Cross-talk cancellation in multi-beam charged particle inspection

Abstract

An improved apparatus and method for enhancing images is disclosed, and more specifically, an apparatus and method for enhancing images by crosstalk cancellation in multi-beam inspection is disclosed. A method for enhancing images includes obtaining a first image signal of a plurality of image signals from a detector of a multi-beam inspection system. The first image signal corresponds to detection signals from a first region of the detector on which electrons of a first secondary electron beam and electrons of a second secondary electron beam are incident. The method further includes reducing, from the first image signal, crosstalk contamination originating from the second secondary electron beam using a relationship between the first image signal and beam intensities associated with the first secondary electron beam and the second secondary electron beam. The method further includes generating, after the reducing, a first image corresponding to the first secondary electron beam.

Description

[0001] Cross Reference to Related Applications

[0002] This application claims priority to U.S. Application 62 / 886,882, filed August 14, 2019, the entire contents of which are incorporated herein by reference. TECHNICAL FIELD

[0003] Embodiments provided herein relate to image enhancement techniques, and more specifically to crosstalk cancellation in multi charged-particle beam inspection. BACKGROUND

[0004] In the manufacturing process of integrated circuits (ICs), unfinished or finished circuit components are inspected to ensure that they are manufactured according to design and are free of defects. Inspection systems such as scanning electron microscopes (SEMs) that utilize optical microscopes or charged-particle (e.g., electron) beam microscopes can be employed. As the physical size of IC components continues to shrink, the accuracy of defect detection and yield become more important. Although multiple electron beams can be used to increase throughput, limitations in the fidelity of the imaging signals received by the charged-particle detector can limit the imaging resolution required for reliable defect detection and analysis, rendering the inspection tool inadequate for its intended purpose. SUMMARY

[0005] Embodiments provided herein disclose a charged-particle beam inspection apparatus, and more specifically, an inspection apparatus that uses multiple charged-particle beams.

[0006] In some embodiments, a method for enhancing an image includes obtaining a first image signal of a plurality of image signals from a detector of a multi-beam inspection system. The first image signal corresponds to detection signals from a first region of the detector, where electrons of a first beam of secondary electrons and electrons of a second beam of secondary electrons are incident on the first region. The method further includes reducing, from the first image signal, crosstalk contamination originating from the second beam of secondary electrons using a relationship between the first image signal and beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons. The method further includes reducing a post-generation of a first image corresponding to the first beam of secondary electrons.

[0007] In some embodiments, an image enhancement device includes a memory storing a set of instructions and at least one processor configured to execute the set of instructions to cause the device to perform obtaining a first image signal of a plurality of image signals from a detector of a multi-beam inspection system. The first image signal corresponds to detection signals from a first region of the detector on which electrons of a first beam of secondary electrons and electrons of a second beam of secondary electrons are incident. The at least one processor is further configured to execute the set of instructions to cause the device to also reduce, from the first image signal, cross-talk contamination originating from the second beam of secondary electrons using a relationship between the first image signal and beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons. The at least one processor is further configured to execute the set of instructions to cause the device to also perform generating, after the reducing, a first image corresponding to the first beam of secondary electrons.

[0008] In some embodiments, a non-transitory computer-readable medium stores a set of instructions executable by at least one processor of a computing device to cause performance of a method for enhancing an image. The method includes obtaining a first image signal of a plurality of image signals from a detector of a multi-beam inspection system. The first image signal corresponds to detection signals from a first region of the detector on which electrons of a first beam of secondary electrons and electrons of a second beam of secondary electrons are incident. The method also includes reducing, from the first image signal, cross-talk contamination originating from the second beam of secondary electrons using a relationship between the first image signal and beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons. The method also includes generating, after the reducing, a first image corresponding to the first beam of secondary electrons.

[0009] In some embodiments, a method for reducing cross-talk contamination in a multi-beam inspection system includes obtaining, by the multi-beam inspection system, a first image and a second image of a first region and a second region, respectively, on a sample. The first image is generated based on first detection signals from a first detection region of the multi-beam inspection system, and the second image is generated based on second detection signals from a second detection region of the multi-beam inspection system. The method also includes determining, by using a first reference image corresponding to the first region, a primary pattern of the first image originating from a first beam of secondary electrons. The method also includes determining, by using a second reference image corresponding to the second region, whether the first image includes an artifact pattern originating from a second beam of secondary electrons. The method also includes determining a relationship between the first detection signals and beam intensities of the first beam of secondary electrons and the second beam of secondary electrons. The method also includes reducing, based on the determined relationship, cross-talk contamination from third detection signals from the first detection region.

[0010] In some embodiments, a method for enhancing an image includes acquiring a first image from a detector of a multi-beam inspection system. The first image is generated based on first detection signals from a first region of the detector, where the first region is on which electrons of a first beam of secondary electrons and electrons of a second beam of secondary electrons are incident. The method further includes reducing, from the first image signal, crosstalk contamination originating from the second beam of secondary electrons based on a relationship between the first image signal and beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons.

[0011] Other advantages of embodiments of the present disclosure will become apparent from the following description, taken in connection with the accompanying drawings, wherein, by way of illustration and example, certain embodiments of the present disclosure are BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 is a schematic diagram illustrating an example electron beam inspection (EBI) system consistent with embodiments of the present disclosure.

[0013] Figure 2 is a schematic diagram illustrating a portion of an example electron beam inspection system that can be Figure 1 consistent with embodiments of the present disclosure.

[0014] Figures 3A-3C are examples of an electron beam image with a first pattern, an electron beam image with a second pattern, and an electron beam image with crosstalk contamination, respectively.

[0015] Figure 4 is a block diagram of an example image enhancement apparatus consistent with embodiments of the present disclosure.

[0016] Figure 5 is a block diagram of an example transform function provider of an example image enhancement apparatus consistent with embodiments of the present disclosure. Figure 1

[0017] Figure 6 is an example of extracting an image patch from an electron beam image consistent with embodiments of the present disclosure.

[0018] Figure 7 is a process flow diagram representing an example method for enhancing an image in a multi-beam inspection system consistent with embodiments of the present disclosure.

[0019] Figure 8 is a process flow diagram representing an example method for generating a transform function that reduces crosstalk contamination from an electron beam detection signal consistent with embodiments of the present disclosure. DETAILED DESCRIPTION

[0020] ​Reference will now be made in detail to the example embodiments, examples of which are illustrated in the accompanying drawings. The following description refers to the accompanying drawings in which the same numbers represent the same or similar elements between the several drawings. The implementations set forth in the following description of example embodiments do not represent all of the implementations consistent with the subject technology. Instead, they are merely examples of apparatuses and methods consistent with aspects related to the disclosure as described in the appended claims. For example, although some embodiments are described in the context of using an electron beam, the present disclosure is not limited thereto. Other types of charged particle beams can be similarly applied. Moreover, other imaging systems can be used, such as optical imaging, light detection, X-ray detection, etc.

[0021] Electronic devices are made up of circuits formed on a piece of silicon called a substrate. Many circuits can be formed together on the same piece of silicon, called an integrated circuit or IC. The size of these circuits has been significantly reduced, so more circuits can be installed on the substrate. For example, an IC chip in a smartphone can be as small as a thumbnail Figure 1 but can include more than 2 billion transistors, each smaller than 1 / 1000th the size of a human hair.

[0022] Manufacturing these tiny ICs is a complex, time-consuming, and expensive process, often involving hundreds of individual steps. Even a mistake in one step can cause a defective finished IC, making it useless. Therefore, one goal of the manufacturing process is to avoid such defects to maximize the number of functional ICs manufactured in the process, i.e., to improve the overall yield of the process.

[0023] One component of improving yield is to monitor the chip manufacturing process to ensure that it is producing a sufficient number of functional integrated circuits. One way to monitor the process is to inspect the chip circuit structures at various stages of their formation. The inspection can be performed using a scanning electron microscope (SEM). The SEM can be used to image, i.e., take a “picture” of these tiny structures. The image can be used to determine whether the structure was formed correctly and in the correct location. If the structure is defective, the process can be adjusted so that the defect is less likely to occur again.

[0024] While multi-beam charged particle beam imaging systems, such as multi-beam SEMs, can be used to increase wafer inspection throughput, the imaging resolution of multi-beam SEMs can be limited by the quality of the imaging signals received and detected by the secondary electron detection system. Secondary charged particle beams, such as electron beams, generated by the interaction of primary beamlets on a sample surface can include secondary electrons having a large energy spread of about 50 eV and a large range of emission angles of about 90° relative to the sample surface normal. Such defocused electron beams can have a large point of incidence on the secondary electron detector. In a conventional multi-beam SEM, the defocused electron beams can be incident on multiple detection elements of the secondary electron detector. In other words, each of the multiple detection elements can receive secondary electrons from the corresponding secondary electron beam and other adjacent beams. Thus, the imaging signal of one detection element can include an intended signal (e.g., a signal corresponding to an image of Figure 3A ) from the corresponding secondary electron beam intended to be incident on the one detection element and one or more cross-talk signals (e.g., a signal corresponding to an image of Figure 3C ) from adjacent electron beams not intended to be incident on the one detection element resulting in a contaminated image (e.g., an image of Figure 3B ). In addition to this, the cross-talk contamination can degrade the fidelity of the imaging signal. Therefore, it is desirable to minimize the cross-talk of the secondary electron beams between the multiple detection elements to improve the imaging resolution.

[0025] To reduce the occurrence of cross-talk, various efforts have been made. However, it is quite challenging to completely remove cross-talk by mechanical and electronic techniques. Furthermore, even with computational methods, it is difficult to detect and remove the cross-talk contamination. Embodiments of the present disclosure can provide techniques to reduce the cross-talk contamination from the detection signal in real-time. In the present disclosure, a plurality of images including a particular image and its surrounding images can be used to obtain a transform function that can remove at least a portion of the cross-talk contamination from the detection signal, such as by using machine learning.

[0026] For clarity, relative dimensions of the components in the drawings can be exaggerated. In the following description of drawings, like or similar components or entities are referred to with like or similar reference numerals, and only the differences with respect to the various embodiments are described. As used herein, the term “or” encompasses all possible combinations, unless otherwise explicitly stated. For example, if a component is stated to include A or B, then unless otherwise explicitly stated or unless impracticable, the component can include A, or B, or A and B. As a second example, if a component is stated to include A, B, or C, then unless otherwise explicitly stated or unless impracticable, the component can include A, or B, or C, or A and B, or A and C, or B and C, or A and B and C.

[0027] Reference is now made to Figure 1which shows a schematic diagram of an exemplary electron beam inspection (EBI) system consistent with embodiments of the present disclosure. As shown Figure 1 The charged-particle beam inspection system 100 includes a main chamber 10, a load lock chamber 20, an electron beam tool 40, and an equipment front end module (EFEM) 30. The electron beam tool 40 is located within the main chamber 10. While the description and drawings are directed to an electron beam, it should be understood that embodiments do not limit the present disclosure to a particular charged-particle.

[0028] The EFEM 30 includes a first load port 30a and a second load port 30b. The EFEM 30 can include one or more additional load ports. The first load port 30a and the second load port 30b receive front opening wafer containers (FOUPs) containing wafers (e.g., semiconductor wafers or wafers made of one or more other materials) or samples to be inspected (wafers and samples are hereinafter collectively referred to as “wafers”). One or more robotic arms (not shown) in the EFEM 30 transport the wafers to the load lock chamber 20.

[0029] The load lock chamber 20 is connected to a load / lock vacuum pump system (not shown) that removes gas molecules in the load lock chamber 20 to reach a first pressure that is lower than atmospheric pressure. After reaching the first pressure, one or more robotic arms (not shown) transport the wafers from the load lock chamber 20 to the main chamber 10. The main chamber 10 is connected to a main chamber vacuum pump system (not shown) that removes gas molecules in the main chamber 10 to reach a second pressure that is lower than the first pressure. After reaching the second pressure, the wafers are subjected to inspection by the electron beam tool 40. In some embodiments, the electron beam tool 40 can include a single-beam inspection tool. In other embodiments, the electron beam tool 40 can include a multi-beam inspection tool.

[0030] The controller 50 can be electrically connected to the electron beam tool 40, as well as to other components. The controller 50 can be a computer configured to perform various controls of the charged-particle beam inspection system 100. The controller 50 can also include processing circuitry configured to perform various signal and image processing functions. While the controller 50 is shown in FIG. 1 as being external to the structure that includes the main chamber 10, the load lock chamber 20, and the EFEM 30, it should be understood that the controller 50 can be part of the structure. Figure 1

[0031] While the present disclosure provides examples of a main chamber 10 that houses an electron beam inspection system, it should be noted that aspects of the present disclosure are not limited in their broadest sense to a chamber that houses an electron beam inspection system. Rather, it can be understood that the principles described above can also be applied to other chambers.

[0032] Reference is now made to Figure 2 ​which shows an example charged particle beam inspection system 100 that can be Figure 1 a schematic diagram of an example electron beam tool 40 showing a portion of an example charged particle beam inspection system 100 consistent with embodiments of the present disclosure. The electron beam tool 40 (also referred to herein as apparatus 40) includes an electron source 101, a gun aperture plate 171 having a gun aperture 103, a pre-sub-beam forming mechanism 172, a condenser lens 110, a source conversion unit 120, a primary projection optical system 130, a sample stage (not shown in Figure 2 The primary projection optical system 130 can include an objective lens 131. The electron detection device 140 can include a plurality of detection elements 140_1, 140_2, and 140_3. A beam separator 160 and a deflection scanning unit 132 can be placed within the main projection optical system 130. It can be appreciated that other well-known components of the apparatus 40 can be added / omitted as appropriate.

[0033] The electron source 101, the gun aperture plate 171, the condenser lens 110, the source conversion unit 120, the beam separator 160, the deflection scanning unit 132, and the main projection optical system 130 can be aligned with a main optical axis 100_1 of the apparatus 100. The imaging system 150 and the electron detection device 140 can be aligned with a second optical axis 150_1 of the apparatus 40.

[0034] The electron source 101 can include a cathode from which primary electrons can be emitted and an extractor or an anode to accelerate the primary electrons to form a primary electron beam 102 that forms a cross (virtual or real) 101s. The primary electron beam 102 can be visualized as being emitted from the cross 101s.

[0035] The source conversion unit 120 can include an imaging element array (not shown in Figure 2 The imaging element array can include a plurality of micro-deflectors or micro-lenses to form a plurality of parallel images (virtual or real) of the cross 101s with a plurality of sub-beams 102. Figure 2 Three sub-beams 102_1, 102_2, and 102_3 are shown as an example, and it should be appreciated that the source conversion unit 120 can handle any number of sub-beams.

[0036] In some embodiments, the source conversion unit 120 can be provided with a beam limiting aperture array and an imaging element array (both not shown). The beam limiting aperture array can include beam limiting apertures. It should be appreciated that any number of apertures can be used as appropriate. The beam limiting apertures can be configured to limit the size of the sub-beams 102_1, 102_2, and 102_3 of the primary electron beam 102. The imaging element array can include imaging deflectors (not shown) configured to deflect the sub-beams 102_1, 102_2, and 102_3 by changing the angle towards the principal optical axis 100_1. In some embodiments, deflectors further away from the principal optical axis 100_1 can deflect the sub-beams to a greater extent. Further, the imaging element array can include multiple layers (not shown), and the deflectors can be provided in separate layers. The deflectors can be configured to be individually controlled independently of one another. In some embodiments, the deflectors can be controlled to adjust the pitch of the probe points (e.g., 102_1S, 102_2S, and 102_3S) formed on the surface of the sample 1. As described herein, the pitch of the probe points can be defined as the distance between two immediately adjacent probe points on the surface of the sample 1.

[0037] The deflectors of the centrally located imaging element array can be aligned with the principal optical axis 100_1 of the electron beam tool 40. Thus, in some embodiments, the central deflectors can be configured to keep the trajectory of the sub-beams 102_1 straight. In some embodiments, the central deflectors can be omitted. However, in some embodiments, the primary electron source 101 can not necessarily be aligned with the center of the source conversion unit 120. Further, it should be appreciated that while the deflectors are shown as being aligned with the principal optical axis 100_1, the deflectors can be offset from the principal optical axis 100_1. Figure 2 A side view of the apparatus 40 is shown in which the sub-beams 102_1 are on the principal optical axis 100_1, but the sub-beams 102_1 can be off the principal optical axis 100_1 when viewed from different sides. That is, in some embodiments, all of the sub-beams 102_1, 102_2, and 102_3 can be off-axis. The off-axis component can be offset with respect to the principal optical axis 100_1.

[0038] The deflection angle of the deflected beamlets can be set based on one or more criteria. In some embodiments, the deflector can deflect the off-axis beamlets radially outward or away (not shown) from the main optical axis 100_1. In some embodiments, the deflector can be configured to deflect the off-axis beamlets radially inward or towards the main optical axis 100_1. The deflection angle of the beamlets can be set such that the beamlets 102_1, 102_2 and 102_3 fall perpendicularly on the sample 1. The image off-axis aberrations due to a lens such as the objective lens 131 can be reduced by adjusting the path of the beamlets through the lens. Thus, the deflection angle of the off-axis beamlets 102_2 and 102_3 can be set such that the probe points 102_2S and 102_3S have small aberrations. The beamlets can be deflected to pass through or close to the front focal point of the objective lens 131, thereby reducing the aberrations of the off-axis probe points 102_2S and 102_3S. In some embodiments, the deflector can be set to make the beamlets 102_1, 102_2 and 102_3 fall perpendicularly on the sample 1, while the probe points 102_1S, 102_2S and 102_3S have small aberrations.

[0039] The condenser lens 110 is configured to focus the primary electron beam 102. The current of the beamlets 102_1, 102_2 and 102_3 downstream of the source conversion unit 120 can be changed by adjusting the focusing power of the condenser lens 110 or by changing the radial dimension of the respective beam limiting aperture within the beam limiting aperture array. The current can be changed by changing the radial dimension of the beam limiting aperture and the focusing power of the condenser lens 110. The condenser lens 110 can be a tunable condenser lens, which can be configured such that the position of its first principal plane is movable. The tunable condenser lens can be configured to be magnetic, which can cause the off-axis beamlets 102_2 and 102_3 to illuminate the source conversion unit 120 with a rotation angle. The rotation angle can change with the focusing power or the position of the first principal plane of the tunable condenser lens. Thus, the condenser lens 110 can be a rotation-free condenser lens, which can be configured to keep the rotation angle constant when the focusing power of the condenser lens 110 changes. In some embodiments, the condenser lens 110 can be a tunable rotation-free condenser lens, wherein the rotation angle is constant when the focusing power and the position of the first principal plane of the condenser lens 110 change.

[0040] The electron beam tool 40 can include a pre-beamlet formation mechanism 172. In some embodiments, the electron source 101 can be configured to emit primary electrons and form a primary electron beam 102. In some embodiments, the gun aperture plate 171 can be configured to block peripheral electrons of the primary electron beam 102 to reduce the Coulomb effect. In some embodiments, the pre-beamlet formation mechanism 172 also cuts the peripheral electrons of the primary electron beam 102 to further reduce the Coulomb effect. The primary electron beam 102 can be trimmed into three primary electron sub-beams 102_1, 102_2, and 102_3 (or any other number of sub-beams) after passing through the pre-beamlet formation mechanism 172. The electron source 101, the gun aperture plate 171, the pre-beamlet formation mechanism 172, and the condenser lens 110 can be aligned with the main optical axis 100_1 of the electron beam tool 40.

[0041] The pre-beamlet formation mechanism 172 can include a Coulomb aperture array. The central aperture (also referred to herein as an on-axis aperture) of the pre-beamlet formation mechanism 172 and the central deflector of the source conversion unit 120 can be aligned with the main optical axis 100_1 of the electron beam tool 40. The pre-beamlet formation mechanism 172 can be provided with a plurality of pre-trimmed apertures (e.g., a Coulomb aperture array). In some embodiments, the pre-beamlet formation mechanism 172 can be positioned between the gun aperture plate 171 and the condenser lens 110. In some embodiments, the pre-beamlet formation mechanism 172 can be positioned between the gun aperture plate 171 and the source conversion unit 120. Figure 2 In some embodiments, the pre-beamlet formation mechanism 172 can be positioned closer to the electron source 101 than the gun aperture plate 171. In some embodiments, the pre-beamlet formation mechanism 172 can be positioned closer to the electron source 101 than the source conversion unit 120. In some embodiments, the pre-beamlet formation mechanism 172 can be positioned closer to the electron source 101 than the condenser lens 110. In some embodiments, the pre-beamlet formation mechanism 172 can be positioned closer to the electron source 101 than the source conversion unit 120 and the condenser lens 110. In some embodiments, the pre-beamlet formation mechanism 172 can be positioned closer to the electron source 101 than the gun aperture plate 171, the source conversion unit 120, and the condenser lens 110. Figure 2 Although three apertures of the pre-beamlet formation mechanism 172 are shown, it should be understood that there can be any number of apertures as appropriate.

[0042] In some embodiments, the pre-beamlet formation mechanism 172 can be positioned below the condenser lens 110. Positioning the pre-beamlet formation mechanism 172 closer to the electron source 101 can more effectively reduce the Coulomb effect. In some embodiments, the gun aperture plate 171 can be omitted when the pre-beamlet formation mechanism 172 can be positioned close enough to the source 101 while still manufacturable.

[0043] The objective lens 131 can be configured to focus the sub-beams 102_1, 102_2 and 102_3 onto the sample 1 for inspection and can form three probe spots 102_1s, 102_2s and 102_3s on the surface of the sample 1. The gun aperture plate 171 can block the peripheral electrons of the unused primary electron beams 102 to reduce the Coulomb interaction effect. The Coulomb interaction effect can enlarge the size of each of the probe spots 102_1s, 102_2s and 102_3s, thus reducing the detection resolution.

[0044] The beam separator 160 can be a Wien filter type beam separator including electrostatic deflectors (both not shown in Figure 2 Fig. 1) that generate an electrostatic dipole field E1 and a magnetic dipole field B1. If applied, the electrostatic dipole field E1 exerts a force on the electrons of the sub-beams 102_1, 102_2 and 102_3 that is equal in magnitude and opposite in direction to the force exerted on the electrons by the magnetic dipole field B1. The sub-beams 102_1, 102_2 and 102_3 can thus pass through the beam separator 160 directly with a zero deflection angle.

[0045] The deflection scanning unit 132 can deflect the sub-beams 102_1, 102_2 and 102_3 to scan the probe spots 102_1s, 102_2s and 102_3s over three small scan areas in a portion of the surface of the sample 1. In response to the incidence of the sub-beams 102_1, 102_2 at the probe spots 102_1, 102_2 and 102_3, three secondary electron beams 102_1se, 102_2se and 102_3se can be emitted from the sample 1. Each of the secondary electron beams 102_1se, 102_2se and 102_3se can contain electrons having an energy distribution including secondary electrons (energy < 50 eV) and backscattered electrons (energy between 50 eV and the landing energy of the sub-beams 102_1, 102_2 and 102_3). The beam separator 160 can direct the secondary electron beams 102_1se, 102_2se and 102_3se towards the secondary imaging system 150. The secondary imaging system 150 can focus the secondary electron beams 102_1se, 102_2se and 102_3se onto the detection elements 140_1, 140_2 and 140_3 of the electron detection device 140. The detection elements 140_1, 140_2 and 140_3 can detect the respective secondary electron beams 102_1se, 102_2se and 102_3se and generate respective signals for constructing images of the respective scan areas of the sample 1.

[0046] In Figure 2By the three probe points 102_1S, 102_2S and 102_3S, three secondary electron beams 102_1se, 102_2se and 102_3se are generated, respectively, and travel along the primary optical axis 100_1 upward toward the electron source 101, successively passing through the objective lens 131 and the deflection scanning unit 132. The three secondary electron beams 102_1se, 102_2se and 102_3se are diverted by a beam separator 160, such as a Wien filter, to enter the secondary imaging system 150 along a secondary optical axis 150_1 of the secondary imaging system 150. The secondary imaging system 150 focuses the three secondary electron beams 102_1se-102_3se onto an electron detection device 140 containing three detection elements 140_1, 140_2 and 140_3. Thus, the electron detection device 140 can simultaneously generate images of the three scan areas scanned by the three probe points 102_1S, 102_2S and 102_3S, respectively. In some embodiments, the electron detection device 140 and the secondary imaging system 150 form one detection unit (not shown). In some embodiments, the electron optical elements on the secondary electron beam path, such as but not limited to the objective lens 131, the deflection scanning unit 132, the beam separator 160, the secondary imaging system 150 and the electron detection device 140, can form one detection system.

[0047] In some embodiments, the controller 50 can include an image processing system including an image acquirer (not shown) and a memory (not shown). The image acquirer can include one or more processors. For example, the image acquirer can include 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 can be communicatively coupled to the electron detection device 140 of the apparatus 40 by a medium, such as an electrical conductor, a fiber optic cable, a portable storage medium, IR, Bluetooth, the Internet, a wireless network, a radio, etc., or a combination thereof. In some embodiments, the image acquirer can receive signals from the electron detection device 140 and can construct images. The image acquirer can thus acquire images of the sample 1. The image acquirer can also perform various post-processing functions, such as generating contours, superimposing indicators on the acquired images, etc. The image acquirer can be configured to perform adjustments of brightness and contrast, etc., of the acquired images. In some embodiments, the memory can 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 can be coupled with the image acquirer and can be used to save the scanned raw image data as raw images and post-processed images.

[0048] In some embodiments, the image acquirer can acquire one or more images of the sample based on one or more imaging signals received from the electron detection device 140. The imaging signals can correspond to a scanning operation for charged particle imaging. The acquired image can be a single image that includes multiple imaging regions or can involve multiple images. The single image can be stored in memory. The single image can be a raw image that can be divided into multiple regions. Each of the regions can include one imaging region that contains a feature of the sample 1. The acquired image can include multiple images of a single imaging region of the sample 1 that are sampled over a time sequence or can include multiple images of different imaging regions of the sample 1. The multiple images can be stored in memory. In some embodiments, the controller 50 can be configured to perform image processing steps on multiple images of the same location of the sample 1.

[0049] In some embodiments, the controller 50 can include a measurement circuit (e.g., an analog-to-digital converter) to obtain a distribution of detected secondary electrons. The electron distribution data collected during a detection time window, in combination with respective scan path data of each primary beamlet 102_1, 102_2, and 102_3 incident on the wafer surface, can be used to reconstruct an image of the wafer structure under inspection. The reconstructed image can be used to reveal various features of the internal or external structure of the sample 1, which can be used to reveal any defects that can be present in the wafer.

[0050] In some embodiments, the controller 50 can control a motorized stage (not shown) to move the sample 1 during the inspection. In some embodiments, the controller 50 can enable the motorized stage to continuously move the sample 1 in one direction at a constant speed. In other embodiments, the controller 50 can enable the motorized stage to vary the movement speed of the sample 1 over time according to steps of the scanning process. In some embodiments, the controller 50 can adjust a configuration of the primary projection optical system 130 or the secondary imaging system 150 based on the images of the secondary electron beams 102_1se, 102_2se, and 102_3se.

[0051] While Figure 2 The electron beam tool 40 is shown using three primary electron beams, but it should be understood that the electron beam tool 40 can use two or a larger number of primary electron beams. The present disclosure does not limit the number of primary electron beams used in the apparatus 40.

[0052] As with the Figure 2As explained, the surface of the sample 1 is simultaneously scanned by the plurality of primary electron beams or sub-beams 102_1, 102_2 and 102_3 at different locations such as the probe points 102_1S, 102_2S and 102_3S. A plurality of secondary electron beams 102_1se, 102_2se and 102_3se emitted from the sample 1 and respectively corresponding to the plurality of primary electron beams 102_1, 102_2 and 102_3 can be detected on respective areas on the electron detection device 140. For example, it can be designed to detect the first secondary electron beam 102_1se on the first detection element 140_1, the second secondary electron beam 102_2se on the second detection element 140_2, and the third secondary electron beam 102_3se on the third detection element 140_3. Defocused secondary electron beams are generated due to, but not limited to, aberrations and imperfections of the beam separation device of the secondary optical system and the electron beam tool 40 during the projection of the secondary electron beams. Such defocused secondary electron beams can be incident on a plurality of detection elements different from the detection element corresponding thereto. Therefore, the imaging signal of one detection element can comprise an intended signal originating from the corresponding secondary electron beam and one or more crosstalk signals originating from adjacent electron beams. Thus, the image constituted by the detected image signal of one detection element can contain a main pattern originating from the intended signal and a ghost pattern originating from the crosstalk signals.

[0053] Reference is now made to Figure 3A-3C which are examples of an electron beam image with a first pattern A, an electron beam image with a second pattern B and an electron beam image with crosstalk contamination. For example, Figure 3A may be an image reconstructed from the electron detection signal from the first detection element 140_1 when no crosstalk signal is incident on the first detection element 140_1 except for the intended secondary electron beam 102_1se. Here, Figure 3A the image from the first detection element 140_1 shown contains the pattern A with a dot shape. Figure 3B may be an image reconstructed from the electron detection signal from the second detection element 140_2 when no crosstalk signal is incident on the second detection element 140_2 except for the intended secondary electron beam 102_2se. Here, Figure 3B the image from the second detection element 140_2 shown contains the pattern B with a cross shape. Figure 3C shows an example of an image generated based on the electron detection signal from the first detection element 140_1 when a crosstalk signal originating from the second secondary electron beam 102_2se in addition to the intended first secondary electron beam 102_1se is incident on the first detection element 140_1. Thus, Figure 3CThe illustrated image from the first detection element 140_1 includes a ghost pattern B as well as a main pattern A. Such ghost patterns in SEM images will lead to higher false defect detection rates. Thus, it is desirable to minimize crosstalk to improve imaging resolution.

[0054] It is difficult to detect and remove crosstalk contamination for a number of reasons, including: (1) it is difficult to detect and determine crosstalk contamination from the detection signal or reconstructed image without external information help; (2) it is difficult to distinguish ghost artifacts (such as ghost patterns) from the main pattern since the ghost pattern and the main pattern often have many common features, such as in the frequency domain or spatial domain; (3) even after the area of the ghost pattern is detected or defined, it is difficult to completely remove the artifact without destroying the original pattern or background image since it requires accurate estimation of the artifact signal and seamless recovery of the original pattern and underlying background image; (4) the reference image, such as Graphic Database System (GDS) format, can only help find the rough location of the ghost pattern, thus, it is still challenging to detect the precise pixel-level location of the ghost pattern; (5) in addition, it is time-consuming and laborious to identify the crosstalk artifact and compute the amount of crosstalk component. Embodiments of the present disclosure can provide techniques to reduce, minimize, or remove crosstalk contamination in the detection signal, which can be performed offline or in batch, or can be performed in real-time.

[0055] Figure 4 is a block diagram of an exemplary image enhancement apparatus consistent with embodiments of the present disclosure. It should be appreciated that in various embodiments, the image enhancement apparatus 400 can be part of a charged particle beam inspection system (e.g., the electron beam inspection system 100 of Figure 1 FIG. 1) or can be separate from the charged particle beam inspection system. In some embodiments, the image enhancement apparatus 400 can be part of the controller 50 and can include an image acquirer, a measurement circuit, or a memory, etc. Further, the image enhancement apparatus 400 can include an image processing system and can include an image acquirer, a memory, etc.

[0056] In some embodiments, as shown in Figure 4 FIG. 4, the image enhancement apparatus 400 can include a detection signal acquirer 410, an analog-to-digital converter (ADC) 420, a crosstalk reducer 430, a transform function provider 440, and an image generator 450. In some embodiments, the image enhancement apparatus 400 receives data from the electron beam tool 40. Figure 2 FIG. 1).

[0057] According to embodiments of the present disclosure, the detection signal acquirer 410 can receive detection signals from the electron detection apparatus 140 of the electron beam tool 40. In some embodiments, the detection signal acquirer 410 can be communicatively coupled to the electron detection apparatus 140 of the electron beam tool 40 through a medium such as an electrical conductor, a fiber optic cable, a portable storage medium, IR, Bluetooth, the Internet, a wireless network, a radio, or the like, or a combination thereof. As Figure 2 shown, the electron detection apparatus 140 can include a plurality of detection elements. Although Figure 2 three detection elements 140_1, 140_2, and 140_3 arranged in a line are shown, it is understood that the electron detection apparatus 140 can include any number of detection elements arranged in any manner. For example, the electron detection apparatus 140 can have m detection elements arranged in a matrix having k rows and n columns. Here, “k” and “n” can be natural numbers equal to or greater than 1, and “m” can be a natural number equal to or greater than 2. Thus, in some embodiments, the detection signal acquirer 410 can receive m detection signals s1 to sm corresponding to the m detection elements 140_1 to 140_m. m It is understood that the acquired detection signals s1 to sm m may be subjected to various signal processing such as amplification, noise removal, etc. The detection signals s1 to sm acquired from the electron detection apparatus 140 can be analog signals. m

[0058] An analog-to-digital converter (ADC) 420 is configured to convert the acquired analog detection signals to digital detection signals s1 to sm. m In the present disclosure, an analog detection signal and a corresponding digital detection signal are denoted by the same reference numerals such as s m In some embodiments, the ADC 420 converts the analog detection signals to digital detection signals s1 to sm m to obtain a distribution of detected secondary electrons. The electron distribution data collected during a detection time window, in combination with respective scan path data of each of the primary beamlets 102_1, 102_2, and 102_3 incident on the wafer surface, can be used to generate an image of the wafer structure under inspection.

[0059] As described above, the detection signals s1 to sm m may include crosstalk signals in addition to the corresponding expected signals. For example, according to Figure 2 ​The first detection signal s1 detected from the first detection element 140_1 can include an intended signal originating from the first secondary electron beam 102_1se and a crosstalk signal originating from the adjacent secondary electron beam 102_2se, 102_3se, or both 102_2se and 102_3se. It should be understood that electrons from any of the second secondary electron beams can reach the first detection element 140_1, although electrons from the secondary electron beam adjacent to the intended signal are more likely to be contaminated. Similarly, the second detection signal s2 detected from the second detection element 140_2 can include an intended signal originating from the second secondary electron beam 102_2se and a crosstalk signal originating from the other secondary electron beam (e.g., the first secondary electron beam 102_1se). Similarly, any detection signal s1 to s m may include an intended signal originating from the corresponding secondary electron beam and a crosstalk signal originating from a secondary electron beam other than the intended secondary electron beam.

[0060] Again referring to Figure 4 , the crosstalk reducer 430 can reduce the crosstalk contamination from the detection signal s1 to s m . According to an embodiment of the present disclosure, the crosstalk reducer 430 can reduce the crosstalk contamination from the detection signal s1 to s m by using a transformation function obtained from the transformation function provider 440, which will be explained in detail with reference to Figure 5 . The principle for reducing the crosstalk contamination of the detection signal s1 to s m is explained mathematically below.

[0061] The detection signal from each detection element is denoted by s i , where i is an index number corresponding to the i-th detection element 140_i. The intensity of each secondary electron beam is denoted by b j , where j is an index number corresponding to the j-th secondary electron beam 102_jse. It is assumed that m secondary electron beams and m detection elements are included in the present disclosure, where m is a natural number equal to or greater than 2. In some embodiments, the number of secondary electron beams and the number of detection elements can be different. The intensities I of the secondary electron beams 102_1se to 102_mse and the detection signal S from the electron detection device 140 can be represented as the following matrix or vector:

[0062] I = (b1…b m ) τ ( Equation 1 )

[0063] S = (s1…s m ) T ( Equation 2 )

[0064] Here, the superscript T in equations 1 and 2 indicates the transpose of a matrix or vector.

[0065] The relationship between the beam intensity I of the secondary electron beams 102_1se to 102_mse and the detection signal S from the electron detection device 140 can be expressed as follows:

[0066]

[0067] Here, R represents each detection signal s from the i-th detection element 140_i. i The responsivity matrix relating the beam intensities of m secondary electron beams 10²_1se to 10²_mse. R includes r 11 to r mm As its elements. Equation 3 shows that the detection signal from any detection element can be expressed as a linear combination of the beam intensities of all secondary electron beams. For example, the first detection signal s1 can be expressed as s1 = r through the combination of the beam intensities of m secondary electron beams 102_1se to 102_mse. 11 *b1+r 12 *b2+···+r 1(m-1) *b (m-1) +r 1m *b m The second detection signal s2 can be expressed as a combination of the beam intensities of m secondary electron beams 102_1se to 102_mse, as s1 = r 21 *b1+r 22 *b2+···+r 2(m-1) *b (m-1) +r 2m *b m Similarly, the i-th detection signal s i s can be expressed as a combination of the beam intensities of m secondary electron beams from 10²_1se to 10²_mse. i =r i1 *b1+r i2 *b2+···+r i(m-1) *b (m-1) +r im * bm .

[0068] In the response matrix R, the crosstalk contribution rate r ij (where i≠j) represents the beam intensity b of the j-th secondary electron beam 102_jse incident on the i-th detection element 140_i. j A portion of the total beam intensity b of the j-th secondary electron beam 102_jse j The ratio of self-contribution rate r iia ratio of a portion of the total beam intensity b i of the i-th secondary electron beam 102_ise incident on the i-th detection element 140_i to the total beam intensity b i of the i-th secondary electron beam 102_ise. It will be appreciated that in an ideal case where no cross-talk occurs, the self-contribution ratio r ii has a value of 1, while the cross-talk contribution ratio r ij for i≠j has a value of 0. However, due to aberrations and imperfections of the electron beam tool 40, the secondary electron beams can be defocused, which can cause cross-talk to occur. Therefore, typically the value of the self-contribution ratio r ii is not equal to 1, and the cross-talk contribution ratio r ij for i≠j is not equal to 0.

[0069] The beam intensities I of the secondary electron beams 102_1se to 102_mse can be obtained by multiplying the inverse responsivity matrix R -1 with both sides of equation 3, and can be represented as follows:

[0070]

[0071] By applying equation 4, cross-talk contamination can be reduced, minimized or removed from the detection signals s1 to s m , and thus beam intensities b1 to b m of the secondary electron beams 102_1se to 102_mse without or with minimal cross-talk contamination can be obtained.

[0072] Designing the responsivity matrix R to be invertible enables the calculation of the beam intensities b1 to b m of the secondary electron beams 102_1se to 102_mse according to equation 4. In some embodiments, the responsivity matrix R can be configured to be invertible by appropriately configuring the size, shape and position of the detection elements with respect to the grid of points onto which the secondary electron beams are projected onto the surface of the electron detection device.

[0073] However, not every responsivity matrix R can be made invertible. Furthermore, even if the responsivity matrix R is invertible, calculating the inverse responsivity matrix R -1 mathematically can be very time consuming. Therefore, reconstructing the SEM image from the detection signals s1 to s m can result in a time-consuming calculation process involving a large amount of data, which can make real-time removal of cross-talk contamination difficult to achieve.

[0074] According to embodiments of the present disclosure, the cross-talk reducer 430 can obtain a transformation function from the transformation function provider 440. Here, the transformation function can correspond to the inverse responsivity matrix R -1 of equation 4. In some embodiments, the transformation function can be R -1transform function. In some embodiments, the transform function can be determined with respect to a certain e-beam tool 40 or EBI system 100, since the cross-talk contamination is unique to the tool or system itself, and not to an error in operating the tool or system. Thus, in accordance with embodiments of the present disclosure, the transform function provider 440 can pre-compute the transform function with images acquired from the particular tool 40 or system 100, and provide the pre-obtained transform function to the cross-talk reducer 430 on demand. In some embodiments, the transform function obtained from the transform function provider 440 can be stored in a storage medium (not shown) and can be accessed by the cross-talk reducer 430. In some embodiments, the transform function provider 440 can update the transform function when new images are acquired from the tool 40 or system 100. In some embodiments, the transform function provider 440 can periodically update the transform function, since the degree or level of defocus due to imperfections inherent in the tool 40 or system 100 can change over time. Thus, the transform function provider 440 can provide the latest transform function to the cross-talk reducer 430 or the storage medium. Details will be provided below with respect to Figure 5 The configuration and function of the transform function provider 440 will be explained in detail.

[0075] Referring again to Figure 4 , the cross-talk reducer 430 can reduce, minimize, or remove the cross-talk contamination of the detection signals s1 to s m by multiplying the detection signals s1 to s m by the transform function shown in Equation 4. In this way, the cross-talk reducer 430 can obtain the beam intensities b1 to b m of the secondary electron beams 102_1se to 102_mse without or with minimal cross-talk contamination. In some embodiments, the cross-talk reducer 430 can be implemented as firmware or other software in the image enhancement apparatus 400.

[0076] In accordance with embodiments of the present disclosure, the image generator 450 can construct a plurality of images corresponding to the plurality of secondary electron beams 102_1se to 102_mse, which are reduced, minimized, or removed from the cross-talk contamination of the detection signals s1 to s mThe images are obtained by eliminating crosstalk contamination. In some embodiments, the multiple images constructed by the image generator 450 may be images of multiple regions scanned by multiple probe points (e.g., 102_1S, 102_2S, and 102_3S) on sample 1, wherein corresponding secondary electron beams (e.g., 102_1se, 102_2se, and 102_3se) are generated. In some embodiments, according to embodiments of the present disclosure, the image generator 450 may generate multiple images at once or sequentially. In some embodiments, the multiple images generated from the image generator 450 may be images without crosstalk artifacts.

[0077] Now for reference Figure 5 It shows embodiments consistent with those of this disclosure. Figure 4 A block diagram of an exemplary transform function provider 440 of an exemplary image enhancement apparatus 400. (See also:) Figure 5 As shown, the transformation function provider 440 may include a pattern identifier 441, an image patch extractor 442, a machine learning network 443, and an information file 444.

[0078] Pattern identifier 441 can receive multiple images IM acquired from a target device (such as, but not limited to, a tool 40 or system 100 that can be used to determine a transformation function). l To IM n In some embodiments, multiple image IMs l To IM n This can include multiple images obtained through a single scan operation on any sample and corresponding to multiple sub-regions of the sample. In some embodiments, the multiple images IM l To IM n This can include multiple sets of images obtained from multiple scanning operations on a sample. Here, each set of images can include multiple images obtained through a corresponding scanning operation and corresponding to multiple sub-regions of the sample. In some embodiments, the multiple images IM l To IM n This can include multiple sets of images obtained from multiple scan operations on multiple samples. Here, each set of images can include multiple images obtained from corresponding scan operations for multiple sub-regions of a corresponding sample.

[0079] According to embodiments of this disclosure, the pattern identifier 441 can identify the input image IM. l To IM nThe pattern identifier 441 can identify which pattern is the main pattern of a particular image and which pattern is a ghosting pattern originating from crosstalk contamination. In some embodiments, the pattern identifier 441 can identify patterns via a reference information file 444, which can provide reference information for a particular image when determining the main pattern and ghosting patterns caused by crosstalk contamination. In some embodiments, the information file 444 may contain an image showing the input image IM. l To IM n The reference image of the main pattern. In some embodiments, the pattern identifier 441 can determine the input image IM. l To IM n The shape and position of the main pattern. For example, information file 444 may contain information related to the input image IM. l To IM n Multiple corresponding reference images are shown, along with the corresponding input image IM if no crosstalk contamination occurs. l To IM n What is the intention? In some embodiments, information file 444 may contain input images IM corresponding to them. l To IM n The reference images are the baseline truth images of the sample sub-regions. In some embodiments, the reference images included in the information file 444 may be in Graphics Database System (GDS) format, Graphics Database System II (GDS II) format, Open Art System Exchange Standard (OASIS) format, Caltech Intermediate Format (CIF), etc. In some embodiments, the reference images included in the information file 441 may include the input images IM corresponding to them. l To IM n The wafer design layout is a sample sub-region. The wafer design layout can be based on a pattern layout used to construct the wafer. The wafer design layout can correspond to one or more photomasks or stencils for transferring features from the photomasks or stencils to the wafer, such as sample 1. In some embodiments, an information file in GDS or OASIS, etc., may include feature information stored in a binary file format, representing planar geometry, text, and other information related to the wafer design layout.

[0080] For the purposes of explanation and for simplicity, the following will be addressed. Figures 3A-3C The electron beam image is used to interpret the pattern marking process. For example, suppose the pattern marker 441 receives a first image IM1 generated based on the first detection signal s1 of the first detection element 140_1 (such as...). Figure 3C As shown in the image, and a second image IM2 (not shown) generated based on the second detection signal s2 of the second detection element 140_2. In this example, the information file 441 may contain a first reference image and a second reference image, which correspond to...Figure 3A and Figure 3B The first and second reference images correspond to the images expected from the first and second detection elements 140_1, 140_2 without crosstalk contamination. The first reference image shows how the sub-region of the sample 1 detected by the first detection element 140_1 is expected to look without crosstalk contamination, and the second reference image shows how the sub-region of the sample 1 detected by the second detection element 140_2 is expected to look without crosstalk contamination. In this example, the pattern identifier 441 can identify the primary and ghost patterns of the first image IM1 with reference to the first and second reference images. For example, the pattern identifier 441 identifies pattern A as the primary pattern originating from the secondary electron beam 102_1se and pattern B as the ghost pattern originating from the adjacent secondary electron beam 102_2se in the first image IM1. Similarly, the pattern identifier 441 can identify pattern B as the primary pattern and pattern A as the ghost pattern in the second image IM2.

[0081] Referring again to Figure 5 Consistent with embodiments of the present disclosure, the image block extractor 442 can extract image blocks from the input images IM l to IM n based on the pattern information identified by the pattern identifier 441. As an example, the process for extracting image blocks from the input images IM l to IM n will be explained with respect to the first image IM1 as shown in Figure 3C . As shown in Figure 6 , the first image contains the primary pattern A and the ghost pattern B. In this example, the image block extractor 442 can extract three types of image blocks, such as but not limited to, a background block P1, a primary pattern block P2, and a ghost pattern block P3. In some embodiments, the background block P1 can be a portion of the first image IM1 and include only regions that do not have features that constitute either the primary pattern A or the ghost pattern B. The primary pattern block P2 can be a portion of the first image IM1 and include only regions that have features that constitute the primary pattern A but not the ghost pattern B. Similarly, the ghost pattern block P3 can be a portion of the first image IM1 and include only regions that have features that constitute the ghost pattern B but not the primary pattern A. In some embodiments, multiple background blocks P1, multiple primary pattern blocks P2, and multiple ghost pattern blocks P3 can be extracted from the first image IM1. Similarly, one or more background blocks, one or more primary pattern blocks, and one or more ghost pattern blocks can be extracted from the second image IM2 (not shown).

[0082] The intensity of the secondary electron beam can vary depending on the external or internal structure of the scanned area of ​​the wafer. It should be noted that the value of the detection signal can typically be constant, or it can remain unchanged in background areas without features constituting patterns (such as patterns A and B), because the wafer region corresponding to the background area may have a uniform structure, material, shape, etc. Therefore, by extracting the background block P1 and other pattern blocks P2 and P3, including no more than one pattern, and using them as input to the machine learning network 443, the computational data can be significantly reduced, and thus cost and time efficiency can be improved when using the machine learning network 443 to infer the transformation function. The inference accuracy of the transformation function can also be improved. Although for example... Figure 6 The image shown, with only two patterns, illustrates the extraction of blocks from an electron beam image; however, it should be understood that embodiments of this disclosure can be applied to images with three or more patterns. l To IM n The scene, and three or more images IM l To IM n Each of the blocks comprises three or more patterns. If three or more patterns exist in the image generated based on the detection signal, three types of blocks can also be extracted, such as background block P1, main pattern block P2, and ghost pattern block P3 (this example is not shown in the figure). If image IM1 contains main pattern A, first ghost pattern B1, and second ghost pattern B2, then ghost pattern block P3 can include two types of ghost pattern blocks P31 and P32. For example, the first type of ghost pattern P31 can be a part of the image and only includes regions with features that constitute the first ghost pattern B1, but not the main pattern A and the second ghost pattern B2. In this way, even if the number of patterns in the image increases, the computational complexity can be reduced by using blocks with no more than one pattern. In some embodiments, for higher accuracy, regions where multiple patterns overlap can be extracted as blocks. For example, when regions with no more than one pattern in the image are limited and therefore the data for computation is short, blocks with two or more patterns can be used for computation.

[0083] According to embodiments of this disclosure, machine learning network 443 can be configured to acquire a transformation function associated with a target device, wherein multiple image IMs are acquired from the target device. l To IM nConsistent with some embodiments of the present disclosure, the machine learning network 443 can receive the image patches P1, P2, and P3 extracted from the image patch extractor 442 as its input for inferring the transformation function. In some embodiments, the machine learning network 443 can obtain information from the received image patches P1, P2, and P3 for inferring the transformation function. In some embodiments, the information for inferring the transformation function can include, but is not limited to, the locations of the extracted image patches and the detection signals corresponding to the locations of the extracted image patches.

[0084] For the use of a simplified example of scanning using only two primary electron sub-beams, such as the first and second electron sub-beams 102_1 and 102_2, it will be explained how to select a patch and how to obtain the secondary electron beam intensity and the detection signal corresponding to the selected patch. The first sub-beam 102_1 and the second sub-beam 102_2 can be used to scan a first probe point 102_1S and a second probe point 102_2S on the surface of the sample 1. The first probe point 102_1S and the second probe point 102_2S can correspond to a first detection element 140_1 and a second detection element 140_2, respectively. During scanning, each sub-beam moves along a certain path on the surface of the sample 1 with respect to the sample 1 over time, preferably according to a predetermined schedule. In some embodiments, the first sub-beam 102_1 and the second sub-beam 102_2 can be designed to move parallel to each other within their corresponding probe points 102_1S and 102_2S during scanning. Thus, at a certain point in time during the time period for scanning, the first sub-beam 102_1 and the second sub-beam 102_2 can be directed to the corresponding probe points 102_1S and 102_2S, respectively. Which location within the corresponding probe points the sub-beams will be incident at a certain point in time during the time period for scanning can be determined in advance and is known to the system.

[0085] During the time period used to scan the first probe spot 102_1S, a first secondary electron beam 102_1se can be emitted from the sample associated with the movement of the first beamlet 102_1. Thus, the beam intensity b1 of the first secondary electron beam 102_1se can comprise the beam intensity values (continuous or discrete) within the time period corresponding to the movement of the first beamlet 102_1 over the first probe spot 102_1S during the same relative time period. The beam intensity value of the beam intensity b1 at a particular time can represent the beam intensity corresponding to the position on the surface of the sample 1 to which the first beamlet 102_1 is directed at the particular time. In some embodiments, if a position within a probe spot is determined, the beam intensity of the position can be obtained by finding the beam intensity value at the time point corresponding to the position from the beam intensity such as b1. It will be appreciated that the beam intensity b2 of the secondary electron beam 102_2se and the detection signals s1 and s2 corresponding to the first and second secondary electron beams 102_1se and 102_2se can be represented in a similar manner. For example, the first detection signal s1 can comprise the detection signal values (continuous or discrete) within the time period corresponding to the time periods used to scan the first probe spot 102_1S and the second probe spot 102_2S. The detection signal value of the first detection signal s1 at a particular time can be a combination of the beam intensity value of the beam intensity b1 and the beam intensity value of the beam intensity b2 corresponding to the relative position within its corresponding probe spot 102_1S and 102_2S at the particular time. In some embodiments, the extracted image block can have a small size such that one beamlet can cover to obtain a particular beam intensity value at a particular time that is constant within the block area at the particular time.

[0086] According to embodiments of the present disclosure, the machine learning network 443 can obtain the position of the background patch P1 in the first image IM1 and the detection signal value corresponding to the position of the background patch P1 as described above. The machine learning network 443 also knows that the detection signal at the background patch P1 is the sum of a portion of the first secondary electron beam 102_1se and a portion of the second secondary electron beam 102_2se corresponding to the position of the background patch P1. Here, the first secondary electron beam 102_1se and the second secondary electron beam 102_2se at the position of the background patch P1 have the same intensity corresponding to the background region of the sample 1. In the same example, the machine learning network 443 can obtain the position of the main pattern patch P2 in the first image IM1 and the detection signal value corresponding to the position of the main pattern patch P2. The machine learning network 443 also knows that the detection signal at the main pattern patch P2 is the sum of a portion of the first secondary electron beam 102_1se and a portion of the second secondary electron beam 102_2se corresponding to the position of the main pattern. Here, at the position of the main pattern patch P, the second secondary electron beam 102_2se can have an intensity corresponding to the background region of the sample 1. Further, the machine learning network 443 can obtain the position of the ghost pattern patch P3 in the first image IM1 and the detection signal value corresponding to the position of the ghost pattern patch P3. The machine learning network 443 also knows that the detection signal at the ghost pattern patch P2 is the sum of a portion of the first secondary electron beam 102_1se and a portion of the second secondary electron beam 102_2se corresponding to the position of the ghost pattern patch P3. Here, at the position for the ghost pattern patch P3, the first secondary electron beam 102_1se can have an intensity corresponding to the background region of the sample. If the secondary electron beams 102_1se and 102_2se have constant values on the background region of the sample 1, the number of variables for inferring the transformation function can be reduced.

[0087] In some embodiments, the machine learning network 443 can be configured to learn the elements of the transformation function using an unsupervised machine learning algorithm. For example, referring to Equation 4, the machine learning network 443 can be configured to learn the elements r -1 to r 11 and the elements b1 to b m of the transformation function (corresponding to R mm ) by using the information obtained from the image patches as input data and by setting conditions for the learning process. For the above example, the transformation function can have four elements r 11 to r 22The signal S can have two elements: a first detection signal s1 from the first detection element 140_1 and a second detection signal s2 from the second detection element 140_2, and the beam intensity I can have two elements b1 and b2, which have the same value corresponding to the background region. Under these input data and conditions, the machine learning network 443 can be configured to infer the elements of the transformation function and the elements of the beam intensity I. In some embodiments, the machine learning network 443 can update the previously learned transformation function and beam intensity using subsequent image patches such as P2 and P3. In some embodiments, the machine learning network 443 can generate a transformation function if conditions for ending the learning process are met. For example, the process can end if the number of iterations becomes equal to a specific value, or if the accuracy of the transformation function reaches a specific level.

[0088] In some embodiments, machine learning network 443 can be configured to learn the elements of a transformation function using a supervised machine learning algorithm. For example, referring to Equation 4, machine learning network 443 can be configured to learn the elements of a transformation function from b1 to b... m Given the machine learning network 443, information obtained from image patches is used as input data to learn the transformation function (corresponding to R). -1 element r 11 to r mm For the example above, the transformation function can have four elements r. 11 to r 22 The signal S may have two elements: a first detection signal s1 from the first detection element 140_1 and a second detection signal s2 from the second detection element 140_2, and provides two elements b1 and b2 for the beam intensity I. Under these inputs and conditions, the machine learning network 443 may be configured to infer elements of the transformation function. In some embodiments, the machine learning network 443 may update the previously learned transformation function by using subsequent image patches such as P2 and P3, as well as beam intensity information. In some embodiments, the information file 444 may include time-varying beam intensity information corresponding to each image patch P1 to P3, and may be provided to the machine learning network 443. In some embodiments, the beam intensity information may be obtained from simulations, previous experiments, etc.

[0089] Although targeting Figures 3A-3C The images shown illustrate the process performed by the transform function provider 440. It should be understood that embodiments of this disclosure can be applied to situations involving the use of three or more images IM1 to IM2. n Each image includes a scene with two or more patterns. It should be noted that, with images IM1 to IM... nAs the number of images and the patterns included in them increase, the computation for obtaining the transform function becomes exponentially more complex. Therefore, pre-compiling a transform function specific to a particular device and system makes it possible to generate reconstructed images from the detected signals in real time.

[0090] Figure 7 This is a process flowchart illustrating an exemplary method for enhancing images in a multi-beam detection system, consistent with embodiments of this disclosure. For illustrative purposes, it will be compared with... Figure 2 Together with electron beam tool 40, a method for enhancing images is described.

[0091] In step S710, a detection signal can be acquired. Step S710 can be performed by, for example, a detection signal acquisition device 410. In some embodiments, the detection signal can be acquired from the electronic detection device 140 of the electron beam tool 40. Figure 2 As shown, the electronic testing device 140 may include multiple testing elements. Although Figure 2 The diagram illustrates an electronic inspection device 140 comprising three detection elements 140_1, 140_2, and 140_3 arranged in a line. However, it should be understood that the electronic inspection device 140 may include any number of detection elements arranged in any manner. For example, the electronic inspection device 140 may have m detection elements arranged in a matrix with k rows and n columns. Here, "k" and "n" can be natural numbers equal to or greater than 1, and "m" can be a natural number equal to or greater than 2. Therefore, in some embodiments, m detection signals s1 to s2 corresponding to the m detection elements 140_1 to 140_m can be acquired. m Detection signal. It should be understood that the acquired detection signals s1 to s2 are... m It can withstand various signal processing methods, such as amplification and noise cancellation. The detection signals s1 to s2 acquired from the electronic detection device 140... m It can be an analog signal.

[0092] In step S720, the acquired analog detection signal can be converted into digital detection signals s1 to s2. m Step S710 can be performed by, for example, an analog-to-digital converter 420. In some embodiments, the analog detection signal can be converted into digital detection signals s1 to s2. m The distribution of detected secondary electrons is obtained. The electron distribution data collected during the detection time window, combined with the corresponding scan path data of each primary sub-beam 102_1, 102_2, and 102_3 incident on the wafer surface, can be used to generate an image of the wafer structure under inspection.

[0093] As described above, the detection signals s1 to s mIt may include crosstalk signals in addition to the corresponding expected signals. In other words, the detection signals s1 to s2 detected from the corresponding detection elements 140_1 to 140_m m This can include the expected signal originating from the corresponding secondary electron beam and crosstalk signals originating from secondary electron beams other than the expected secondary electron beam. Therefore, in step S730, the detection signals s1 to s2 can be... m Reduce, minimize, or remove crosstalk pollution. Step S710 can be performed by, for example, a crosstalk reducer 430. According to embodiments of this disclosure, the detection signals s1 to s2 can be reduced by using a transformation function. m Crosstalk pollution can be mitigated by, for example, through... Figure 8 The process shown is obtained and by Figure 5 The transformation function provider 440 performs the transformation. It is used to reduce the signal from detection signal s1 to s2. m The principle of crosstalk pollution has been explained with reference to equations 1 to 4 above, so it will be omitted here.

[0094] According to embodiments of this disclosure, the transformation function can correspond to the inverse response matrix R of Equation 4. -1 In some embodiments, by detecting signals s1 to s2 m Multiply by the transformation function shown in Equation 4 to reduce, minimize, or remove the detection signals s1 to s2. m Crosstalk contamination. In this way, beam intensities b1 to b2_mse of secondary electron beams with little or no crosstalk contamination can be obtained. m .

[0095] In step S740, detection signals s1 to s2 can be generated by reducing, minimizing, or removing them. m Multiple images corresponding to multiple secondary electron beams 102_1se to 102_mse are obtained to eliminate crosstalk contamination, consistent with embodiments of this disclosure. Step S710 can be performed by, for example, an image generator 440. In some embodiments, the generated multiple images can be images of multiple regions scanned by multiple probe points (e.g., 102_1S, 102_2S, and 102_3S) on sample 1, wherein corresponding secondary electron beams (e.g., 102_1se, 102_2se, and 102_3se) are generated in these regions. According to embodiments of this disclosure, multiple images can be generated one at a time or sequentially. In some embodiments, the generated multiple images can be images without crosstalk artifacts.

[0096] Although this disclosure has been described in conjunction with embodiments of reducing or removing crosstalk contamination in detection signals, it will be understood that, according to embodiments of this disclosure, crosstalk contamination in images generated based on detection signals can also be reduced or removed. For example, this can be done based on detection signals s1 to s2 detected from corresponding detection elements 140_1 to 140_m. m Multiple images are generated, and each of the multiple images may include a desired pattern derived from a corresponding secondary electron beam and a ghost pattern derived from a secondary electron beam other than the desired secondary electron beam. In this example, crosstalk contamination such as ghost patterns can be reduced or removed from the multiple images based on a transformation function. In some embodiments, the transformation relationship between the detection signal and the pixel values ​​of the image can be modified, for example, by... Figure 8 The process shown is obtained and by Figure 5 The transformation function provider 440 executes a transformation function, and the modified transformation function can be used to reduce or remove ghosting patterns from an image. In some embodiments, this can be achieved through... Figure 8 A process similar to the one shown is used to obtain the transform function, allowing the transform function to be directly applied to an image. In some embodiments, a machine learning network can be trained on the transform function in a form directly applicable to an image, such as pixel values, to reduce or remove crosstalk contamination. Here, the transform function can still be based on the relationship between the detection signal and the beam intensity of the secondary electron beam. In some embodiments, a method for enhancing an image may include acquiring a first image from a detector of a multi-beam inspection system, wherein the first image is generated based on a first detection signal from a first region of the detector, wherein electrons from a first secondary electron beam and electrons from a second secondary electron beam are incident on the first region, and reducing crosstalk image originating from the second secondary electron beam from the first image based on the relationship between the first detection signal and the beam intensities associated with the first and second secondary electron beams.

[0097] Figure 8 This is a process flow diagram illustrating an exemplary method for generating a transform function that reduces crosstalk contamination from an electron beam detection signal, consistent with embodiments of this disclosure. In some embodiments, it is possible to... Figure 7 Step S730 uses by Figure 8 The transformation function generated by the process.

[0098] In step S810, multiple images IM can be acquired from the target device (such as, but not limited to, a tool 40 or system 100 that can determine the transformation function). l To IM n Step S810 can be performed by, for example, a pattern identifier 441. In some embodiments, multiple images IM l To IM nThis can include multiple images obtained through a single scan operation on any sample and corresponding to multiple sub-regions of the sample. In some embodiments, the multiple images IM l To IM n This can include multiple sets of images obtained from multiple scanning operations on a sample. Here, each set of images can include multiple images obtained through a corresponding scanning operation and corresponding to multiple sub-regions of the sample. In some embodiments, the multiple images IM l To IM n This can include multiple sets of images obtained from multiple scanning operations on multiple samples. Here, each set of images can include multiple images obtained through a corresponding scanning operation and corresponding to multiple sub-regions of a corresponding sample.

[0099] According to an embodiment of this disclosure, in step S820, the input image IM can be identified. l To IM n The pattern. Step S820 can be performed by, for example, a pattern identifier 441, etc. It can be determined which pattern is the main pattern of a specific image and which pattern is a ghosting pattern originating from crosstalk contamination. In some embodiments, this can be done by referring to an information file (e.g., Figure 4 Information file 444 is used to identify patterns, and this information file can provide reference information for a specific image when determining the main pattern and ghosting pattern caused by crosstalk contamination. In some embodiments, information file 444 may include an image showing the input image IM. l To IM n The main pattern image. In some embodiments, the input image IM can be determined. l To IM n The shape and position of the main pattern. For example, information file 444 may contain information related to the input image IM. l To IM n Multiple corresponding reference images are shown, along with the corresponding input image IM if no crosstalk contamination occurs. l To IM n What is the intention? In some embodiments, information files 444 may contain input images IM. l To IM n The corresponding baseline truth image of the sample sub-region. In some embodiments, the images included in information file 444 may be in Graphics Database System (GDS) format, Graphics Database System II (GDS II) format, Open Art System Exchange Standard (OASIS) format, Caltech Intermediate Format (CIF), etc. In some embodiments, the images included in information file 441 may include their input images IM. l To IM nA wafer design layout corresponding to the sample sub-region. The wafer design layout can be based on a pattern layout used to construct the wafer. The wafer design layout can correspond to one or more photomasks or reticles used to transfer features from the photomask or reticle to the wafer, e.g., sample 1. In some embodiments, the information file in GDS or OASIS, etc., can include feature information stored in a binary file format representing planar geometry, text, and other information related to the wafer design layout.

[0100] For illustrative purposes and simplicity, the process of step S820 will be explained with respect to an electron beam image of Figures 3A-3C For illustrative purposes and simplicity, the process of step S820 will be explained with respect to an electron beam image of Figure 3C For illustrative purposes and simplicity, the process of step S820 will be explained with respect to an electron beam image of Figure 3A For illustrative purposes and simplicity, the process of step S820 will be explained with respect to an electron beam image of Figure 3B For illustrative purposes and simplicity, the process of step S820 will be explained with respect to an electron beam image of

[0101] In step S830, consistent with embodiments of the present disclosure, image blocks can be extracted from the input image IM l to IM n Step S830 can be performed by, e.g., an image block extractor 442, etc. As an example, the process of step S830 will be explained with respect to the first image IM1 as shown in Figure 3C As an example, the process of step S830 will be explained with respect to the first image IM1 as shown in Figure 6As shown, the first image contains a main pattern A and a ghost pattern B. In this example, three types of image blocks can be extracted, such as, but not limited to, a background block P1, a main pattern block P2, and a ghost pattern block P3. In some embodiments, the background block P1 may be part of the first image IM1 and includes only areas that do not have features constituting the main pattern A or the ghost pattern B. The main pattern block P2 may be part of the first image IM1 and includes only areas that have features constituting the main pattern A but not the ghost pattern B. Similarly, the ghost pattern block P3 may be part of the first image IM1 and includes only areas that have features constituting the ghost pattern B but not the main pattern A. In some embodiments, multiple background blocks P1, multiple main pattern blocks P2, and multiple ghost pattern blocks P3 can be extracted from the first image IM1. Similarly, one or more background blocks, one or more main pattern blocks, and one or more ghost pattern blocks can be extracted from the second image IM2 (not shown).

[0102] Despite being for such Figure 6 The image shown, with only two patterns, illustrates the extraction of blocks from an electron beam image; however, it should be understood that embodiments of this disclosure can be applied to images with three or more patterns. l To IM n The scene is such that each image includes three or more patterns. If there are three or more patterns in the image generated based on the detection signal, three types of blocks can also be extracted, such as background block P1, main pattern block P2, and ghost pattern block P3. If image IM1 contains main pattern A, first ghost pattern B1, and second ghost pattern B2, then ghost pattern block P3 can include two types of ghost pattern blocks P31 and P32. For example, the first type of ghost pattern P31 can be a part of the image and only includes regions with features that constitute the first ghost pattern B1, but not the main pattern A and the second ghost pattern B2. In this way, even if the number of patterns in the image increases, the computational complexity can be reduced by using blocks with no more than one pattern. In some embodiments, for higher accuracy, regions where multiple patterns overlap can be extracted as blocks. For example, when regions with no more than one pattern in the image are limited and therefore the data for computation is short, blocks with two or more patterns can be used for computation.

[0103] In step S840, multiple images IM1 to IM1 can be generated and obtained from them. n The transformation function associated with the target device. Step S840 can be, for example, by... Figure 5The transformation function can be inferred by receiving the extracted image patches P1, P2, and P3 from step S830 as its input. In some embodiments, information for inferring the transformation function can be obtained from the received image patches P1, P2, and P3. In some embodiments, the information for inferring the transformation function can include, but is not limited to, the positions of the extracted image patches and the detected signals corresponding to the positions of the extracted image patches.

[0104] For simplicity, the process of step S830 will be explained by way of example using only two primary electron sub-beams, such as the first sub-beam and the electron sub-beams 102_1 and 102_2, for scanning. As mentioned above, it should be understood that if the relative positions within the probe point or the corresponding detection element are determined, then the beam intensity and the detection signal at the relative positions can be determined by finding the beam intensity values and the detection signal values at the time points corresponding to the relative positions from the beam intensity such as b1 and the detection signal such as s1. According to embodiments of the present disclosure, the position of the background patch P1 in the first image IM1 can be obtained, as well as the detected signal values corresponding to the position of the background patch P1 as mentioned above. Here, the detected signal at the background patch P1 can be the sum of a portion of the first secondary electron beam 102_1 se and a portion of the second secondary electron beam 102_2 se corresponding to the position of the background patch P1. In some embodiments, the first secondary electron beam 102_1 se and the second secondary electron beam 102_2 se at the position of the background patch P1 can have the same intensity corresponding to the background region of the sample 1. In the same example, the position of the main pattern patch P2 in the first image IM1 and the detected signal values corresponding to the position of the main pattern patch P2 can be obtained. Here, the detected signal at the main pattern patch P2 can be the sum of a portion of the first secondary electron beam 102_1 se and a portion of the second secondary electron beam 102_2 se corresponding to the position of the main pattern patch P2. In some embodiments, the second secondary electron beam 102_2 se at the position of the main pattern patch P can have an intensity corresponding to the background region of the sample 1. The position of the ghost pattern patch P3 in the first image IM1 and the detected signal values corresponding to the position of the ghost pattern patch P3 can be obtained. Here, the detected signal at the ghost pattern patch P2 can be the sum of a portion of the first secondary electron beam 102_1 se and a portion of the second secondary electron beam 102_2 se corresponding to the position of the ghost pattern patch P3. In some embodiments, the first secondary electron beam 102_1 se at the position of the ghost pattern patch P3 can have an intensity corresponding to the background region of the sample. If the secondary electron beams 102_1 se and 102_2 se have constant values over the background region of the sample 1, then the number of variables for inferring the transformation function can be reduced.

[0105] In some embodiments, unsupervised machine learning algorithms can be used to learn the transformation function. For example, referring to Equation 4, the transformation function (corresponding to R) can be learned by using information obtained from image patches as input data and by setting conditions for the learning process. -1 element r 11 to r mm and elements b1 to b m For the example above, the transformation function can have four elements r. 11 to r 22 The signal S can have two elements: a first detection signal s1 from the first detection element 140_1 and a second detection signal s2 from the second detection element 140_2, and the beam intensity I can have two elements b1 and b2, which have the same value corresponding to the background region. Under these input data and conditions, the elements of the transformation function and the elements of the beam intensity I can be inferred using a machine learning algorithm. In some embodiments, the previously learned transformation function and beam intensity can be updated by using subsequent image patches such as P2 and P3. In some embodiments, the transformation function can be generated if a condition for ending the learning process is met. For example, the process can end if the number of iterations becomes equal to a certain value, or if the accuracy of the transformation function reaches a certain level.

[0106] In some embodiments, supervised machine learning algorithms can be used to learn the transformation function. For example, referring to Equation 4, in elements b1 to b m Given that the machine learning algorithm is known, information obtained from image patches is used as input data to learn the transformation function (corresponding to R). -1 element r 11 to r mm For the example above, the transformation function can have four elements r. 11 to r 22 The signal S may have two elements: a first detection signal s1 from the first detection element 140_1 and a second detection signal s2 from the second detection element 140_2, and provides two elements b1 and b2 for the beam intensity I. Under these inputs and conditions, the machine learning network 443 can be configured to infer elements of the transformation function. In some embodiments, the previously learned transformation function can be updated by using subsequent image patches such as P2 and P3, along with beam intensity information. In some embodiments, the time-varying beam intensity information corresponding to each image patch P1 to P3 can be stored in a storage medium and provided to the machine learning algorithm. In some embodiments, the beam intensity information can be obtained from simulations, previous experiments, etc.

[0107] Although targeting Figures 3A-3C The image shown explains Figure 8of the present disclosure can be applied to processes involving the use of three or more images IM1 to IM n and each image includes a scene of two or more patterns. It should be noted that as the number of images IM1 to IM n The computation for obtaining the transform function becomes exponentially complex with the increase in the number of images and the patterns included in the images. Therefore, pre-computing the transform function specific to a particular device and a particular system enables the generation of reconstructed images from detection signals in real-time.

[0108] In some embodiments, the transform function can be determined with respect to a particular e-beam tool 40 or EBI system 100, as the cross-talk contamination is specific to the tool or system itself and not an error in operating the tool or system. Therefore, in accordance with embodiments of the present disclosure, the transform function can be pre-computed with images acquired from the particular tool 40 or system 100 and provided on-demand to reconstruct images from detection signals. In some embodiments, the transform function can be stored in a storage medium (not shown) and can be accessed when needed. In some embodiments, the transform function can be updated when new images are acquired from the tool 40 or system 100. In some embodiments, the transform function can be periodically updated as the degree or level of defocus due to imperfections inherent in the tool 40 or system 100 can change over time. Therefore, the latest transform function can be used to reduce or eliminate the cross-talk contamination.

[0109] Embodiments can be further described using the following clauses:

[0110] 1. A method for enhancing an image, the method comprising:

[0111] acquiring a first image signal of a plurality of image signals from a detector of a multi-beam inspection system, wherein the first image signal corresponds to detection signals from a first region of the detector on which electrons of a first beam of secondary electrons and electrons of a second beam of secondary electrons are incident;

[0112] reducing, from the first image signal, cross-talk contamination originating from the second beam of secondary electrons using a relationship between the first image signal and beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons; and

[0113] generating, after the reducing, a first image corresponding to the first beam of secondary electrons.

[0114] 2. The method of clause 1, wherein a third beam of secondary electrons is also incident on the first region of the detector, and wherein the reducing the cross-talk contamination comprises reducing, from the first image signal, cross-talk contamination originating from the second beam of secondary electrons and the third beam of secondary electrons using a relationship between the first image signal and beam intensities associated with the first beam of secondary electrons, the second beam of secondary electrons, and the third beam of secondary electrons.

[0115] 3. The method of clause 1 or 2, wherein the relationship is in the form of a transformation matrix, and reducing the crosstalk contamination is performed by multiplying the transformation matrix by the first image signal.

[0116] 4. The method of any of clauses 1-3, wherein reducing the crosstalk contamination is performed in real-time.

[0117] 5. The method of any of clauses 1-4, further comprising:

[0118] obtaining the relationship based on a plurality of previous images obtained by the multi-beam inspection system prior to acquiring the first image.

[0119] 6. The method of clause 5, wherein obtaining the relationship further comprises:

[0120] identifying an expected pattern and a non-expected pattern in each of the plurality of previous images;

[0121] extracting a first block covering a region outside of the expected pattern and the non-expected pattern, a second block covering the expected pattern, and a third block covering the non-expected pattern from each of the plurality of images; and generating the relationship using a machine learning network based on the first block, the second block, and the third block.

[0122] 7. The method of clause 6, wherein identifying the expected pattern and the non-expected pattern is performed by using a reference image containing information for determining the expected pattern of the plurality of previous images.

[0123] 8. The method of clause 7, wherein the reference image is a Graphic Database System (GDS) format, a Graphic Database System II (GDS II) format, an Open Artwork System Interchange Standard (OASIS) format, or a California Institute of Technology Intermediate Format (CIF).

[0124] 9. The method of any of clauses 6-8, wherein the non-expected pattern comprises a plurality of non-expected patterns.

[0125] 10. The method of any of clauses 1-9, wherein a portion of the electrons of the second beam of secondary electrons are incident on the first region of the detector.

[0126] 11. The method of any of clauses 1-10, further comprising:

[0127] acquiring a second image signal of the plurality of image signals, wherein the second image signal corresponds to detection signals from a second region of the detector, wherein the electrons of the first beam of secondary electrons and the electrons of the second beam of secondary electrons are incident on the second region;

[0128] reducing, from the second image signal, crosstalk contamination originating from the first secondary electron beam using a relationship between the second image signal and beam intensities associated with the first secondary electron beam and the second secondary electron beam; and

[0129] reducing, from the second image signal, crosstalk contamination originating from the first secondary electron beam using a relationship between the second image signal and beam intensities associated with the first secondary electron beam and the second secondary electron beam; and

[0130] 12. An image enhancement apparatus comprising:

[0131] a memory storing a set of instructions; and at least one processor configured to execute the set of instructions to cause the apparatus to perform:

[0132] reducing, from the second image signal, crosstalk contamination originating from the first secondary electron beam using a relationship between the second image signal and beam intensities associated with the first secondary electron beam and the second secondary electron beam; and

[0133] 13. The apparatus of clause 12, wherein a third secondary electron beam also impinges on the first region of the detector, and wherein reducing the crosstalk contamination comprises reducing, from the first image signal, crosstalk contamination originating from the second secondary electron beam and the third secondary electron beam using a relationship between the first image signal and beam intensities associated with the first secondary electron beam, the second secondary electron beam, and the third secondary electron beam.

[0134] 14. The apparatus of clause 12 or 13, wherein the relationship is in the form of a transformation matrix, and reducing the crosstalk contamination is performed by multiplying the first image signal by the transformation matrix.

[0135] 15. The apparatus of any of clauses 12 to 14, wherein reducing the crosstalk contamination is performed in real-time.

[0136] 16. The apparatus of any of clauses 12 to 15, wherein prior to acquiring the first image, the at least one processor is configured to execute the set of instructions to cause the apparatus to further perform:

[0137] obtaining the relationship based on a plurality of previous images acquired by the multi-beam inspection system.

[0138] 17. The apparatus of clause 16, wherein obtaining the relationship further comprises:

[0139] identifying an expected pattern and an unexpected pattern in each of the plurality of previous images;

[0140] extracting, from each of the plurality of images, a first block covering a region outside of the expected pattern and the unexpected pattern, a second block covering the expected pattern, and a third block covering the unexpected pattern; and generating the relationship using a machine learning network based on the first block, the second block, and the third block.

[0141] 18. The apparatus of clause 17, wherein identifying the expected pattern and the unexpected pattern is performed using a reference image containing information for determining the expected pattern for a plurality of previous images.

[0142] 19. The apparatus of clause 18, wherein the reference image is a Graph Database System (GDS) format, a Graph Database System II (GDS II) format, an Open Art System Interchange Standard (OASIS) format, or a California Institute of Technology Intermediate Format (CIF).

[0143] 20. The apparatus of any one of clauses 16-19, wherein the unexpected pattern comprises a plurality of unexpected patterns.

[0144] 21. The apparatus of any one of clauses 12-20, wherein a portion of the electrons of the second beam of secondary electrons are incident on the first region of the detector.

[0145] 22. The apparatus of any one of clauses 12-21, wherein the at least one processor is configured to execute the set of instructions to cause the apparatus to further perform:

[0146] obtaining a second image signal of the plurality of image signals, wherein the second image signal corresponds to detection signals from a second region of the detector, wherein the electrons of the first beam of secondary electrons and the electrons of the second beam of secondary electrons are incident on the second region;

[0147] reducing, from the second image signal, crosstalk contamination originating from the first beam of secondary electrons using a relationship between the second image signal and the beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons; and

[0148] generating, after the reducing, a second image corresponding to the second beam of secondary electrons.

[0149] 23. A non-transitory computer-readable medium storing a set of instructions executable by at least one processor of a computing device to cause performance of a method for enhancing an image, the method comprising:

[0150] obtaining a first image signal of a plurality of image signals from a detector of a multi-beam inspection system, wherein the first image signal corresponds to detection signals from a first region of the detector, wherein the electrons of the first beam of secondary electrons and the electrons of the second beam of secondary electrons are incident on the first region;

[0151] reducing crosstalk contamination from the second beam of secondary electrons from the first image signal using a relationship between the first image signal and beam intensities associated with the first and second beams of secondary electrons; and

[0152] generating a first image corresponding to the first beam of secondary electrons after the reducing.

[0153] 24. The computer readable medium of clause 23, wherein a third beam of secondary electrons also impinges on the first region of the detector, and wherein reducing crosstalk contamination comprises reducing crosstalk contamination from the second and third beams of secondary electrons from the first image signal using a relationship between the first image signal and beam intensities associated with the first, second, and third beams of secondary electrons.

[0154] 25. The apparatus of clause 23 or 24, wherein the relationship is in the form of a transformation matrix, and reducing crosstalk contamination is performed by multiplying the transformation matrix by the first image signal.

[0155] 26. The computer readable medium of any of clauses 23 to 25, wherein reducing crosstalk contamination is performed in real time.

[0156] 27. The computer readable medium of any of clauses 23 to 26, wherein the set of instructions, prior to acquiring the first image, are executable by the at least one processor of the computing device to cause the computing device to further perform:

[0157] obtaining the relationship based on a plurality of previous images acquired by the multi-beam inspection system.

[0158] 28. The computer readable medium of clause 27, wherein obtaining the relationship further comprises:

[0159] identifying expected patterns and unexpected patterns in each of the plurality of previous images; extracting a first block covering regions outside of the expected patterns and unexpected patterns, a second block covering the expected patterns, and a third block covering the unexpected patterns from each of the plurality of images; and generating the relationship using a machine learning network based on the first, second, and third blocks.

[0160] 29. The computer readable medium of clause 28, wherein identifying the expected patterns and unexpected patterns is performed by using a reference image containing information for determining the expected patterns of the plurality of previous images.

[0161] 30. The computer readable medium of clause 29, wherein the reference image is a Graphic Database System (GDS) format, a Graphic Database System II (GDS II) format, an Open Artwork System Interchange Standard (OASIS) format, or a California Institute of Technology Intermediate Format (CIF).

[0162] 31. The computer-readable medium of any of clauses 28-30, wherein the unintended pattern comprises a plurality of unintended patterns.

[0163] 32. The computer-readable medium of any of clauses 23-31, wherein a portion of the electrons of the second beam of secondary electrons are incident on the first region of the detector.

[0164] 33. The computer-readable medium of any of clauses 23-32, wherein the set of instructions is executable by the at least one processor of the computing device to cause the computing device to further perform:

[0165] obtaining a second image signal of the plurality of image signals, wherein the second image signal corresponds to detection signals from a second region of the detector, wherein the second region is incident with electrons of the first beam of secondary electrons and electrons of the second beam of secondary electrons;

[0166] reducing, from the second image signal, crosstalk contamination originating from the first beam of secondary electrons using a relationship between the second image signal and beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons; and

[0167] generating, after the reducing, a second image corresponding to the second beam of secondary electrons.

[0168] 34. A method for reducing crosstalk contamination in a multi-beam inspection system, the method comprising:

[0169] obtaining, by the multi-beam inspection system, a first image and a second image of a first region and a second region on a sample, respectively, wherein the first image is generated based on first detection signals from a first detection region of the multi-beam inspection system, and the second image is generated based on second detection signals from a second detection region of the multi-beam inspection system;

[0170] determining, by using a first reference image corresponding to the first region, a primary pattern of the first image originating from a first beam of secondary electrons;

[0171] determining, by using a second reference image corresponding to the second region, whether the first image includes an artifact pattern originating from a second beam of secondary electrons;

[0172] determining a relationship between the first detection signals and beam intensities of the first beam of secondary electrons and the second beam of secondary electrons; and

[0173] reducing, based on the determined relationship, crosstalk contamination of third detection signals from the first detection region.

[0174] 35. The method of clause 34, wherein the reducing crosstalk contamination is performed in real-time after the obtaining the third detection signals.

[0175] 36. The method of clause 34 or 35, wherein determining the relationship is performed based on analyzing image tiles of the first image, wherein the image tiles include a first image tile covering a region outside of the primary pattern and the ghost pattern, a second image tile covering the primary pattern, and a third image tile covering the ghost pattern.

[0176] 37. A method for enhancing an image, the method comprising:

[0177] obtaining a first image from a detector of a multi-beam inspection system, wherein the first image corresponds to first detection signals from a first region of the detector, wherein electrons of a first beam of secondary electrons and electrons of a second beam of secondary electrons are incident on the first region; and

[0178] reducing a cross-talk image originating from the second beam of secondary electrons from the first image based on a relationship between the first detection signals and beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons.

[0179] 38. The method of clause 37, wherein the relationship is obtained by:

[0180] obtaining a first previous image and a second previous image obtained by the multi-beam inspection system;

[0181] identifying, based on a reference image containing information for determining an expected pattern of the first and second previous images, the expected pattern and an unexpected pattern in each of the first and second previous images;

[0182] extracting, from each of the first and second previous images, a first tile covering a region outside of the expected pattern and the unexpected pattern, a second tile covering the expected pattern, and a third tile covering the unexpected pattern; and

[0183] generating the relationship using a machine learning network based on the first tile, the second tile, and the third tile.

[0184] 39. The method of clause 38, wherein the reference image is a Graphic Database System (GDS) format, a Graphic Database System II (GDS II) format, an Open Artwork System Interchange Standard (OASIS) format, or a California Institute of Technology Intermediate Format (CIF).

[0185] A storage medium storing instructions for a controller (e.g., Figure 1instructions of the processor of the controller 50) to perform image inspection, image acquisition, stage positioning, beam focusing, electric field adjustment, beam bending, condenser lens adjustment, activating a charged particle source, beam deflection, etc. Common forms of non-transitory media include, for example, a floppy disk, a flexible disk, a hard disk, a solid-state drive, a magnetic tape, or any other magnetic data storage medium, a Compact Disc Read Only Memory (CD-ROM), any other optical data storage medium, any physical medium with patterns of holes, a Random Access Memory (RAM), a Programmable Read Only Memory (PROM), and Erasable Programmable Read Only Memory (EPROM), a FLASH- EPROM or any other flash memory, a Non-Volatile Random Access Memory (NVRAM), a cache, a register, any other storage chips or cartridges, and networked versions of the same.

[0186] It is to be understood that the embodiments of the present disclosure are not limited to the exact structures described above and shown in the drawings, and that various modifications and changes can be made without departing from the scope thereof. The present disclosure has been described in connection with various embodiments, and it will be apparent to those skilled in the art that variations in the foregoing details are possible without departing from the scope of the application as disclosed herein. The description and examples are intended to be illustrative, and the true scope and spirit of the present application is indicated by the appended claims.

[0187] The above description is intended to be illustrative, and not restrictive. Therefore, it will be apparent to persons skilled in the art that modifications can be made to the described embodiments without departing from the scope of the claims as set forth below.

Claims

1. A method for enhancing an image, the method comprising: acquiring, from a detector of a multi-beam inspection system, a first image signal of a plurality of image signals, wherein the first image signal corresponds to detection signals from a first area of the detector, wherein electrons of a first beam of secondary electrons and electrons of a second beam of secondary electrons are incident on the first area; reducing, from the first image signal, crosstalk contamination originating from the second beam of secondary electrons using a relationship between the first image signal and beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons; and generating, after the reducing, a first image corresponding to the first beam of secondary electrons.

2. An image enhancement apparatus comprising: a memory storing a set of instructions; and at least one processor configured to execute the set of instructions to cause the apparatus to perform: acquiring, from a detector of a multi-beam inspection system, a first image signal of a plurality of image signals, wherein the first image signal corresponds to detection signals from a first area of the detector, wherein electrons of a first beam of secondary electrons and electrons of a second beam of secondary electrons are incident on the first area; reducing, from the first image signal, crosstalk contamination originating from the second beam of secondary electrons using a relationship between the first image signal and beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons; and generating, after the reducing, a first image corresponding to the first beam of secondary electrons.

3. The apparatus of claim 2, wherein a third beam of secondary electrons is also incident on the first area of the detector, and wherein reducing crosstalk contamination comprises reducing, from the first image signal, crosstalk contamination originating from the second beam of secondary electrons and the third beam of secondary electrons using a relationship between the first image signal and beam intensities associated with the first beam of secondary electrons, the second beam of secondary electrons, and the third beam of secondary electrons.

4. The apparatus of claim 2, wherein the relationship is in the form of a transformation matrix, and reducing crosstalk contamination is performed by multiplying the transformation matrix by the first image signal.

5. The apparatus of claim 2, wherein reducing the crosstalk contamination is performed in real-time.

6. The apparatus of claim 2, wherein, prior to acquiring the first image, the at least one processor is configured to execute the set of instructions to cause the apparatus to further perform: obtaining the relationship based on a plurality of previous images acquired by the multi-beam inspection system.

7. The apparatus of claim 6, wherein obtaining the relationship further comprises: identifying an expected pattern and an unexpected pattern in each of the plurality of previous images; extracting, from each of the plurality of images, a first block covering an area outside of the expected pattern and the unexpected pattern, a second block covering the expected pattern, and a third block covering the unexpected pattern; and generating the relationship using a machine learning network based on the first block, the second block, and the third block. ​ ​ ​ 8. The apparatus of claim 7, wherein identifying the expected pattern and the unexpected pattern is performed using a reference image, the reference image containing information used to determine an expected pattern for the plurality of previous images.

9. The apparatus of claim 8, wherein the reference image is a Graphic Database System (GDS) format, a Graphic Database System II (GDSII) format, an Open Artwork System Exchange Standard (OASIS) format, or a Caltech Intermediate Format (CIF).

10. The apparatus of claim 7, wherein the unexpected pattern comprises a plurality of unexpected patterns.

11. The apparatus of claim 2, wherein a portion of electrons of the second beam of secondary electrons are incident on the first region of the detector.

12. The apparatus of claim 2, wherein the at least one processor is configured to execute the set of instructions to cause the apparatus to further perform: obtaining a second image signal of the plurality of image signals, wherein the second image signal corresponds to detection signals from a second region of the detector on which electrons of the first beam of secondary electrons and the second beam of secondary electrons are incident; reducing, from the second image signal, crosstalk contamination originating from the first beam of secondary electrons using a relationship between the second image signal and beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons; and generating, after the reducing, a second image corresponding to the second beam of secondary electrons.

13. A non-transitory computer-readable medium storing a set of instructions executable by at least one processor of a computing device to cause performance of a method for enhancing images, the method comprising: obtaining a first image signal of a plurality of image signals from a detector of a multi-beam inspection system, wherein the first image signal corresponds to detection signals from a first region of the detector on which electrons of a first beam of secondary electrons and a second beam of secondary electrons are incident; reducing, from the first image signal, crosstalk contamination originating from the second beam of secondary electrons using a relationship between the first image signal and beam intensities associated with the first beam of secondary electrons and the second beam of secondary electrons; and generating, after the reducing, a first image corresponding to the first beam of secondary electrons.

14. The computer-readable medium of claim 13, wherein a third beam of secondary electrons is also incident on the first region of the detector, and wherein reducing crosstalk contamination comprises reducing, from the first image signal, crosstalk contamination originating from the second beam of secondary electrons and the third beam of secondary electrons using a relationship between the first image signal and beam intensities associated with the first beam of secondary electrons, the second beam of secondary electrons, and the third beam of secondary electrons.

15. The computer-readable medium of claim 13, wherein the relationship is in the form of a transformation matrix, and reducing crosstalk contamination is performed by multiplying the transformation matrix by the first image signal. ​ ​

Images