Defect detection on the sample

Abstract

A method and system for detecting defects on a sample are provided. One system generates a final reference image by calculating different candidate reference images from different composites of sample images generated by an inspection subsystem and compositing the different parts of the candidate reference images without altering them. The final reference image is then used for defect detection, which may be single or double detection. The embodiment is particularly useful for defect detection in areas of a sample containing only repetitive device patterns that are not resolvable, such as cell regions, but can also be used for inspecting other types of regions.

Description

Technical Field

[0001] The present invention generally relates to methods and systems for detecting defects on a sample in semiconductor technology.

Background Art

[0002] The following explanations and examples are not admitted to be prior art for the reason that they are included in this section.

[0003] Manufacturing semiconductor devices such as logic devices and memory devices typically involves processing a sample such as a semiconductor wafer using a number of semiconductor manufacturing processes to form various features and multiple levels of the semiconductor device. For example, lithography is a semiconductor manufacturing process that typically includes transferring a pattern to a resist disposed on a semiconductor wafer. Additional examples of semiconductor manufacturing processes include, but are not limited to, chemical mechanical polishing, etching, deposition, and ion implantation. Multiple semiconductor devices may be fabricated in an array on a single semiconductor wafer and then diced into individual semiconductor devices.

[0004] Inspection using either optical or electron beam imaging is an important technique for debugging semiconductor manufacturing processes, monitoring process variations, and improving production yields in the semiconductor industry. As the scale of modern integrated circuits (ICs) continues to decrease and the complexity of the manufacturing process increases, inspection is becoming increasingly difficult.

[0005] In each processing step performed on a semiconductor wafer, the same circuit pattern is printed on each die. Most wafer inspection systems utilize this fact to detect defects on the wafer using relatively simple comparisons between dies. However, the printed circuitry on each die may contain many areas where patterned features are repeated in the x or y direction, such as DRAM, SRAM, or FLASH regions. This type of region is generally called an "array region" (the remaining regions are called "random regions" or "logic regions"). To achieve better sensitivity, advanced inspection systems employ different strategies for inspecting array regions and random or logic regions.

[0006] Array detection algorithms are designed to achieve relatively high sensitivity to DRAM cell regions by utilizing the reproducibility of cell regions. For example, inspection systems configured for array region inspection often perform cell-to-cell comparisons, where images of different cells within an array region on the same die are subtracted from each other, and the difference is examined for defects. This array inspection strategy can achieve much higher sensitivity in the array region than random inspection (which is usually performed by subtracting the image of one die from the image of another die) because it avoids noise caused by variability between dies.

[0007] Many array defect detection methods currently in use divide the entire care area into several units called cells. All cells are placed in independent groups, and each group is used to generate one reference cell, which forms the basis of the final reference image. Current array defect detection methods typically assume that all cells within each group are identical except for some random noise. The grayscale at the same location in all cells fluctuates around noise-free ground truth. However, because reflectivity from different wafer regions varies considerably, this assumption of a uniform grayscale may not always be correct. For example, relatively strong leakage from the edge regions of a cell can result in relatively low-frequency grayscale fluctuations, which may ultimately lead to relatively low inspection sensitivity. Furthermore, current array detection methods treat each care area as a single detection region. By grouping different locations together, it may be difficult to achieve a proper balance between sensitivity and noise control. In settings where the highest sensitivity can be achieved in the inner regions, a large number of noises may be generated from the edge / corner regions.

[0008] Based on the results of two detection trials, the random defect detection algorithm demonstrates superior performance compared to relying on a single detection. This dual detection capability is also beneficial for reducing newsworthiness in array defect detection methods, although it is not currently used in array defect detection methods. [Prior art documents] [Patent Documents]

[0009] [Patent Document 1] U.S. Patent Application Publication No. 2017 / 0140516 [Patent Document 2] U.S. Patent Application Publication No. 2021 / 0334989 [Overview of the Initiative] [Problems that the invention aims to solve]

[0010] Therefore, it is advantageous to develop systems and methods for testing samples that do not have one or more of the above-mentioned drawbacks. [Means for solving the problem]

[0011] The following descriptions of various embodiments should not be construed as limiting the subject matter of the invention in the attached claims.

[0012] One embodiment relates to a system configured to detect defects on a sample. The system includes an inspection subsystem configured to generate an image of the sample, which includes a test image and two or more other images corresponding to the test image. The system also includes a computer subsystem configured to calculate first and second candidate reference images from at least two different composites of the test image and the two or more other images. The computer subsystem is also configured to select at least a portion of the first candidate reference image corresponding to a first portion of the test image and a portion of the second candidate reference image corresponding to a second portion of the test image. Furthermore, the computer subsystem is configured to synthesize the selected portions of the first and second candidate reference images without altering the selected portions of the first and second candidate reference images to generate a final reference image. The computer subsystem is further configured to generate a difference image by comparing the test image and the final reference image and to detect defects in the test image by applying a defect detection method to the difference image. The system may be further configured as described herein.

[0013] Another embodiment relates to a computer implementation method for detecting defects on a sample. The method includes acquiring an image of the sample, which includes a test image and two or more other images corresponding to the test image, generated by an inspection subsystem. The method also includes the calculation, selection, synthesis, generation, and detection steps described above. Acquisition, calculation, selection, synthesis, generation, and detection are performed by a computer subsystem coupled to the inspection subsystem. Each step of the method described above may be performed as further described herein. In addition, embodiments of the method described above may include any other steps of any other method described herein. The method described above may be performed by any of the systems described herein.

[0014] Another embodiment relates to a non-temporary computer-readable medium which stores program instructions that cause a computer implementation method for detecting defects on a sample to be executed on a computer system. The computer implementation method includes the steps of the method described above. The computer-readable medium may be further configured as described herein. The steps of the computer implementation method may be performed as further described herein. In addition, the computer implementation method on which the program instructions can be executed may include any other steps of any other method described herein.

[0015] Further advantages of the present invention will become apparent to those skilled in the art through the following detailed description of preferred embodiments and the accompanying drawings. [Brief explanation of the drawing]

[0016] [Figure 1] This is a schematic diagram showing a side view of an embodiment of the system configured as described herein. [Figure 1a] This is a schematic diagram showing a side view of an embodiment of the system configured as described herein. [Figure 2]A schematic diagram showing a plan view of an example of a test image, a corresponding reference image, and a difference image generated therefrom, which are generated by a currently used inspection method, and an embodiment of a final reference image generated as described herein and a difference image generated therefrom. [Figure 3] A schematic diagram showing a plan view of an example of an image generated for a sample and various embodiments for dividing the image into a test cell and two or more other cells adjacent to the test cell. [Figure 4] A flowchart showing an embodiment of steps that can be performed to detect defects on a sample. [Figure 5] A schematic diagram showing a plan view of an example of a job including image frames of different cell regions on a sample. [Figure 6] A schematic diagram showing a plan view of an embodiment of different candidate reference images in which different portions corresponding to different portions of a test image are selected, combined without being changed, thereby generating a final reference image. [Figure 7] A schematic diagram showing a plan view of an embodiment of a care area divided into different segments and a plot showing an embodiment of individually detecting defects in each of the different segments. [Figure 8] A block diagram showing an embodiment of a non - transient computer - readable medium storing program instructions for causing a computer system to execute the computer - implemented method described herein.

Best Mode for Carrying Out the Invention

[0017] The present invention is capable of various changes and alternative forms. However, specific embodiments thereof are shown by way of example in the drawings and are described in detail herein. The drawings may not be to scale. However, it should be understood that the drawings and their detailed description are not intended to limit the present invention to the specific forms disclosed, but rather the intention is to cover all changes, equivalents, and alternatives falling within the spirit and scope of the present invention as defined by the appended claims.

[0018] Turning our attention to the drawings, it should be noted that they are not drawn to a consistent scale. In particular, the scale of some of the elements in the drawings is greatly exaggerated to emphasize the characteristics of those elements. Also, note that the drawings are not drawn to the same scale. Elements shown in two or more drawings that can be similarly constructed are indicated using the same reference number. Unless otherwise specified herein, any of the elements described and illustrated may include any suitable commercially available elements.

[0019] Generally, the embodiments described herein are configured to detect defects on a sample. The embodiments described herein provide novel reference image generation and detection methods, which are particularly useful in non-pattern detection areas. Current methods used for defect detection in non-pattern array (NPA) areas assume that the entire inspection area has a uniform grayscale and that only wafer / imaging noise causes variations in the gray level. However, the presence of color variations due to the reflectivity of different materials and structures overturns this assumption. If color variations are not properly handled, they lead to additional noise, which can sometimes become louder than the target defect (DOI) signal, potentially reducing the sensitivity of NPA defect detection. Based on these findings, the steps further described herein are proposed to improve defect detection performance in non-pattern areas.

[0020] As used herein, the term “newsans” (sometimes used interchangeably with “newsans defect”) is generally defined as an event detected on a sample that is not actually a defect on the sample. Non-defective newsans may be detected as events due to non-defect noise sources on the sample (e.g., particles of metal wiring on the sample, signals from underlying layers or materials on the sample, line edge roughness (LER), relatively small critical dimension (CD) variations in patterned features, thickness variations, etc.), and / or events due to boundary elements in the inspection system itself or its configuration used for inspection.

[0021] The terms “First” and “Second” as used herein are used solely for ease of distinguishing between different things, and these terms do not imply any other implication of the embodiments described herein.

[0022] In some embodiments, the sample is a wafer. The wafer may include any wafer known in semiconductor technology. In some embodiments, this specification may describe one or more wafers, but embodiments are not limited to the samples that can be used. For example, embodiments described herein may be used for samples such as reticles, flat panels, personal computer (PC) boards, and other semiconductor samples.

[0023] Figure 1 shows one embodiment of a system configured to detect defects on a sample. This system includes an inspection subsystem 100. In the embodiment shown in Figure 1, the inspection subsystem is configured as an optical inspection subsystem. However, in other embodiments described herein, the inspection subsystem is configured as an electron beam or charged particle beam-based inspection subsystem. Generally, the inspection subsystem described herein includes at least an energy source and a detector. The energy source is configured to generate energy directed towards the sample. The detector is configured to detect energy from the sample and generate an output in accordance with the detected energy.

[0024] In a light-based inspection subsystem, the energy directed towards the sample includes light, and the energy detected from the sample also includes light. For example, in the embodiment of the system shown in Figure 1, the inspection subsystem includes an illumination subsystem configured to direct light onto the sample 14. The illumination subsystem includes at least one light source. For example, as shown in Figure 1, the illumination subsystem includes a light source 16. The illumination subsystem is configured to direct light onto the sample at one or more angles of incidence, which may include one or more oblique angles and / or one or more normal angles. For example, as shown in Figure 1, light from the light source 16 passes through an optical element 18 and then through a lens 20 to be directed onto the sample 14 at an oblique angle of incidence. The oblique angle of incidence may include any suitable angle of incidence, which may vary depending, for example, on the characteristics of the sample and on the defects detected on the sample.

[0025] The illumination subsystem may be configured to direct light onto the sample at different angles of incidence and at different times. For example, the inspection subsystem may be configured to modify one or more properties of one or more elements of the illumination subsystem so that light can be directed onto the sample at angles of incidence different from those shown in Figure 1. In such an example, the inspection subsystem may be configured to move the light source 16, optical element 18, and lens 20 so that light can be directed onto the sample at different oblique or normal (or nearly normal) angles of incidence.

[0026] The inspection subsystem may be configured to direct light onto the sample simultaneously at two or more angles of incidence. For example, the illumination subsystem may include two or more illumination channels, one of which may include a light source 16, an optical element 18, and a lens 20, as shown in Figure 1, and another illumination channel (not shown) may include similar elements, which may be different or identical in configuration, or may include at least one light source and, optionally, one or more other components as further described herein. When such light is directed onto the sample simultaneously with other light, one or more properties (e.g., wavelength, polarization, etc.) of the light directed onto the sample at different angles of incidence may be different, and as a result, the light produced by the illumination of the sample at different angles of incidence may be distinguishable from one another in the detector.

[0027] The illumination subsystem may include only one light source (e.g., light source 16 shown in Figure 1), and the light from the light source may be split into different optical paths (e.g., based on wavelength, polarization, etc.) by optical elements (not shown) of the illumination subsystem. Each of the light from the different optical paths may then be directed to the sample. Multiple illumination channels may be configured to direct light to the sample simultaneously or at different times (e.g., sequentially illuminating the sample using different illumination channels). In another example, the same illumination channel may be configured to direct light to the sample with different characteristics at different times. For example, optical element 18 may be configured as a spectral filter, and the characteristics of the spectral filter may be changed in various different ways (e.g., by swapping one spectral filter with another), so that different wavelengths of light can be directed to the sample at different times. The illumination subsystem may have any other suitable configuration known in the art to direct light with different or the same characteristics to the sample sequentially or simultaneously at different or the same angle of incidence.

[0028] Light source 16 may include a broadband plasma (BBP) light source. Thus, the light generated by the light source and directed towards the sample may include broadband light. However, the light source may include any other suitable light source, such as a laser. The laser may include any suitable laser known in the art and may be configured to generate light at any suitable wavelength known in the art. In addition, the laser may be configured to generate monochromatic or nearly monochromatic light. For example, the laser may be a narrowband laser. The light source may also include a polychromatic light source that generates light at multiple discrete wavelengths or wavelength bands.

[0029] Light from the optical element 18 can be focused onto the sample 14 by the lens 20. Although the lens 20 is shown as a single refractive optical element in Figure 1, in practice the lens 20 may include multiple refractive optical elements and / or reflective optical elements, which, in combination, focus the light from the optical elements onto the sample. The illumination subsystem shown in Figure 1 and described herein may include any other suitable optical elements (not shown). Examples of such optical elements include, but are not limited to, polarization components, spectral filters, spatial filters, reflective optical elements, apodizers, beam splitters, and apertures, and may include any suitable optical elements well known in the art. In addition, the system may be configured to change one or more elements of the illumination subsystem based on the type of illumination used for inspection.

[0030] The inspection subsystem may also include a scanning subsystem configured to change the position on the sample from which light is directed and detected, and possibly to scan the light on the sample. For example, the inspection subsystem may include a stage 22 on which the sample 14 is placed during inspection. The scanning subsystem may include any suitable mechanical and / or robotic assembly (including the stage 22) configured to move the sample, thereby directing light to different positions on the sample from which it can be detected. In addition, or alternatively, the inspection subsystem may be configured such that one or more optical elements of the inspection subsystem perform any scanning of light on the sample, thereby directing light to different positions on the sample from which it can be detected. The light may scan the sample in any suitable way, for example, a meandering path or a helical path.

[0031] The inspection subsystem may further include one or more detection channels. At least one of the detection channels includes a detector configured to detect light from a sample due to illumination of the sample by the system and to produce an output in response to the detected light. For example, the detection subsystem shown in Figure 1 includes two detection channels, one formed by a light collector 24, an element 26, and a detector 28, and the other formed by a light collector 30, an element 32, and a detector 34. As shown in Figure 1, the two detection channels are configured to collect and detect light at different focusing angles. In some cases, both detection channels are configured to detect scattered light, and the detection channels are configured to detect light scattered from the sample at different angles. However, one or more of the detection channels may be configured to detect another type of light from the sample (e.g., reflected light).

[0032] As further shown in Figure 1, both detection channels are shown positioned within the plane of the paper, and the illumination subsystem is also shown positioned within the plane of the paper. Therefore, both detection channels are positioned within the plane of incidence (e.g., at the center). However, one or more detection channels may be positioned outside the plane of incidence. For example, a detection channel formed by the concentrator 30, element 32, and detector 34 may be configured to collect and detect light scattered outside the plane of incidence. Such a detection channel may therefore be generally called a “side” channel, and such a side channel may be positioned at the center in a plane substantially perpendicular to the plane of incidence.

[0033] The inspection subsystem may include a different number of detection channels than those shown in Figure 1 (e.g., only one detection channel, or two or more detection channels). In such an example, the detection channel formed by the condenser 30, element 32, and detector 34 may form one side channel as described above, and the inspection subsystem may include an additional detection channel (not shown) formed as another side channel located on the opposite side of the incident plane. Thus, the inspection subsystem may include a detection channel which includes the condenser 24, element 26, and detector 28, positioned at the center of the incident plane and configured to collect and detect light at a scattering angle perpendicular or nearly perpendicular to the sample plane. Thus, this detection channel may generally be called the “top” channel, and the inspection subsystem may also include two or more side channels configured as described above. In this way, the inspection subsystem may include at least three channels (i.e., one top channel and two side channels), each of the at least three channels having its own condenser, each condenser configured to collect light at a different scattering angle than each of the other condensers.

[0034] Furthermore, as described above, each detection channel included in the inspection subsystem may be configured to detect scattered light. Therefore, the inspection subsystem shown in Figure 1 may be configured for dark-field (DF) inspection of a sample. However, the inspection subsystem may further, or alternatively, include detection channels configured for bright-field (BF) inspection of a sample. In other words, the inspection subsystem may include at least one detection channel configured to detect light specularly reflected from the sample. Therefore, the inspection subsystems described herein may be configured for DF inspection only, BF inspection only, or both DF and BF inspections. While each concentrator is shown as a single refractive optical element in Figure 1, it should be understood that each concentrator may include one or more refractive optical elements and / or one or more reflective optical elements.

[0035] One or more detection channels may include any suitable detector known in the art. For example, detectors may include photomultiplier tubes (PMTs), charge-coupled devices (CCDs), and time-delay integral (TDI) cameras. Detectors may also include non-imaging detectors or imaging detectors. If the detectors are non-imaging detectors, each detector may be configured to detect certain properties of scattered light, such as intensity, but not necessarily as a function of position in the imaging plane. Thus, the output produced by each detector included in each detection channel may be a signal or data, but not necessarily an image signal or image data. In such cases, a computer subsystem, such as computer subsystem 36, which forms part of the inspection system together with the inspection subsystem, may be configured to generate an image of the sample from the non-imaging output of the detectors. However, in other cases, the detectors may be configured as imaging detectors configured to generate imaging signals or image data. Thus, the inspection subsystem may be configured to generate images in many ways.

[0036] Note that Figure 1 is provided herein to generally illustrate the configuration of a test subsystem that may be included in embodiments of the system described herein. Obviously, the configuration of the test subsystem described herein can be modified to optimize the performance of the test subsystem, as is typically done when designing a commercially available test system. In addition, the system described herein can be implemented using an existing test system (for example, by adding the functions described herein to an existing test system), such as the 29xx / 39xx series tools commercially available from KLA Corporation in Milpitas, California. In such a system, the method described herein may be offered as an optional feature of the test system (for example, in addition to other functions of the test system). Alternatively, the test system described herein can be designed "from scratch" to provide a completely new test system.

[0037] The computer subsystem 36 may be coupled in any suitable way (for example, via one or more transmission media which may include wired and / or wireless transmission media) to the detector of the inspection subsystem, so that the computer subsystem can receive the output generated by the detector. The computer subsystem 36 may be configured to perform several functions using the output of the detector, as further described herein. The computer subsystem coupled to the inspection subsystem may be further configured as described herein.

[0038] Computer subsystems coupled to the inspection subsystem (and other computer subsystems described herein) may also be referred to herein as computer systems. Each of the computer subsystems or systems described herein can take various forms, including personal computer systems, image computers, mainframe computer systems, workstations, network appliances, internet appliances, or other devices. Generally, the term “computer system” can be broadly defined to encompass any device having one or more processors, where processors execute instructions from a memory medium. A computer subsystem or system may also include any suitable processors well known in the art, such as parallel processors. In addition, a computer subsystem or system may include a computer platform with high-speed processing and software, either as a standalone or networked tool.

[0039] If a system includes two or more computer subsystems, different computer subsystems may be coupled to one another, and images, data, information, instructions, etc., may be transmitted between them. For example, computer subsystem 36 may be coupled to computer system 102 by any suitable transmission medium, as shown by the dashed line in Figure 1, and the transmission medium may include any suitable wired and / or wireless transmission medium known in the art. Two or more such computer subsystems may be effectively coupled by a shared computer-readable storage medium (not shown).

[0040] Although the inspection subsystem has been described above as an optical or light-based inspection subsystem, in another embodiment, the inspection subsystem is configured as an electron beam-based inspection subsystem. In an electron beam inspection subsystem, the energy directed at the sample includes electrons, and the energy detected from the sample includes electrons. In one such embodiment shown in Figure 1a, the inspection subsystem includes an electron column 122, and the system includes a computer subsystem 124 coupled to the inspection subsystem. The computer subsystem 124 may be configured as described above. In addition, such an inspection subsystem may be coupled to one or more other computer subsystems in the same manner as described above and in Figure 1.

[0041] As also shown in Figure 1a, the electron column includes an electron beam source 126 configured to generate electrons focused on a sample 128 by one or more elements 130. The electron beam source may include, for example, a cathode source or an emitter tip, and the one or more elements 130 may include, for example, a gun lens, an anode, a beam limiting aperture, a gate valve, a beam current selecting aperture, an objective lens, and a scanning subsystem, all of which may include any suitable elements known in the art.

[0042] Electrons returning from the sample (e.g., secondary electrons) can be focused onto the detector 134 by one or more elements 132. One or more elements 132 may include, for example, a scanning subsystem, which may be the same scanning subsystem as that included in element 130.

[0043] The electronic column may include any other suitable elements well known in the art. In addition, the electronic column may be further constructed as described in U.S. Patent No. 8,664,594 issued to Jiang et al. on April 4, 2014, U.S. Patent No. 8,692,204 issued to Kojima et al. on April 8, 2014, U.S. Patent No. 8,698,093 issued to Gubbens et al. on April 15, 2014, and U.S. Patent No. 8,716,662 issued to MacDonald et al. on May 6, 2014, which are incorporated by reference as if they were fully discussed herein.

[0044] In Figure 1a, the electron column is shown configured such that electrons are directed towards the sample at an oblique incidence angle and scattered from the sample at another oblique angle; however, the electron beam can be directed towards the sample at any suitable angle and scattered from the sample. In addition, the electron beam inspection subsystem can be configured to generate output to the sample using multiple modes (e.g., using different illumination angles, focusing angles, etc.), as further described herein. The multiple modes of the electron beam inspection subsystem may differ with respect to any output generation parameters of the inspection subsystem.

[0045] The computer subsystem 124 may be coupled to the detector 134 as described above. The detector may detect electrons returning from the surface of the sample, thereby forming an electron beam image (or other output to the sample) of the sample. The electron beam image may include any suitable electron beam image. The computer subsystem 124 may be configured to perform any step described herein. The system including the inspection subsystem shown in Figure 1a may be further configured as described herein.

[0046] Figure 1a is provided herein to generally illustrate the configuration of an electron beam inspection subsystem that may be included in the embodiments described herein. Similar to the optical inspection subsystem described above, the configuration of the electron beam inspection subsystem described herein may be modified to optimize the performance of the inspection subsystem, which is typically done when designing a commercially available inspection system. In addition, the systems described herein may be implemented using existing inspection systems, such as commercially available tools from KLA (for example, by adding the functionality described herein to an existing inspection system). For some such systems, the methods described herein may be offered as optional features of the system (for example, in addition to other functionality of the system). Alternatively, the systems described herein may be designed "from scratch" to provide a completely new system.

[0047] Although the inspection subsystem is described above as an optical or electron beam inspection subsystem, the inspection subsystem may also be an ion beam inspection subsystem. Such an inspection subsystem may be configured as shown in Figure 1a, except that the electron beam source can be replaced with any suitable ion beam source known in the art. In addition, the inspection subsystem may include any other suitable ion beam system, such as those included in commercially available focused ion beam (FIB) systems, helium ion microscope (HIM) systems, and secondary ion mass spectrometry (SIMS) systems.

[0048] The inspection subsystem described herein may be configured to generate an output of a sample, such as an image, in multiple modes. Generally, a “mode” is defined by the values ​​of the parameters of the inspection subsystem used to generate the output of the sample and / or an image (or an output used to generate an image of the sample). Thus, modes may differ with respect to at least one value of the parameters of the inspection subsystem (except the position on the sample where the output is generated). For example, in an optical subsystem, different modes may use light of different wavelengths for illumination. Modes may differ with respect to the illumination wavelength, as further described herein (e.g., by using different light sources, different spectral filters, etc., for different modes). In another example, different modes may use different illumination channels of the optical subsystem. For example, as described above, an optical subsystem may include two or more illumination channels. Thus, different illumination channels may be used in different modes. Modes may further, or alternatively, differ with respect to one or more focusing / detecting parameters of the optical subsystem. Modes may differ with respect to any one or more modifiable parameters of the inspection subsystem (e.g., illumination polarization, angle, wavelength, etc.; detection polarization, angle, wavelength, etc.). The inspection subsystem may be configured to scan the sample simultaneously using multiple modes, for example, to scan the sample in different modes during the same or different scans.

[0049] Similarly, the output generated by the electron beam subsystem may include an output, such as an image, generated by the electron beam subsystem using two or more different values ​​of the electron beam subsystem's parameters. Multiple modes of the electron beam subsystem may be defined by the values ​​of the electron beam subsystem's parameters used to generate the output and / or image to the sample. Thus, modes may differ with respect to at least one value of the electron beam parameter of the electron beam subsystem. For example, different modes may use different incident angles for illumination.

[0050] As described above, the inspection subsystem is configured to scan energy (e.g., light, electrons, etc.) through the physical deformation of the sample, thereby generating an output for the physical deformation of the sample. Thus, the inspection subsystem can be configured as a “real” subsystem rather than a “virtual” subsystem. However, the storage medium (not shown) and computer subsystem 102 shown in Figure 1 can be configured as a “virtual” system. In particular, the storage medium and computer subsystem can be configured as a “virtual” inspection system as described in U.S. Patent No. 8,126,255 issued to Bhaskar et al. on 28 February 2012 and U.S. Patent No. 9,222,895 issued to Duffy et al. on 29 December 2015, both of which have been assigned to the assignees of the present invention, and both of these patents are invoked by reference as if they were entirely discussed herein. The embodiments described herein can be further configured as described in those patents.

[0051] Non-pattern array (NPA) reference generation is a crucial part of array inspection. NPA inspection methods are used to perform array inspection on pattern regions that are highly repetitive and unresolved. Therefore, in NPA inspection, no patterns exist in the optical image (or other image). One problem affecting the sensitivity of NPA inspection is color variation (CV). CV may appear as relatively low-frequency grayscale fluctuations in the input NPA image, particularly near corner or edge regions. CV poses a significant challenge to NPA inspection. A relatively large number of nuances is another major challenge for NPA inspection. Embodiments described herein provide novel methods for addressing both of the above challenges through novel reference generation methods for handling CV and novel detection strategies for suppressing nuances.

[0052] Figure 2 illustrates these challenges and how the embodiments described herein are designed to overcome them. In this figure, Image 200 is a test image generated on a sample by the optical inspection subsystem described herein. Reference Image 202 is a reference image generated by a method currently used for NPA defect detection, and difference Image 204 is generated by subtracting Reference Image 202 from Test Image 200. As can be seen by comparing Reference Image 202 and Test Image 200, Reference Image 202 is much quieter near the left and right edges of the cell regions in the Reference Image (bright areas in the Reference Image) than near the same edges in the Test Image. These differences between the test image and the reference image can occur, for example, if the test image contains noise in some areas, such as near the edges, and the reference image generation fails to account for such noise.

[0053] These differences between the reference image and the test image appear in the difference image as dark areas located near the left and right edges of the cell regions in the difference image. These noise differences between the reference image and the test image may then be falsely detected as defects when a defect detection method is applied to such a difference image. Raising the defect detection threshold to avoid detecting this noise may result in the absence of DOIs present in the test image. Both of these situations are unfavorable for many obvious reasons.

[0054] In contrast, reference image 206 shows an embodiment of a reference image that can be generated by the embodiments described herein. Difference image 208 is generated by subtracting reference image 206 from test image 200. As can be seen by comparing reference image 206 and test image 200, both reference image 206 and test image 200 contain similar noise near the left and right edges of the cell regions (bright areas) in the test image. These similarities in noise characteristics are made possible by performing the reference generation described herein in such a way that the noise of the test image is better reproduced in the reference image, thereby resulting in a much less noisy difference image, as can be seen by comparing difference image 208 with difference image 204. As a result, defect detection performed on difference image 204 and difference image 208 using the same parameters (e.g., thresholds) detects far less noise in difference image 208 compared to difference image 204. In addition, because the difference image 208 has relatively low noise characteristics, a more sensitive inspection can be performed on the difference image 208 without detecting an excessive number of news signals, and at the same time, it may be possible to detect more DOIs, including DOIs with relatively weak signals.

[0055] One embodiment of a system configured to detect defects on a sample includes an inspection subsystem configured to generate an image of the sample, which may include any of the inspection subsystems described herein. The image includes, but is not limited to, a test image and two or more other images corresponding to the test image. The test image and two or more other images may be generated in areas on the sample that contain only unresolved repeating device patterns. In other words, the images used in the embodiments described herein do not include images of device patterns formed in areas on the sample where the image is generated. In one such example, the test image 200 is an image of two cell regions on the sample that contain formed patterned features, but none of the patterned features are resolved in this image. While the embodiments described herein are particularly suited and advantageous for NPA inspection, these embodiments may also be applied to inspections performed using images in which the patterned features are resolved.

[0056] The terms “image” and “image frame” are used interchangeably herein. Generally, an “image frame” is defined as a collection of pixels in an image that are processed in batches for purposes such as defect detection. Therefore, the size of an image or image frame may vary depending on the specific characteristics of the inspection subsystem or computer subsystem. In some inspection use cases, an image may actually consist of multiple image frames, but the embodiments described herein are not inherently limited by the size of the image that can be processed.

[0057] As used herein, the term "job" is defined as a number of image frames processed in batches by a computer subsystem to detect defects on a sample. Typically, the entire image generated during the inspection process cannot be processed together (for example, if possible, it is usually not time-efficient or cost-efficient). Therefore, the entire image is divided into jobs of image frames that can be processed in batches to detect defects in a less expensive and faster manner.

[0058] Image acquisition may include generating images using an inspection subsystem configured as described herein. This image acquisition may be performed when the computer subsystem and the inspection subsystem are combined into a single tool, and, in some cases, when defect detection as described herein is being performed on the tool, and / or when images are being generated. Otherwise, the computer subsystem may acquire images from another method, system, or storage medium. For example, the computer subsystem and the inspection subsystem may or may not be combined into a single tool, and the inspection subsystem, the computer subsystem, or another computer subsystem may store images generated by the inspection subsystem. The computer subsystem may then acquire images from the storage medium on which the images are stored. This image acquisition may be performed when the computer subsystem is performing the steps described herein outside of the tool, and / or after all (or at least some) of the images have been generated. Each of the above elements may be configured as further described and shown herein.

[0059] In one embodiment, the computer subsystem is configured to divide an image corresponding to a care area on a sample into a test image corresponding to a test cell and two or more other images corresponding to two or more other cells adjacent to the test cell. Each care area may be divided into a number of cells, similar to currently used NPA defect detection methods. However, instead of placing all cells into non-overlapping groups as in currently used methods and systems, the computer subsystem may form groups for each cell containing a small number of adjacent cells. By grouping cells in this way, a CV correction method for NPA reference generation may be provided. By reorganizing the cell groupings, sharper difference images may be achieved.

[0060] Figure 3 shows an example of an image for a care area, and various embodiments of dividing the care area image into a test image of the test cell and other images corresponding to other cells adjacent to the test cell. In particular, image 300 is a virtual image of the care area. This image is merely a single grayscale value image and therefore does not actually represent an image that may be generated by the inspection subsystem configured as described herein, but it accurately represents the NPA image in that none of the patterned device features formed on the sample within the care area are resolved in the image. Multiple identical care areas may exist on the sample, and image 300 represents an image of only one such case.

[0061] Each care area can be evenly divided into a small number of cells, either horizontally or vertically. As shown in Figure 3, if the care area is divided horizontally, then the test cell image may become image portion 302, and the images of other cells adjacent to the test cell may become image portions 304, 306, 308, and 310. If the care area is divided vertically, then the test cell image may become image portion 312, and the images of other cells adjacent to the test cell may become image portions 314, 316, 318, and 320. Thus, adjacent cells are the cells to the left and / or right of a target cell (if the care area is divided horizontally), or the cells above and / or below the target cell (if the care area is divided vertically). The computer subsystem may form groups of adjacent cells for each test cell target. In other words, if the inspection is performed on a test cell corresponding to image portion 304, a different group of image portions may be selected as the cells adjacent to that test cell.

[0062] Therefore, as shown in Figure 3, cells adjacent to a test cell do not necessarily have to be adjacent to the test cell; they may be separated by one or more other cells. However, cells adjacent to a test cell may be selected as adjacent cells. In addition, cells adjacent to a test cell may include cells on both sides of the test cell, but this is also not mandatory. Furthermore, although the number of cells adjacent to a test cell is shown as four cells in Figure 3, the number of adjacent cells may also vary considerably and depend on many factors, such as how much the color change differs between cells and the actual amount of image data that can be processed simultaneously by the computer subsystem. Different numbers of adjacent cells may be selected for cells in the horizontal and vertical directions. The same care area image may be divided along the horizontal and vertical directions, and the resulting test cells may be processed individually. For example, reference generation and defect detection may be performed on test cell 302 and its adjacent cells 304, 306, 308, and 310, and reference generation and defect detection may also be performed individually on test cell 312 and its adjacent cells 314, 316, 318, and 320.

[0063] In some embodiments, the computer subsystem is configured to apply CV correction to the test image and two or more other images before calculating the first and second candidate reference images as further described herein. For example, the first step may be to add CV correction to the reference generation process. One CV correction may be applied to all cells in one group before creating the reference cells. For example, CV correction may be applied to cells 304, 306, 308, and 310, and then a reference cell image is generated using them. After this step, the CVs are essentially integrated into the reference image that is ultimately subtracted from the test image, thereby generating a sharper final difference image. Thus, the embodiments described herein provide a CV correction method for NPA reference generation. A sharper difference image is achieved by reorganizing the cell groupings and correcting the CVs.

[0064] CV can result in each cell having a different gray level, and without correction, cell-based noise reduction capabilities are limited. One particularly suitable method for performing CV correction is to set the gray level of one cell to a variable x i This is the assumption. Next, a special function f(x i │p1,p2,…) are several unknown parameters p n This can be created using [x]. This function can be applied to each pixel position. The parameter is the difference between the target cell and the cell after the function operation, [x] for each cell. t -f(x i This can be calculated by minimizing (│p1,p2,…)). Applying this function to the corresponding cells corrects the CV. In this way, the computer subsystem can fit the grayscale changes within each cell. The basic fitting function can be, for example, a segmented linear function. Once the fitting is complete, the fitting result f(x i │p1,p2,…) become the local reference images within each cell. Since the fitted reference images substantially match the test images, the cross-validations embedded in the raw test images are removed or significantly reduced.

[0065] As mentioned above, CVs are incorporated into the reference cell image. One reason for this is the aforementioned CV correction. In particular, the objective of the improved reference generation described herein is to obtain sharper difference images. Previously, CV correction could not be performed or was not performed, meaning that CVs remained in the difference image, which could be considered too noisy for defect detection. However, by performing CV correction on the image used to generate the reference image, the CV correction can be incorporated into all the generated reference images and subsequently used to generate sharper difference images. In other words, CVs are handled intrinsically in the reference image generation process. Thus, the CVs of the reference image correct the CVs of the test image. CV correction can be intrinsically incorporated into reference image generation in all the reference images generated as described herein.

[0066] The embodiments described herein further improve CV processing capabilities by generating multiple candidate reference images for each test image and mixing them as further described herein. The computer subsystem is configured to compute the first and second candidate reference images from different combinations of at least two images from the test image and two or more other images. For example, as shown in step 400 of Figure 4, the computer subsystem may compute candidate reference images. The computation of candidate reference images described herein may be performed by combining two or more images using, for example, linear combination, median calculation, averaging, or other currently used methods for generating computed references (CRs). Computing candidate reference images from two or more images in this way may result in images having different characteristics from one or more input images. In other words, pixels in the computed candidate reference image may have different image characteristics, such as grayscale values, than the corresponding pixels in each image used to generate the computed candidate reference image. Thus, candidate reference image generation is fundamentally different from final reference image generation, which is further described herein.

[0067] In many cases, two or more candidate reference images may be calculated for each test image. In other words, a candidate reference image cannot be used for more than one test image in a job. Instead, after grouping the images as described further herein, the computer subsystem creates at least two candidate reference images for each test image in each job. However, in other cases, a candidate reference image may be reused for more than one test image.

[0068] In one embodiment, a first distinct composite of a test image and at least two other images used to compute a first candidate reference image includes a portion of the image generated only in the x-direction cross-section of the sample. In another such embodiment, a second distinct composite of a test image and at least two other images used to compute a second candidate reference image includes an additional portion of the image generated only in the y-direction cross-section of the sample. Currently used NPA defect detection methods always generate references in a row-independent manner (i.e., generate references using images generated along the x-direction). Embodiments described herein may also employ the same strategy to form a single candidate reference image. In addition, a similar operation may also be applied in a column-independent manner to generate another candidate reference image (i.e., generate a candidate reference image using images generated along the y-direction). Generating two such candidate reference images provides many important advantages, which will be further described herein.

[0069] In further embodiments, one of at least two different composites of a test image and two or more other images, used to compute one of the first and second candidate reference images, includes all images generated from two or more dies on the sample in the image job. For example, an additional reference may be generated by taking all dies into consideration. In this context, “all dies” means all dies included in a job, and a job is the smallest unit of image processing. A job may include all dies in at most one die row, or at least three dies. Figure 5 shows an example of a job that includes multiple cell regions for inspection. In particular, job 500 includes cell regions 502, 504, 506, and 508, each of which may be included in a different die on the sample. In this embodiment, the candidate reference image may be generated from all images of all cell regions included in job 500.

[0070] Thus, some candidate reference images may be generated from a single die (e.g., row-independent and column-independent references), while one or more other candidate reference images may be generated from two or more die images (e.g., candidate reference images generated from "all dies"). Generally, in the embodiments described herein, at least two candidate reference images are generated, and the final candidate reference image may be created from those two images. However, in many use cases, it is advantageous to generate all three types of candidate reference images described herein: row-independent, column-independent, and multi-die candidate reference images. Two or more of these candidate reference images may be used to generate the final reference image. One or more of these candidate reference images may also be used as a second, or additional, final reference image when double detection is being performed. Such embodiments are described in more detail herein.

[0071] In currently used NPA defect detection, reference generation occurs within each individual frame. In other words, reference generation for frame 3 does not require any input from frame 2 or frame 4. In contrast, in the embodiments described herein, both frame 3 and all other frames may be used to generate the reference image for frame 3. Using these images to generate candidate reference images can be advantageous, for example, when a corner region of a cell has a unique grayscale value relative to the rest of the cell region. Such unique grayscale values ​​may make it difficult to find a well-matching reference image for each corner within each cell region. However, finding a good match from other corners of other cell regions can be much easier. Using all other frames in the job may create better candidate reference images, which can then be used to create a better final reference image.

[0072] The computer subsystem is also configured to select at least a portion of a first candidate reference image corresponding to a first portion of the test image and a portion of a second candidate reference image corresponding to a second portion of the test image. For example, as shown in step 402 of Figure 4, the computer subsystem may be configured to select different portions of candidate reference images corresponding to different portions of the test image. As further described herein, embodiments are configured to provide a better reference image, which can be used to generate a better difference image, thereby enabling less noise detection and / or more sensitive inspection. To this end, a better reference is a reference image that better matches the test image, particularly with respect to noise in the test image. For example, if the noise in the reference image substantially matches the noise in the test image, then when the reference image is subtracted from the test image, the noise in the test image is canceled out and prevented from being detected as noise and / or interfering with DOI detection. Thus, the objective of the selection step is to find the best-matching reference image from the set of candidate reference images for each portion of the test image, which can be done in several ways further described herein.

[0073] In one embodiment, selection involves identifying which portions of the first and second candidate reference images best match different portions of the test image. Figure 6 shows one embodiment of how this can be done. In this embodiment, test image 600 is an example of a test image that may be generated for a care area on a sample. Similar to the test image shown in Figure 2, this test image includes a central portion 606, which is relatively less noisy compared to portions 602 near the left edge and portions 604 near the right edge. These portions are relatively noisier than the central portion, for example, due to color changes that may occur near the edges of the care area image. Furthermore, as mentioned above, if the reference image used with the test image for defect detection does not contain the same level of noise as the test image, problems arise in defect detection. Therefore, a novel reference image generation method described herein has been devised.

[0074] The selection step may include dividing the test image into an array of blocks, as shown in Figure 608. Each block may have the same predetermined characteristics, such as dimensions, and they are simply used to divide the image into smaller, individually evaluable parts, as will be further described herein. The number of blocks into which the image is divided may vary considerably and may be determined based on many factors, such as the initial image size and the rate at which noise changes throughout the image (smaller blocks are more appropriate if the noise changes relatively rapidly, as different values ​​of noise can be captured on the same or similar scale as the noise change).

[0075] As described above, the portion of the candidate reference image that best matches a different part of the test image can be identified by the computer subsystem, meaning that the correspondence between the test image portion and the reference image portion can be relaxed, even if each test image portion is processed individually. In other words, the portion of the candidate reference image that best matches a part of the test image does not have to be limited to only the portion of the candidate reference image that has the same in-image position as the test image portion.

[0076] To illustrate this concept, consider two example candidate reference images 614 and 616, which may be any of the candidate reference images generated by any of the methods described herein. Each of these candidate reference images may be divided into blocks in the same way as the test image, indicated by black lines superimposed on the image. The identification step may begin with the test image block in the upper row of the block column 610. This test image block may be compared with each block of candidate reference image 614 to determine which block in this candidate reference image best matches the test image block. The candidate reference image block that best matches the test image block may be saved, and the same comparison step may then be performed on the blocks of candidate reference image 616 to determine whether a better-matching block is included in that candidate reference image. The best-matching block in either candidate reference image 614 or 616 may then be identified and saved to be included in the final reference image at the location of the test image block.

[0077] In this way, for each test image block, all blocks of all (or at least some) of the candidate reference images can be considered to find the most matching image block. The same process can then be performed for the next block in the test image. In some cases, the same block in one of the candidate reference images may be identified as the most matching block for two or more test image blocks. Allowing any block in any of the candidate reference images to be used as the most matching block in a test image block may result in a better reference image than if only blocks located within the same image were considered. This method may provide a better final reference image when the test image and each candidate reference image exhibit noise with substantially different spatial or other characteristics.

[0078] In another embodiment, the selection includes identifying which portions of the first and second candidate reference images best match the corresponding portions of the test image. Thus, unlike the embodiments described above, for each test image block, only candidate reference image blocks having the same position within the same image may be considered. This embodiment may be faster than the embodiments described above because the number of blocks considered for any one test image block is limited to the number of candidate reference images generated. However, unlike the embodiments described above, this embodiment is not flexible or exhaustive with respect to the blocks considered and may therefore be more suitable in situations where it is known or expected that at least one of the candidate reference images contains noise having similar spatial characteristics and possibly other characteristics to the test image.

[0079] To illustrate this concept, image 608, obtained by dividing test image 600 into blocks, may be used again along with candidate reference images 614 and 616. The identification step may begin with the test image block located in the upper row of column 610 of the blocks. This block is compared with the image block located in the upper row of column 618 of candidate reference image 614 and the image block located in the upper row of column 626 of candidate reference image 616 to determine which of these blocks best matches the test image block. If no other candidate reference images have been generated for this test image, those other candidate reference image blocks may not be considered for this test image block. As shown in images 608, 614, and 616, the image block located in the upper row of column 618 of candidate reference image 614 has a much higher degree of match with the test image block located in the upper row of column 610 than the image block in the same position in candidate reference image 616. The best-matching candidate reference image block identified in this step may then be saved to generate the final reference image, as will be further described herein.

[0080] Next, the identification step may proceed using a test block located in the second row from the top of column 610 of the block. This block is compared with the image block located in the second row from the top of column 618 of candidate reference image 614, and the image block located in the second row from the top of column 626 of candidate reference image 616, to determine which of these two blocks best matches the test image block. Again, if no other candidate reference images have been generated for this test image, those other candidate reference image blocks may not be considered for this test image block. As shown in images 608, 614, and 616, the image block located in the second row from the top of column 618 of candidate reference image 614 has a better match with the test image block located in the second row from the top of column 610 than the image block in the same position in candidate reference image 616. The best-matching candidate reference image block identified in this step may then be saved to generate the final reference image, as will be further described herein. The identification step may then proceed for all remaining test blocks in image 608.

[0081] Regardless of how the comparison is performed on the images shown in Figure 6, different candidate reference images contain different noise than the test image. In particular, the noise at the left and right edges of candidate reference image 614 (indicated by the dark areas of the image) extends further into the candidate reference image than the noise at the left and right edges of the test image (also indicated by the dark areas of the test image). Specifically, columns 618, 620, 622, and 624 of the block in candidate reference image 614 contain noise, while only columns 610 and 612 of the block in test image 608 contain noise. In contrast, candidate reference image 616 contains little to no noise at its left and right edges (indicated by the absence of any relatively dark areas in the candidate reference image). Specifically, columns 626 and 628 of candidate reference image 616 contain little to no noise, while the corresponding columns 610 and 612 of image 608 clearly contain noise. Therefore, regardless of which candidate reference image is used with the test image for defect detection, either reference image may produce significant noise detection or other problems in defect detection. However, by selectively identifying portions of the candidate reference images to include in the final reference image based on how well they match the test image, the final reference image 634 may be generated with noise that clearly indicates a better match with the test image 600.

[0082] The computer subsystem is further configured to generate a final reference image by compositing selected portions of the first and second candidate reference images without modifying the selected portions of the first and second candidate reference images. As shown in step 404 of Figure 4, the computer subsystem can generate a first final reference image by compositing the selected portions, which may be the sole final reference image in the case of single detection, or one of two final reference images in the case of double detection. For example, the final reference image 634 shown in Figure 6 may be generated by compositing selected portions of the first candidate reference image 614 and the second candidate reference image 616 without modifying the selected portions of these images.

[0083] In particular, as shown in Figure 6, for the image block in column 610 of image 608, the image block in column 618 of candidate reference image 614 matches much better than the image block in column 626 of candidate reference image 616. Therefore, the image block in column 636 of final reference image 634 could be the image block in column 618. Similarly, for the image block in column 612 of image 608, the image block in column 620 of candidate reference image 614 matches much better than the image block in column 628 of candidate reference image 616. Therefore, the image block in column 642 of final reference image 634 could be the image block in column 620.

[0084] In contrast, the image blocks in columns 622 and 624 of candidate reference image 614 do not match the corresponding columns in the image blocks of image 608 very well. However, the image blocks in columns 630 and 632 of candidate reference image 616 match the corresponding columns in the image blocks of image 608 much better. Therefore, these image blocks can be included in the final reference image 634 as columns 638 and 640, respectively. The remaining four central columns of image 608 match the four central columns of both candidate reference images very well. Therefore, the four central columns of either candidate reference image can be used as the four central columns of the final reference image 634.

[0085] Thus, creating a final reference image by mixing multiple candidate reference images can be done in a block-based manner. The image is divided into relatively small blocks. In each block, the test image is compared with individual reference images. The reference image block that best matches the test image block is selected to be included in the final reference image.

[0086] Furthermore, as described above, the synthesis step is performed without modifying the selected portions of the first and second candidate reference images to generate the final reference image. Therefore, this reference image generation differs significantly from other computed reference (CR) generation methods. In particular, the final reference image generation described herein can be considered a kind of "pick and place" operation, where different portions of the candidate reference images are identified and selected as described above, and then placed in the positions of the selected portions of the test image. Subsequently, the selected portions of the different candidate reference images can be "stitched together" into the final reference image. However, these steps are performed without modifying the image itself. In other words, the selected portions of the image are not modified when the images are actually synthesized.

[0087] In contrast, current CR generation methods involve combining two or more images to produce a CR, which in turn involves combining images in such a way that the resulting image data differs from the data of the original images. For example, a CR can be generated using a linear combination operation, and one or more pixels in the resulting CR may differ from the same one or more pixels in all the images used to generate the CR. Such alteration of at least some pixels in an image is, in fact, the essence of currently used CR methods. In particular, the goal of currently used CR methods is usually to produce a reference image that is as noiseless as possible, which is essentially a "defect-free" image.

[0088] This objective differs fundamentally from the objective of the synthesis step described herein, which is to generate a final reference image that is as similar as possible to the test image in terms of noise. This is shown in test image 600 in Figure 6, which contains relatively large noise near the left and right edges of the image, as will be further described herein. If a difference image of the test image is generated using a reference image with significantly less noise, such as candidate reference image 616, the noise at the left and right edges of test image 600 may be detected as defects. In contrast, the final reference image 634, which may be generated by the embodiments described herein, may contain relatively large noise with the same or substantially the same characteristics as the test image. Thus, although this final reference image is certainly not noise-free, when a difference image is generated by subtracting it from the test image, a substantially less noisy difference image is produced, and the DOIs, including DOIs with relatively low signals, can be detected using this difference image, and not much noise is detected. Therefore, the embodiments described herein generate a final reference image that is fundamentally different from the CRs currently generated and used, thereby providing a significant advantage to the embodiments described herein.

[0089] Instead of mixing multiple candidate reference images to generate different final reference images as described above, the comparison step described above may be performed for each portion of the test image to determine which candidate reference image best matches the test image. In this way, after multiple candidate reference images have been generated for each test image, the optimal reference image can be selected from among the multiple reference images. For example, in some cases other than the case shown in Figure 6, where portions of both candidate reference images clearly do not match the corresponding portions of the test image, it is possible that one of the candidate reference images generated by one of the various methods described herein has noise characteristics substantially similar to those of the test image. In particular, if image 634 is a third candidate reference image and not the final reference image as described above, then clearly image 634 matches the noise characteristics of test image 600 substantially better than both of the other candidate reference images 614 and 616. Therefore, the synthesis of different portions of different candidate reference images may not be necessary, since a suitable final candidate reference image already exists in candidate reference image 634.

[0090] In some embodiments, the computer subsystem is configured to compute a third candidate reference image from different composites and select at least a portion of the third candidate reference image corresponding to a third portion of the test image, and the composite step includes compositing the selected portions of the first, second, and third candidate reference images without modifying the selected portions of the first, second, and third candidate reference images to generate a final reference image. For example, a single final reference image may be constructed from three candidate reference images for each test image. One of the references is generated in a row-independent manner, another in a column-independent manner, and the third is generated using two or more dies, thereby giving the resulting final reference image a three-dimensional (3D) property. Thus, embodiments described herein may be called 3DNPA. The three candidate reference image generation and compositing approaches described herein offer advantages to embodiments described herein. For example, better noise reduction can be obtained by generating three candidate reference images from row-independent, column-independent, and multi-die methods, and then compositing all three images. These steps can be performed as described above, except for the use of more candidate reference images.

[0091] In one embodiment, following the synthesis step, the computer subsystem is configured to apply color change correction to the synthesis result. For example, compositing different parts of different candidate reference images into a final reference image does not in itself involve modifying selected parts of the different candidate reference images, but one or more additional image processing steps may be performed on the final reference image, which is then used for defect detection. As described above, after the CV correction is applied to the candidate reference images, a step for generating the final reference image may be performed, meaning that the generated final reference image will have essentially the same CV correction. Therefore, performing an additional CV step on the final reference image may not be necessary, and whether or not this step is performed may be evaluated on a case-by-case basis. In any case, the CV correction applied to the generated final reference image may be performed as described above.

[0092] The computer subsystem is further configured to generate a difference image by comparing a test image with a final reference image. For example, as shown in step 408 of Figure 4, the computer subsystem may be configured to generate a difference image 1 by subtracting the final reference image 1 from the test image. The generation of the difference image can be carried out in any suitable method known in the art. In particular, subtracting the final reference image from the test image can be carried out in any suitable method.

[0093] The computer subsystem is further configured to detect defects in a test image by applying a defect detection method to the difference image. For example, as shown in step 412 of Figure 4, the computer subsystem may be configured to perform detection using difference image 1 as input. Perhaps in the simplest implementation, defect detection in step 412 may involve comparing pixels or signals in the difference image with a threshold, determining any pixel or signal with a value exceeding the threshold as a defect, and determining any pixel or signal with a value not exceeding the threshold as not a defect. However, many more complex defect detection methods have been developed in the art and may be used in step 412. In other words, the difference image generated by the embodiments described herein may be used for defect detection in the same way that other difference images are used for defect detection. In addition, the defect detection method may include any defect detection method well known in the art, e.g., the MCAT defect detection method, which is a defect detection algorithm used in some inspection systems commercially available from KLA, or other suitable commercially available defect detection methods and / or algorithms. The output of this step may be the final defect result 418 in the case of a single detection method.

[0094] The embodiments described herein may be configured for both single detection mode and double detection mode. Both detection modes may be based on multiple reference generation performed as described herein. In single detection mode, only one final reference image is required and is used to generate a single difference image for each test image. Defect detection is then performed using the single difference image. Two or more final reference images may be generated for any one test image and used for double detection. For example, the embodiments described herein enable double detection for NPA defect detection, which is not supported by currently used NPA defect detection methods. To enable double detection, two final reference images may be generated from candidate reference images instead of one.

[0095] Several embodiments may be configured and used for such defect detection. In one embodiment, a computer subsystem is configured to generate an additional final reference image by repeatedly selecting and combining. For example, by using the steps described herein, two equally clear but different reference images may be constructed based on three candidate reference images. Thus, the embodiments described herein advantageously provide a dual reference generation method for news reduction.

[0096] In one such embodiment, step 402 shown in Figure 4 may be performed twice, once to select different portions of a candidate reference image for a first final reference image, and at another time to select another different portion of a candidate reference image for a second final reference image. Both selection steps may be performed in one or more of the methods described above. In addition, the selection steps performed for each final reference image may be performed in the same or different ways. As shown in step 406 of Figure 4, the computer subsystem may synthesize the selected portions to generate a second final reference image. In particular, the selection step may generate a first set of selected portions used to generate the first final reference image and a second set of selected portions used to generate the second final reference image. The selected portions of the first and second sets must contain at least several different portions of any one portion of the test image, otherwise the second final reference image would be redundant. The selected portions of the first set may be used to generate the first final reference image in step 404, and the selected portions of the second set may be used to generate the second final reference image in step 406. The synthesis step may otherwise be performed as further described herein.

[0097] In another embodiment, the computer subsystem is configured to select one of the first and second candidate reference images as an additional final reference image. For example, rather than generating two final reference images from the synthesis of different parts of different candidate reference images, one final reference image may be generated in this way, and the other final reference image may simply be one of the candidate reference images. The best candidate reference image to use as the second final reference image may be selected as further described herein.

[0098] In further embodiments, the computer subsystem is configured to select additional candidate reference images as additional final reference images. For example, additional candidate reference images may be available for consideration as second final reference images and / or may be specially generated. In such an example, the computer subsystem may be configured to acquire images of adjacent care areas on either the same die or adjacent dies. Thus, such images may not be CR in that they may not be generated by combining two or more images. Instead, such images may simply be images generated by the inspection subsystem, which may perform one or more image processing steps, such as CV correction and high-pass filtering. Such images may then be compared with one or more of the test images and / or other candidate reference images to determine which one or two images are best suited for defect detection. In any case, multiple candidate reference images may be generated for each test image, and the computer subsystem may select the two best reference images, rather than the single best reference image, for use in defect detection.

[0099] In another such example, the two final reference images may be obtained from a selection process performed on three generated candidate reference images. In particular, the top two best candidate reference images may be selected for double detection. With special care, both of the two selected reference images can be of good quality. For example, in the corner region of a cell area as shown in Figure 5, the best reference image is usually obtained from a third reference image (e.g., a die-to-die operational reference). The next best reference image may be of significantly lower quality in the corner region (i.e., significantly different from the same region in the test image). In this case, the computer subsystem can force the two reference images selected for double detection to be of the same type, meaning they are all obtained from the best multi-die candidate reference images. Thus, the embodiments described herein can ensure that double detection has relatively high sensitivity.

[0100] Dual final reference image generation can activate dual detection in NPA defect detection. In any of the embodiments described above, the computer subsystem may be configured to generate an additional difference image by comparing the test image with the additional final reference image, and to detect defects in the test image by applying a defect detection method to the additional difference image. Regardless of which additional final reference image is generated and selected for use in the embodiments described herein, the computer subsystem may also generate a difference image 2 by subtracting the final reference image 2 from the test image, as shown in step 410 of Figure 4, and performing detection with the difference image 2 as input, as shown in step 414. These steps may be performed as further described herein.

[0101] A defect detection method determines that a defect exists at a certain location in a test image only if the defect detection method detects the defect at the corresponding location in the difference image and the additional difference image. For example, as shown in step 416 of Figure 4, defects detected in both detection steps 412 and 414 can be combined into a single set of defects detected in both comparisons. The combined set of defects can then be analyzed to generate the final defect result 418 shown in Figure 4, which includes only the defects detected at the corresponding locations in the first and second difference images. In particular, the computer subsystem compares the defects detected in the difference images, designating any commonalities between defects as actual defects and any differences between defects as nuances or artifacts. The computer subsystem can compare the results of different defect detection steps performed using different difference images in any suitable way, thereby comparing the results at corresponding locations in the different difference images and determining whether the detected defects are actual defects or artifacts or nuances. Thus, the defect detection method is preferably configured to report only if a defect is detected at the same location or at the same pixel in the two difference images. Thus, embodiments described herein can perform dual detection, which is useful for suppressing nuance detection. In other words, CR dual detection can minimize the influence of inherent nuances or artifacts in the reference image on the defect detection result through mediation.

[0102] The embodiments described herein also advantageously provide a novel segment-based detection strategy that suppresses nuance detection. In one embodiment, the computer subsystem is configured to divide an image corresponding to a care area of ​​a cell region on a sample into a test image corresponding to a test cell and two or more other images corresponding to two or more other cells adjacent to the test cell, and the defect detection method divides the care area into different segments of the cell region and detects defects individually in each of the different segments. Thus, the embodiments described herein may add image segmentation and cluster different locations into separate detection flows. The computer subsystem may divide the entire image into different regions, the dimensions of each region may be determined based on nuance differences. In the detection phase, each region may have independent parameters to adapt to different nuance control requirements.

[0103] In one embodiment, the different segments include a first segment for the upper and lower edges of the cell area, a second segment for the left and right edges of the cell area, a third segment for the center of the cell area, and a fourth segment for the corners of the cell area. One such embodiment of segment-based detection for improving news reduction is shown in Figure 7. In this embodiment, the overall image of the care area 700 is divided into different segments. The segments include a central segment 700a, a left / right edge segment 700b, an upper / lower edge segment 700c, and a corner segment 700d. Each case of this care area on the sample can be divided into these four segments.

[0104] Each step described herein may be performed collectively or independently for each segment. For example, in the case of a test image corresponding to the entire care area, the final reference image may be generated as described herein, and the difference image may be generated for the entire care area by subtracting the final reference image from the test image. The difference image may then be divided into parts corresponding to each of the different segments, and defect detection may be performed individually for each part of the difference image.

[0105] As an example, in plot 702 shown in Figure 7, different noise clouds can be generated individually and independently for each of the different segments within the care area. Specifically, noise cloud 702a may be generated only for the central segment 700a, noise cloud 702b may be generated only for the left edge / right edge segment 700b, noise cloud 702c may be generated only for the upper edge / lower edge segment 700c, and noise cloud 702d may be generated only for the corner segment 700d. Defect detection can then be performed individually for each noise cloud. In the detection phase, each region may have independent parameters to adapt to different noise suppression requirements. In this way, individual noise clouds can be formed for each segment using individual parameters, achieving an optimal balance between DOI detection and noise suppression.

[0106] A multi-noise cloud plot, as shown in Figure 7, can be generated for each test image. In other words, the noise cloud shown in plot 702 is for only one test image, and similar plots can be generated for other test images. In addition, the noise cloud shown in plot 702 is based on a difference image generated for one test image. Individual noise cloud plots may be generated for other difference images generated for the same test image (e.g., double detection where multiple difference images are generated for a single test image).

[0107] Any computer subsystem described herein can generate inspection results, which may include the results of any of the steps described herein. The inspection results may include information about detected defects, such as defect ID, size, detection score, location of the bounding boxes of the detected defects, class label or ID, or any other suitable information known in the art. The defect results may be generated by the computer subsystem in any suitable manner. The defect results may have any suitable form or format, such as a standard file type. By generating and storing the results, the computer subsystem and / or other systems or methods may use the results to perform one or more functions on a sample or other samples of the same type.

[0108] A computer subsystem may be configured to store information about detected defects in any suitable computer-readable storage medium. The information may be stored together with any of the results described herein, or in any manner well known in the art. The storage medium may include any storage medium described herein, or any other suitable storage medium well known in the art. After the information is stored, it may be accessed within the storage medium, used by any of the methods or system embodiments described herein, formatted for display to a user, and used by other software modules, methods, or systems.

[0109] The results and information generated by performing inspections on a sample can be used in various ways by the embodiments and / or other systems and methods described herein. Such functions include, but are not limited to, modifying processes, such as manufacturing processes or steps, that have been performed or will be performed on the sample under inspection or on other samples, in a feedback or feedforward manner. For example, a computer subsystem may be configured to determine one or more changes to processes that have been performed or will be performed on the sample inspected as described herein, based on the detected defects. Changes to processes may include any appropriate changes to one or more parameters of the process. The computer subsystem preferably determines these changes so that defects can be reduced or prevented on other samples on which the modified process is performed, so that defects can be corrected or eliminated on the sample in another process performed on the sample, so that defects can be compensated for in another process performed on the sample, and so on. The computer subsystem may determine such changes in any appropriate way known in the art.

[0110] These changes may then be transmitted to a semiconductor manufacturing system (not shown) or a storage medium (not shown) accessible to both the computer subsystem and the semiconductor manufacturing system. The semiconductor manufacturing system may or may not be part of the system embodiments described herein. For example, the computer subsystem and / or inspection subsystem described herein may be coupled to the semiconductor manufacturing system via one or more common elements, such as a housing, power supply, sample processing device, or mechanism. The semiconductor manufacturing system may include any semiconductor manufacturing system well known in the art, such as lithography tools, etching tools, chemical mechanical polishing (CMP) tools, and deposition tools.

[0111] The embodiments described herein offer several advantages over other methods and systems for detecting defects on a sample. For example, the embodiments described herein provide better reference generation to reduce inspection noise in the presence of color changes and offer novel detection strategies to further suppress noise. In addition, the embodiments described herein provide novel reference generation methods and detection strategies in non-patterned areas, thereby providing better difference image generation with less noise, thereby improving sensitivity in the presence of color changes and enhancing noise suppression capabilities.

[0112] The advantages described herein are realized by several important new features of the embodiments described herein. For example, the embodiments described herein allow for a new degree of freedom to simultaneously generate two or three candidate reference images for any one test image. This capability is novel for both unresolved device pattern features and resolvable patterns. In the case of resolvable repeating cell regions, unlike the embodiments described herein, currently used inspection systems and methods can typically only generate reference images along the repeating direction. The multi-candidate reference image mixing process described herein may also be useful for inspecting unresolved pattern regions with other types of regions. In addition, the embodiments described herein support both single detection using the best single final reference image and dual detection using the best two reference images generated by any of the methods described herein. The embodiments described herein are not limited to any particular defect detection method and may be used to generate input to any suitable defect detection method known in the art.

[0113] Each of the above-described embodiments of the system can be combined as a single embodiment. In other words, unless otherwise specified herein, no system embodiment is mutually exclusive with any other system embodiment.

[0114] Another embodiment relates to a computer-implemented method for detecting defects on a sample. This method includes acquiring an image of the sample generated by an inspection subsystem, the image including a test image and two or more other images corresponding to the test image. The method also includes calculating first and second reference images from at least two different composites of the test image and the two or more other images (step 400 shown in Figure 4). The method also includes selecting at least a portion of a first candidate reference image corresponding to a first portion of the test image and a portion of a second candidate reference image corresponding to a second portion of the test image (step 402 in Figure 4). In addition, the method includes compositing the selected portions of the first and second candidate reference images without modifying the selected portions of the first and second candidate reference images to generate a final reference image (step 404 in Figure 4). The method also includes generating a difference image by comparing the test image and the final reference image (step 408 in Figure 4). Furthermore, the method includes detecting defects in the test image by applying a defect detection method to the difference image (step 412 shown in Figure 4). The steps of acquisition, calculation, selection, synthesis, generation, and detection are performed by a computer subsystem coupled to an inspection subsystem, which may be configured according to any embodiment described herein.

[0115] Each step of the method may be performed as further described herein. The method may include any other steps that can be performed by the inspection subsystem and / or computer subsystem described herein. In addition, the above-described method may be performed by any system embodiment described herein.

[0116] Additional embodiments relate to a non-temporary computer-readable medium storing program instructions executable on a computer system for performing a computer implementation method for detecting defects on a sample. One such embodiment is shown in Figure 8. In particular, as shown in Figure 8, the non-temporary computer-readable medium 800 includes program instructions 802 executable on a computer system 804. The computer implementation method may include any step of any method described herein.

[0117] Program instructions 802 implementing a method such as those described herein may be stored in a computer-readable medium 800. The computer-readable medium may be a storage medium such as a magnetic or optical disk, a magnetic tape, or any other suitable non-temporary computer-readable medium known in the art.

[0118] Program instructions can be implemented in any of the following ways, including, in particular, procedure-based techniques, component-based techniques, and / or object-oriented techniques. For example, program instructions can be implemented using ActiveX controls, C++ objects, JavaBeans, Microsoft Foundation Classes ("MFC"), SSE (Streaming SIMD Extensions), or other techniques or methodologies as needed.

[0119] The computer system 804 may be configured according to any of the embodiments described herein.

[0120] Further modifications and alternative embodiments of various aspects of the present invention will be apparent to those skilled in the art in consideration of this description. For example, methods and systems for detecting defects on a sample are provided. Therefore, this description should be interpreted as merely illustrative and aims to disclose to those skilled in the art a general way of carrying out the present invention. It should be understood that the forms of the present invention shown and described herein should be taken as preferred embodiments at present. Elements and materials may be replaced with those illustrated and described herein, parts and processes may be reversed, and certain attributes of the present invention may be used independently, all of which will be apparent to those skilled in the art in consideration of this description of the present invention. Modifications may be made in elements described herein without departing from the spirit and scope of the present invention as set forth in the following claims.

Claims

1. A system configured to detect defects on a sample, An inspection subsystem configured to generate an image of a sample including a test image and two or more other images corresponding to the test image, A computer subsystem, The first and second candidate reference images are calculated from the test image and at least two different composites of the two or more other images. Select at least a portion of the first candidate reference image corresponding to the first portion of the test image and a portion of the second candidate reference image corresponding to the second portion of the test image. The selected portions of the first and second candidate reference images are combined without changing the selected portions of the first and second candidate reference images to generate a final reference image. A difference image is generated by comparing the test image with the final reference image, and, By applying the defect detection method to the difference image, defects in the test image are detected. A computer subsystem configured as follows, A system equipped with these features.

2. The system according to claim 1, wherein the test image and the two or more other images are generated in a region on the sample that includes only unresolved repeating device patterns.

3. The system according to claim 1, wherein the computer subsystem is further configured to divide the image corresponding to the care area on the sample into the test image corresponding to the test cell and the two or more other images corresponding to two or more other cells adjacent to the test cell.

4. The system according to claim 1, wherein, prior to the calculation, the computer subsystem is further configured to apply color change correction to the test image and the two or more other images.

5. The system according to claim 1, wherein, following the synthesis, the computer subsystem is further configured to apply color change correction to the result of the synthesis.

6. The system according to claim 1, wherein the first different composite of the test image and at least two of the two or more other images used to calculate the first candidate reference image includes a portion of the image generated only in the x-direction cross-section of the sample.

7. The system according to claim 6, wherein the second different composite of the test image and at least two of the two or more other images used to calculate the second candidate reference image includes an additional portion of the image generated only in the y-direction transverse of the sample.

8. The system according to claim 1, wherein one of the test image and at least two of the other images used to compute one of the first and second candidate reference images includes all of the images generated on two or more dies on the sample in the image job.

9. The system according to claim 1, wherein the computer subsystem is further configured to calculate a third candidate reference image from the different composites and to select at least a portion of the third candidate reference image corresponding to a third portion of the test image, the composite comprising compositing the selected portions of the first, second, and third candidate reference images without modifying the selected portions of the first, second, and third candidate reference images to generate the final reference image.

10. The system according to claim 1, wherein the selection includes identifying which portion of the first and second candidate reference images best matches which portion of the test image.

11. The system according to claim 1, wherein the selection includes identifying which portion of the first and second candidate reference images best matches the corresponding portion of the test image.

12. The system according to claim 1, wherein the computer subsystem is further configured to generate an additional final reference image by repeating the selection and synthesis, generate an additional difference image by comparing the test image with the additional final reference image, and detect defects in the test image by applying the defect detection method to the additional difference image, the defect detection method determines that a defect exists at a certain location in the test image only if the defect detection method detects the defect at a corresponding location in the difference image and the additional difference image.

13. The system according to claim 1, wherein the computer subsystem is further configured to select one of the first and second candidate reference images as an additional final reference image, generate an additional difference image by comparing the test image with the additional final reference image, and detect a defect in the test image by applying the defect detection method to the additional difference image, the defect detection method determines that a defect exists at a certain location in the test image only if the defect detection method detects the defect at a corresponding location in the difference image and the additional difference image.

14. The system according to claim 1, wherein the computer subsystem is further configured to select an additional candidate reference image as an additional final reference image, generate an additional difference image by comparing the test image with the additional final reference image, and detect a defect in the test image by applying the defect detection method to the additional difference image, the defect detection method determines that a defect exists at a certain location in the test image only if the defect detection method detects the defect at a corresponding location in the difference image and the additional difference image.

15. The system according to claim 1, wherein the computer subsystem is further configured to divide the image corresponding to the care area of ​​the cell region on the sample into the test image corresponding to the test cell and the two or more other images corresponding to two or more other cells adjacent to the test cell, and the defect detection method divides the care area into different segments of the cell region and detects defects individually for each of the different segments.

16. The system according to claim 15, wherein the different segments include a first segment for the upper and lower edges of the cell region, a second segment for the left and right edges of the cell region, a third segment for the center of the cell region, and a fourth segment for the corners of the cell region.

17. The system according to claim 1, wherein the inspection subsystem is further configured as an optical-based inspection subsystem.

18. The system according to claim 1, wherein the inspection subsystem is further configured as an electron beam-based inspection subsystem.

19. A non-temporary computer-readable medium that stores program instructions for executing a computer implementation method for detecting defects on a sample, which is executable on a computer system, and the computer implementation method is The inspection subsystem generates an image of the sample, which includes a test image and two or more other images corresponding to the test image. The first and second candidate reference images are calculated from the test image and at least two different composites of the two or more other images. Select at least a portion of the first candidate reference image corresponding to the first portion of the test image and a portion of the second candidate reference image corresponding to the second portion of the test image. The selected portions of the first and second candidate reference images are combined without changing the selected portions of the first and second candidate reference images to generate a final reference image. A difference image is generated by comparing the test image with the final reference image, and, By applying the defect detection method to the difference image, defects in the test image are detected. Non-temporary computer-readable media, including [specific examples of such media].

20. A computer implementation method for detecting defects on a sample, The inspection subsystem generates an image of the sample, which includes a test image and two or more other images corresponding to the test image. The first and second candidate reference images are calculated from the test image and at least two different composites of the two or more other images. Select at least a portion of the first candidate reference image corresponding to the first portion of the test image and a portion of the second candidate reference image corresponding to the second portion of the test image. The selected portions of the first and second candidate reference images are combined without changing the selected portions of the first and second candidate reference images to generate a final reference image. A difference image is generated by comparing the test image with the final reference image, and, Defects in the test image are detected by applying a defect detection method to the difference image, and the acquisition, calculation, selection, synthesis, generation, and detection are performed by a computer subsystem coupled to the inspection subsystem. Computer implementation methods, including those mentioned above.

Images