e.g., the mask) MA relative to the path of the radiation beam B. The patterning apparatus (e.g., the mask) MA and the substrate W can be aligned using patterning apparatus alignment marks M1, M2 and substrate alignment marks P1, P2.

[0118] The described device LPA can be used in at least one of the following modes: step mode, scan mode, and stationary mode.

[0119] In step mode, the support structure (e.g., patterning apparatus stage) MT and substrate stage WT remain substantially stationary while the entire pattern imparted by the radiation beam is projected onto the target portion C at once (i.e., single static exposure). The substrate stage WT is then offset in the X and / or Y directions so that different target portions C can be exposed.

[0120] In scanning mode, as the pattern imparted by the radiation beam is projected onto the target portion C (i.e., single dynamic exposure), the support structure (e.g., patterning apparatus stage) MT and the substrate stage WT are scanned synchronously. The velocity and direction of the substrate stage WT relative to the support structure (e.g., patterning apparatus stage) MT can be determined by the magnification (reduction) and image inversion characteristics of the projection system PS.

[0121] In static mode, while the pattern imparted by the radiation beam is projected onto the target portion C, the support structure (e.g., the patterning apparatus stage) MT remains essentially stationary, thus holding the programmable patterning apparatus in place, and the substrate stage WT is moved or scanned. In this mode, a pulsed radiation source is typically employed, and the programmable patterning apparatus is updated as needed after each movement of the substrate stage WT or between consecutive radiation pulses during scanning. This operating mode can be readily applied to maskless lithography, which utilizes a programmable patterning apparatus (such as a programmable mirror array of the type cited above).

[0122] Figure 12 This is a detailed view of a photolithography projection apparatus according to an embodiment.

[0123] As shown, the LPA may include a source collector module SO, an irradiation system IL, and a projection system PS. The source collector module SO is constructed and arranged such that a vacuum environment can be maintained within the enclosed structure 220 of the source collector module SO. The EUV radiation emitting plasma 210 may be formed by a plasma source generated by a discharge. EUV radiation may be generated by a gas or vapor, such as xenon, lithium vapor, or tin vapor, wherein a very hot plasma 210 is created to emit radiation in the EUV range of the electromagnetic spectrum. The very hot plasma 210 is created, for example, by a discharge that results in at least partially ionized plasma. To efficiently generate radiation, a partial pressure of, for example, 10 Pa of Xe, Li, Sn vapor, or any other suitable gas or vapor may be required. In an embodiment, a plasma that excites tin (Sn) is provided to generate EUV radiation.

[0124] Radiation emitted by the thermal plasma 210 is transferred from the source chamber 211 to the collector chamber 212 via an optional gas barrier or contaminant trap 230 (also referred to in some cases as a contaminant barrier or fin trap) located in or behind an opening in the source chamber 211. The contaminant trap 230 may include a channel structure. The contaminant trap 230 may also include a gas barrier or a combination of a gas barrier and a channel structure. As known in the art, the contaminant trap or contaminant barrier 230 further indicated herein includes at least a channel structure.

[0125] Collector chamber 211 may include a radiation collector CO, which may be a so-called grazing incidence collector. The radiation collector CO has an upstream radiation collector side 251 and a downstream radiation collector side 252. Radiation traversing the collector CO may be reflected from a grating spectral filter 240 to be focused along the optical axis indicated by the dotted dashed line 'O' into a virtual source point IF. The virtual source point IF is generally referred to as the intermediate focus, and the source collector module is arranged such that the intermediate focus IF is located at or near the opening 221 in the enclosed structure 220. The virtual source point IF is an image of the radiative emission plasma 210.

[0126] Subsequently, a radiation traversal illumination system IL, which may include a faceted field mirror assembly 22 and a faceted pupil mirror assembly 24, is arranged to provide a desired angular distribution of the radiation beam 21 at the patterning apparatus MA and a desired uniformity of radiation intensity at the patterning apparatus MA. As the radiation beam 21 is reflected at the patterning apparatus MA held by the support structure MT, a patterned beam 26 is formed, and the patterned beam 26 is imaged by the projection system PS via reflective elements 28 and 30 onto the substrate W held by the substrate stage WT.

[0127] More components than shown may typically be present in the illumination optics unit IL and the projection system PS. Depending on the type of lithography equipment, the grating spectral filter 240 may optionally be present. Furthermore, more mirrors than shown in the figures may be present, for example, with… Figure 12 In contrast, the projection system PS may contain 1 to 6 additional reflective elements.

[0128] like Figure 12 The collector optics CO illustrated is depicted as a nested collector with grazing incidence reflectors 253, 254, and 255, serving only as an example of a collector (or collector mirror). The grazing incidence reflectors 253, 254, and 255 are arranged symmetrically about the optical axis O, and this type of collector optics CO can be used in combination with a plasma source generated by a discharge, commonly referred to as a DPP source.

[0129] Figure 13This is a detailed view of the source collector module SO of the photolithography projection apparatus LPA according to an embodiment.

[0130] The source-collector module SO can be part of the LPA radiation system. The laser LA can be arranged to deposit laser energy into a fuel, such as xenon (Xe), tin (Sn), or lithium (Li), thereby creating a highly ionized plasma 210 with electron temperatures of tens of eV. Energetic radiation generated during the deexcitation and recombination of these ions is emitted from the plasma, collected by a near-normally incident collector optics CO, and focused onto an opening 221 in the enclosed structure 220.

[0131] The concepts disclosed in this paper can be used to simulate or mathematically model any general-purpose imaging system for imaging subwavelength features, and may be particularly useful for emerging imaging techniques capable of producing increasingly shorter wavelengths. Emerging techniques already in use include EUV (Extreme Ultraviolet) and DUV lithography, which can produce wavelengths of 193 nm using ArF lasers and even 157 nm using fluorine lasers. Moreover, EUV lithography can produce wavelengths in the 20 to 50 nm range by using synchrotrons or by bombarding materials (solid-state or plasma) with high-energy electrons to generate photons in this range.

[0132] The embodiments of this disclosure may be further described by the following terms:

[0133] 1. A non-transient computer-readable medium having instructions recorded thereon, the instructions being executed by a computer to perform a method for evaluating a selected set of patterns, the method comprising:

[0134] Obtain (i) a first pattern set generated by the pattern selection process, (ii) first pattern data associated with the first pattern set, (iii) characteristic data associated with the first pattern data, and (iv) second pattern data associated with the second pattern set;

[0135] A machine learning model is trained based on feature data associated with a first pattern, and the machine learning model is configured to predict pattern data input into the machine learning model.

[0136] By feeding the second pattern set into a trained machine learning model, predicted second pattern data of the second pattern set is generated; and

[0137] The first pattern set is evaluated by comparing the second pattern data with the predicted second pattern data.

[0138] 2. The medium according to Clause 1, wherein obtaining the first pattern data includes:

[0139] A first contour or first image is generated by executing a reference model, which is configured to simulate the patterning process using a first set of patterns as input.

[0140] 3. The medium according to Clause 1, wherein obtaining the first pattern data and the second pattern data includes:

[0141] A contour or image is obtained from a measurement image of a patterned substrate comprising a first pattern set and a second pattern set.

[0142] 4. The medium according to Clause 1, wherein the first pattern set includes a first plurality of patterns of the design layout, and the second pattern set includes a second plurality of patterns of the design layout.

[0143] 5. The medium according to Clause 4, wherein the first set of patterns is a subset of the second set of patterns.

[0144] 6. The medium according to Clause 1, wherein the second pattern data includes a second contour or second image generated by performing a reference model using a second pattern set as input, wherein the reference model is configured to simulate the patterning process.

[0145] 7. The medium according to Clause 1, wherein the second pattern set includes more patterns than the first pattern set.

[0146] 8. The medium according to Clause 1, wherein the second pattern set comprises a full chip layout.

[0147] 9. The medium according to Clause 1, wherein the characteristic data includes data of a standard metric generated from the first pattern data, the standard metric being configured to quantify one or more physical characteristics of the pattern.

[0148] 10. The medium pursuant to Clause 9, wherein the standard measures include:

[0149] Standard metrics are placed at the edges, located at multiple positions along the contour of the first pattern data;

[0150] Critical dimension (CD) standard metric is configured to measure the CD value of the first pattern set;

[0151] Configured as a standard metric for measuring lines in the first pattern set;

[0152] A standard metric configured to measure the space between features of a first set of patterns;

[0153] Configured as a standard metric for measuring tip-to-tip structures; and / or

[0154] It is configured as a standard metric to measure the difference between the profile predicted by the measurement model and the design profile.

[0155] 11. The medium according to Clause 1, wherein the evaluation of the first pattern set includes:

[0156] Calculate the difference between the second pattern data and the predicted second pattern data.

[0157] 12. The medium pursuant to Clause 11, wherein the evaluation of the first pattern set includes:

[0158] Determine whether the difference violates a difference threshold; and

[0159] In response to the fact that the difference does not violate the difference threshold, the first pattern set is classified as an acceptable performance measurement.

[0160] 13. The medium pursuant to Clause 12, wherein the evaluation of the first pattern set includes:

[0161] The absolute pattern overlay is determined as a function of the absolute error, which is associated with a trained machine learning model trained using a first set of patterns.

[0162] 14. The medium pursuant to Clause 12, wherein the evaluation of the first pattern set includes:

[0163] The relative pattern coverage is determined as a function of the relative error, which is a comparison between a first error range associated with a trained machine learning model trained using a first set of patterns and a second error range associated with another set of patterns.

[0164] 15. The medium pursuant to Clause 12 also includes:

[0165] Based on the assessment, a risk pattern is identified within the design layout, which is associated with model prediction errors that violate the expected error threshold.

[0166] Supplement the first pattern set with risk patterns.

[0167] 16. The medium pursuant to Clause 12 also includes:

[0168] Based on the first set of patterns, identify a list of patterns to be checked by the measurement tool.

[0169] 17. The medium according to Clause 16, wherein the measuring instrument determines a measuring value including at least one of the following: critical dimensions, overlap, and edge placement errors associated with a first set of patterns patterned on a substrate.

[0170] 18. The medium pursuant to Clause 12 also includes:

[0171] Identify the locations in the second pattern set corresponding to violations of the difference threshold.

[0172] The first pattern set is supplemented with one or more patterns associated with the identified location, the supplemented first pattern set having higher pattern coverage compared to the first pattern set; and

[0173] Use the supplementary first set of patterns to train another machine learning model.

[0174] 19. The medium pursuant to Clause 1, wherein the machine learning model is a convolutional neural network.

[0175] 20. The medium according to Clause 2, wherein the reference model includes one or more models characterizing the patterning process.

[0176] 21. The medium pursuant to Clause 20, wherein the reference model includes a source model, an optical device model, a resist model, an etching model, or a combination thereof.

[0177] 22. The medium pursuant to Clause 21, wherein the first pattern data, the second pattern data, and the predicted second pattern data comprise at least one of the following:

[0178] From the spatial image or contour extracted from it,

[0179] From the extracted mask image or outline,

[0180] The resist image or resist outline extracted from it; and

[0181] Etched images or outlines extracted from them.

[0182] 23. The medium pursuant to Clause 22, wherein the reference model is a calibrated non-machine learning model.

[0183] 24. The medium pursuant to Clause 1 also includes:

[0184] Improvements to the patterning process are identified through a trained machine learning model.

[0185] 25. The medium pursuant to Clause 24, wherein the identified improvements include:

[0186] By using a trained machine learning model to simulate the patterning process, the optical proximity effect correction of the mask pattern associated with the patterning process is determined;

[0187] By using a trained machine learning model to simulate the patterning process, source mask optimizations associated with the patterning process are determined; and / or

[0188] By using a trained machine learning model to simulate the patterning process, the pattern fidelity matching between the pattern printed on the substrate and the pattern of the design layout is improved.

[0189] 26. A method for evaluating a selected set of patterns, the method comprising:

[0190] Obtain (i) a first pattern set generated by the pattern selection process, (ii) first pattern data associated with the first pattern set, (iii) characteristic data associated with the first pattern data, and (iv) second pattern data associated with the second pattern set;

[0191] A machine learning model is trained based on feature data associated with a first pattern, and the machine learning model is configured to predict pattern data input into the machine learning model.

[0192] By feeding the second pattern set into a trained machine learning model, predicted second pattern data of the second pattern set is generated; and

[0193] The first pattern set is evaluated by comparing the second pattern data with the predicted second pattern data.

[0194] 27. The method according to Clause 26, wherein obtaining the first pattern data includes:

[0195] A first contour or first image is generated by executing a reference model, which is configured to simulate the pattern formation process using a first set of patterns as input.

[0196] 28. The method according to Clause 26, wherein obtaining the first pattern data and the second pattern data includes:

[0197] A contour or image is obtained from a measurement image of a patterned substrate comprising a first pattern set and a second pattern set.

[0198] 29. The method according to Clause 26, wherein the first pattern set includes a first plurality of patterns of the design layout, and the second pattern set includes a second plurality of patterns of the design layout.

[0199] 30. The method according to Clause 29, wherein the first set of patterns is a subset of the second set of patterns.

[0200] 31. The method according to Clause 26, wherein obtaining the second pattern data includes:

[0201] A second contour or second image is generated by executing a reference model, which is configured to simulate the patterning process using a second set of patterns as input.

[0202] 32. The method according to Clause 26, wherein the second pattern set includes more patterns than the first pattern set.

[0203] 33. The method according to Clause 26, wherein the second pattern set includes a full chip layout.

[0204] 34. The method according to Clause 26, wherein the characteristic data includes standard measures of one or more physical characteristics of a pattern configured to quantize the pattern, generated from the first pattern data.

[0205] 35. The method according to Clause 34, wherein the standard measures include:

[0206] Standard metrics are placed at the edges, located at multiple positions along the contour of the first pattern data;

[0207] Critical dimension (CD) standard metric is configured to measure the CD value of the first pattern set;

[0208] Configured as a standard metric for measuring lines in the first pattern set;

[0209] A standard metric configured to measure the space between features of a first set of patterns;

[0210] Configured as a standard metric for measuring tip-to-tip structures; and / or

[0211] It is configured as a standard metric to measure the difference between the profile predicted by the measurement model and the design profile.

[0212] 36. The method according to Clause 26, wherein evaluating the first pattern set includes:

[0213] Calculate the difference between the second pattern data and the predicted second pattern data.

[0214] 37. The method according to Clause 36, wherein evaluating the first pattern set includes:

[0215] Determine whether the difference violates a difference threshold; and

[0216] In response to the fact that the difference does not violate the difference threshold, the first pattern set is classified as an acceptable performance measurement.

[0217] 38. The method according to Clause 37, wherein evaluating the first pattern set includes:

[0218] The absolute pattern overlay is determined as a function of the absolute error, which is associated with a trained machine learning model trained using a first set of patterns.

[0219] 39. The method according to Clause 37, wherein evaluating the first pattern set includes:

[0220] The relative pattern coverage is determined as a function of the relative error, which is a comparison between a first error range associated with a trained machine learning model trained using a first set of patterns and a second error range associated with another set of patterns.

[0221] 40. The method according to Clause 37, wherein evaluating the first pattern set includes:

[0222] Determine the risk pattern within the design layout, which is associated with model prediction errors that violate the expected error threshold;

[0223] Supplement the first pattern set with risk patterns.

[0224] 41. The method pursuant to Clause 37 also includes:

[0225] Based on the first set of patterns, identify a list of patterns to be checked by the measurement tool.

[0226] 42. The method according to Clause 41, wherein the measuring instrument determines a measuring value including at least one of the following: critical dimensions, overlap, and edge placement errors associated with a first set of patterns patterned on a substrate.

[0227] 43. The method pursuant to Clause 37 also includes:

[0228] Identify the locations in the second pattern set corresponding to violations of the difference threshold.

[0229] The first pattern set is supplemented with one or more patterns associated with the identified location, the supplemented first pattern set having higher pattern coverage than the first pattern set; and

[0230] Use the supplementary first set of patterns to train another machine learning model.

[0231] 44. The method according to Clause 26, wherein the machine learning model is a convolutional neural network.

[0232] 45. The method according to Clause 27, wherein the reference model comprises one or more models characterizing the patterning process.

[0233] 46. ​​The method according to Clause 45, wherein the reference model includes a source model, an optical device model, a resist model, an etching model, or a combination thereof.

[0234] 47. The method according to Clause 46, wherein the first pattern data and the second pattern data include at least one of the following:

[0235] From the spatial image or contour extracted therefrom,

[0236] From the extracted mask image or outline,

[0237] The resist image or resist outline extracted from it; and

[0238] Etched images or outlines extracted from them.

[0239] 48. The method according to Clause 47, wherein the reference model is a non-machine learning model.

[0240] 49. The method pursuant to Clause 26 also includes:

[0241] Improvements to the patterning process are identified through a trained machine learning model.

[0242] 50. The method according to Clause 49, wherein the identified improvements include:

[0243] Optical proximity correction of the mask pattern associated with the patterning process is determined by using a trained machine learning model to simulate the patterning process.

[0244] By using a trained machine learning model to simulate the patterning process, source mask optimizations associated with the patterning process are determined; and / or

[0245] By using a trained machine learning model to simulate the patterning process, the pattern fidelity matching between the pattern printed on the substrate and the pattern of the design layout is improved.

[0246] While the concepts disclosed herein can be used for imaging on substrates such as silicon wafers, it should be understood that the disclosed concepts can be used with any type of lithography imaging system, such as those used for imaging on substrates other than silicon wafers. The description herein is intended to be illustrative and not restrictive. Therefore, it will be apparent to those skilled in the art that modifications can be made as described without departing from the scope of the claims set forth below.

Claims

1. A non-transient computer-readable medium having instructions recorded thereon, the instructions being executed by a computer to perform a method for evaluating a selected set of patterns, the method comprising: Obtain (i) a first pattern set generated by the pattern selection process, (ii) first pattern data associated with the first pattern set, (iii) characteristic data associated with the first pattern data, and (iv) second pattern data associated with the second pattern set; A machine learning model is trained based on the feature data associated with the first pattern, the machine learning model being configured to predict pattern data of the pattern input into the machine learning model; By inputting the second pattern set into a trained machine learning model, predicted second pattern data of the second pattern set is generated; as well as The first pattern set is evaluated by comparing the second pattern data with the predicted second pattern data.

2. The medium according to claim 1, wherein obtaining the first pattern data comprises: A first contour or first image is generated by executing a reference model, which is configured to simulate the patterning process using the first set of patterns as input.

3. The medium according to claim 1, wherein obtaining the first pattern data and the second pattern data comprises: A contour or image is obtained from a measurement image of a patterned substrate including the first pattern set and the second pattern set.

4. The medium according to claim 1, wherein the first pattern set is a subset of the second pattern set.

5. The medium of claim 2, wherein the second pattern data comprises a second contour or second image generated by performing the reference model using the second pattern set as input, wherein the reference model is configured to simulate the patterning process.

6. The medium of claim 1, wherein the characteristic data includes data of a standard metric derived from the first pattern data, the standard metric being configured to quantify one or more physical characteristics of the pattern.

7. The medium according to claim 6, wherein the standard metric comprises: The edge placement standard measure is located at multiple locations along the contour of the first pattern data; A critical size metric is configured to measure critical size values ​​of the first pattern set; Configured as a standard metric for measuring lines in the first pattern set; A standard metric configured to measure the space between features of the first pattern set; It is configured as a standard metric for measuring tip-to-tip structures; and / or It is configured as a standard metric to measure the difference between the profile predicted by the measurement model and the design profile.

8. The medium of claim 1, wherein evaluating the first pattern set comprises: The absolute pattern overlay is determined as a function of the absolute error, which is associated with the trained machine learning model trained using the first pattern set.

9. The medium of claim 1, wherein evaluating the first pattern set comprises: The relative pattern overlay is determined as a function of relative error, which is a comparison between a first error range associated with the trained machine learning model trained using the first pattern set and a second error range associated with another pattern set.

10. The medium according to claim 1, further comprising: Based on the assessment, a risk pattern is determined within the design layout, which is associated with the model prediction error that violates the error threshold; The risk pattern is used to supplement the first pattern set.

11. The medium according to claim 1, further comprising: Based on the assessment, a list of patterns to be checked by the measurement tools is identified.

12. The medium according to claim 1, further comprising: Identify the locations where a threshold violation of the second pattern set corresponds to the difference between the second pattern data and the predicted second pattern data. The first pattern set is supplemented with one or more patterns associated with the identified location, the supplemented first pattern set having higher pattern coverage compared to the first pattern set; and Use the supplementary first pattern set to train another machine learning model.

13. The medium of claim 2, wherein the reference model comprises one or more models characterizing the patterning process, and wherein the reference model comprises one or more of a source model, an optical device model, a resist model, and an etching model.

14. The medium of claim 1, wherein the first pattern data, the second pattern data, and the predicted second pattern data comprise at least one of the following: From the spatial image or contour extracted therefrom, From the extracted mask image or outline, The resist image or resist outline extracted from it; and Etched images or outlines extracted from them.

15. The medium of claim 2, wherein the reference model is a calibrated non-machine learning model.

16. The medium according to claim 1, wherein the method further comprises: By using the trained machine learning model to simulate the patterning process, an optical proximity effect correction for the mask pattern associated with the patterning process is determined; By using the trained machine learning model to simulate the patterning process, source mask optimizations associated with the patterning process are determined; and / or By using the trained machine learning model to simulate the patterning process, the pattern fidelity matching between the pattern printed on the substrate and the pattern of the design layout is improved.

Images