Measurement in the presence of CMOS underarray (CuA) structures using machine learning and physical modeling.

Abstract

The system may include a controller including a processor configured to execute program instructions, the program instructions causing the processor to execute a measurement recipe as follows: generate a conversion model for converting full-loop optical measurement data to short-loop optical measurement data, the short-loop optical measurement data including optical measurement data of a periodic memory array structure, the full-loop optical measurement data including optical measurement data of a complementary metal oxide semiconductor (CMOS) under-array (CuA) device, the CuA device including a CMOS structure under a replica of the periodic memory array structure, generate a measurement model for determining measurements of the CuA device, receive full-loop optical measurement data of the CuA device on a test sample, convert the full-loop optical measurement data to short-loop optical measurement data using the conversion model, and determine the values ​​of measurements of the periodic memory array structure on a test sample using the measurement model.

Description

Technical Field

[0001] The present disclosure generally relates to optical measurement, and more specifically to optical measurement of a memory structure including a buried CMOS structure.

Background Art

[0002] As one approach to meet the requirement of improving the performance while maintaining or reducing the physical size of a memory device (e.g., a 3D memory device), fabricating a CMOS circuit (e.g., a logic circuit) under a memory array structure can be mentioned. This approach is generally referred to as complementary metal oxide semiconductor (CMOS) under array (CuA) technology.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Patent Document 3

Patent Document 4

Patent Document 5

Patent Document 6

Patent Document 7

Patent Document 8

Patent Document 9

Patent Document 10

Patent Document 11

[0004] [Non-Patent Document 1] Germer, et al., “Intercomparison between optical and x-ray scatterometry measurements of FinFET structures” Proc.SPIE, v.8681, p.86810Q (2013) [Non-Patent Document 2] Kline,et al. “X-ray scattering critical dimensional metrology using a compact x-ray source for next generation semiconductor devices.”Journal of Micro / Nanolithography,MEMS,and MOEMS 16.1(2017) [Overview of the project] [Problems that the invention aims to solve]

[0005] However, CuA technology presents unique challenges specific to measurement systems used in process control, as the underlying CMOS circuitry can affect the measurement of the memory array structure. Therefore, there is a need to develop systems and methods to address this challenge. [Means for solving the problem]

[0006] A system is disclosed according to one or more exemplary embodiments of the present disclosure. In an embodiment, the system comprises a controller including one or more processors configured to execute a program instruction, which causes one or more processors to execute a measurement recipe by: generating a conversion model for converting full-loop optical measurement data to short-loop optical measurement data, wherein the short-loop optical measurement data includes optical measurement data of a periodic memory array structure, and the full-loop optical measurement data includes optical measurement data of a complementary metal oxide semiconductor (CMOS) under-array (CuA) device, wherein the CuA device includes a CMOS structure under a replica of the periodic memory array structure; generating a measurement model for determining one or more measurements of the CuA device based on the short-loop optical measurement data; receiving full-loop optical measurement data of the CuA device on one or more test samples; using the conversion model to convert the full-loop optical measurement data of the CuA device on one or more test samples to short-loop optical measurement data of the CuA device on one or more test samples; and using the measurement model to determine the values ​​of one or more measurements of the periodic memory array structure on one or more test samples based on the short-loop optical measurement data of the CuA device.

[0007] A system is disclosed according to one or more exemplary embodiments of this disclosure. In an embodiment, the system comprises an optical characterization system. In an embodiment, the system comprises a controller communicatively coupled to an optical characterization system and a reference characterization system, the controller comprising one or more processors configured to execute program instructions, the program instructions causing one or more processors to execute a measurement recipe by: generating a conversion model for converting full-loop optical measurement data to short-loop optical measurement data, the short-loop optical measurement data comprising optical measurement data of a periodic memory array structure, the full-loop optical measurement data comprising optical measurement data of a complementary metal oxide semiconductor (CMOS) under-array (CuA) device, the CuA device comprising a CMOS structure under a replica of a periodic memory array structure; generating a measurement model for determining one or more measurements of the CuA device based on the short-loop optical measurement data; receiving full-loop optical measurement data of the CuA device on one or more test samples; converting the full-loop optical measurement data of the CuA device on one or more test samples to short-loop optical measurement data of the CuA device on one or more test samples using the conversion model; and determining the values ​​of one or more measurements of the periodic memory array structure on one or more test samples using the measurement model based on the short-loop optical measurement data of the CuA device.

[0008] A method is disclosed in accordance with one or more exemplary embodiments of the present disclosure. In an embodiment, the method includes generating a conversion model for converting full-loop optical measurement data to short-loop optical measurement data, where the short-loop optical measurement data includes optical measurement data of a periodic memory array structure, the full-loop optical measurement data includes optical measurement data of a complementary metal oxide semiconductor (CMOS) under array (CuA) device, and the CuA device includes a CMOS structure under a replica of the periodic memory array structure. In an embodiment, the method includes generating a measurement model for determining one or more measurements of the CuA device based on the short-loop optical measurement data. In an embodiment, the method includes generating full-loop optical measurement data of the CuA device on one or more test samples. In an embodiment, the method includes using the conversion model to convert the full-loop optical measurement data of the CuA device on one or more test samples to short-loop optical measurement data of the CuA device on one or more test samples. In an embodiment, the method includes determining the values of one or more measurements of the periodic memory array structure on one or more test samples using the measurement model with the short-loop optical measurement data of the CuA device.

[0009] It should be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not necessarily restrictive of the claimed invention. The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the general description, serve to explain the principles of the invention.

[0010] Those skilled in the art will better understand many advantages of the present disclosure by referring to the accompanying drawings.

Brief Description of the Drawings

[0011] [Figure 1A] It is a block diagram of a measurement system according to one or more embodiments of the present disclosure. [Figure 1B] A simplified schematic diagram of a characteristic evaluation subsystem configured as an optical characteristic evaluation subsystem according to one or more embodiments of the present disclosure. [Figure 1C] A simplified schematic diagram of a characteristic evaluation subsystem configured as an X-ray characteristic evaluation subsystem according to one or more embodiments of the present disclosure. [Figure 1D] A simplified schematic diagram of a characteristic evaluation subsystem configured as a particle beam characteristic evaluation subsystem according to one or more embodiments of the present disclosure. [Figure 2] A simplified schematic diagram of a CuA device according to one or more embodiments of the present disclosure. [Figure 3] A flowchart showing steps executed in a method for evaluating the characteristics of a CuA device according to one or more embodiments of the present disclosure.

Embodiments for Carrying Out the Invention

[0012] Here, refer in detail to the subject matter of the disclosure shown in the accompanying drawings. The present disclosure is specifically illustrated and described in a particular embodiment and with respect to its particular features. The embodiments described herein are considered to be illustrative rather than limiting. It should be readily apparent to those skilled in the art that various changes and modifications can be made in form and detail without departing from the spirit and scope of the present disclosure.

[0013] [[ID=:24]] Embodiments of the present disclosure relate to a system and method for optical measurement of a CuA device based on converting full-loop optical measurement data of a fully fabricated complementary metal oxide semiconductor (CMOS) under array (CuA) device into short-loop optical measurement data associated with a simulation of optical measurement data without a buried CMOS structure, and then generating measurements based on the short-loop optical measurement data.

[0014] A CuA structure (e.g., a CuA memory structure) typically includes logic circuits (e.g., CMOS logic circuits) physically placed beneath a replica of a memory array structure (e.g., a three-dimensional (3D) memory stack or a 3D NAND structure). As used herein, the term CuA structure may encompass a wide range of logic and memory array structure designs. Therefore, this disclosure is not limited to any particular CuA architecture.

[0015] Optical measurement is commonly used in semiconductor process control because it can offer relatively high measurement throughput and is generally non-destructive. In optical measurement, a sample is illuminated with light, and measurements are generated based on the corresponding light emitted from the sample. In optical measurement of subsurface features, the light typically needs to propagate through at least the top of the sample to reach the subsurface feature of interest. Therefore, optical measurement systems typically utilize wavelengths of light selected to propagate through the target structure with relatively low absorption.

[0016] On the other hand, in the case of CuA structures, incident light interacts with both the memory array structure and the embedded CMOS structure, which can impair the ability to generate independent measurements of the memory array structure. This specification argues that existing optical measurement techniques may become insufficient for accurately characterizing CuA structures, especially as devices become smaller and more complex. For example, some techniques may depend on the wavelength of light within the transparent window of the memory array structure in question, which may be absorbed at least partially by the underlying logic circuitry. For instance, some logic circuitry may utilize a polysilicon layer that absorbs light with wavelengths greater than approximately 450 nanometers (nm). In this case, optical measurements at wavelengths lower than approximately 450 nm may generate independent measurements of the memory array structure. However, such techniques may be limited to CuA designs selected to incorporate the aforementioned absorbing materials, may have limited sensitivity to deep structures, and may be limited in terms of broadband optical measurement methods where multi-wavelength data are advantageous. As another example, some techniques rely on supervised learning of artificial neural networks using optical measurement data with labels generated through additional measurement methods. However, these methods may have various limitations, including, but are not limited to, the requirement of high sampling of ground truth reference data, the considerable time required to generate enough labels for training, performance limitations for deeply embedded structures, insensitivity to process changes, and generally inapplicability to CuA structures outside the training dataset.

[0017] In some embodiments, a measurement model is generated that is suitable for receiving and generating measurements of a memory array structure based solely on received short-loop optical measurement data. In some embodiments, a conversion model is generated that converts full-loop optical measurement data to short-loop optical measurement data or otherwise transforms it. Thus, full-loop optical measurement data can be generated and converted to short-loop optical measurement data during the production process, so that measurements of the fabricated CuA device can be generated based on the short-loop measurement model.

[0018] Physical-based measurement models based solely on memory array structures (e.g., short-loop measurement models) offer numerous advantages over physical-based measurement models based on full CuA devices (e.g., full-loop measurement models), including, but not limited to, shallower sample depths required for simulation, simpler and / or fewer interfaces with the model, comparable computational resources (or simulation time with a given set of computational resources), and greater strength of the modeled results. Thus, the systems and methods disclosed herein may be superior to alternatives that utilize physical-based models based on full-loop optical measurement data.

[0019] A system and method for characterizing CuA structures according to one or more embodiments of the present disclosure will now be described in more detail with reference to Figures 1A to 3.

[0020] Figure 1A is a block diagram of a measurement system 100 according to one or more embodiments of the present disclosure. In some embodiments, the measurement system 100 comprises a characterization subsystem 102 that generates measurement data of a sample 104 using optical techniques, and a controller 106 that generates one or more measurements based on the measurement data.

[0021] The characterization subsystem 102 may include any components or combinations of components suitable for generating measurement data for sample 104.

[0022] In some embodiments, the characterization subsystem 102 includes an optical characterization subsystem 102 that generates measurement data based on the interaction between the sample 104 and light. For example, the characterization subsystem 102 may include, but is not limited to, a spectroscopic ellipsometer (SE), an SE with illumination at multiple angles, an SE that measures Müller matrix elements (e.g., using a rotational compensator), a single-wavelength ellipsometer, a beam profile ellipsometer (angle-resolved ellipsometer), a beam profile reflectometer (angle-resolved reflectometer), a broadband reflectometer (spectrophotometer), a single-wavelength reflectometer, an angle-resolved reflectometer, an imaging system, a scattermeter (e.g., a speckle analyzer), or a combination thereof.

[0023] In some embodiments, the characterization subsystem 102 includes an X-ray characterization subsystem 102 for generating measurement data based on the interaction between the sample 104 and X-rays. For example, the characterization subsystem 102 may be, but is not limited to, a small-angle X-ray scattering (SAXS) system or an X-ray reflection scattering measurement (SXR) system.

[0024] In some embodiments, the characterization subsystem 102 includes a particle beam characterization subsystem 102 that generates measurement data based on the interaction between a particle beam, such as but not limited to an electron beam (e-beam), an ion beam, or a neutral particle beam, and a sample 104.

[0025] In some embodiments, the characterization subsystem 102 provides multiple types of measurements. In some embodiments, the measurement system 100 includes multiple measurement subsystems 102, each providing one or more different combinations of measurements. Furthermore, the measurement system 100 may be provided as a single tool or as multiple tools. A single tool providing multiple measurement configurations is generally described in Patent Document 1, published April 26, 2011, which is incorporated herein by reference in its entirety. Multiple tools and structural analysis are generally described in Patent Document 2, published January 13, 2009, which is incorporated herein by reference in its entirety.

[0026] Furthermore, Patent Document 3, published on October 29, 2019, entitled "Model based optical measurements of semiconductor structures with anisotropic dielectric permittivity," Patent Document 4, published on February 7, 2023, entitled "Scatterometry based methods and systems for measurement of strain in semiconductor structures," Patent Document 5, published on June 15, 2021, entitled "Measurement models of nanowire semiconductor structures based on re-useable sub-structures," Patent Document 6, published on January 17, 2023, entitled "Measuring thin films on grating and bandgap on grating," Patent Document 7, published on October 26, 2021, entitled "Measurement methodology of advanced nanostructures," and Patent Document 8, published on October 6, 2020, entitled "Visualization of three-dimensional semiconductor structures," are all incorporated herein by reference in their entirety.

[0027] In some embodiments, the controller 106 comprises one or more processors 108 configured to execute a set of program instructions held in memory 110 or a memory device, the program instructions causing the processors 108 to perform various operations.

[0028] One or more processors 108 of the controller 106 may include any processor or processing element known in the art. For the purposes of this disclosure, the terms “processor” or “processing element” may be broadly defined to include any device having one or more processing or logic elements (e.g., one or more microprocessor devices, one or more application-specific integrated circuit (ASIC) devices, one or more field-programmable gate arrays (FPGAs), or one or more digital signal processors (DSPs)). In this sense, one or more processors 108 may include any device configured to execute algorithms and / or instructions (e.g., program instructions stored in memory). In some embodiments, one or more processors 108 may be embodied as a desktop computer, a mainframe computer system, a workstation, an image computer, a parallel processor, a network computer, or any other computer system configured to execute a program that operates or is configured to operate with the characterization subsystem 102, as described throughout this disclosure. Furthermore, different subsystems of the measurement system 100 may include processors or logic elements suitable for performing at least some of the steps described in this disclosure. Therefore, the above description should not be construed as an limitation to embodiments of the present disclosure, but rather as an example. Furthermore, the steps described throughout the present disclosure may be performed by a single controller, or alternatively, by multiple controllers. In addition, controller 106 may include one or more controllers housed in a common housing or multiple housings. In this way, any controller or combination of controllers can be separately packaged as modules suitable for integration into the measurement system 100.

[0029] The memory 110 may include any storage medium known in the art that is suitable for storing program instructions executable by one or more associated processors 108. For example, the memory 110 may include a non-temporary storage medium. Another example of the memory 110 may include, but is not limited to, read-only memory (ROM), random access memory (RAM), magnetic or optical memory devices (e.g., disks), magnetic tapes, and solid-state drives. It should be further noted that the memory 110 may be housed in a common controller housing together with one or more processors 108. In some embodiments, the memory 110 may be located remotely from the physical locations of one or more processors 108 and the controller 106. For example, one or more processors 108 of the controller 106 may access remote memory (e.g., a server) accessible via a network (e.g., the Internet, and an intranet).

[0030] The controller 106 may be communicatively coupled to any component or combination of components of the measurement system 100. In some embodiments, the controller 106 may receive data (e.g., measurement data) from one or more components of the measurement system 100. In some embodiments, the controller 106 may control one or more components of the measurement system 100 via drive signals. More generally, the controller 106 may perform any of the steps described herein.

[0031] In some embodiments, the controller 106 generates one or more measurements of sample 104 based at least in part on measurement data generated by the characterization subsystem 102. Measurement of the target parameter may involve numerous algorithms that can be performed by the controller 106. For example, the optical interaction between sample 104 and the incident beam can be modeled using an electromagnetic (EM) solver, and algorithms such as exact coupled-wave analysis (RCWA), finite element method (FEM), method of moments, surface integral, volume integral, or finite-difference time-domain (FDTD) methods may be used, but are not limited to these. Sample 104 may be modeled (e.g., parameterized) using a geometric engine, a process modeling engine, or a combination of both. The use of process modeling is generally described in Patent Document 9, published September 8, 2020, which is incorporated herein by reference in its entirety. Geometric engines are implemented, for example, in KLA Corporation's AcuShape software.

[0032] Controller 106 may analyze the collected measurement data using, but is not limited to, libraries, fast order reduction models, regression, statistical methods (see, for example, “Statistical model-based metrology” in Patent Document 10 by S. Pandev et al.), machine learning algorithms (e.g., more generally, neural networks, support vector machines (SVM), principal component analysis (PCA), independent component analysis (ICA), local linear embeddings (LLE), dimensionality reduction techniques), sparse representation techniques, Fourier transform techniques, wavelet transform techniques, or any appropriate combination of data fitting and / or optimization techniques such as Kalman filtering. Statistical model-based metrology is generally described in Patent Document 10, published on October 16, 2018, which is incorporated herein by reference in its entirety. Controller 106 may analyze the collected measurement data using algorithms that do not involve modeling, optimization, and / or fitting. The characterization of patterned wafers is generally described in Patent Document 11, published on December 10, 2019, which is incorporated herein by reference in its entirety. In some embodiments, the controller 106 utilizes one or more algorithms to facilitate matching from the same or different types of tools (e.g., different instances or configurations of the characterization subsystem 102).

[0033] The controller 106 may be designed to provide efficient performance through any appropriate technique such as parallelization, computational distribution, load balancing, multi-service support, or dynamic load optimization, but is not limited to these. Furthermore, the controller 106 may perform any step using any type or combination of configurations such as dedicated hardware (e.g., FPGA), software, or firmware, but is not limited to these.

[0034] Furthermore, the controller 106 may generate any type of measurement of sample 104 (or a portion thereof) based at least in part on measurement data from the characterization subsystem 102. In some embodiments, the controller 106 generates measurement measurements such as overlay measurements, limit dimension (CD) measurements, shape measurements (e.g., height measurements, inclination measurements, or sidewall angle measurements), stress measurements, composition measurements, bandgap measurements, electrical property measurements, or process condition measurements (e.g., focus and / or dose conditions, resist state, partial pressure, temperature, or focusing model). In some embodiments, the controller 106 generates inspection measurements in which one or more defects on sample 104 are identified or classified.

[0035] The measurement system 100 and any of its components (e.g., the characterization subsystem 102, or the controller 106) may be configured to implement a recipe (e.g., a measurement recipe) which can define various configuration parameters and / or steps performed in a measurement or a series of measurements.

[0036] For example, a recipe may include various embodiments of the design of sample 104 (e.g., the design of the CuA device 202 on sample 104). Such embodiments include, but are not limited to, the layout of features on one or more sample layers, the size of features, or the pitch of features. Another example of a recipe may include, but are not limited to, illumination parameters such as illumination wavelength, illumination pupil distribution (e.g., the distribution of illumination angles and the associated intensity of illumination at those angles), polarization of incident illumination, illumination spatial distribution, or sample height. Another example of a recipe may include, but are not limited to, collection parameters such as collection pupil distribution (e.g., the desired distribution of angular light from the sample used for measurement and the associated filtered intensity at those angles), collection field aperture settings for selecting the portion of the sample under consideration, polarization of collected light, or wavelength filters. Another example of a recipe may include various processing steps (e.g., those performed by controller 106 to generate measurements based on the measurement data generated according to the recipe).

[0037] Referring now to Figure 2, which is a simplified schematic diagram of a CuA device 202 according to one or more embodiments of the present disclosure. The CuA device 202 may comprise a memory array structure 204 and various CMOS structures 206 (e.g., logic structures) located beneath a replica of the memory array structure 204. For example, full-loop optical measurement data and short-loop measurement data may have the same set of memory array structures 204 (e.g., replicas of the memory array structure 204). This may be associated with a design of experiments (DOE) with known variations.

[0038] The memory array structure 204 may include any number or type of structures suitable for forming a memory array. For example, the memory array structure 204 may include, but is not limited to, a 3D NAND structure formed from patterned features 208 in a multilayer stack 210. Furthermore, such a memory array structure 204 is typically a periodic structure having periodicity along one or more dimensions.

[0039] The CMOS structure 206 may include any number or type of structures fabricated under a replica of the memory array structure 204. For example, the CMOS structure 206 may, but does not necessarily, be suitable for controlling and / or powering the memory array structure 204. In this way, a memory device can be formed by the combination of the CMOS structure 206 and the memory array structure 204 (e.g., a 3D memory device). Furthermore, the CMOS structure 206 may have a spatially variable distribution such that the number and / or design of its constituent features may not be periodic across the entire CuA device 202. Thus, the CMOS structure 206 can generally be described as aperiodic. However, it should be noted that the CMOS structure 206 may exhibit local periodicity in some regions.

[0040] Furthermore, the memory array structure 204 and / or CMOS structure 206 can generally have any design, and therefore the term CuA device 202 as used herein is not limited to any particular design. For example, the CuA device 202 may, but is not limited to, include an intervening layer between the memory array structure 204 and the CMOS structure 206, such as a source layer 212 (e.g., a polysilicon source layer). In another example not shown, the CuA device 202 may include an intervening layer between the CMOS structure 206 and the substrate 214.

[0041] Referring here to Figure 3, a technique for characterizing a CuA device 202 or a portion thereof, according to one or more embodiments of the present disclosure, will be described in more detail.

[0042] At various stages of the fabrication process, it may be desirable to generate measurements of the structural components of the CuA device 202 (e.g., CMOS structure 206 and / or memory array structure 204). Such measurements may include, but are not limited to, measurement or defect measurements (e.g., inspection measurements). Measurement may include, but are not limited to, overlay measurements, critical dimension (CD) measurements, shape measurements (e.g., height measurements, inclination measurements, or sidewall angle measurements), stress measurements, composition measurements, bandgap measurements, electrical property measurements, or process condition measurements (e.g., focus and / or dose conditions, resist state, partial pressure, temperature, or focusing model). Inspection measurements may include, but are not limited to, identification and / or characterization of defects in the fabrication process (e.g., unwanted features, missing features, or features with inappropriate shape or location). Furthermore, such measurements may be used for a wide range of purposes, including, but not limited to, process control, placement, or performance estimation of the fabricated CuA device 202.

[0043] Measurements may be generated after any processing step for fabricating the CuA device 202. For example, measurements may be generated after the fabrication of the CMOS structure 206 and / or after the fabrication of the memory array structure 204 for forming the complete CuA device 202. For convenience of explanation, measurements of the complete CuA device 202, including both the memory array structure 204 and the underlying CMOS structure 206, are referred to herein as “full-loop” measurements.

[0044] In general, this specification assumes that, depending on the interaction between the illumination beam 114 and the sample 104, a single processing step measurement may provide information about any features fabricated on the sample 104. Thus, performing independent measurements of newly fabricated features can be difficult. For example, full-loop measurements generally provide information about both the memory array structure 204 and the underlying CMOS structure 206, or may be influenced by both, which can limit or impair the ability to generate independent measurements of the memory array structure 204.

[0045] In some embodiments, measurements of various test structures may be generated to support the generation of independent measurements of specific features. For example, a measurement of a test structure including a memory array structure 204 without the corresponding embedded CMOS structure 206 is referred to herein as a “short-loop” measurement.

[0046] Furthermore, measurements at any processing step can generally be produced using any suitable technique, including but not limited to optical, X-ray, or particle-based techniques. However, different measurement techniques may have different trade-offs. For example, optical measurement techniques can generally provide non-destructive measurements with high measurement throughput, but may have limited resolution or be restricted to certain types of structures (e.g., periodic structures) based on the corresponding analysis or modeling step. Therefore, optical measurements are commonly used when throughput is particularly critical. As another example, X-ray and / or particle-based techniques may offer higher resolution than some optical techniques, but may have relatively lower throughput and / or may be destructive measurements. Consequently, such techniques are commonly used for reference measurements.

[0047] However, this specification does not consider it feasible or desirable in all applications to generate every possible type of measurement at every measurement step. In such cases, depending on the available data, various techniques may be used to generate measurements of a specific structure (e.g., independent measurements of the memory array structure 204).

[0048] Figure 3 is a flowchart illustrating the steps performed in a method 300 for characterizing a CuA device 202 according to one or more embodiments of the present disclosure. The applicant notes that embodiments and enabling techniques described herein with respect to the measurement system 100 should be construed to extend to method 300 as well. For example, any of the steps associated with method 300 may be implemented by the controller 106 and / or the characterization subsystem 102 of the measurement system 100. However, it should be further noted that method 300 is not limited to the architecture of the measurement system 100.

[0049] In this specification, Method 300 is considered suitable for applications where generating independent measurements of a memory array structure 204 is desirable, but is not limited thereto. Here, both full-loop and short-loop measurements are available for learning purposes, but full-loop data only is preferred for process control. One objective of this technique is to construct a measurement model suitable for receiving and generating measurements of the memory array structure 204 based solely on received short-loop optical measurement data (e.g., data from the optical characterization subsystem 102). Another objective of this technique is to construct a conversion model that converts full-loop optical measurement data to short-loop optical measurement data or otherwise transforms it. Thus, full-loop optical measurement data can be generated during the production process and converted to short-loop optical measurement data so that measurements of the fabricated CuA device 202 can be generated based on the short-loop measurement model.

[0050] A physically based measurement model based solely on the memory array structure 204 (e.g., a short-loop measurement model) may offer many advantages over a physically based measurement model based on a full CuA device 202 (e.g., a full-loop measurement model), including, but not limited to, a shallower sample depth required for simulation, simpler and / or fewer interfaces with the model, equivalent computational resources required (or simulation time with a given set of computational resources), or greater strength of the modeled results. Thus, Method 300 may be superior to alternatives that utilize physically based models based on full-loop optical measurement data.

[0051] In some embodiments, method 300 includes a step 302 of generating a conversion model that converts full-loop optical measurement data into short-loop optical measurement data. The short-loop optical measurement data includes optical measurement data of a periodic memory array structure 204, while the full-loop optical measurement data includes optical measurement data of a CuA device 202 having a CMOS structure 206 beneath a replica of the periodic memory array structure 204. In this way, the conversion model can remove the effect of the CMOS structure 206 from the full-loop optical measurement data for simulation of the optical measurement data that would be generated in the absence of the CMOS structure 206.

[0052] In some embodiments, the transformation model includes a machine learning model, which may utilize any suitable type of machine learning technique known in the art, including but not limited to supervised, unsupervised, or reinforcement learning techniques, and may include any combination of such learning techniques. In some embodiments, the machine learning model is a neural network model.

[0053] The machine learning model can then generate a relationship between full-loop and short-loop optical measurement data that demonstrates the effect of the CMOS structure 206. Once trained, the machine learning model can be used to generate short-loop optical measurement data from received full-loop measurement data. As a result, the machine learning model can identify and remove the effect of the CMOS structure 206 on the optical measurement data.

[0054] Machine learning models can be trained on any suitable dataset. In some embodiments, the training data may include experimentally generated short-loop and full-loop optical measurement data, as well as reference data providing ground truth measurements of the parameter of interest.

[0055] In some embodiments, step 302 includes generating short-loop optical measurement data for a set of periodic memory array structures 204 on one or more training samples, generating short-loop reference data for a set of periodic memory array structures 204 on one or more training samples, generating full-loop optical measurement data for a set of CuA devices 202 on one or more training samples, and generating full-loop reference data for a set of CuA devices 202 on one or more training samples.

[0056] For example, short-loop and full-loop optical measurement data can be generated using the optical characterization subsystem 102. Alternatively, reference data can be generated by a high-resolution characterization subsystem 102, including but not limited to the X-ray characterization subsystem 102 or the particle beam characterization subsystem 102. Non-limiting examples include, but are not limited to, transmission electron microscope (TEM) data, scanning electron microscope (SEM) data (e.g., metric SEM (CD-SEM) data or electron beam SEM (EB-SEM) data), SAXS data (e.g., transmission SAXS (T-SAXS) data or CD-SAXS data), X-ray photoelectron spectroscopy (XPS) data, or X-ray diffraction (XRD) data. CD-SAXS data may be particularly suitable for use as reference data, but this disclosure is not limited thereto.

[0057] Step 302 may then involve training a transformation model with full-loop and short-loop optical measurement data, as well as full-loop and short-loop reference data. In this way, the training dataset may include ground truth measurements associated with the CMOS structures 206, thereby allowing for the identification of the contributions of these structures to the full-loop optical measurement data and the removal of these contributions from the generated short-loop optical measurement data.

[0058] In some embodiments, method 300 includes step 304 of generating a measurement model for determining one or more measurements of a CuA device based on short-loop optical measurement data.

[0059] Any suitable measurement model can be generated in step 304.

[0060] In some embodiments, the measurement model constructed in step 304 includes a physical-based model in which the properties of the CuA device 202 relate to the measurement under consideration through a model of the interaction between one or more illumination beams 114 and the properties of the constituent features. Examples include, but are not limited to, RCWA models, FEM models, method of moments models, surface integral models, volume integral models, or FDTD models, and any suitable physical-based measurement model may be used.

[0061] In some embodiments, the measurement model constructed in step 304 includes an additional machine learning model, which may be any suitable type of machine learning technique known in the art, and may include, but not limited to, any combination of, supervised, unsupervised, or reinforcement learning techniques. In some embodiments, the additional machine learning model is a neural network model. In this configuration, the additional machine learning model can identify patterns between short-loop optical measurement data and the measurement under consideration.

[0062] Additional machine learning models can be trained on any suitable training dataset. In some embodiments, additional machine learning models are trained on experimentally generated short-loop optical measurement data (e.g., generated by the optical characterization subsystem 102) and experimentally generated short-loop reference data (e.g., generated by the X-ray characterization subsystem 102, or the particle beam characterization subsystem 102, etc.).

[0063] In some embodiments, additional machine learning models are generated based on synthetic optical measurement data generated by the physical-based model as described above. For example, Method 300 may include the step of generating synthetic measurement data using a physical-based measurement model having a set of parameters describing the synthetic CuA device 202. Method 300 may then include the step of training an additional machine learning model to generate one or more measurements based on the training data, where the training data includes at least one of optical measurement data for one or more training samples, or synthetic data describing the synthetic CuA device 202. Thus, the physical-based model may be used to provide supplementary training data beyond the experimental data (e.g., optical measurement data and / or reference data) associated with the training and / or test samples described above.

[0064] For example, the synthetic measurement data may include a synthesized equivalent of optical measurement data (which may be generated, for example, by the optical characterization subsystem 102) based on various combinations of geometric and dispersion parameters of the CuA device 202. For instance, different combinations of geometric and dispersion parameters of the memory array structure 204, as well as different effective medium models of the CMOS structure 206, may be provided to a physical-based measurement model as input to generate the synthetic measurement data as output. Thus, this synthetic measurement data can be characterized as a synthetic design of experiments (DOE) suitable for providing training data for machine learning models.

[0065] The measurement model constructed in step 304 can then be used during the fabrication process to generate measurements of additional samples (e.g., test samples) with unknown properties.

[0066] In some embodiments, the method 300 includes step 306 of generating full-loop optical measurement data of a CuA device 202 on one or more test samples (for example, by an optical characterization subsystem 102).

[0067] In some embodiments, method 300 includes step 308 of converting full-loop optical measurement data of CuA device 202 on one or more test samples to short-loop optical measurement data of CuA device 202 on one or more test samples using a conversion model (e.g., from step 302).

[0068] In some embodiments, Method 300 includes step 310 of determining one or more measurement values ​​of a memory array structure 204 on one or more test samples using a measurement model based on short-loop optical measurement data of a CuA device 202. Any suitable measurement of the memory array structure 204 can be generated, including, but not limited to, overlay measurement, limit dimension (CD) measurement, shape measurement (e.g., height measurement, inclination measurement, or sidewall angle measurement), stress measurement, composition measurement, bandgap measurement, electrical property measurement, process condition measurement (e.g., focus and / or dose conditions, resist state, partial pressure, temperature, or focusing model), defect identification, or defect classification.

[0069] Measurements of the memory array structure 204 on the test sample can then be used for various purposes. In some embodiments, the measurements are used to estimate the process control, placement, and / or performance of the CuA device 202 on the test sample. For example, the measurements may be used to generate modifiable values ​​for one or more process tools (e.g., scanners or steppers) in feedback and / or feedforward processes.

[0070] With reference to Figures 1B to 1D, various non-limiting configurations of the characterization subsystem 102 according to one or more embodiments of the present disclosure will be described in more detail.

[0071] In some embodiments, the characterization subsystem 102 is an optical measurement subsystem that generates measurement data based on the interaction between the sample 104 and light. Figure 1B is a simplified schematic diagram of the characterization subsystem 102 configured as an optical characterization subsystem 102 according to one or more embodiments of the present disclosure. For example, the characterization subsystem 102 may include, but is not limited to, a spectroscopic ellipsometer (SE), a multi-angle illumination SE, an SE that measures Müller matrix elements (e.g., using a rotational compensator), a single-wavelength ellipsometer, a beam profile ellipsometer (e.g., an angle-resolved ellipsometer), a beam profile reflectometer (e.g., an angle-resolved reflectometer), a broadband reflectance spectrometer (e.g., a spectrophotometer), a single-wavelength reflectometer, an angle-resolved reflectometer, an imaging system, a scattermeter (e.g., a speckle analyzer), or a combination thereof.

[0072] In some embodiments, the characterization subsystem 102 includes an illumination source 112 configured to produce at least one illumination beam 114. The illumination from the illumination source 112 may include, but not limited to, ultraviolet (UV), visible, or infrared (IR) radiation, one or more selected wavelengths of light. For example, the characterization subsystem 102 may include one or more apertures in the illumination pupil plane to split the illumination from the illumination source 112 into one or more illumination beams 114 or illumination lobes. In this regard, the characterization subsystem 102 may provide dipole illumination, or orthogonal illumination, etc. Furthermore, the spatial profiles of one or more illumination beams 114 on the sample 104 may be controlled by a field plane aperture to have any selected spatial profile.

[0073] The illumination source 112 may include any type of illumination source suitable for providing at least one illumination beam 114. In some embodiments, the illumination source 112 is a laser source. For example, the illumination source 112 may include, but is not limited to, one or more narrowband laser sources, broadband laser sources, supercontinuum laser sources, or white light laser sources. In some embodiments, the illumination source 112 includes a laser sustained plasma (LSP) source. For example, the illumination source 112 may include, but is not limited to, an LSP lamp, LSP bulb, or LSP chamber, which preferably includes one or more elements capable of emitting broadband illumination when excited to a plasma state by a laser source. In some embodiments, the illumination source 112 includes a lamp source. In some embodiments, the illumination source 112 may include, but is not limited to, an arc lamp, a discharge lamp, or an electrodeless lamp.

[0074] The illumination source 112 may provide one or more illumination beams 114 using free-space technology and / or optical fibers.

[0075] In some embodiments, the characterization subsystem 102 directs the illumination beam 114 to the sample 104 via an illumination path 118 through at least one illumination lens 116 (e.g., an objective lens). The illumination path 118 directs the illumination beam 114 to the sample 104 and may include one or more optical components suitable for modifying and / or adjusting the illumination beam 114. In some embodiments, the illumination path 118 includes one or more illumination path optics 120 for shaping or otherwise controlling the illumination beam 114. For example, the illumination path optics 120 may include, but are not limited to, one or more field diaphragms, one or more pupil diaphragms, one or more polarizers, one or more filters, one or more beam splitters, one or more diffusers, one or more homogenizers, one or more apodizers, one or more beam shapers, or one or more mirrors (e.g., static mirrors, translationally movable mirrors, or scanning mirrors).

[0076] The characterization subsystem 102 may position the sample 104 for measurement using any suitable technique. In some embodiments, as shown in Figure 1B, the characterization subsystem 102 includes a sample stage 122 which includes one or more actuators (e.g., linear actuators, tip / tilt actuators, or rotary actuators) for positioning the sample 104 relative to the illumination beam 114. In some embodiments, although not explicitly shown, the characterization subsystem 102 includes a beam scanning optical system (e.g., galvanometer mirrors, or scanning prisms) for adjusting the position and / or scanning one or more illumination beams 114.

[0077] In some embodiments, the characterization subsystem 102 includes at least one collecting lens 124 for capturing light or other radiation emitted from the sample 104, referred to in this detail as collected light 126, and directing this collected light 126 to one or more detectors 128 via a collecting path 130. The collecting path 130 may include one or more optical elements suitable for modifying and / or adjusting the collected light 126 from the sample 104. In some embodiments, the collecting path 130 includes one or more collecting path optics 132 for shaping or otherwise controlling the collected light 126. For example, the collecting path optics 132 may include, but are not limited to, one or more field diaphragms, one or more pupil diaphragms, one or more polarizers, one or more filters, one or more beam splitters, one or more diffusers, one or more homogenizers, one or more apodizers, one or more beam shapers, or one or more mirrors (e.g., static mirrors, translationally movable mirrors, or scanning mirrors).

[0078] The characterization subsystem 102 may generally include any number or type of detectors 128. For example, the characterization subsystem 102 may include, but is not limited to, a photodiode, avalanche photodiode, or single-photon detector, at least one single-pixel detector 128. Another example is the characterization subsystem 102, which may include, but is not limited to, a charge-coupled device (CCD) or complementary metal-oxide-semiconductor (CMOS) device, a line detector, or a time-delay integral (TDI) detector, at least one multi-pixel detector 128.

[0079] The detector 128 can be positioned at any selected location within the collection path 130. In some embodiments, the characterization subsystem 102 includes the detector 128 on the field of view (e.g., the plane conjugate to the sample 104) to generate an image of the sample 104. In some embodiments, the characterization subsystem 102 includes the detector 128 on the pupil plane (e.g., the diffraction plane) to generate a pupil image. In this regard, the pupil image may correspond to the angular distribution of light from the sample 104 captured by the detector 128. For example, the diffraction order related to the diffraction of the illumination beam 114 from the sample 104 may be imaged on the pupil plane or otherwise observed. Generally, the detector 128 can capture any combination of reflected (or transmitted), scattered, or diffracted light from the sample 104.

[0080] The illumination path 118 and collection path 130 of the characterization subsystem 102 can be oriented in a wide range of configurations. For example, as shown in Figure 1B, the illumination path 118 and collection path 130 can include non-overlapping optical paths. In some embodiments, though not explicitly shown, the characterization subsystem 102 may include an oriented beam splitter such that a common objective lens directs the illumination beam 114 towards the sample 104 while simultaneously capturing the collected light 126.

[0081] Figure 1C is a simplified schematic diagram of a characterization subsystem 102 configured as an X-ray characterization subsystem 102 according to one or more embodiments of the present disclosure. Such a characterization subsystem 102 may, but is not limited to, a small-angle X-ray scattermeter (SAXR) or a soft X-ray reflectometer (SXR). X-ray characterization systems and related measurement techniques are generally described in Patent Document 12, published April 19, 2011; Patent Document 13, published February 6, 2018; Patent Document 14, published July 3, 2018; Patent Document 15, published June 18, 2019; Patent Document 16, published July 16, 2019; Patent Document 17, published September 15, 2020; Non-Patent Document 1; Non-Patent Document 2; Patent Document 18, published May 17, 2022; and Patent Document 19, published July 8, 2021, which are incorporated herein by reference in their entirety.

[0082] In some embodiments, the illumination source 112 is an X-ray source configured to produce an X-ray illumination beam 114 having any particle energy (e.g., soft X-rays or hard X-rays). The characterization subsystem 102 may include any combination of components suitable for capturing the associated acquired signal 134, which may include, but are not limited to, X-ray emission, optical emission, or particle emission.

[0083] For example, the characterization subsystem 102 may include an X-ray illumination lens 116 suitable for collimating or focusing the X-ray illumination beam 114, and a collection path lens (not shown) suitable for collecting, collimating, and / or focusing the collected signal 134 from the sample 104. Furthermore, the characterization subsystem 102 may include, but is not limited to, various illumination path optics (not shown) and / or collection path optics (not shown), such as a specular X-ray optical system including an X-ray collimating mirror, an oblique incidence ellipsoidal mirror, a polycapillary optical system including a hollow capillary X-ray waveguide, a multilayer optical system, or a system, or any combination thereof. In embodiments, the characterization subsystem 102 may include, but is not limited to, an X-ray detector 128, such as an X-ray monochromator (e.g., a crystalline monochromator such as a Loxley-Tanner-Bowen monochromator), an X-ray aperture, an X-ray beam diaphragm, or a diffraction optical system (e.g., a zone plate).

[0084] Figure 1D is a simplified schematic diagram of a characterization subsystem 102 configured as a particle beam characterization subsystem 102 according to one or more embodiments of the present disclosure.

[0085] In one embodiment, the illumination source 112 includes a particle source (e.g., an electron beam source or an ion beam source), thereby the illumination beam 114 includes a particle beam (e.g., an electron beam or a particle beam). The illumination source 112 may include any particle source known in the art that is suitable for generating the particle illumination beam 114. For example, the illumination source 112 may include, but is not limited to, an electron gun or an ion gun. In another embodiment, the illumination source 112 is configured to provide a particle beam with adjustable energy. For example, the illumination source 112 including an electron source may provide, but is not limited to, an acceleration voltage in the range of 0.1 kilovolts (kV) to 30 kV. As another example, the illumination source 112 including an ion source may provide, but is not necessarily limited to, an ion beam with energy in the range of 1 kiloelectron volt (keV) to 50 keV.

[0086] In another embodiment, the illumination path 118 includes one or more particle focusing elements (e.g., an illumination lens 116 or a collecting lens 124). For example, one or more particle focusing elements may include, but are not limited to, a single particle focusing element or one or more particle focusing elements forming a composite system. In another embodiment, one or more particle focusing elements include an illumination lens 116 configured to guide the particle illumination beam 114 to the sample 104. Furthermore, one or more particle focusing elements may include, but are not limited to, electrostatic lenses, magnetic lenses, single-potential lenses, or double-potential lenses, or any type of electron lens known in the art.

[0087] In another embodiment, the characterization subsystem 102 includes one or more particle detectors 128 for imaging or otherwise detecting particles emitted from the sample 104. For example, the detector 128 may include an electron collector (e.g., a secondary electron collector, or a backscatter electron detector). In another example, the detector 128 may include a photon detector (e.g., a photodetector, an X-ray detector, or a scintillation element coupled to a photomultiplier tube (PMT) detector) for detecting electrons and / or photons from the sample surface.

[0088] The subject matter described herein may also refer to different components that are contained within or connected to other components. It should be understood that such illustrated architectures are merely illustrative, and that numerous other architectures achieving the same functionality are, in fact, possible. Conceptually, any arrangement of components to achieve the same functionality is effectively “related” to achieve the desired functionality. Therefore, any two components combined herein to achieve a particular functionality, regardless of the architecture or components in between, can be considered “related” to achieve the desired functionality. Similarly, any two such related components can also be considered “connected” or “linked” to each other to achieve the desired functionality, and any two such related components can also be considered “linkable” to each other to achieve the desired functionality. Specific examples of linkable components include, but are not limited to, physically interactable and / or physically interacting components, as well as / or wirelessly interactable and / or wirelessly interacting components, and / or logically interactable and / or logically interacting components.

[0089] Many of the present disclosure and its associated advantages are to be understood from the foregoing description, and it will be clear that various modifications can be made to the shape, structure, and arrangement of the components without departing from the disclosed subject matter or without impairing any of its important advantages. The shapes described are for illustrative purposes only, and the claims below are intended to encompass and include such modifications. Furthermore, it should be understood that the present invention is defined by the appended claims.

Claims

1. It is a system, A controller including one or more processors configured to execute program instructions, wherein the program instructions are to be executed by the one or more processors The method involves generating a conversion model for converting full-loop optical measurement data into short-loop optical measurement data, wherein the short-loop optical measurement data includes optical measurement data of a periodic memory array structure, and the full-loop optical measurement data includes optical measurement data of a complementary metal oxide semiconductor (CMOS) under-array (CuA) device, wherein the CuA device includes a CMOS structure under a replica of the periodic memory array structure. To generate a measurement model for determining one or more measurements of the CuA device based on the short-loop optical measurement data, Receiving full-loop optical measurement data of CuA devices on one or more test samples, Using the conversion model, the full-loop optical measurement data of the CuA device on one or more test samples is converted to short-loop optical measurement data of the CuA device on one or more test samples. The short-loop optical measurement data of the CuA device is used to determine the values ​​of one or more measurements of the periodic memory array structure on one or more test samples using the measurement model, The controller executes the measurement recipe. A system equipped with these features.

2. Generating the conversion model for converting the full-loop optical measurement data into the short-loop optical measurement data is, To generate short-loop optical measurement data of a set of periodic memory array structures on one or more training samples, To generate short-loop reference data of the periodic memory array structure of the set on one or more training samples, To generate full-loop optical measurement data for a set of CuA devices on one or more training samples, To generate full-loop reference data for the set of CuA devices on the 1 or more training samples, The transformation model is generated based on the short-loop optical measurement data of the set of periodic memory array structures on one or more training samples, the short-loop reference data of the set of periodic memory array structures on one or more training samples, the full-loop optical measurement data of the set of CuA devices on one or more training samples, and the full-loop optical measurement data of the set of CuA devices on one or more training samples. The system according to claim 1, including the following:

3. The system according to claim 1, wherein the conversion model is a supervised machine learning model.

4. The system according to claim 1, wherein the conversion model is a neural network.

5. The system according to claim 1, wherein the transformation model is an unsupervised machine learning model.

6. The measurement model is based on the interaction of light and matter with the periodic memory array structure, according to the system of claim 1.

7. The aforementioned measurement model, Exact coupled wave analysis (RCWA) model, finite element method (FEM) model, moment method model, surface integral model, volume integral model, or finite difference time-domain (FDTD) model The system according to claim 6, which utilizes at least one of the following.

8. The system according to claim 1, wherein the measurement model is a supervised machine learning model.

9. The system according to claim 1, wherein the measurement model is an unsupervised machine learning model.

10. It is a system, Optical characterization system, A controller communicatively coupled to the optical characterization system and the reference characterization system, the controller including one or more processors configured to execute program instructions, the program instructions to the one or more processors, The method involves generating a conversion model for converting full-loop optical measurement data into short-loop optical measurement data, wherein the short-loop optical measurement data includes optical measurement data of a periodic memory array structure, and the full-loop optical measurement data includes optical measurement data of a complementary metal oxide semiconductor (CMOS) under-array (CuA) device, wherein the CuA device includes a CMOS structure under a replica of the periodic memory array structure. To generate a measurement model for determining one or more measurements of the CuA device based on the short-loop optical measurement data, Receiving full-loop optical measurement data of CuA devices on one or more test samples, Using the conversion model, the full-loop optical measurement data of the CuA device on one or more test samples is converted to short-loop optical measurement data of the CuA device on one or more test samples. The short-loop optical measurement data of the CuA device is used to determine the values ​​of one or more measurements of the periodic memory array structure on one or more test samples using the measurement model, A controller that executes the measurement recipe, A system equipped with these features.

11. Reference Characterization System Furthermore, Generating the conversion model for converting the full-loop optical measurement data into the short-loop optical measurement data is, The optical characterization system generates short-loop optical measurement data of a set of periodic memory array structures on one or more training samples, The reference characteristic evaluation system generates short-loop reference data of the periodic memory array structure of the set on one or more training samples, The optical characterization system generates full-loop optical measurement data for a set of CuA devices on one or more training samples, The reference characteristic evaluation system generates full-loop reference data for the set of CuA devices on one or more training samples, The transformation model is generated based on the short-loop optical measurement data of the set of periodic memory array structures on one or more training samples, the short-loop reference data of the set of periodic memory array structures on one or more training samples, the full-loop optical measurement data of the set of CuA devices on one or more training samples, and the full-loop optical measurement data of the set of CuA devices on one or more training samples. including, The system according to claim 10.

12. The aforementioned optical characterization system is At least one of the following: a polarization analyzer, a reflectometer, or a scattermeter. The system according to claim 10, comprising:

13. The aforementioned reference characteristic evaluation system is At least one of the following: an X-ray characterization system or a particle beam characterization system. The system according to claim 10, comprising:

14. It is a method, The method involves generating a conversion model for converting full-loop optical measurement data into short-loop optical measurement data, wherein the short-loop optical measurement data includes optical measurement data of a periodic memory array structure, and the full-loop optical measurement data includes optical measurement data of a complementary metal oxide semiconductor (CMOS) under-array (CuA) device, wherein the CuA device includes a CMOS structure under a replica of the periodic memory array structure. To generate a measurement model for determining one or more measurements of the CuA device based on the short-loop optical measurement data, To generate full-loop optical measurement data for CuA devices on one or more test samples, Using the conversion model, the full-loop optical measurement data of the CuA device on one or more test samples is converted to short-loop optical measurement data of the CuA device on one or more test samples. The short-loop optical measurement data of the CuA device is used to determine the values ​​of one or more measurements of the periodic memory array structure on one or more test samples using the measurement model, Methods that include...

15. Generating the conversion model for converting the full-loop optical measurement data into the short-loop optical measurement data is, To generate short-loop optical measurement data of a set of periodic memory array structures on one or more training samples, To generate short-loop reference data of the periodic memory array structure of the set on one or more training samples, To generate full-loop optical measurement data for a set of CuA devices on one or more training samples, To generate full-loop reference data for the set of CuA devices on the 1 or more training samples, The transformation model is generated based on the short-loop optical measurement data of the set of periodic memory array structures on one or more training samples, the short-loop reference data of the set of periodic memory array structures on one or more training samples, the full-loop optical measurement data of the set of CuA devices on one or more training samples, and the full-loop optical measurement data of the set of CuA devices on one or more training samples. The method according to claim 14, including the method described in claim 14.

16. The method according to claim 14, wherein the transformation model is a supervised machine learning model.

17. The method according to claim 16, wherein the conversion model is a neural network.

18. The method according to claim 14, wherein the transformation model is an unsupervised machine learning model.

19. The method according to claim 14, wherein the measurement model is based on the interaction of light and matter with the periodic memory array structure.

20. The aforementioned measurement model, Exact coupled wave analysis (RCWA) model, finite element method (FEM) model, moment method model, surface integral model, volume integral model, or finite difference time-domain (FDTD) model The method according to claim 19, wherein at least one of the following is used.

21. The method according to claim 14, wherein the measurement model is a supervised machine learning model.

22. The method according to claim 14, wherein the measurement model is an unsupervised machine learning model.

Images