Training model to generate predictive data

Abstract

A method of training a generator model includes: generating prediction data based on first measurement data using the generator model, where the first measurement data and the prediction data are usable to form an image of a sample; pairing a subset of the first measurement data with a subset of the prediction data, the subset corresponding to a location within a sample image formable from the first measurement data and the prediction data; evaluating, using a discriminator, the likelihood that the predicted data comes from the same data distribution as second measurement data measured from the sample after the etching process; and training the generator model based on correlations of pairs corresponding to the same location relative to correlations of pairs corresponding to different locations and the likelihood evaluated by the discriminator, the correlations being correlations between the paired data subsets.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to European application No. 22186636.1 filed on July 25, 2022 and European application No. 22196424.0 filed on September 19, 2022, the entire contents of which are incorporated herein by reference. Technical Field

[0003] Embodiments herein relate to generating prediction data for a sample after an etching process based on measurement data of the sample after an exposure process but before an etching process, eg, allowing optimization of parameters of the sample before etching. Background Art

[0004] When manufacturing semiconductor integrated circuit (IC) chips, undesirable pattern defects on the substrate (i.e., wafer) or mask are inevitably present during the manufacturing process due to optical effects and accidental particles, thereby reducing yield. Therefore, monitoring the extent of undesirable pattern defects is an important process in IC chip manufacturing. More generally, inspecting and / or measuring the surface of a substrate or other object / material during and / or after its manufacture is an important process.

[0005] Pattern inspection devices with charged particle beams have been used to inspect objects (which may be referred to as samples), such as detecting pattern defects. These devices typically use electron microscopy techniques, such as scanning electron microscopes (SEMs). In an SEM, a primary electron beam with relatively high energy electrons targets a final deceleration step so that it lands on a sample with a relatively low landing energy. The electron beam is focused as a probe point on the sample. The interaction between the material structure at the probe point and the landing electrons from the electron beam results in the emission of signal electrons, such as secondary electrons, backscattered electrons, or Auger electrons, from the surface. Signal electrons can be emitted from the material structure of the sample. By scanning the primary electron beam as a probe point on the sample surface, signal electrons can be emitted across the surface of the sample. By collecting these emitted signal electrons from the sample surface, the pattern inspection device can obtain an image representing the material structure features of the sample surface.

[0006] It may be desirable to scan a sample at different processing stages of the sample. For example, the sample may be scanned after a lithography exposure process has been performed on the sample but before a subsequent etching process. This allows the measurement of a so-called after-development inspection (ADI) image of the sample, sometimes also referred to as a after-lithography inspection image. The sample may then be scanned after the etching process. This allows the formation of a so-called after-etch inspection (AEI) image of the sample. The data from these images may be used, for example, to optimize sample parameters before etching and / or to improve the lithography processing of subsequent substrates. The process of scanning the sample may damage the sample, for example, by damaging the resist material of the sample. The scanning process may also reduce the throughput of the sample evaluation method.

[0007] There is a general need to increase the throughput of sample evaluation methods and / or reduce the damage caused by sample evaluation methods and / or improve accuracy. Summary of the invention

[0008] It is an object of the present disclosure to provide embodiments that support increasing the throughput of a sample evaluation method and / or reducing the damage caused by a sample evaluation method.

[0009] According to one aspect of the present invention, a method for training a generator model is provided, which processes first measurement data measured from a sample before an etching process to generate prediction data for predicting the sample after the etching process, the method comprising: using the generator model to generate prediction data based on the first measurement data, wherein the first measurement data and the prediction data can be used to form an image of the sample; pairing a subset of the first measurement data with a subset of the prediction data, the subsets corresponding to positions within the sample image that can be formed based on the measurement data and the prediction data; using a discriminator to evaluate the likelihood that the prediction data comes from the same data distribution as second measurement data measured from the sample at a different position after the etching process; and training the generator model based on: the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions and the likelihood evaluated by the discriminator, the correlation being the correlation between the paired data subsets.

[0010] According to one aspect of the present invention, a generator model training device is provided for training a generator model, which processes first measurement data measured from a sample before an etching process to generate prediction data for predicting the sample after the etching process, the device comprising: a processor, configured to: generate prediction data based on the first measurement data using the generator model, wherein the first measurement data and the prediction data can be used to form an image of the sample; pair a subset of the first measurement data with a subset of the prediction data, the subsets corresponding to positions within the sample image that can be formed based on the first measurement data and the prediction data; use a discriminator to evaluate the likelihood that the prediction data comes from the same data distribution as second measurement data measured from a sample at a different position after the etching process; and train the generator model based on: the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions and the likelihood evaluated by the discriminator, the correlation being the correlation between the paired data subsets.

[0011] According to one aspect of the present invention, a computer-readable medium is provided, which stores instructions configured to control a processor to train a generator model, which processes first measurement data measured from a sample before an etching process to generate prediction data for predicting the sample after the etching process. The computer-readable medium stores instructions configured to control the processor to: generate prediction data based on the first measurement data using the generator model, wherein the first measurement data and the prediction data can be used to form an image of the sample; pair a subset of the first measurement data with a subset of the prediction data, which subsets correspond to positions within the sample image that can be formed based on the measurement data and the prediction data; use a discriminator to evaluate the likelihood that the prediction data comes from the same data distribution as second measurement data measured from a sample at a different position after the etching process; and train the generator model based on: the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions and the likelihood evaluated by the discriminator that the correlation is the correlation between the paired data subsets.

[0012] According to one aspect of the present invention, a method for training a generator model is provided, which processes paired measurement data measured from a sample after an etching process (the sample has been measured before the etching process) to generate hypothesis data of the sample after a simulated etching process if the sample has not been measured before the etching process. The method includes: using the generator model to generate hypothesis data based on the paired measurement data, wherein the paired measurement data and the hypothesis data can be used to form an image of the sample; using a discriminator to evaluate the possibility that the hypothesis data comes from the same data distribution as the true measurement data measured from the sample after the etching process, the sample has not been measured before the etching process; and training the generator model based on: a function indicating the degree of correlation between the paired measurement data and the hypothesis data, and the possibility evaluated by the discriminator.

[0013] According to one aspect of the present invention, a generator model training device is provided for training a generator model, which processes paired measurement data measured from a sample after an etching process, wherein the sample has been measured before the etching process, to generate hypothesis data of the sample after the simulated etching process if the sample has not been measured before the etching process. The device includes: a processor configured to: generate hypothesis data based on the paired measurement data using the generator model, wherein the paired measurement data and the hypothesis data can be used to form an image of the sample; use a discriminator to evaluate the possibility that the hypothesis data comes from the same data distribution as the real measurement data measured from the sample after the etching process, wherein the sample has not been measured before the etching process; and train the generator model based on the following function: a function indicating the level of correlation between the paired measurement data and the hypothesis data, and the possibility evaluated by the discriminator.

[0014] According to one aspect of the present invention, a computer-readable medium is provided, which stores instructions configured to control a processor to train a generator model, which processes paired measurement data measured from a sample after an etching process (the sample has been measured before the etching process) to generate hypothesis data of the sample after a simulated etching process if the sample has not been measured before the etching process. The computer-readable medium stores instructions configured to control the processor to: use the generator model to generate hypothesis data based on the paired measurement data, wherein the paired measurement data and the hypothesis data can be used to form an image of the sample; use a discriminator to evaluate the possibility that the hypothesis data comes from the same data distribution as the real measurement data measured from the sample after the etching process, the sample has not been measured before the etching process; and train the generator model based on: a function indicating the level of correlation between the paired measurement data and the hypothesis data, and the possibility evaluated by the discriminator. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The above and other aspects of the present disclosure will become more apparent through the description of exemplary embodiments with reference to the attached drawings.

[0016] Figure 1 is a schematic diagram illustrating an exemplary charged particle beam inspection system.

[0017] Figure 2 is shown as Figure 1 Schematic diagram of an exemplary multi-beam charged particle evaluation apparatus that is a portion of an exemplary charged particle beam inspection system.

[0018] Figure 3 is a schematic diagram of an exemplary single-beam electron optical array.

[0019] Figure 4Depicted are images of the sample taken at different locations on the sample before and after the etching process.

[0020] Figure 5 Depicted are pairings of an image patch of a sample before the etching process and a predicted image patch of the same sample after the etching process.

[0021] The schematic diagrams and views show the components described below. However, the components depicted in the figures are not drawn to scale. DETAILED DESCRIPTION

[0022] Reference will now be made in detail to exemplary embodiments, examples of which are shown in the accompanying drawings. The following description refers to the accompanying drawings, wherein, unless otherwise specified, the same numbers in different figures represent the same or similar elements. The implementations set forth in the following description of exemplary embodiments do not represent all implementations consistent with the present invention. Instead, they are merely examples of devices and methods consistent with various aspects related to the present invention as described in the appended claims.

[0023] The increased computing power of electronic devices has led to a reduction in the physical size of the devices, which can be achieved by significantly increasing the packing density of circuit components such as transistors, capacitors, diodes, etc. on IC chips. The increase in resolution makes it possible to manufacture smaller structures. For example, the IC chip of a smartphone, which is only the size of a thumbnail and will be launched in 2019 or earlier, may contain more than 2 billion transistors, each of which is less than one thousandth the size of a human hair. It is not surprising that semiconductor IC manufacturing is a complex and time-consuming process involving hundreds of individual steps. Even if a step goes wrong, it may greatly affect the function of the final product. Just one "fatal defect" can cause device failure. The goal of the manufacturing process is to improve the yield of the entire process. For example, to achieve a yield of 75% for a 50-step process (where a step can represent the number of layers formed on a wafer), the yield of each step must be greater than 99.4%. If the yield of each step is 95%, the yield of the entire process will be as low as 7%.

[0024] While high process yield is desirable in IC chip fabrication plants, it is also essential to maintain high substrate (i.e., wafer) throughput (defined as the number of substrates processed per hour). High process yield and high substrate throughput may be affected by the presence of defects. This is especially true if operator intervention is required to detect the defect. Therefore, high throughput detection and identification of micro- and nano-scale defects by inspection equipment such as scanning electron microscopes (SEMs) is critical to maintaining high yields and low costs.

[0025] The SEM includes a scanning device and a detector device. The scanning device includes an irradiation device, which includes an electron source for generating primary electrons, and a projection device for scanning a sample (e.g., a substrate) with one or more focused primary electron beams. The irradiation device or irradiation system and the projection device or projection system can at least be collectively referred to as an electron optical system or device. The primary electrons interact with the sample and generate secondary electrons. The detection device captures secondary electrons from the sample while scanning the sample so that the SEM can create an image of the sample scanning area. For high-throughput inspection, some inspection devices use multiple focused beams of primary electrons, i.e., multi-beams. The component beams of the multi-beams can be referred to as sub-beams or sub-beams. The multi-beams can scan different parts of the sample simultaneously. Therefore, a multi-beam inspection device can inspect samples at a much higher speed than a single-beam inspection device.

[0026] The following describes a known implementation of a multi-beam inspection device.

[0027] Although the description and drawings are directed to electron optical systems, it should be appreciated that these embodiments are not intended to limit the present disclosure to specific charged particles. Therefore, references to electrons throughout this document may be more generally considered to be references to charged particles, which are not necessarily electrons.

[0028] Reference now Figure 1 , which is a schematic diagram illustrating an exemplary charged particle beam inspection system 100 , which may also be referred to as a charged particle beam evaluation system or simply an evaluation system. Figure 1 The charged particle beam inspection system 100 includes a main chamber 10, a load lock chamber 20, an electron beam system 40, an equipment front end module (EFEM) 30, and a controller 50. The electron beam system 40 is located in the main chamber 10.

[0029] The EFEM 30 includes a first loading port 30a and a second loading port 30b. The EFEM 30 may include additional loading ports (multiple). The first loading port 30a and the second loading port 30b may, for example, receive a substrate front opening pod (FOUP) containing a substrate (e.g., a semiconductor substrate or a substrate made of other (multiple) materials) or a sample to be inspected (substrate, wafer and sample are collectively referred to as "sample" hereinafter). One or more robotic arms (not shown) in the EFEM 30 transport the sample to the load lock chamber 20.

[0030] The load lock chamber 20 is used to remove the gas around the sample. This will produce a vacuum, that is, the local gas pressure is lower than the pressure of the surrounding environment. The load lock chamber 20 can be connected to a load lock vacuum pump system (not shown), which removes gas particles in the load lock chamber 20. The operation of the load lock vacuum pump system enables the load lock chamber to reach a first pressure lower than atmospheric pressure. After reaching the first pressure, one or more robot arms (not shown) transport the sample from the load lock chamber 20 to the main chamber 10. The main chamber 10 is connected to the main chamber vacuum pump system (not shown). The main chamber vacuum pump system removes gas particles in the main chamber 10 so that the pressure around the sample reaches a second pressure lower than the first pressure. After reaching the second pressure, the sample is transported to the electron beam system, which can be inspected by the electron beam system. The electron beam system 40 may include a multi-beam electron optical device.

[0031] The controller 50 is electrically connected to the electron beam system 40. The controller 50 may be a processor (such as a computer) configured to control the charged particle beam inspection apparatus 100. The controller 50 may also include a processing circuit system configured to perform various signal and image processing functions. Figure 1 The controller 50 is shown as being located outside the structure including the main chamber 10, the load lock chamber 20 and the EFEM 30, but it is understood that the controller 50 can be part of the structure. The controller 50 can be located in one of the constituent elements of the charged particle beam inspection device, or can be distributed on at least two constituent elements. Although the present disclosure provides an example of a main chamber 10 that accommodates an electron beam system, it should be noted that the aspects of the present disclosure in its broadest sense are not limited to a chamber that accommodates an electron beam system. On the contrary, it is understood that the aforementioned principles can also be applied to the arrangement of other equipment and other devices operating at a second pressure.

[0032] Reference now Figure 2 , which is a schematic diagram showing an exemplary electron beam system 40, including a multi-beam electron optical system 41, which is Figure 1 1 and 2. The electron beam system 40 includes an electron source 201 and a projection device 230. The electron beam system 40 also includes an electric stage 209 and a sample holder 207. The electron source 201 and the projection device 230 can be collectively referred to as an electron optical system 41 or an electron optical device array. The sample holder 207 is supported by the electric stage 209 so as to hold a sample 208 (e.g., a substrate or a mask) for inspection. The multi-beam electron optical system 41 also includes a detector 240 (e.g., an electronic detection device).

[0033] The electron source 201 may include a cathode (not shown) and an extractor or anode (not shown). During operation, the electron source 201 is configured to emit electrons from the cathode as primary electrons. The extractor and / or anode extract or accelerate the primary electrons to form a primary electron beam 202.

[0034] The projection device 230 is configured to convert the primary electron beam 202 into a plurality of sub-beams 211, 212, 213 and direct each sub-beam onto the sample 208. Although only three sub-beams are shown for simplicity, there may be tens, hundreds, thousands, tens of thousands or hundreds of thousands of sub-beams. A sub-beam may be referred to as a beamlet.

[0035] The controller 50 may be connected to Figure 1 The controller 50 may be used to control various parts of the charged particle beam inspection device 100, such as the electron source 201, the detector 240, the projection device 230 and the electric stage 209. The controller 50 may perform various image and signal processing functions. The controller 50 may also generate various control signals to control the operation of the charged particle beam inspection device (including the charged particle multi-beam device).

[0036] The projection device 230 may be configured to focus the sub-beams 211, 212, and 213 onto the sample 208 for inspection, and may form three detection points 221, 222, and 223 on the surface of the sample 208. The projection device 230 may be configured to deflect the primary sub-beams 211, 212, and 213 to scan the detection points 221, 222, and 223 across respective scanning areas in a portion of the surface of the sample 208. In response to the primary sub-beams 211, 212, and 213 being incident on the detection points 221, 222, and 223 on the sample 208, electrons are generated from the sample 208, which include secondary electrons and backscattered electrons, which may be referred to as signal particles. The electron energy of the secondary electrons is typically less than or equal to 50 eV. The energy of the actual secondary electrons may be less than 5 eV, but any electron with an energy lower than 50 eV is typically considered a secondary electron. The electron energy of the backscattered electrons is generally between 0 eV and the landing energy of the primary beamlets 211, 212, and 213. Since the detected electrons with energy less than 50 eV are generally regarded as secondary electrons, a portion of the actual backscattered electrons will be regarded as secondary electrons.

[0037] The detector 240 is configured to detect signal particles (such as secondary electrons and / or backscattered electrons) and generate corresponding signals, which are sent to the signal processing system 280, for example to construct an image of the corresponding scanned area of ​​the sample 208. The detector 240 may be incorporated into the projection device 230.

[0038] The signal processing system 280 may include circuitry (not shown) configured to process signals from the detector 240 to form an image. The signal processing system 280 may also be referred to as an image processing system. The signal processing system may be incorporated into a component of the electron beam system 40, such as the detector 240 (e.g., Figure 2 As shown). However, the signal processing system 280 can be incorporated into any number of components of the inspection device 100 or the electron beam system 40, such as as part of the projection device 230 or the controller 50. The signal processing system 280 may include an image acquirer (not shown) and a storage device (not shown). For example, the signal processing system may include a processor, a computer, a server, a mainframe host, a terminal, a personal computer, any type of mobile computing device, etc., or a combination thereof. The image acquirer may include at least part of the processing function of the controller. Therefore, the image acquirer may include at least one or more processors. The image acquirer may be communicatively coupled to the detector 240 to allow signal communication, such as an electrical conductor, a fiber optic cable, a portable storage medium, IR, Bluetooth, the Internet, a wireless network, radio, etc., or a combination thereof. The image acquirer may receive a signal from the detector 240, may process the data included in the signal, and may construct an image therefrom. Therefore, the image acquirer may acquire an image of the sample 208. The image acquirer may also perform various post-processing functions, such as generating a contour, superimposing an indicator on the acquired image, etc. The image acquirer may be configured to perform adjustments such as brightness and contrast on the acquired image. The memory may be a storage medium such as a hard disk, a flash drive, cloud storage, random access memory (RAM), other types of computer-readable memory, etc. The memory may be coupled to the image acquirer and may be used to save the scanned raw image data as raw images and post-processed images.

[0039] The signal processing system 280 may include a measurement circuit system (e.g., an analog-to-digital converter) to obtain the distribution of the detected secondary electrons. The electron distribution data collected during the detection time window can be used in conjunction with the corresponding scan path data of each primary sub-beam 211, 212, and 213 incident on the sample surface to reconstruct an image of the inspected sample structure. The reconstructed image can be used to reveal various features of the internal or external structure of the sample 208. Therefore, the reconstructed image can be used to reveal any defects that may exist in the sample. The above-mentioned functions of the signal processing system 280 can be performed in the controller 50, or shared between the signal processing system 280 and the controller 50 for ease of use.

[0040] The controller 50 may control the motorized stage 209 to move the sample 208 during the inspection of the sample 208. The controller 50 may cause the motorized stage 209 to move the sample 208 in one direction (preferably continuously, for example, at a constant speed) at least during the inspection of the sample, which may be referred to as a type of scan. The controller 50 may control the movement of the motorized stage 209 so that it changes the movement speed of the sample 208 according to various parameters. For example, the controller 50 may control the speed of the stage (including its direction) according to the inspection steps of the scanning process and / or the characteristics of the scanning of the scanning process, such as disclosed in EPA 21171877.0 filed on May 3, 2021, which is hereby incorporated herein at least in part with respect to the combined stepping and scanning strategy of the stage.

[0041] Known multi-beam systems, such as the above-mentioned electron beam system 40 and charged particle beam inspection device 100, have been disclosed in US2020118784, US20200203116, US2019 / 0259570 and US2019 / 0259564, which are incorporated herein by reference.

[0042] The electron beam system 40 may include a projection assembly to adjust the charge accumulated on the sample 208 by irradiating the sample.

[0043] Figure 3 is a schematic diagram of an exemplary single beam electron beam system 41'' according to one embodiment. Figure 3 As shown, in one embodiment, the electron beam system includes a sample holder 207 supported by a motorized stage 209 to hold a sample 208 to be inspected. The electron beam system includes an electron source 201. The electron beam system also includes a gun aperture 122, a beam limiting aperture 125, a focusing lens 126, a device array aperture 135, an objective lens assembly 132, and an electron detector 144. In some embodiments, the objective lens assembly 132 can be a modified swinging objective delayed immersion lens (SORIL), which includes a pole piece 132a, a control electrode 132b, a deflector 132c, and an excitation coil 132d. An aperture for the electron beam to pass through is formed in the control electrode 132b. The control electrode 132b forms a facing surface 72, as described in more detail below.

[0044] During imaging, the electron beam emitted from source 201 can pass through gun aperture 122, beam limiting aperture 125, focusing lens 126, and focused by the modified SORIL lens into a detection spot, and then impinge on the surface of sample 208. The detection spot can be scanned across the surface of sample 208 by deflector 132c or other deflectors in the SORIL lens. Secondary electrons emitted from the sample surface can be collected by electron detector 144 to form an image of the area of ​​interest on sample 208.

[0045] The concentrator and illumination optics of the electron optical system 41 may include or be supplemented by an electromagnetic quadrupole electron lens. Figure 3 As shown, the electron optical system 41 may include a first quadrupole lens 148 and a second quadrupole lens 158. In one embodiment, the quadrupole lens is used to control the electron beam. For example, the first quadrupole lens 148 may be controlled to adjust the beam current, and the second quadrupole lens 158 may be controlled to adjust the beam spot size and beam shape.

[0046] As described in the introduction of the specification, the sample evaluation method can be used to evaluate the degree of undesirable dominant defects in the sample. Such methods may involve scanning the sample (or at least a portion of the sample) at one or more stages of the process of forming a pattern. When manufacturing an IC chip, the manufacturing process may involve a photolithography exposure process. The photolithography exposure process may include irradiating the sample (i.e., substrate) with radiation. For example, a resist (e.g., resist) may be irradiated. The manufacturing process may include an etching process. The etching process may include etching the irradiated portion or the unirradiated portion of the resist.

[0047] The sample can be scanned after the exposure process and before the etching process, the so-called post-development inspection (ADI). The scan can generate information about the degree of unwanted dominant defects in the sample. This information can be used to form an image of the sample. It may not be necessary to form an image. For example, the information (from which an image can be formed) can be used in a subsequent processing step without actually generating an image. The data set formed by scanning the sample after the exposure process and before the etching process can be referred to as an ADI image. Additionally or alternatively, the sample can be scanned after the etching process (and therefore also after the exposure process), the so-called post-etching inspection (AEI). This can generate data about the degree of unwanted dominant defects in the sample. This data can be used to form an image. This data can be referred to as an AEI image. It may not be necessary to generate a visual image. On the contrary, data that can be used to form an image can be used in a subsequent processing step without actually generating an image.

[0048] A method for training a generator model is disclosed. The generator model is configured to process first measurement data 60 to generate prediction data 81. The first measurement data 60 is measured from a sample 208 before an etching process. The etching process is a process of etching the sample 208, such as etching a resist layer of the sample 208. In one embodiment, the first measurement data includes data that can be used to form an ADI image. The prediction data 81 predicts the sample 208 after the etching process. The prediction data 81 can be data that can be used to form an AEI image.

[0049] In one embodiment, the method is used to train a deep learning model (i.e., a generator model) that can map between ADI and AEI SEM images. In one embodiment, the method includes measuring first measurement data 60 from a sample 208. For example, the first measurement data 60 can be measured from the sample 208 by scanning the sample 208 using a charged particle beam inspection system 100 (e.g., SEM). Alternatively, the first measurement data 60 may have been measured from the sample 208. The method can use the first measurement data 60 that has been measured from the sample 208.

[0050] Figure 4 is a diagram showing a data set used in a method according to an embodiment of the present invention. Figure 4 The data set shown on the left corresponds to the first measurement data 60. Figure 4 As shown, in one embodiment, the first measurement data 60 includes one or more ADI images 61-63. Figure 4 In the example shown, ADI images 61-63 show contact hole 65 of sample 208. Figure 4 As shown, contact hole 65 can have a generally circular shape when viewed from above. Sample 208 can have additional or alternative features. For example, sample 208 can have linear features.

[0051] In one embodiment, the method includes generating, using a generator module, predicted data 81 based on the first measurement data 60. The first measurement data 60 and the predicted data 81 may be used to form an image of the sample 208.

[0052] Figure 5 is a diagram showing an ADI image 61 of the first measurement data 60 and a predicted AEI image that can be formed based on the predicted data 81. Figure 5 In one embodiment, the method includes pairing the subset 67-69 of the first measurement data 60 with the subset 87 of the prediction data 81. Figure 5 As shown, subsets 67 - 69 , 87 correspond to locations within an image of the sample 208 that may be formed based on the first measured data 60 and the predicted data 81 .

[0053] pass Figure 5 The data subsets 67-69, 87 shown can form three possible pairs. The first pair can be formed between subset 67 and subset 87. The two subsets 67, 87 correspond to the same position of the sample 208, as shown from Figure 581. The comparison between the ADI image 61 shown and the AEI image formed from the predicted data 81 is shown. In contrast, subsets 68, 69 correspond to different locations than subset 87. A possible second pair may be formed by subset 68 and subset 87. A possible third pair may be formed by subset 69 and subset 87. Each pair consists of a subset of the first measured data 60 and a subset of the predicted data 81. Some pairs (such as the first pair of subsets 67, 87) correspond to the same location of the sample 208. Other pairs (such as the second and third pairs described above) correspond to different locations of the sample 208.

[0054] In one embodiment, the method includes training a generator model based on the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions. Correlation is the correlation between two subsets of a pair of data. For example, in one embodiment, the method includes determining the correlation between a subset 67 from the first measurement data 60 and a subset 87 from the prediction data 81. The correlation can be determined by calculating the cross entropy of the paired data subsets. In one embodiment, the method includes determining the correlation of other paired data subsets corresponding to different positions, such as the second pair of subsets 68, 87 and the third pair of subsets 69, 87. The determination of the correlation can include determining the cross entropy of each pair.

[0055] In one embodiment, the generator model is trained to increase the correlation of pairs corresponding to the same location relative to the correlation of pairs corresponding to different locations. The generator model is trained to generally increase the similarity between the ADI image 61 from the first measurement data 60 and the predicted AEI image formed from the prediction data 81. In practice, the sample 208 is changed during the etching process. Therefore, it is expected that there will be some differences between the ADI image 61 and the predicted AEI image formed from the prediction data 81. However, in general, it is expected that the measured ADI image 61 and the predicted AEI image will show the same physical structure, for example, the same features (e.g., contact holes or linear features) of the sample 208.

[0056] like Figure 5 As shown, the contact hole 65 visible in the ADI image 61 may be similar to the post-etch contact hole 75 in the predicted AEI image. Feature parameters and dimensions may differ between the two images. For example, the critical dimension (CD) may be different. However, the general shape of the physical structure may remain the same. By training the generator model based on the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions, it can be expected that the generator model will be improved.

[0057] By generating predicted data 81, it may not be necessary to scan sample 208 to form an AEI image corresponding to the position of the measured ADI image 61-63. It is expected that embodiments of the present invention will reduce damage to sample 208. It is not necessary to measure the ADI image and the AEI image at the same position of sample 208. This is desirable because inspection of sample 208 would otherwise damage the sample. For example, ADI may have a destructive effect on the resist, and this effect will affect the measurement results after the etching process. It is expected that embodiments of the present invention will improve the accuracy of forming patterns on sample 208.

[0058] The data subsets 67-69, 87 may be referred to as tiles. In one embodiment, the method includes comparing tiles from the ADI image 61 and the generated AEI image. In one embodiment, the method includes setting the ADI tiles corresponding to the AEI tiles as positive examples. For example, the first pair of subsets 67, 87 is considered a positive example because the ADI tile 67 corresponds to the AEI tile 87 because they correspond to the same location of the sample 208. In one embodiment, all other tile pairs are set as negative examples. For example, the second pair of subsets 68, 87 and the third pair of subsets 69, 87 may be set as negative examples. In one embodiment, the method includes reducing or minimizing the cross entropy of the positive examples.

[0059] In one embodiment, the method includes using a discriminator to evaluate the likelihood that the predicted data 81 comes from the same data distribution as the second measured data 70. The second measured data 70 is measured from the sample 208 after the etching process. The second measured data 70 is shown in Figure 4 . The second measurement data 70 can be used to form AEI images 71-73. AEI images 71-73 show views of the contact hole 75 after etching. In one embodiment, the method includes measuring second measurement data 70 from the sample 208. For example, the second measurement data 70 can be measured by scanning the sample 208 after the etching process using a charged particle beam inspection system 100 (e.g., SEM). Alternatively, the second measurement data 70 may have been previously measured.

[0060] In one embodiment, the present invention uses a generative adversarial network (GAN). The discriminator is configured to evaluate the likelihood that the predicted AEI image formed by the predicted data 81 coincides with the true AEI image 71-73 formed by the second measurement data 70. In one embodiment, the method includes training the generator model based on the likelihood evaluated by the discriminator. For example, the generator model can be trained to increase the likelihood evaluated by the discriminator. By taking into account the evaluation made by the discriminator, it can be expected that the generator model will be improved.

[0061] In one embodiment, the output of the generator model is constrained to produce images from the same distribution as the AEI images 71-73 in the training set (or to produce a data set that can be used to form an image). The training set may include the ADI images 61-63 of the first measurement data 60 and the AEI images 71-73 of the second measurement data 70. In one embodiment, the discriminator network is used to take the prediction data 81 as input and output a value, such as a value between 0 and 1. In one embodiment, the larger the value output by the discriminator, the more the image formed by the prediction data 81 looks like it is from the desired distribution, that is, the AEI images 71-73 of the second measurement data 70.

[0062] In one embodiment, the generator model is trained based on a calculated patch-wise contrast loss and a discriminator. One embodiment of the present invention is expected to allow the generator model to be effective over a wider range of possible mappings between ADI images and AEI images. For example, the mapping between the first measurement data 60 and the second measurement data 70 may be irreversible. This means, for example, that there is no one-to-one correspondence between the ADI image 61 from the first measurement data 60 and the corresponding image of the same position of the sample 208 after the etching process. As an example, contact holes of different sizes of the sample 208 before the etching process may have the same size after the etching process. Therefore, it is impossible to map back from the AEI image to determine the size of the contact hole before the etching process. This is an example of an irreversible mapping. One embodiment of the present invention is expected to allow accurate prediction data to be generated without relying on the reversibility of the ADI / AEI mapping.

[0063] In one embodiment, the method includes calculating one or more parameter values ​​of one or more parameters of a feature of the sample 208 based on the predicted data 81 and the second measured data 70. For example, in one embodiment, the parameters may include one or more of the following: CD, local critical dimension uniformity (LCDU), local edge placement error (LEPE), line edge roughness, and line width roughness. LCDU is related to the uniformity of CD values ​​of features such as contact holes or linear features. CD values ​​can be calculated locally and their standard deviations calculated to determine LCDU. LEPE is related to the placement of feature edges. LEPE can be a combination of CD and overlay. For linear features, line edge roughness is related to the uniformity of line edge positions. Line edge roughness can be a measure of how straight the edge of a linear feature is. Line width roughness is related to the uniformity of the width of a linear feature along its length. When training the generator model, one or more key performance indicators, such as these parameters, can be considered.

[0064] In one embodiment, the method includes comparing one or more parameter values ​​calculated from the predicted data 81 with one or more parameter values ​​calculated from the second measured data 70. In one embodiment, the evaluation performed by the discriminator depends on the comparison of the one or more parameter values. The discriminator can receive parameter values ​​so as to force a matching distribution of the parameters. For example, the discriminator can consider the possibility that the CD calculated from the predicted data 81 comes from the distribution of CD values ​​calculated from the second measured data 70. It is contemplated that one embodiment of the present invention improves a generator model of an image applied to a sample 208 during lithography.

[0065] In one embodiment, the smaller the difference between one or more parameter values ​​calculated from the predicted data 81 and one or more parameter values ​​calculated from the second measured data 70, the greater the likelihood assessed by the discriminator. Such a smaller difference may indicate that the predicted data 81 matches the second measured data 70 well. A larger difference may indicate that the predicted data 81 is not very realistic.

[0066] An embodiment of the present invention is expected to reduce the requirement for reversibility of ADI / AEI mapping. In practice, ADI / AEI mapping is unlikely to be completely reversible. An embodiment of the present invention is expected to achieve more accurate edge deviation prediction in a wider application space. An embodiment of the present invention is expected to improve lithography processes.

[0067] In one embodiment, the generator model is trained by minimizing the following loss L:

[0068] L(G,D,X,Y)=L GAN (G,D,X,Y)+αL patch (G,H,X)

[0069] The total loss function L is L GAN and L patch The sum of the two contribution losses. L GAN It is related to how well the predicted AEI image fits the measured AEI image in the same data distribution. GAN refers to the loss function corresponding to the evaluation performed by the discriminator. patch Related to the similarity between the measured ADI image and the predicted AEI image. patch is the loss function of the block-wise contrastive loss. α is a parameter that can be controlled to control the extent to which the generator model is trained based on increasing the similarity between the ADI image and the AEI image and the extent to which the generator model is trained based on improving the similarity between the predicted AEI image and the measured AEI image. A larger value of α means that the generator model is trained more based on increasing the similarity between the measured ADI image and the predicted AEI image.

[0070] The loss function is a function of G, D, X, Y, and H. G refers to the generator model. D refers to the discriminator model. X refers to the first measurement data 60. Y refers to the second measurement data 70. H refers to a network that extracts features from the first measurement data 60 so as to compress the first measurement data 60 into relevant features of the sample 208.

[0071] The contribution loss function can be defined mathematically as follows:

[0072]

[0073] in

[0074]

[0075] E refers to the expected value of the data distribution. The z component is related to the compressed representation of the tile, ie the data subset from the first measured data 60 and the predicted data 81 .

[0076] In one embodiment, the generator model includes an encoder and a decoder. As indicated above, in one embodiment, the method includes determining the cross entropy of the extracted features of the paired data subsets so as to determine the correlation between the paired data subsets. In one embodiment, the method includes encoding the paired data subsets using the encoder of the generator model so that the features can be extracted.

[0077] As shown above, in one embodiment, the block-wise contrast loss is calculated by summing over multiple blocks. In one embodiment, the block-wise contrast loss function is determined by summing over multiple layers of sample 208. Sample 208 may include multiple layers. Each layer may include a set of features.

[0078] In one embodiment, a method of processing first measurement data 60 measured from a sample 208 before an etching process is provided to generate prediction data 81 predicting the sample 208 after the etching process. In one embodiment, the method includes using a generator model to generate the prediction data 81 based on the first measurement data 60. In one embodiment, the generator model has been trained by the above method.

[0079] In one embodiment, the method includes applying a deep learning model (i.e., a generator model) that can map between ADI and AEI SEM images. In one embodiment, a given ADI SEM image is converted to a corresponding predicted AEI SEM image. Alternatively, a generator image may not be required. In one embodiment, the method includes converting given ADI data into corresponding predicted AEI data.

[0080] As described above, the prediction data can be generated without using measure-etch-measure (MEM) data. MEM data is data measured when both ADI and AEI are performed at relevant locations. In an alternative embodiment, a method of applying corrections to MEM data using unpaired image mapping is provided.

[0081] As described above, a block-wise contrastive loss and a discriminator may be used. In one embodiment, these techniques or, alternatively, the cycleGAN technique may be used in conjunction with MEM data.

[0082] In one embodiment, there is a method for training a generator model that processes paired measurement data measured from a sample after an etching process, which sample has been measured before the etching process, to generate hypothetical data of the sample after a simulated etching process if the sample has not been measured before the etching process.

[0083] In one embodiment, the method includes generating paired measurement data. In one embodiment, the method includes inspecting a portion of the sample after the development process and before any etching process (e.g., generating an ADI image). The inspection can be performed by an electron beam. In one embodiment, the charged particle beam inspection system 100 is used to inspect the sample with one or more charged particle beams (e.g., an electron beam). The method may also include inspecting the same portion of the sample after the etching process (e.g., generating an AEI image). The ADI image and the AEI image may correspond to the same position of the sample. The data measured after the etching process is paired measurement data.

[0084] During the process of inspecting the sample before the etching process, the sample may be damaged. In particular, the (multiple) electron beams used to inspect the sample may damage the resist. The damage affects the paired measurement data recorded after the etching process. Therefore, if the sample is not inspected before the etching process, the paired measurement data will be different from the data measured after etching. Of course, there may be some random variation in the measurement data. This random variation can be accounted for by repeating the measurements and taking the average. However, the damage caused to the sample by the inspection performed after development and before etching is a difference that remains even when the repeated measurements are averaged. Due to the damage to the resist caused by ADI, the AEI pattern is affected and not completely true.

[0085] In an alternative embodiment, the paired measurement data has been previously generated. The method can be performed using such previously provided paired measurement data. Therefore, the step of generating the paired measurement data can be omitted.

[0086] In one embodiment, the method includes training a network (such as a cycleGAN or a network using a block-wise contrastive loss and a discriminator) to learn how to map AEI data from MEM data (i.e., paired measurement data subjected to inspection after development) to AEI data that was not corrupted during ADI (e.g., hypothetical data simulated for a sample after an etching process if the sample was not measured before the etching process).

[0087] Alternatively, such a network may have been previously trained. The already trained network may be provided and used to perform the mapping. The training step may be omitted.

[0088] In one embodiment, the method includes applying the trained model to new MEM data. The output (i.e., hypothetical data) can be used for related research. For example, the output can be used to monitor whether the etching process has any changes (e.g., drift) over time. The output can be used to monitor one or more other aspects of the etching process, such as any defects.

[0089] The present invention may be embodied as a training scheme that uses MEM experimental data while reducing / avoiding the effects of ADI damage on AEI data. One embodiment of the present invention is expected to produce hypothetical data that is closer to reality than the actual patterns examined during MEM experiments (i.e., closer to patterns that have not undergone ADI).

[0090] In one embodiment, the method includes generating hypothetical data based on paired measurement data using a generator model. The paired measurement data can be used to form an image of the sample, i.e., an AEI image. The hypothetical data can be used to form an image of the sample, i.e., a hypothetical AEI image. There is no need to actually generate an image. In one embodiment, the data can be retained in a non-image form. Alternatively, an image can be generated and displayed.

[0091] In one embodiment, the method includes using a discriminator to assess the likelihood that hypothesized data is from the same data distribution as true measurement data measured from a sample after an etching process, where the sample has not been measured before the etching process.

[0092] In one embodiment, the method includes generating real measurement data. The real measurement data can be used to form an image of the sample, i.e., an AEI image (or multiple images of one or more locations of the sample). Alternatively, the real measurement data may have been generated before. The step of generating the real measurement data can be omitted.

[0093] In one embodiment, the true measurement data corresponds to one or more locations of the sample that are physically similar (e.g., have similar features) to the locations corresponding to the paired measurement data. For example, if the paired measurement data can be used to form an image of a contact hole, then desirably, the true measurement data corresponds to the locations of the contact hole having a similar location.

[0094] In one embodiment, the true measurement data corresponds to one or more locations of the sample that are located near the location corresponding to the paired measurement data. For example, desirably, the true measurement data corresponds to a location adjacent to the location corresponding to the paired measurement data. This can help explain any systematic variations in the data measured at different locations of the sample (which can be referred to as a fingerprint). For example, there may be systematic effects related to the distance of the location from the center of the sample. In one embodiment, the true measurement data corresponds to one or more locations that are at a distance from the center of the sample that is similar to the distance of the paired measurement data from the center of the sample.

[0095] In one embodiment, the true measurement data corresponds to one or more locations of the sample that are surrounded by a pattern density similar to the pattern density corresponding to the location of the paired measurement data. For example, the location of the paired measurement data may be surrounded by contact holes in a regular hexagonal pattern. Desirably, the true measurement data similarly corresponds to one or more locations of the sample that are surrounded by regularly spaced contact holes.

[0096] In one embodiment, the method includes training a generator model based on the likelihood assessed by the discriminator. In one embodiment, the generator model is trained to increase the likelihood assessed by the discriminator. This helps to increase the closeness between the generated hypothesis data and the data of the sample position without ADI4 measurement.

[0097] As described above, in one embodiment, a contrastive loss may be used. In one embodiment, the method includes training a generator model based on a function indicating a level of correlation between paired measurement data and hypothesis data. In one embodiment, the method includes pairing a subset of paired measurement data with a subset of hypothesis data, the subset corresponding to locations within a sample image that can be formed from the paired measurement data and hypothesis data. The function (indicating a level of correlation between paired measurement data and hypothesis data) is a correlation of pairs corresponding to the same location relative to a correlation of pairs corresponding to different locations, the correlation being the correlation between the paired data subsets.

[0098] In one embodiment, the generator model is trained to increase the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions. In one embodiment, the method includes determining the cross entropy of the extracted features of the paired data subsets to determine the correlation between the paired data subsets. In one embodiment, the method includes encoding the paired data subsets using an encoder of the generator model so that the features can be extracted.

[0099] Alternatively, as described above, in one embodiment, cycleGAN can be used. In one embodiment, the method includes using an inverse generator model to generate simulated paired data based on the hypothesis data, wherein the simulated paired data can be used to form an image of the sample. The function (indicating the level of correlation between the paired measurement data and the hypothesis data) is the similarity between the paired measurement data and the simulated paired data. In one embodiment, the generator model and the inverse generator model are trained to increase the similarity between the paired measurement data and the simulated paired data.

[0100] In one embodiment, a processor device is provided, including a processor configured to perform the above method. For example, the processor can be configured to perform the method of training the generator model. Additionally or alternatively, the processor can be configured to perform the method of applying the generator model.

[0101] In some embodiments, an evaluation method is provided that includes generating a sample map according to any of the above methods. The evaluation method includes inspecting the sample 208 using the generated sample map to locate one or more features of interest. The evaluation method may include evaluating the extent to which one or more features of interest contain defects. This may be achieved by comparing an image of the feature of interest with a reference image elsewhere on the sample, on other samples, or in a database.

[0102] In some embodiments, the inspection is performed by directing one or more charged particle beams (e.g., electrons) onto the sample 208 and detecting one or more charged particles (e.g., electrons) emitted from the sample 208. The inspection may be performed using the method described above with reference to Figure 1-5 Any of the electron optical arrangements described or any other arrangement suitable for examining a sample using a charged particle beam. The examination may also be performed using other techniques, such as optical techniques based on electromagnetic radiation.

[0103] References to upper and lower, upward and downward, above and below, etc. should be understood to refer to directions parallel to the upstream and downstream directions (typically but not always perpendicular) of the electron beam or beams impinging on the sample 208. Thus, references to upstream and downstream are intended to refer to the direction of the beam path independent of any existing gravitational field.

[0104] The embodiments described herein may take the form of a series of aperture arrays or electron optical elements arranged in an array along a beam or multi-beam path. Such electron optical elements may be electrostatic. In one embodiment, all electron optical elements (e.g., the last electron optical element in the sub-beam path from the beam-limiting aperture array to the sample before) may be electrostatic and / or may be in the form of an aperture array or a plate array. In some arrangements, one or more electron optical elements are manufactured as a micro-electromechanical system (MEMS) (i.e., using MEMS manufacturing technology). The electron optical element may have a magnetic element and an electrostatic element. For example, a composite array lens may have a macroscopic magnetic lens that surrounds a multi-beam path, wherein an upper pole plate and a lower pole plate are provided in the magnetic lens and arranged along the multi-beam path. The pole plate may be an aperture array for the beam path of the multi-beam. Electrodes may be present above, below, or between the pole plates to control and optimize the electromagnetic field of the composite lens array.

[0105] An evaluation tool or evaluation system according to the present disclosure may include a device for qualitatively evaluating a sample (e.g., pass / fail), a device for quantitatively measuring a sample (e.g., the size of a feature), or a device for generating an image of a sample map. Examples of evaluation tools or systems are inspection tools (e.g., for identifying defects), review tools (e.g., for classifying defects), and metrology tools, or a tool capable of performing any combination of evaluation functions associated with an inspection tool, review tool, or metrology tool (e.g., a metrology inspection tool).

[0106] Reference to a system or element of a component or component that can be controlled to manipulate a charged particle beam in a certain manner includes configuring a controller or control system or control unit to control the component to manipulate the charged particle beam in the manner described, and optionally using other controllers or devices (e.g., voltage sources) to control the component to manipulate the charged particle beam in this manner. For example, a voltage source can be electrically connected to one or more components to apply a potential to the component under the control of a controller or control system or control unit, such as applying a potential to electrodes that control lens array 250 and objective lens array 241. Using one or more controllers, control systems or control units to control the actuation of the component, an actuable component (such as a stage) can be controllable to actuate and thus move relative to another component (such as a beam path).

[0107] The function provided by the controller or control system or control unit can be implemented by a computer. Any suitable combination of components can be used to provide the required functions, including, for example, CPU, RAM, SSD, mainboard, network connection, firmware, software and / or other components known in the art that allow the required computing operations to be performed. The required computing operations can be defined by one or more computer programs. One or more computer programs can be provided in the form of a medium (optionally a non-transient medium) storing computer-readable instructions. When a computer reads a computer-readable instruction, the computer executes the required method steps. The computer can be composed of an independent unit or a distributed computing system, which has a plurality of different computers interconnected via a network.

[0108] The terms "beamlet" and "beamlet" are used interchangeably herein and are both to be understood to include any radiation beam derived from a parent radiation beam by dividing or splitting the parent radiation beam. The term "manipulator" is used to include any element that affects a beamlet or beamlet path, such as a lens or deflector. References to elements aligned along a beam path or beamlet path are to be understood as positioning the corresponding element along the beam path or beamlet path. References to optics are to be understood as electron optics.

[0109] The method of the present invention can be performed by a computer system including one or more computers. The computer for implementing the present invention may include one or more processors, including a general-purpose CPU, a graphics processing unit (GPU), a field programmable gate array (FPGA), an application-specific integrated circuit (ASIC) or other special-purpose processors. As mentioned above, in some cases, a specific type of processor can provide advantages in reducing costs and / or increasing processing speed, and the method of the present invention can adapt to the use of a specific processor type. Some steps of the method of the present invention involve parallel computing, which tends to be implemented on a processor (such as a GPU) capable of parallel computing.

[0110] As used herein, the term "image" is intended to refer to any array of values ​​where each value is associated with a sample at a location and the arrangement of the values ​​in the array corresponds to the spatial arrangement of the sampled locations. An image may comprise a single layer or multiple layers. In the case of a multilayer image, each layer (which may also be referred to as a channel) represents a different sample at a location. The term "pixel" is intended to refer to a single value of an array, or, in the case of a multilayer image, a set of values ​​corresponding to a single location.

[0111] Embodiments of the present disclosure are defined in the following numbered clauses.

[0112] 1. A method for training a generator model, the generator model processing first measurement data measured from a sample before an etching process to generate prediction data of the sample after the etching process, the method comprising:

[0113] generating the predicted data based on the first measured data using the generator model, wherein the first measured data and the predicted data can be used to form an image of the sample;

[0114] pairing a subset of the first measured data with a subset of the predicted data, the subsets corresponding to locations within the image of the sample that can be formed from the measured data and the predicted data;

[0115] evaluating, using a discriminator, a likelihood that the predicted data is from the same data distribution as second measurement data measured from a sample at a different location after an etching process; and

[0116] The generator model is trained based on the following:

[0117] the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions, said correlation being the correlation between paired subsets of said data, and

[0118] The likelihood evaluated by the discriminator.

[0119] 2. A method according to clause 1, wherein the generator model is trained so as to increase the correlation of the pairs corresponding to the same position relative to the correlation of the pairs corresponding to different positions.

[0120] 3. A method according to clause 1 or 2, wherein the generator model is trained to increase the likelihood evaluated by the discriminator.

[0121] 4. A method according to any of the preceding clauses, comprising:

[0122] Calculating one or more parameter values ​​of one or more parameters of a characteristic of the sample based on the predicted data and the second measured data; and

[0123] comparing the one or more parameter values ​​calculated from the predicted data with the one or more parameter values ​​calculated from the second measured data,

[0124] wherein the evaluation performed by the discriminator depends on the comparison of the one or more parameter values.

[0125] 5. The method of clause 4, wherein the parameters include one or more of: critical dimension, local critical dimension uniformity, local edge placement error, line edge roughness, and line width roughness.

[0126] 6. A method according to clause 4 or 5, wherein the smaller the difference between the one or more parameter values ​​calculated based on the predicted data and the one or more parameter values ​​calculated based on the second measured data, the greater the possibility evaluated by the discriminator.

[0127] 7. A method according to any of the preceding clauses, wherein the second measurement data corresponds to a different location than the first measurement data.

[0128] 8. A method according to any of the preceding clauses, wherein the mapping between the first measurement data and the second measurement data is irreversible.

[0129] 9. A method according to any of the preceding clauses, wherein the generator model comprises an encoder and a decoder.

[0130] 10. The method of clause 9, comprising determining a cross entropy of extracted features of paired subsets of the data in order to determine the correlation between paired subsets of the data.

[0131] 11. The method of clause 10, comprising encoding a subset of the paired data with the encoder so that the features can be extracted.

[0132] 12. The method according to any of the preceding clauses, wherein the first measurement data is measured from the sample before an etching process and after a photolithography exposure process.

[0133] 13. A method of processing first measurement data measured from a sample before an etching process to generate prediction data for predicting the sample after an etching process, the method comprising:

[0134] The prediction data is generated based on the first measurement data using a generator model, wherein the generator model has been trained by a method as described in any of the preceding clauses.

[0135] 14. A processing device comprising:

[0136] A processor configured to execute the method described in any one of the preceding clauses.

[0137] 15. A computer program comprising instructions configured to control a processor to perform the method of any of the preceding clauses.

[0138] 16. A generator model training apparatus for training a generator model, the generator model processing first measurement data measured from a sample before an etching process to generate prediction data of the sample after the etching process, the apparatus comprising:

[0139] The processor is configured as:

[0140] The predicted data is generated based on the first measured data using the generator model, wherein the

[0141] The first measurement data and the predicted data can be used to form an image of the sample;

[0142] pairing a subset of the first measured data with a subset of the predicted data, the subsets corresponding to locations within the image of the sample that can be formed from the first measured data and the predicted data;

[0144] Use the discriminator to evaluate the predicted data from samples at different locations after and after the etching process.

[0145] likelihood of a data distribution being the same as the second measurement data; and

[0146] The generator model is trained according to the following:

[0147] The correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions, the correlation between the paired subsets of the data, and the likelihood evaluated by the discriminator.

[0148] 17. A generator model training apparatus according to clause 16, wherein the processor is configured to train the generator model so as to increase the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions.

[0149] 18. A generator model training apparatus according to clause 16 or 17, wherein the processor is configured to train the generator model so as to increase the likelihood evaluated by the discriminator.

[0150] 19. A generator model training apparatus according to any one of clauses 16 to 18, wherein the processor is configured to:

[0151] Calculate one or more parameter values ​​of one or more characteristic parameters of the sample according to the predicted data and the second measured data; and

[0152] comparing the one or more parameter values ​​calculated from the predicted data with the one or more parameter values ​​calculated from the second measured data,

[0153] wherein said evaluation by said discriminator is dependent upon said comparison of said one or more parameter values.

[0154] 20. A generator model training apparatus according to clause 19, wherein the parameters include one or more of the following: critical dimension, local critical dimension uniformity, local edge placement error, line edge roughness, and line width roughness.

[0155] 21. A generator model training device according to clause 19 or 20, wherein the smaller the difference between the one or more parameter values ​​calculated based on the predicted data and the one or more parameter values ​​calculated based on the second measured data, the greater the possibility evaluated by the discriminator.

[0156] 22. A generator model training apparatus according to any one of clauses 16 to 21, wherein the second measurement data corresponds to a different location from the first measurement data.

[0157] 23. A generator model training device according to any one of clauses 16 to 22, wherein the mapping between the first measurement data and the second measurement data is irreversible.

[0158] 24. A generator model training apparatus according to any one of clauses 16 to 23, wherein the generator model comprises an encoder and a decoder.

[0159] 25. A generator model training device according to clause 24, wherein the processor is configured to determine the cross entropy of the extracted features of the paired subsets of the data in order to determine the correlation between the paired subsets of the data.

[0160] 26. A generator model training device according to clause 25, wherein the processor is configured to encode a subset of the paired data using the encoder so as to enable feature extraction.

[0161] 27. A generator model training apparatus according to any one of clauses 16 to 26, wherein the first measurement data is measured from the sample before an etching process and after a photolithography exposure process.

[0162] 28. A prediction data generating device for processing first measurement data measured from a sample before an etching process to generate prediction data of the sample after the etching process, the device comprising:

[0163] A processor configured to:

[0164] The prediction data is generated based on the first measurement data using a generator model, wherein the generator model has been trained by the method of any one of clauses 1-12.

[0165] 29. A computer readable medium storing instructions configured to control a processor to train a generator model, the generator model processing first measurement data measured from a sample before an etching process to generate prediction data of the sample after the etching process, the computer readable medium storing instructions configured to control the processor:

[0166] generating the predicted data based on the first measured data using the generator model, wherein the first measured data and the predicted data can be used to form an image of the sample;

[0167] pairing a subset of the first measured data with a subset of the predicted data, the subsets corresponding to locations within the image of the sample that can be formed from the measured data and the predicted data;

[0168] using a discriminator to assess the likelihood that the predicted data is from the same data distribution as second measurement data measured from a sample at a different location after an etching process; and

[0169] The generator model is trained according to the following:

[0170] the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions, said correlation being said correlation between paired subsets of said data, and

[0171] The likelihood evaluated by the discriminator.

[0172] 30. A computer-readable medium according to clause 29, storing instructions configured to control the processor to train the generator model so as to increase the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions.

[0173] 31. A computer-readable medium according to clause 29 or 30, storing instructions configured to control the processor to train the generator model so as to increase the likelihood evaluated by the discriminator.

[0174] 32. A computer readable medium according to any of clauses 29 to 31, storing instructions configured to control the processor to:

[0175] Calculate one or more parameter values ​​of one or more characteristic parameters of the sample according to the predicted data and the second measured data; and

[0176] comparing the one or more parameter values ​​calculated from the predicted data with the one or more parameter values ​​calculated from the second measured data,

[0177] wherein said evaluation by said discriminator is dependent upon said comparison of said one or more parameter values.

[0178] 33. The computer-readable medium of clause 32, wherein the parameters include one or more of: critical dimension, local critical dimension uniformity, local edge placement error, line edge roughness, and line width roughness.

[0179] 34. A computer-readable medium according to clause 32 or 33, wherein the smaller the difference between the one or more parameter values ​​calculated based on the predicted data and the one or more parameter values ​​calculated based on the second measured data, the greater the likelihood assessed by the discriminator.

[0180] 35. The computer-readable medium of any of clauses 29 to 34, wherein the second measurement data corresponds to a different location than the first measurement data.

[0181] 36. The computer-readable medium of any of clauses 29 to 35, wherein the mapping between the first measurement data and the second measurement data is irreversible.

[0182] 37. A computer-readable medium according to any of clauses 29 to 36, wherein the generator model comprises an encoder and a decoder.

[0183] 38. A computer-readable medium according to clause 37, storing instructions configured to control the processor to determine the cross entropy of the extracted features of the paired subsets of the data in order to determine the correlation between the paired subsets of the data.

[0184] 39. A computer-readable medium according to clause 38, storing instructions configured to control the processor to encode a subset of the paired data using the encoder so as to enable extraction of the feature.

[0185] 40. The computer-readable medium of any one of clauses 29 to 39, wherein the first measurement data is measured from the sample before an etching process and after a photolithography exposure process.

[0186] 41. A method of training a generator model, the generator model processing paired measurement data measured from a sample after an etching process, the sample having been measured before the etching process, to generate hypothetical data simulating the sample after the etching process if the sample had not been measured before the etching process, the method comprising:

[0187] generating the hypothesis data based on the paired measurement data using the generator model, wherein the paired measurement data and the hypothesis data can be used to form an image of the sample;

[0188] using a discriminator to assess the likelihood that the hypothesized data are from the same data distribution as true measured data measured from a sample after an etching process, the sample having not been measured before the etching process; and

[0189] The generator model is trained based on:

[0190] a function indicating the level of correlation between the paired measurement data and the hypothesis data, and

[0191] The likelihood evaluated by the discriminator.

[0192] 42. A method according to clause 41, comprising:

[0193] pairing a subset of the paired measurement data with a subset of the hypothesis data, the subset corresponding to locations within an image of the sample that can be formed from the paired measurement data and the hypothesis data;

[0194] Wherein the function is the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions, the correlation is the correlation between paired subsets of the data.

[0195] 43. A method according to clause 42, wherein the generator model is trained to increase the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions.

[0196] 44. A method according to clause 42 or 43, comprising determining the cross entropy of the extracted features of the paired subsets of the data in order to determine the correlation between the paired subsets of the data.

[0197] 45. The method of clause 44, comprising encoding a subset of the paired data with an encoder of the generator model so that the features can be extracted.

[0198] 46. ​​A method according to clause 41, comprising:

[0199] generating simulated paired data based on the hypothetical data using an inverse generator model, wherein the simulated paired data can be used to form an image of the sample;

[0200] The function is the similarity between the paired measurement data and the simulated paired data.

[0201] 47. A method according to clause 46, wherein the generator model and the inverse generator model are trained to increase the similarity between the paired measurement data and the simulated paired data.

[0202] 48. A method according to any one of clauses 41 to 47, wherein the generator model is trained to increase the likelihood assessed by the discriminator.

[0203] 49. A method according to any of clauses 41 to 48, wherein the paired measurement data is measured from a sample after an etching process, the sample having been previously measured before the etching process and after a lithography exposure process.

[0204] 50. A method of processing paired measurement data measured from a sample after an etching process, the sample having been measured before the etching process, to generate hypothetical data simulating the sample after the etching process if the sample had not been measured before the etching process, the method comprising:

[0205] The hypothesis data is generated based on the paired measurement data using a generator model, wherein the generator model has been trained by the method of any of clauses 41-49.

[0206] 51. A processing device comprising:

[0207] A processor configured to perform the method of any of clauses 41 to 50.

[0208] 52. A computer program comprising instructions configured to control a processor to perform the method of any of clauses 41-51.

[0209] 53. A generator model training apparatus for training a generator model, the generator model processing paired measurement data measured from a sample after an etching process, the sample having been measured before the etching process, to generate hypothetical data simulating the sample after the etching process if the sample had not been measured before the etching process, the apparatus comprising:

[0210] A processor configured to:

[0211] The hypothesis data is generated based on the paired measurement data using the generator model, wherein the

[0212] Paired measurement data and the hypothesized data can be used to form an image of the sample;

[0213] evaluating, using a discriminator, a likelihood that the hypothesized data is from the same data distribution as true measured data measured from a sample after an etching process, the sample having not been measured before the etching process;

[0214] as well as

[0215] The generator model is trained based on:

[0216] a function indicating a level of correlation between the paired measurement data and the hypothesis data, and the likelihood evaluated by the discriminator.

[0217] 54. The generator model training apparatus of clause 53, wherein the processor is configured to:

[0218] pairing a subset of the paired measurement data with a subset of the hypothesis data, the subset corresponding to locations within the image of the sample that can be formed from the paired measurement data and the hypothesis data;

[0219] Wherein the function is the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions, the correlation is the correlation between paired subsets of the data.

[0220] 55. A generator model training apparatus according to clause 54, wherein the processor is configured to train the generator model so as to increase the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions.

[0221] 56. A generator model training device according to clause 54 or 55, wherein the processor is configured to determine the cross entropy of the extracted features of the paired subsets of the data in order to determine the correlation between the paired subsets of the data.

[0222] 57. A generator model training device according to clause 56, wherein the processor is configured to encode a subset of the paired data using an encoder of the generator model so as to be able to extract the features.

[0223] 58. The generator model training apparatus of clause 53, wherein the processor is configured to:

[0224] generating simulated paired data based on the hypothetical data using an inverse generator model, wherein the simulated paired data can be used to form an image of the sample;

[0225] The function is the similarity between the paired measurement data and the simulated paired data.

[0226] 59. A generator model training device according to clause 58, wherein the generator model and the inverse generator model are trained to increase the similarity between the paired measurement data and the simulated paired data.

[0227] 60. A generator model training apparatus according to any one of clauses 53 to 59, wherein the processor is configured to train the generator model so as to increase the likelihood evaluated by the discriminator.

[0228] 61. A generator model training apparatus according to any one of clauses 53 to 60, wherein the paired measurement data is measured from a sample after an etching process, the sample having been previously measured before the etching process and after a photolithography exposure process.

[0229] 62. A hypothetical data generating apparatus for processing paired measurement data measured from a sample after an etching process, the sample having been measured before the etching process, to generate hypothetical data simulating the sample after the etching process if the sample had not been measured before the etching process, the apparatus comprising:

[0230] The processor is configured as:

[0231] The hypothesis data is generated based on the paired measurement data using a generator model, wherein the generator model has been trained by the method of any of clauses 41-49.

[0232] 63. A computer readable medium storing instructions configured to control a processor to train a generator model, the generator model processing paired measurement data measured from a sample after an etching process, the sample having been measured before the etching process, to generate hypothetical data simulating the sample after the etching process if the sample had not been measured before the etching process, the computer readable medium storing instructions configured to control the processor to:

[0233] generating the hypothesis data based on the paired measurement data using the generator model, wherein the paired measurement data and the hypothesis data can be used to form an image of the sample;

[0234] using a discriminator to assess the likelihood that the hypothesized data are from the same data distribution as true measured data measured from a sample after an etching process, the sample having not been measured before the etching process; and

[0235] The generator model is trained based on:

[0236] a function indicating the level of correlation between the paired measurement data and the hypothesis data, and

[0237] The likelihood evaluated by the discriminator.

[0238] 64. The computer-readable medium of clause 63, storing instructions configured to control the processor to pair a subset of the paired measurement data with a subset of the hypothesis data, the subset corresponding to locations within the image of the sample that can be formed from the paired measurement data and the hypothesis data;

[0239] Wherein the function is the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions, the correlation is the correlation between paired subsets of the data.

[0240] 65. A computer-readable medium according to clause 64, storing instructions configured to control the processor to train the generator model so as to increase the correlation of pairs corresponding to the same position relative to the correlation of pairs corresponding to different positions.

[0241] 66. A computer-readable medium according to clause 64 or 65, storing instructions configured to control the processor to determine the cross entropy of extracted features of the paired subsets of the data in order to determine the correlation between the paired subsets of the data.

[0242] 67. A computer-readable medium according to clause 66, storing instructions configured to control the processor to encode a subset of the paired data using an encoder of the generator model so as to extract the features.

[0243] 68. The computer-readable medium of clause 63, storing instructions configured to control the processor to generate simulated paired data based on the hypothetical data using an inverse generator model, wherein the simulated paired data can be used to form an image of the sample;

[0244] The function is the similarity between the paired measurement data and the simulated paired data.

[0245] 69. The computer-readable medium of clause 68, wherein the generator model and the inverse generator model are trained to increase the similarity between the paired measurement data and the simulated paired data.

[0246] 70. A computer-readable medium according to any one of clauses 63 to 69, storing instructions configured to control the processor to train the generator model so as to increase the likelihood evaluated by the discriminator.

[0247] 71. The computer-readable medium of any of clauses 63 to 70, wherein the paired measurement data is measured from a sample after an etching process, the sample having been previously measured before the etching process and after a photolithography exposure process.

[0248] The computer for implementing the present invention may be physical or virtual. The computer for implementing the present invention may be a server, a client, or a workstation. Multiple computers for implementing the present invention may be distributed and interconnected via a local area network (LAN) or a wide area network (WAN). The results of the method of the present invention may be displayed to a user or stored in any suitable storage medium. The present invention may be embodied in a non-transient computer-readable storage medium that stores instructions for executing the method of the present invention. The present invention may be embodied in a computer system that includes one or more processors and a memory or stores instructions for executing the method of the present invention.

[0249] Although the present invention has been described in conjunction with various embodiments, other embodiments of the present invention will be apparent to those skilled in the art from the specification and practice of the invention disclosed herein. The specification and examples should be considered as exemplary only, and the true scope and spirit of the present invention are indicated by the following claims.

Claims

1. A method for training a generator model, the generator model processing first measurement data measured from a sample before an etching process to generate prediction data of the sample after the etching process, the method comprising: generating the predicted data based on the first measured data using the generator model, wherein the first measured data and the predicted data can be used to form an image of the sample; pairing a subset of the first measured data with a subset of the predicted data, the subsets corresponding to locations within the image of the sample that can be formed from the measured data and the predicted data; evaluating, using a discriminator, a likelihood that the predicted data is from the same data distribution as second measurement data measured from a sample at a different location after an etching process; and The generator model is trained based on the following: a correlation of the pairs corresponding to the same position relative to a correlation of the pairs corresponding to different positions, the correlation being the correlation between the paired subsets of the data, and The likelihood evaluated by the discriminator.

2. The method of claim 1 , wherein the generator model is trained so as to increase the relevance of the pairs corresponding to the same position relative to the relevance of the pairs corresponding to different positions.

3. The method of claim 1 , wherein the generator model is trained to increase the likelihood evaluated by the discriminator.

4. The method according to claim 1, comprising: Calculate one or more parameter values ​​of one or more parameters of the characteristic of the sample based on the predicted data and the second measured data; as well as comparing the one or more parameter values ​​calculated from the predicted data with the one or more parameter values ​​calculated from the second measured data, wherein the evaluation performed by the discriminator depends on the comparison of the one or more parameter values. 5 . The method of claim 4 , wherein the parameters include one or more of: critical dimension, local critical dimension uniformity, local edge placement error, line edge roughness, and line width roughness.

6. A method according to claim 4, wherein the smaller the difference between the one or more parameter values ​​calculated based on the predicted data and the one or more parameter values ​​calculated based on the second measured data, the greater the possibility evaluated by the discriminator. The method of claim 1 , wherein the second measurement data corresponds to a different location than the first measurement data. The method according to claim 1 , wherein the mapping between the first measurement data and the second measurement data is irreversible.

9. The method of claim 1, wherein the generator model comprises an encoder and a decoder.

10. The method of claim 9, comprising determining the cross entropy of the extracted features of the paired subsets of the data in order to determine the correlation between the paired subsets of the data.

11. The method of claim 10, comprising encoding a subset of the paired data with the encoder so that the features can be extracted. 12 . The method of claim 1 , wherein the first measurement data is measured from the sample before an etching process and after a photolithography exposure process.

13. A method of processing first measurement data measured from a sample before an etching process to generate prediction data for predicting the sample after an etching process, the method comprising: The prediction data is generated based on the first measurement data using a generator model, wherein the generator model has been trained by the method of any of the preceding claims.

14. A processing device comprising: A processor configured to execute the method according to any one of the preceding claims.

15. A computer program comprising instructions configured to control a processor to perform the method according to any one of the preceding claims.

Images