System and method for optical wafer characterization with image upsampling

Abstract

In broadband plasma (BBP) wafer inspection, sensitivity can be improved by making pixels smaller, but throughput is sacrificed. [Solution] The system comprises a processing unit that is communicatively connected to a detector array of an optical wafer characterization system. The processing unit includes the steps of: acquiring one or more target images of a target location on the wafer from the detector array; applying a noise reduction filter to at least one or more target images; determining one or more difference images from one or more reference images and one or more target images; and generating one or more upsampled images by upsampling one or more difference images. One or more wafer defects are detected in one or more difference images or one or more upsampled images.

Description

Technical Field

[0001] The present invention generally relates to specimen characteristic evaluation and inspection, and more specifically to an optical wafer characteristic evaluation system and method involving image upsampling.

Background Art

[0002] Cross - reference to related applications This application claims the priority of U.S. Provisional Application No. 63 / 191,845, filed on May 21, 2021, the entire disclosure of which is incorporated herein by reference.

[0003] The demand for electronic logic and memory devices with ever - smaller footprints and features presents a wide range of manufacturing challenges that go beyond manufacturing at the desired scale. For example, in broadband plasma (BBP) wafer inspection, by making pixels smaller, sensitivity can be improved, but throughput is sacrificed. Thus, a compromise between sensitivity and throughput is required. In another example, BBP wafer inspection may involve obtaining a difference image by subtracting the gray levels of a target image and a reference image, but the capacity of the filter applied to the difference image may be limited in order to improve the signal - to - noise ratio (SNR) of the defect of interest. In another example, aliasing noise may be introduced into the image depending on the type and / or amount of sampling in the inspection mode used in BBP wafer inspection. <​​​​​​​​​​​​​​​​​​​​​​​​ Therefore, it is desirable to provide a system and method that overcomes the shortcomings of the above-mentioned prior methods. [Means for solving the problem]

[0006] A system according to one or more embodiments of the present disclosure is disclosed. In one exemplary embodiment, the system comprises a processing unit communicatively connected to a detector array of an optical wafer characterization system. In another exemplary embodiment, the processing unit is configured to acquire one or more target images of a target location on a wafer from the detector array. In another exemplary embodiment, the processing unit is configured to determine one or more difference images from one or more reference images and one or more target images. In another exemplary embodiment, the processing unit is configured to generate one or more upsampled images by upsampling one or more difference images. In another exemplary embodiment, one or more wafer defects are detectable in one or more upsampled images.

[0007] Methods according to one or more embodiments of this disclosure are disclosed. In one exemplary embodiment, the method may include, not limited to, a processing unit acquiring one or more target images of a target location on a wafer from a detector array of an optical wafer characterization system. In another exemplary embodiment, the method may include, not limited to, a processing unit determining one or more difference images from one or more reference images and one or more target images. In another exemplary embodiment, the method may include, not limited to, a processing unit generating one or more upsampled images by upsampling one or more difference images. In another exemplary embodiment, one or more wafer defects are detectable in one or more upsampled images.

[0008] Disclosed are optical wafer characterization systems according to one or more embodiments of the present disclosure. In one exemplary embodiment, the optical wafer characterization system comprises a light source configured to generate and emit light. In another exemplary embodiment, the optical wafer characterization system comprises an illumination arm having one or more illumination optics configured to irradiate a wafer placed on a stage assembly with light received from the light source. In another exemplary embodiment, the optical wafer characterization system comprises a focusing arm having one or more focusing optics. In another exemplary embodiment, the optical wafer characterization system comprises a detector array having one or more sensors. In another exemplary embodiment, one or more focusing optics are configured to image the wafer onto the detector array. In another exemplary embodiment, the optical wafer characterization system comprises a processing unit communicatively connected to the detector array. In another exemplary embodiment, the processing unit is configured to acquire one or more target images of a target location on the wafer from the detector array. In another exemplary embodiment, the processing unit is configured to determine one or more difference images from one or more reference images and one or more target images. In another exemplary embodiment, the processing unit is configured to generate one or more upsampled images by upsampling one or more difference images. In another exemplary embodiment, one or more wafer defects are detectable in one or more upsampled images.

[0009] A system according to one or more embodiments of the present disclosure is disclosed. In one exemplary embodiment, the system comprises a processing unit communicatively connected to a detector array of an optical wafer characterization system. In another exemplary embodiment, the processing unit is configured to acquire one or more target images of a target location on a wafer from the detector array. In another exemplary embodiment, the processing unit is configured to apply a denoising filter to at least one or more target images. In another exemplary embodiment, the processing unit is configured to determine one or more difference images from one or more reference images and one or more target images after denoising at least one or more target images. In another exemplary embodiment, one or more wafer defects are detectable in one or more difference images.

[0010] Methods according to one or more embodiments of this disclosure are disclosed. In one exemplary embodiment, the method may, non-limitingly, include a processing unit acquiring one or more target images of a target location on a wafer from a detector array of an optical wafer characterization system. In another exemplary embodiment, the method may, non-limitingly, include a processing unit applying a denoising filter to at least one or more target images. In another exemplary embodiment, the method may, non-limitingly, include a processing unit determining one or more difference images from one or more reference images and one or more target images after denoising at least one or more target images. In another exemplary embodiment, one or more wafer defects are detectable in one or more difference images.

[0011] Disclosed are optical wafer characterization systems according to one or more embodiments of the present disclosure. In one exemplary embodiment, the optical wafer characterization system comprises a light source configured to generate and emit light. In another exemplary embodiment, the optical wafer characterization system comprises an illumination arm having one or more illumination optics configured to irradiate a wafer placed on a stage assembly with light received from the light source. In another exemplary embodiment, the optical wafer characterization system comprises a focusing arm having one or more focusing optics. In another exemplary embodiment, the optical wafer characterization system comprises a detector array having one or more sensors. In another exemplary embodiment, one or more focusing optics are configured to image the wafer onto the detector array. In another exemplary embodiment, the optical wafer characterization system comprises a processing unit communicatively connected to the detector array. In another exemplary embodiment, the processing unit is configured to acquire one or more target images of a target location on the wafer from the detector array. In another exemplary embodiment, the processing unit is configured to apply a noise reduction filter to at least one or more target images. In another exemplary embodiment, the processing unit is configured to determine one or more difference images from one or more reference images and one or more target images after denoising at least one or more target images. In another exemplary embodiment, one or more wafer defects are detectable in one or more difference images.

[0012] The above summary and the following detailed description are for illustrative and explanatory purposes only and do not necessarily limit the claimed invention. The accompanying drawings incorporated herein and forming part thereof illustrate embodiments of the present invention and, together with the summary, are used to explain the principles of the present invention. [Brief explanation of the drawing]

[0013] Many of the advantages of this disclosure can be better understood by those skilled in the art by referring to the attached drawings below. [Figure 1]Figure 1 shows a simplified block diagram of a system for optical wafer characterization with image upsampling according to one or more embodiments of the present disclosure. [Figure 2] Figure 2 shows a simplified schematic diagram of a system for optical wafer characterization with image upsampling according to one or more embodiments of the present disclosure. [Figure 3] Figure 3 is a flowchart illustrating the steps performed in a method for optical wafer characterization with image upsampling according to one or more embodiments of the present disclosure. [Figure 4] Figure 4 shows a block diagram of a method for optical wafer characterization with image upsampling according to one or more embodiments of the present disclosure. [Figure 5A] Figure 5A is a graph showing the separation of optical information from noise according to one or more embodiments of the present disclosure. [Figure 5B] Figure 5B is a graph showing the separation of optical information from noise according to one or more embodiments of the present disclosure. [Figure 6A] Figure 6A is a circular image showing the transfer function of a finite impulse response (FIR) filter according to one or more embodiments of the present disclosure. [Figure 6B] Figure 6B is a comparative graph of the transfer functions of FIR filters via x-space normalized frequencies according to one or more embodiments of the present disclosure. [Figure 7] Figure 7 is an image sequence showing an upsampling process using frequency domain zero padding according to one or more embodiments of the present disclosure. [Figure 8A] Figure 8A is an original bright-field (BF) image according to one or more embodiments of the present disclosure. [Figure 8B] Figure 8B is a regenerated BF image according to one or more embodiments of the present disclosure. [Figure 9A] Figure 9A is a graph comparing upsampling filters, difference filters, and combinations of upsampling and difference filters according to one or more embodiments of the present disclosure. [Figure 9B] Figure 9B is a graph comparing upsampling filters, difference filters, and combinations of upsampling and difference filters according to one or more embodiments of the present disclosure. [Figure 10A] Figure 10A shows an unblurred image according to one or more embodiments of the present disclosure. [Figure 10B] Figure 10B shows an image blurred by 2 pixels according to one or more embodiments of the present disclosure. [Figure 11] Figure 11 is a graph comparing the interpolation reconstruction error before blurring, after blurring by one pixel, and after blurring by two pixels, according to one or more embodiments of the present disclosure. [Figure 12A] Figure 12A shows an aliased image according to one or more embodiments of the present disclosure. [Figure 12B] Figure 12B shows an anti-aliased image according to one or more embodiments of the present disclosure. [Figure 13] Figure 13 is a graph showing the improvement in the mean signal-to-noise ratio (SNR) by comparing the SNR when anti-aliased with the improvement in the number of defects for the anti-aliased image, according to one or more embodiments of the present disclosure. [Figure 14] Figure 14 is a flowchart illustrating the steps performed in an optical wafer characterization method with image upsampling according to one or more embodiments of the present disclosure. [Figure 15] Figure 15 shows a simplified schematic diagram of a system for optical wafer characterization with image upsampling according to one or more embodiments of the present disclosure. [Modes for carrying out the invention]

[0014] This disclosure is illustrated and described in detail with respect to specific embodiments and their specific features. The embodiments described herein are intended to be illustrative and not limiting. It will be readily apparent to those skilled in the art that various changes and modifications can be made in form and detail without departing from the gist and scope of this disclosure.

[0015] The subject matter of this disclosure, as shown in the attached drawings, is referenced in detail below.

[0016] The increasing demand for electronic logic and memory devices with smaller footprints and features presents a wide range of manufacturing challenges that go beyond simply producing them at the desired scale.

[0017] For example, in broadband plasma (BBP) wafer inspection, reducing pixel size can improve sensitivity, but at the expense of throughput (e.g., high sensitivity, low throughput). Conversely, increasing pixel size can improve throughput, but at the expense of sensitivity (e.g., high throughput, low sensitivity). Therefore, a compromise between sensitivity and throughput is ultimately necessary.

[0018] In another example, BBP wafer inspection may include obtaining a difference image by subtracting the gray levels of the target image from a reference image, but the capacity of the filter applied to the difference image may be limited in order to improve the signal-to-noise ratio (SNR) of defects of interest. For the purposes of this disclosure, the phrase "filtering the difference image" may be understood to mean that a two-dimensional kernel is convolved with the image as the square of the size of the detector pixels.

[0019] In this example, the filter, i.e., "diff-filter" or "dif-filter," can only be applied to the difference image. It cannot be applied to the target and reference images because the images may change significantly, potentially degrading the image alignment algorithm. Therefore, it is conceivable that the benefits of diff-filter / dif-filter cannot be obtained for the target and reference images, even though noise in the target and reference images may potentially affect wafer inspection quality by influencing two-dimensional (2D) detection algorithms (e.g., multi-die automatic threshold (MDAT) algorithms) and attributes used in nuisance event filters (NEF) during BBP inspection. Furthermore, while diff-filter / dif-filter can be applied in the central processing unit (CPU) of the imaging computer in the BBP inspection system, in this case, the maximum filter that can be used without reducing the throughput of the imaging computer may be limited.

[0020] As another example, depending on the type and / or amount of sampling in the inspection mode used during BBP wafer inspection, aliasing noise may be introduced into the image.

[0021] Referring here to Figures 1A to 15, an optical wafer characterization system and method with image upsampling according to one or more embodiments of the present disclosure will be described.

[0022] Embodiments of this disclosure relate to optical wafer characterization systems and methods involving image upsampling. Embodiments of this disclosure further relate to improving the trade-off between throughput and sensitivity by using larger pixels through upsampling to achieve equivalent low pixel sensitivity. Embodiments of this disclosure further relate to spatial frequency filters or band-limited filters (BLFs) that can be used to improve the signal-to-noise ratio (SNR) of optical images imaged by a detector. Embodiments of this disclosure relate to anti-aliasing techniques.

[0023] Figures 1 and 2 show optical wafer characterization systems 100 and 200, respectively, according to one or more embodiments of the present disclosure. In general, systems 100 and / or system 200 may include any characterization systems known in the art, including, but not limited to, inspection or measurement systems. For example, systems 100 and / or system 200 may include a broadband plasma (BBP) optical inspection system. Thus, systems 100 and / or system 200 may be configured to perform inspection or optical measurement on a sample. However, it should be noted that the above description is not intended to limit the present disclosure but should be understood as illustrative only. Various embodiments and components described with respect to system 100 in Figure 1 should be understood to apply to system 200 in Figure 2, and vice versa.

[0024] In one embodiment, the system 100 comprises a light source 102 configured to generate and emit light, an illumination arm 104 configured to receive light, a focusing arm 106 configured to direct light to a detector array 108, and a controller 110 having one or more processors 112 and memory 114. For example, the light source 102 may, non-limitedly, include a broadband plasma (BBP) light source. For example, the light source 102 may, non-limitedly, include a laser-supported plasma (LSP) irradiation source 102.

[0025] As used throughout this disclosure, the term “sample” generally refers to a substrate made of a semiconductor or non-semiconductor material (e.g., a wafer, reticle / photomask, etc.). For example, the semiconductor or non-semiconductor material may, not limited to, single-crystal silicon, gallium arsenide, or indium phosphide. A sample may comprise one or more layers. For example, such layers may, not limited to, resists, dielectric materials, conductive materials, or semiconductor materials. Various types of such layers are known in the art, and as used herein, the term “sample” is intended to include samples on which any type of such layer may be formed. One or more layers formed on a sample may or may not be patterned. For example, a sample may comprise multiple dies, each having a repeatedly patterned feature. In another example, each die may comprise multiple fields. The formation and processing of the material of the layers may result in a final completed device. Various types of devices may be formed on a sample, and as used herein, the term “sample” is intended to include samples on which any type of device known in the art has been manufactured. Furthermore, for the purposes of this disclosure, the terms, sample, and wafer should be understood to be synonymous.

[0026] In embodiments, the sample 116 is placed on a stage assembly 118, which facilitates the movement of the sample 116. The stage assembly 118 may, in no particular way, include any stage assembly 118 known in the art, including XY stages, R-θ stages, and the like. In embodiments, the stage assembly 118 is adjustable in height during inspection or imaging so that focus on the sample 116 is maintained. Generally, the stage assembly 118 may, in no particular way, include one or more linear stages suitable for selectively moving the sample 116 linearly along one or more linear directions (e.g., x, y, and / or z directions). Furthermore, the stage assembly 118 may, in no particular way, include one or more rotating stages suitable for selectively rotating the sample 116 along a rotational direction. In another example, the stage assembly 118 may, in no particular way, include rotating and linear stages suitable for selectively moving the sample 116 linearly and / or rotating the sample 116 along a rotational direction.

[0027] In an embodiment, the illumination arm 104 is configured to guide light 120 from the light source 102 onto the sample 116. The illumination arm 104 may include any number and / or types of optical elements known in the art. For example, the illumination arm 104 may, non-limitingly, include one or more optical elements 122, a beam splitter 124, and an objective lens 126. For example, one or more optical elements 122 may include any optical element or combination of optical elements known in the art, non-limitingly including one or more mirrors, one or more lenses, one or more polarizers, one or more gratings, one or more filters, or one or more beam splitters. Thus, the illumination arm 104 may be configured to focal the light 120 from the light source 102 onto the surface of the sample 116.

[0028] In embodiments, the focusing arm 106 is configured to focus light reflected, dispersed, diffracted, and / or emitted from the sample 116. The focusing arm 106 may include any number and / or types of optical components known in the art. For example, the focusing arm 106 may, non-limitingly, include one or more optical elements 128. For example, one or more optical elements 128 may include, non-limitingly, one or more mirrors, one or more lenses, one or more polarizers, one or more gratings, one or more filters, or one or more beam splitters, or any combination of optical elements known in the art.

[0029] In the embodiment, the focusing arm 106 is configured to guide and / or focus light from the sample 116 to the sensor or detector 130 of the detector array 108. The detector 130 and the detector array 108 may include any sensor and detector assembly known in the art. The detector 130 may, in no particular way, include a charge-coupled device (CCD) detector, a complementary metal-oxide-semiconductor (CMOS) detector, a time-delay integral (TDI) detector, a photomultiplier tube (PMT), or an avalanche photodiode (APD). Furthermore, the detector 130 may, in no particular way, include a line sensor or an electron-impact line sensor.

[0030] In one embodiment, the detector array 108 is communicatively connected to the processing unit 132. For example, the processing unit 132 may, non-limited, include a field-programmable gate array (FPGA). In another example, the processing unit 132 may, non-limited, include an application-specific integrated circuit, a graphics processing unit (GPU), or a central processing unit (CPU). Here, the processing unit 132 may include one of the above examples or a combination thereof.

[0031] In this embodiment, the processing unit 132 is communicatively connected to a controller 110 having one or more processors 112 and memory 114. One or more processors 112 may be communicatively connected to memory 114. In this case, one or more processors 112 are configured to execute program instruction sets stored in memory 114. However, it should be noted that the processing unit 132 is not an independent component but may be incorporated into the controller 110. Also, it should be noted that the detector array 108 may be communicatively connected to a controller 110 having one or more processors 112 and memory 114. In this case, the processing unit 132 may be incorporated into the detector array 108 or the controller 110, and the processing unit 132 may not be a component of the system 100 (for example, the controller 110 may be configured to perform the functions of the processing unit 132).

[0032] In an embodiment, one or more processors 112 are configured to analyze the output of the detector array 108. In an embodiment, a set of program instructions is configured to cause one or more processors 112 to analyze one or more features of the sample 116. In an embodiment, a set of program instructions is configured to cause one or more processors 112 to modify one or more features of the system 100 to maintain focusing on the sample 116 and / or the detector 130. For example, one or more processors 112 may be configured to adjust the objective lens 126 or one or more optical elements 122 so that light 120 from the light source 102 is focused onto the surface of the sample 116. In another example, one or more processors 112 may be configured to adjust the objective lens 126 and / or one or more optical elements 128 so that illumination from the surface of the sample 116 is focused and the focused illumination is focused onto the detector 130.

[0033] System 100 can be configured in any optical configuration known in the art, including but not limited to dark-field configurations and bright-field orientations.

[0034] It should be noted that one or more components of system 100 may be communicated to various other components of system 100 in any manner known in the art. For example, the light source 102, the detector array 108, the processing unit 132, the controller 110, and / or one or more processors 112 may be communicated to each other and to other components by wired (e.g., copper wire or fiber optic cable) or wireless (e.g., RF connection, IR connection, WiMAX®, Bluetooth®, 3G, 4G, 4G LTE, or 5G).

[0035] In one embodiment, the system 200 comprises a light source 102, an illumination arm 104, a focusing arm 106, a detector array 108, and a controller 110 having one or more processors 112 and memory 114. The system 200 may be a system 200 arranged within a reflectance measurement method and / or polarization analysis method configuration. It should be noted that the system 200 may include any type of measurement system known in the art.

[0036] In one embodiment, light 120 from a light source 102 is guided to a sample 116 by an illumination arm 104. In another embodiment, the system 200 focuses the emitted light from the sample 116 via a focusing arm 106. The illumination arm 104 may include one or more beam adjustment elements 202 suitable for deforming and / or adjusting the light 120. For example, one or more beam adjustment elements 202 may, but are not limited to, one or more polarizers, one or more filters, one or more beam splitters, one or more diffusers, one or more homogenizers, one or more apodizers, one or more beam shapers, or one or more lenses.

[0037] In an embodiment, the illumination arm 104 may utilize a first focusing element 204 to focus and / or guide light 120 onto a sample 116 placed on a stage assembly 118. In an embodiment, the focusing arm 106 may have a second focusing element 206 to focus the light emitted from the sample 116. The focusing arm 106 may further have one or more focusing beam adjustment elements 208 to guide and / or deform the illumination focused by the second focusing element 206. For example, one or more focusing beam adjustment elements 208 may, but are not limited to, one or more lenses, one or more filters, one or more polarizers, or one or more phase plates.

[0038] In one embodiment, the detector array 108 is configured to image the synchrotron radiation emitted from the sample 116 through the focusing arm 106. For example, the detector assembly or array 108 may receive synchrotron radiation reflected or diffused (e.g., specular or diffuse reflection) from the sample 116. In another example, the detector array 108 may receive synchrotron radiation generated by the sample 116 (e.g., emission associated with the absorption of light 120). It should be noted that the detector array 108 may include any sensors and detector assemblies known in the art. The sensors may, but are not limited to, include CCD detectors, CMOS detectors, TDI detectors, PMTs, APDs, and the like.

[0039] System 200 may consist of any type of measuring tool known in the art. Such tools include, but are not limited to, a spectroscopic ellipsometer having one or more illumination angles, a spectroscopic ellipsometer for measuring Müller matrix elements (e.g., using a rotational compensator), a single-wavelength ellipsometer, an angle-resolved ellipsometer (e.g., a beam-profile ellipsometer), a spectrophotometer, a single-wavelength reflectometer, an angle-resolved reflectometer (e.g., a beam-profile reflectometer), an imaging system, a pupil imaging system, a spectrophotometer, or a scatorometer.

[0040] Suitable inspection / measurement tools for implementation in the various embodiments of this disclosure are described below, and these are incorporated herein by reference in their entirety. The inventions include: "Split Field Inspection System Using Small Catadioptric Objectives," U.S. Patent Application Publication No. 2009 / 0180176, published on July 16, 2009; "Beam Delivery System for Laser Dark-Field Illumination in a Catadioptric Optical System," U.S. Patent Application Publication No. 2007 / 0002465, published on January 4, 2007; "Ultra-broadband UV Microscope Imaging System with Wide Range Zoom Capability," U.S. Patent No. 5,999,310, issued on December 7, 1999; "Surface Inspection System Using Laser Line Illumination with Two Dimensional Imaging," U.S. Patent No. 7,525,649, issued on April 28, 2009; and the invention by Wang et al. titled "Dynamically Adjustable Semiconductor Metrology." U.S. Patent Application Publication No. 2013 / 0114085, published on May 9, 2013, is titled "Focused Beam Spectroscopic Ellipsometry Method and System" by Piwonka-Corle et al., and is U.S. Patent No. 5,608,526, issued on March 4, 1997, is titled "Apparatus for Analyzing Multi-Layer Thin Film Stacks on Semiconductors" by Rosencwaig et al., and is U.S. Patent No. 6,297,880, issued on October 2, 2001.

[0041] In embodiments, systems 100, 200 may be configured as “standalone tools,” which are understood herein to be tools not physically connected to process tools. In other embodiments, such as inspection or measurement systems, systems 100, 200 may be connected to process tools (not shown) by a transmission medium which may include wired and / or wireless components. For example, process tools may include any process tools known in the art. Examples include lithography tools, etching tools, deposition tools, polishing tools, plating tools, cleaning tools, and ion implantation tools. The results of inspections or measurements performed by systems 100, 200 as described herein may be used to modify process or process tool parameters using feedback control techniques, feedforward control techniques, and / or in-situ control techniques. Process or process tool parameters may be modified manually or automatically.

[0042] Embodiments of systems 100 and 200 may be further configured as described herein. Furthermore, systems 100 and 200 may be configured to perform any other step(s) of any embodiment(s) of the method described herein.

[0043] Figure 3 shows a flowchart of a method or process 300 for optical wafer characterization with image upsampling according to one or more embodiments of the present disclosure. Figure 4 shows a block diagram 400 of a method or process 300 for optical wafer characterization with image upsampling according to one or more embodiments of the present disclosure. One or more steps of the method or process 300 may utilize information such as one or more images from system 100 and / or system 200. One or more steps of the method or process 300 may be performed on or by one or more components of system 100 and / or system 200. For example, one or more steps of the method or process 300 may be performed by processing unit 132 and / or controller 110.

[0044] In step 302, one or more reference images are acquired. In step 304, one or more target images are acquired. For example, one or more target images may be images of a location (e.g., a field) on a die on wafer 116 under test, imaged by the detector array 108, and one or more reference images may be images of the same location (e.g., a field) from a nearby die (e.g., neighboring, adjacent, or within an interval distance, etc.) on wafer 116, imaged by the detector array 108. In another example, one or more target images may be images of a location (e.g., a field) on a die on wafer 116 under test, imaged by the detector array 108, and one or more reference images may be historical data of the same location (e.g., a field) from a die on another wafer 116. For example, one or more target images may be acquired at a preferred pixel size. The acquired image 402 may include one or more reference images and one or more target images. However, please note that here, one or more target images will be shown as acquired image 402 (for example, if one or more reference images are historical data).

[0045] In a non-limiting example, broadband plasma (BBP) inspection systems 100, 200 are automated inspection microscopes for inspecting defects in semiconductor wafers 116. The sample 116 may be placed on an operating stage 118 that moves the sample 116 continuously. The sample 116 can be imaged onto a detector array 108 by a detector 130. The imaging optical system of systems 100, 200 may include a zoom lens. For example, the zoom lens may adjust the magnification from the sample 116 to the detector array 108. For example, at a high magnification (e.g., 600x), the image of one detector pixel on the sample 116 may be 1 / 600th the actual size of the pixel on the detector 130, and the field of view of the detector 130 may be 1 / 600th of the detector array 108. The detector 130 may operate in TDI mode. For example, photogenerated charge carriers on the CCD detector 130 can be swept in by an applied electrolysis in sync with the movement of the image of the sample 116 across the CCD detector 130. The charges can be digitized when they reach the edge of the detector array 108.

[0046] Systems 100 and 200 may have a laser-excited plasma light source 102 that performs irradiation in a wide spectrum. The illumination optical system in the illumination arm 104 may illuminate a part of the sample 116. For example, the illumination optical system may insert a wavelength filter and / or a polarizer into the illumination arm 104. The irradiation field on the sample 116 may be fixed and independent of the imaging zoom setting of the condenser arm 106. For example, the irradiation field on the sample 116 may be at least as large as the field of view of the detector 130 at the minimum magnification. It should be noted that since the range of the BBP light source 102 is too wide to image all the generated light onto the field of view of the detector 130 without sacrificing light intensity, a zoom lens is not provided in the illumination optical system, so the irradiation field can be fixed. Therefore, the light beam F [photons / s / m 2 reflected from the sample 116 may be independent of the imaging magnification.

[0047] Regarding the pixel size p at the sample 116, the number of photons n reaching each pixel of the detector 130 per second is expressed as n = η * F * p 2 where η represents the transmission efficiency of the imaging systems 100 and 200 and the detector 130, and p is the physical dimension of the pixel on the detector array 108 divided by the magnification in the condenser arm 106.

[0048] For example, when TDI is clocked such that a row of pixels is digitized at a frequency f L , the TDI integration time for each pixel is T int = n x / f L where n x represents the number of pixels on the detector array 108 in the scanning direction (e.g., the x direction). Further, the speed v of the stage assembly 118 is v = p * f L . Further, when the height of the detector 130 in the y direction is n x pixels, the detector 130 has a band height h swath = p * n xThe strip region of wafer 116 can be scanned. Furthermore, the inspection throughput (e.g., measured per wafer inspected per unit time) is T put =(v*h swath ) / A=(p 2 *f L *n y ) / A, where A is the inspection area on sample 116. Furthermore, n is the number of photons focused by detector 130 pixels. v is, n v =η*F*p 2 *T int =(η*F*p 2 *n x ) / f L This is shown. When combining the above, throughput relates to the number of photons in the detector 130 pixels, as shown in Equation 1 below. T put =(p 2 *f L *n y ) / A=(p 4 *F*n x *n y ) / (A*n y ) formula 1

[0049] Throughput T put This increases depending on the pixel size on sample 116 with or without filtering. Here, the number of photons per pixel is constant, and the throughput T put p is the fourth power of the pixel length dimension on sample 116. 4 It is proportional to the throughput T. put Assuming that is constant, the number of photons per pixel is p 4 It is proportional to the throughput T. This relationship holds until the detector is saturated (for example, the potential well of the CCD is filled). However, when a filter is applied to the image, the throughput T put ga p 2 It decreases in proportion to . Here, the short noise variable is kept constant.

[0050] It should be noted that all acquired images 402 (e.g., one or more target images and / or one or more reference images) may be upsampled before applying the denoising filter 404 and / or generating one or more difference images 406. Upsampling will be described in detail herein.

[0051] In step 306, a noise reduction filter 404 is applied to the acquired image. For example, the frequency plane (f x ,f y A filter 404 that rejects spatial frequencies exceeding 2*NA / λmin in any direction in the above can be used to denoise one or more reference images and / or one or more target images.

[0052] The spatial resolution of system 100 may depend on the proximity of two points that can be combined and resolved. In both dark-field and / or bright-field imaging systems, the optical image has a maximum spatial frequency of f in any direction. max The bandwidth is limited in the spatial frequency plane so as not to exceed =2*NA / λmin. In the formula, NA is the numerical aperture of the imaging optical system on the sample 116 side, λmin is the shortest or minimum wavelength in the spectrum of the detected light, and the unit of spatial frequency f is the period per unit length. Note that when NA is the numerical aperture on the detector side, the spatial frequency is the round-trip distance in the detector array plane. Furthermore, note that when NA is the numerical aperture of the object in this specification, the spatial frequency is the round-trip distance in the object (wafer) plane.

[0053] One of the advantages of this disclosure is a spatial frequency filter or band-limited filter (BLF). A BLF can pass all spatial frequencies within a half-shape 2*NA / λmin circle (e.g., transfer function amplitude = 1, phase = 0) and block all frequencies outside the circle (e.g., transfer function amplitude = 1). With a BLF, a noise-free optical image may remain unchanged (e.g., no blurring or distortion occurs). Therefore, a BLF may not adversely affect image alignment and / or image processing algorithms. On the other hand, a BLF can improve the signal-to-noise ratio (SNR) of an image by filtering out some of the noise in the image. Therefore, a spatial frequency filter can be understood as being configured to remove some of the noise in one or more target images and / or one or more reference images acquired by the detector array 108, thereby suppressing changes in the noise-free image.

[0054] Since detector 130 detects integer photons that follow a Poison distribution, the image contains shot noise. For a given light intensity, the average number of photons is m, which can be a fraction. Detector 130 has P(n)=e -m *m n An integer n photons can be randomly detected with a probability of / n!. Number of photons in a 130-pixel detector (σ 2 The random variation of =m) is uncorrelated between pixels. Therefore, shot noise has a white spectrum (e.g., its spatial spectral density is constant), and the spectral portion of this white noise outside a circle of radius 2*NA / λmin can be filtered out by the BLF without affecting the optical image.

[0055] Readout noise can generate broadband noise. For example, readout noise can be generated by electronic devices during A / D conversion of the output of detector array 108. BLF can filter out the spectral portion of readout noise outside a circle with radius 2*NA / λmin.

[0056] In a non-limiting example where the pixel pitch is p and the detector array 108 has N × N pixels, the detector array 108 can sample a (1 / p) × (1 / p) square region in the spatial frequency plane. If the gray level in the field of view is uniform, the shot noise can be uniformly distributed in this plane, and the BLF can pass the shot noise within a circle of radius 2 * NA / λmin. The noise variation is reduced by the fraction shown in Equation 2 below. Filtered shot noise variable / Unfiltered shot noise variable = 4π(p*NA) 2 / λ 2 min = π / 4 * (p / Δ Nyquist ) 2 formula 2 In the formula Δ Nyquist = 1 / 2f min This is true.

[0057] p < Δ Nyquist In this case, the image is oversampled, and the larger the oversampling factor, the greater the signal-to-noise advantage of the filter. When the image is oversampled, shot noise and the optical image can be more easily separated. The SNR can be improved by oversampling, which filters out the prominent shot noise from the optical image. This improvement in SNR may be achieved at the expense of speed. Note that the filter reduces shot noise by approximately 20% even near critical sampling.

[0058] Figures 5A and 5B show graphs 500 and 510, respectively, illustrating the separation of optical information 502 from noise 504 (e.g., shot noise, readout noise, etc.). In the non-limiting examples shown in Figures 5A and 5B, the optical information is contained in a disk with radius 4πNA / λmin and filter value 1, while the noise 504 is outside the disk and may be given a filter value of 0. Note that the more oversampled the configuration (e.g., comparing graph 510 with grid size (1 / t2) × (1 / t2) to graph 500 with grid size (1 / t1) × (1 / t1)), the greater the percentage of shot noise removed.

[0059] In some embodiments, BLF can be performed using one of two methods: finite impulse response (FIR) or fast Fourier transform (FFT).

[0060] When performed using FIR, FIR is a straightforward convolution with filter coefficient h. The filtered image can be calculated as shown in Equation 3.

number

[0061] The ideal transfer function is shown in Equation 4.

number

[0062] In Equation 4,

number

number

number

number

[0063] In Equation 4, the weight function w(f x ,f y As shown in Equation 6, the target g(f) is around the discontinuity during the transition from the passband to the stopband. x ,f y This could contribute to relaxing the requirements to conform to ).

number

[0064] In one non-restrictive example, ∈ in Equation 6 is set to 0.05. Note that by not confining the transfer function within ±5% of the radius of the circular passband, passband penetration outside the ±5% margin and attenuation in the stopband can be improved.

[0065] When performed using FFT, the Fourier transform is an experimentally obtained value, and as shown in Equation 7, the filter is limited to symmetrical values, thus possessing the same symmetry. h(x,y)=h(-x,y)=h(x,-y)=h(y,x) Equation 7

[0066] Figure 6A shows Image 600 of circles illustrating the transfer function of a 15×15 FIR filter designed for a 30 nanometer (nm) pixel pitch, 0.9 NA, and a minimum wavelength of 190 nm. In Image 600, the spatial frequency axis f x and f y is, f max It is normalized by =2*NA / λmin

[0067] Figure 6B shows the transfer function of a 15×15 FIR filter designed for a 30 nanometer (nm) pixel pitch, 0.9 NA, and a minimum wavelength of 190 nm, through the normalized x spatial frequency (e.g., through the zero y spatial frequency region, or f). y Graph 610 compares this with (=0). In Graph 610, the spatial frequency axis is f max It is normalized by =2*NA / λmin

[0068] The implementation using FFT is shown in Equations 8 and 9. Image out =F -1 {h(f x ,f y )FImage in} expression 8

number

[0069] In equations 8 and 9, F represents FFT, -1 This shows the inverse FFT, f x and f y This indicates a spatial frequency variable (e.g., a transformation variable), and f r represents the radial spatial frequency. The dimensionless variable ∈ can be considerably smaller than 1 (e.g., 0.05), and can define the width of a smooth cosine transition from the passband to the stopband. Here, a smooth transition is defined as Image out Note that this can reduce ringing (e.g., vibration) at (x,y).

[0070] In equations 8 and 9, the FFT F results in a circle where the selected image frequencies fall within the circle, and the unselected image frequencies and / or noise fall outside the circle. Therefore, the inverse FFT F -1 Noise can be removed before the calculation.

[0071] When comparing BLF performed using FIR and BLF performed using FFT, it should be understood that when the input image size is large (e.g., 1024 x 1024 pixels) and the FIR filter size is small (e.g., 20 x 20 pixels), the BLF performed using FFT approximates the ideal transfer function more closely than the BLF performed using FIR. Therefore, higher accuracy can be achieved when using FFT.

[0072] However, in terms of computational throughput, FIR-based implementation may be preferable. Note that the processing unit 132 may apply FIR-based implementation (e.g., a 15×15 FIR filter) to raw images acquired by one or more detectors 130 (for example, if the processing unit 132 is located between the detector array 108 and the controller 110 in the system 100, or within the detector array 108, and is configured to receive raw images before transmission to the controller 110).

[0073] It should be noted that different filters may be required for different SNR and / or NA values. Therefore, embodiments of this disclosure relate to a dynamic optical setup in which filters can be adjusted according to the SNR and / or NA value.

[0074] Embodiments of this disclosure demonstrate the implementation of a noise reduction filter, but it should be noted that the noise reduction filter 404 may be omitted from the method or process 300. Therefore, the above description should be understood as illustrative only and not as limiting to this disclosure.

[0075] In step 308, one or more difference images 406 are determined from one or more acquired images. For example, one or more difference images 406 may be determined by transforming one or more target images with one or more reference images. For example, one or more difference images 406 may be determined by subtracting one or more reference images from one or more target images.

[0076] In step 310, one or more difference images 406 are upsampled to generate one or more upsampled images 408. For example, one or more wafer defects may be detectable within one or more upsampled images 408.

[0077] It should be noted that smaller pixel sizes on the wafer (e.g., high magnification) are advantageous in terms of sensitivity. Multiple different regions in the integrated circuit design on the wafer under test can be inspected with different sensitivity settings. Image mapping to design features can be completed by assigning a region ID to each image pixel. Generally, integrated circuit design features can be 1 / 2 to 1 / 3 the size of the pixels in the optical image. Therefore, the smaller the pixels, the more useful the mapping to design features becomes.

[0078] Smaller pixel sizes are advantageous for sensitivity, but larger pixel sizes (e.g., lower magnification) are preferable for throughput. Embodiments of this disclosure relate to improving the trade-off between throughput and sensitivity by acquiring an image with the maximum pixel size that does not undersample, and then algorithmically upsampling that image. Note that upsampling may make it easier to find signal peaks and increase signal intensity. Also note that upsampling may make it easier to fit the image to the actual design than when smaller pixels are used.

[0079] To avoid undersampling (e.g., aliasing), the pixel pitch should satisfy the inequality p < λmin / 4*NA. For example, the preferred pixel size on wafer 116 is the largest pixel size that satisfies the inequality p < λmin / 4*NA with a 5% margin.

[0080] As a non-restrictive example of upsampling, upsampling is performed using the Whittaker-Shannon interpolation equation, as shown in Equation 10.

number

[0081] In equation 10, p is Image in (m,n) indicates the sampling interval, where integers m and n are the column and row indices of the image to be interpolated. out (x,y) is the interpolated output, and (x,y) is the center coordinate of the pixel in the output image. The variables x, y, and p have the same distance unit and are upsampled Image. out The pixels at (x,y) are smaller than p. The normalized sinc function can be defined as sinc(z) = sin(πz) / πz.

[0082] The interpolation error approaches zero as the integer N approaches infinity. For the purposes of this disclosure, one preferred computational throughput and interpolation accuracy is achieved at N=6. However, any suitable interpolation is usable for the purposes of this disclosure as long as p < λmin / 4*NA holds, which guarantees that the optical image is oversampled according to the Nyquist theorem.

[0083] In another non-limiting example of upsampling, applying image interpolation or upsampling to a subset of image data that satisfies the Nyquist criterion (e.g., in a sequence of analog images, images must be sampled at a rate of at least twice the highest frequency) using only images from defects that have passed the detection algorithm may result in improved defect signals within the defect patch, which can then be used for further post-processing filtering or improved feature extraction for nuisance and / or binning. For example, it may be preferable to use a large pixel size that does not meet the Nyquist criterion to achieve target inspection throughput or to achieve appropriate light levels in a small light-dark-field optical mode.

[0084] Figure 7 shows a sequence of images 700, 710, 720, and 730 illustrating the upsampling process using frequency-domain zero padding. In a non-limiting example, upsampling is performed on the Image in Take the two-dimensional FFT F of (for example, from image 700 to image 710), multiply or zero-padding the transformed array (from image 710 to image 720), and take the inverse FFT F of the padded array (for example, from image 720 to image 730). -1 This can be done by taking [a specific action]. The zero-padding spectrum can be inversely transformed. The result includes interpolated values ​​that can be controlled within the range of zero-padding in the frequency domain. For example, 2x upsampling (e.g., interpolating equidistant points between the original pixels) is achieved with 2x zero-padding.

[0085] Sampling of a defective signal is a statistical process, and any single acquisition may result in some improvement in SNR due to upsampling. Further improvement in the defective SNR can be achieved with diff filters designed to improve defect energy while reducing noise. For example, a diff filter may be applied before upsampling. Therefore, a diff filter and / or denoising filter that perform FIR and / or FFT to improve the upsampled pixels compared to smaller pixels may be combined with upsampling. This improves the trade-off between throughput and sensitivity. For example, upsampling with a diff filter may result in improved noise. However, it should be noted that while a diff filter can degrade the defective SNR if not properly designed, upsampling does not degrade the defective SNR.

[0086] Figures 8A and 8B show images 800 and 810 of the bright-field (BF) images, respectively. In Figure 8A, image 800 includes the actual 36nm BF image. In Figure 8B, image 810 includes the reconstructed 36nm BF image by upsampling the 50nm image with an upsampling factor of 50 / 36 = 1.39. A comparison of Figures 8A and 8B shows that the 36nm BF image is reconstructed with high accuracy, as shown in the cross-sectional comparison. Under selective conditions, the actual 36nm pixels may show a 15% improvement over 50nm pixels, and 36nm upsampling may show a 10% improvement over 50nm pixels.

[0087] Figures 9A and 9B show graphs 900 and 910 of the SNR improvement (%) for various defects by comparing upsampling (e.g., Upsample), diff-filter / dif-filter (e.g., AutoDiff), and a combination of upsampling and diff-filter / dif-filter (e.g., Combination). As shown in Figures 9A and 9B, the combination of upsampling and diff-filter / dif-filter achieves a greater improvement than either upsampling or diff-filter / dif-filter alone. Under selection conditions, the improvement from upsampling averages 7 to 8%, the improvement from diff-filter / dif-filter averages 18 to 28%, and the improvement from the combination of diff-filter / dif-filter and upsampling is 6 to 9% higher.

[0088] As an attempt to further improve the defect SNR, anti-aliasing pre-filtering may be performed for aliasing-enabled inspection modes in field-unit inspection. Here, a field can be understood as a comparable region, such as a repeating logical pattern area within a die. Image aliasing is an error resulting from the digitization of an image at a rate insufficient to image all image frequencies passing through an optical system. Anti-aliasing can be a filtering process used in digital image and signal processing to limit the frequency content in an analog image or signal before digitization. In this case, the filter must be designed to meet the Nyquist criterion and not produce aliasing.

[0089] A method for suppressing aliasing can be found in U.S. Patent No. 8,947,521, issued on February 3, 2015, which is incorporated in its entirety herein. In a non-limiting example of the application of such a method, a TDI-type method for anti-aliasing pre-filters may include intentionally introducing blur in the image in the x and / or y directions across one or more detectors 130. Blur can be introduced into the image on the TDI before readout by introducing a small error in the TDI clock, stage speed, or image magnification. For an N-pixel wide TDI, a 2 / N error in the nominal TDI magnification introduces a blur of 2 pixels into the final image. This blur can be modeled as a convolution by a square function in the image domain, or equivalently, as a multiplication by a sinc function in the spatial frequency domain. The resulting sinc pre-filter may include side lobes. As a result, some aliasing energy leaks into the passband.

[0090] Equation 11 shows the imaging region synthesis equation including pixelation and image blurring.

number

[0091] The corresponding modulation transfer function in the frequency domain is shown in Equation 12.

number

[0092] Note that the blur required to minimize aliasing energy and generate spatial frequencies cut off at half the sampling frequency may be approximately 2 pixels. For example, for a given undersampling pixel rate r, the frequency components must be limited to r / 2. Therefore, a blur of 2 pixels is required.

[0093] Figures 10A and 10B show images 1000 and 1010, respectively. In Figure 10A, image 1000 is not blurred. In Figure 10B, image 1010 is anti-aliased with a blur of 2 pixels. A comparison of Figures 9A and 9B shows that the high-frequency spectral X component is significantly attenuated by the blur of 2 pixels.

[0094] Figure 11 is a graph 1100 showing the interpolation reconstruction error by comparing the error percentage (%) relative to the interpolation distance between an image without blur (no filtering), an image with 1 pixel of blur, and an image with 2 pixels of blur. As shown in Figure 11, 1 pixel of blur may be insufficient for suppressing aliasing noise, while 2 pixels of blur generally results in an error of <5%, which is considered minimal aliasing.

[0095] Figures 12A and 12B show images 1200 and 1210, respectively, relating to the number of defects. Figure 12A shows the aliasing target (e.g., without pre-filtering). Figure 12B shows the anti-aliasing target. A comparison of Figures 12A and 12B shows that when field-level inspection is performed, the SNR is improved by approximately 2 times in field-level inspection with aliasing compared to field-level inspection without pre-filtering. Under selective conditions, when anti-aliasing is applied during field-level inspection, the defect signal can be improved by an average of 14 to 20%.

[0096] Figure 13 is a graph showing the average SNR improvement, comparing the number of defects with the improvement in SNR under anti-aliasing compared to aliasing, expressed as a percentage. As shown in Figure 13, under selection conditions, an average improvement of 117% was observed when anti-aliasing pre-filtering was applied to the image.

[0097] Here, it should be noted that in the case of die-unit inspection, the average defect signal and the SNR in the case of anti-aliasing may be slightly degraded, but the improvements achieved by anti-aliasing techniques in field-unit inspection will offset and / or outweigh such slight degradation.

[0098] In step 312, one or more upsampled images 408 pass through an anomaly (or defect) detector. The anomaly detector 410 looks for pixel values ​​that are significantly higher or lower than the background. For example, the anomaly detector 410 may look for pixel values ​​that exceed a set threshold that is above or below the background. For example, the anomaly detector 410 may adjust or remove upsampled image pixel values ​​that exceed a set threshold that is above or below the background. Note that, if there are no defects, the target and reference images are preferably identical and their difference is zero. However, adjacent dies are often not identical, and the difference image is not zero.

[0099] Therefore, a minimum magnification is selected to accommodate the oversampling of the optical image, and by upsampling the image, a smaller pixel size is obtained. In particular, this improves the signal-to-shot noise ratio, provided that the optical image does not saturate the detector 130. Furthermore, this is advantageous when the illumination intensity is not sufficient to utilize the well depth, i.e., the dynamic range of the detector array 108.

[0100] In step 314, one or more control signals are determined to make one or more adjustments to one or more optical inspection components. For example, one or more control signals may be determined based on one or more upsampled images. In another example, one or more control signals may be determined based on one or more difference images (for example, instead of, or in addition to, determining one or more control signals based on one or more upsampled images). Generally, one or more control signals may be configured to selectively adjust one or more features of one or more components of system 100 and / or system 200.

[0101] The embodiments and implementation techniques described above for System 100 and / or System 200 should be understood to apply to Method or Process 300. Therefore, the steps of Method or Process 300 may be performed in System 100 and / or System 200, and Method or Process 300 may further include one or more steps required or suggested by the architecture of System 100 and / or System 200. However, it will be understood that Method or Process 300 is not limited to the architecture of System 100 and / or System 200, and that one or more steps or parts thereof of Method or Process 300 may be performed in alternative system components and / or architectures. Furthermore, the steps of Method or Process 300 may be performed in any order unless otherwise specified herein.

[0102] Figure 14 shows a method or process 1400 according to one or more embodiments of the present disclosure. Figure 15 shows a system 1500 for performing one or more steps of the method or process 1400 according to one or more embodiments of the present disclosure. The method or process 1400 may utilize information from the system 1500. The method or process 1400 may be performed on or by one or more components of the system 1500.

[0103] In step 1402, one or more semiconductor manufacturing processes are performed. One or more semiconductor manufacturing processes may be performed on sample 116 by one or more semiconductor manufacturing tools 1502. For example, one or more semiconductor manufacturing tools 1502 may include any microelectronics manufacturing process tools known in the art. This may include, but is not limited to, lithography tools and / or plasma process tools (e.g., etching tools, deposition tools, polishing tools, scanners, etc.).

[0104] In step 1404, optical wafer characterization with image upsampling is performed. For example, one or more steps of the method or process 300 may be performed by system 100 and / or system 200, as described throughout this disclosure.

[0105] In step 1406, one or more control signals are determined to make one or more adjustments to one or more manufacturing processes. For example, one or more control signals may be determined based on one or more upsampled images generated by one or more steps of the method or process 300. In another example, one or more control signals may be determined based on one or more difference images generated by one or more steps of the method or process 300 (for example, instead of, or in addition to, determining one or more control signals based on one or more upsampled images). Generally, one or more control signals may be configured to selectively adjust one or more features of one or more components of system 100 and / or system 1500.

[0106] One or more semiconductor manufacturing tools 1502 can be communicated to a controller 1504 having one or more processors 112 and memory 114.

[0107] Controller 1504 may be communicatively connected to system 100 and / or system 200. Here, controller 1504 may be isolated from controller 110, or may be communicatively connected to controller 110 directly or via a third-party server. Furthermore, it should be noted here that controller 110 may be shared by one or more semiconductor manufacturing tools 1502, system 100, and / or system 200 (for example, so that controller 1504 is not required).

[0108] The controller 1504 may be configured to generate one or more control signals configured to adjust one or more characteristics of one or more process tools in a feedforward or feedback loop based on inputs received from system 100 and / or system 200. The controller 1504 may be connected to the user interface 1506.

[0109] One or more processors 112 may include any one or more processing elements known in the Art. In that sense, one or more processors 112 may include any microprocessor-type device configured to execute software algorithms and / or instructions. One or more processors 112 may be a desktop computer, a mainframe computer system, a workstation, an image computer, a parallel processor, or other computer system (e.g., a networked computer) and may be configured to execute a program configured to operate system 100 as described throughout this disclosure. It should be understood that the steps described throughout this disclosure may be performed by a single computer system or by multiple computer systems. Furthermore, the steps described throughout this disclosure may be performed by any one or more of the one or more processors 112. In general, the term “processor” may be broadly defined to include any device having one or more processing elements that execute program instructions obtained from memory 114. Furthermore, various subsystems of system 100 (e.g., detector array 108, processing unit 132, user interface 1506, etc.) may include, or may contain, processors or logic elements suitable for performing at least a portion of the steps described throughout this disclosure. Therefore, the above statements should be understood as examples only, and not as limitations on this disclosure.

[0110] The memory 114 may include any storage medium known in the art that is suitable for storing program instructions executable by the associated one or more processors 112 and data received from the measurement subsystem. For example, the memory 114 may include a non-temporary storage medium. For example, the memory 114 may include, but is not limited to, a read-only memory (ROM), random access memory (RAM), magnetic or optical storage devices (e.g., disks), magnetic tape, a solid drive, and the like. Furthermore, the memory 114 may be housed together with one or more processors 112 in a common controller housing. In an alternative embodiment, the memory 114 may be located remotely from the physical locations of the processors 112, controller 110, controller 1504, etc. In another embodiment, the memory 114 holds program instructions, which may cause one or more processors 112 to perform the various steps described throughout this disclosure.

[0111] In one embodiment, the user interface 1506 is communicably connected to the controller 1504. In another embodiment, the user interface 1506 may include, but is not limited to, one or more desktops, tablets, smartphones, smartwatches, etc. In yet another embodiment, the user interface 1506 has a display. The display is used to show data from system 100 to the user. The display of the user interface 1506 may include any display known in the art. Examples of such displays may include, but is not limited to, liquid crystal displays (LCDs), organic light-emitting diode (OLED) displays, or CRT displays. Those skilled in the art will understand that any display device that can be integrated with the user interface 1506 is suitable for implementation in this disclosure. In yet another embodiment, the user may input selections and / or commands in response to data displayed to the user through the user input device of the user interface 1506.

[0112] The embodiments and implementations described herein with respect to System 1500 should be understood to apply to Method or Process 1400. Thus, the steps of Method or Process 1400 may be performed by System 1500, and Method or Process 1400 may further include one or more steps required or suggested by the architecture of System 1500. However, it will be understood that Method or Process 1400 is not limited to the architecture of System 1500, and one or more steps or parts thereof of Method or Process 1400 may be performed by alternative system components and / or architectures. Furthermore, the steps of Method or Process 1400 may be performed in any order unless otherwise specified herein.

[0113] It should be noted that methods or processes 300, 1400 are not limited to the steps and / or substeps described herein. The number of steps and / or substeps included in methods or processes 300, 1400 may be more or less than the number described herein. The steps and / or substeps of methods or processes 300, 1400 may be performed simultaneously. The steps and / or substeps of methods or processes 300, 1400 may be performed sequentially, including in the order described herein or in a different order. Therefore, the above description should be understood as illustrative only and not as limiting the scope of this disclosure.

[0114] Therefore, this disclosure offers many advantages over conventional systems and methods. The advantages of this disclosure relate to optical wafer characterization systems and methods with image upsampling. The advantages of this disclosure further relate to improving the trade-off between throughput and sensitivity by using larger pixels through upsampling to achieve equivalent small-pixel sensitivity. The advantages of this disclosure further relate to spatial frequency filters or band-limited filters (BLFs) that can be used to improve the signal-to-noise ratio (SNR) of the optical image formed by the detector. The advantages of this disclosure further relate to anti-aliasing techniques.

[0115] Those skilled in the art will understand that the various components (e.g., operations), devices, objects, and associated descriptions described herein are provided as examples for conceptual clarity, and that various configuration modifications are anticipated. Accordingly, the specific examples and associated descriptions described herein are intended to represent a more general classification. Specifically, the use of any specific example is intended to represent a classification, and should not be taken as an limitation that it does not include specific components (e.g., operations), devices, and objects.

[0116] Those skilled in the art will understand that there are various means (e.g., hardware, software, and / or firmware) by which the processes and / or systems and / or other technologies described herein can be implemented, and that the appropriate means will depend on the context in which the processes and / or systems and / or other technologies are used. For example, if the implementer determines that speed and accuracy are critical, they will primarily choose hardware and / or firmware-based means; if flexibility is critical, they will primarily choose software-based implementations; and if neither is critical, they will choose some combination of hardware, software, and / or firmware. Thus, among the various potential means of implementing the processes and / or devices and / or other technologies described herein, none is inherently superior to the others, and the choice of means is a matter of selection, depending on the context in which the means are most important and the implementer's specific concerns (e.g., speed, flexibility, or predictability).

[0117] The above description is provided to enable those skilled in the art to implement and use the invention as presented in the context of specific uses and conditions. The directional terms used herein, such as “top,” “bottom,” “up,” “down,” “upward,” “upward,” “downward,” “downward,” and “downward,” are intended to indicate relative positions for descriptive purposes and do not specify an absolute coordinate system. It will be apparent to those skilled in the art that various modifications can be made to the described embodiments. Furthermore, the general principles set forth herein can be applied to other embodiments. Thus, the invention is not intended to be limited to the specific embodiments illustrated and described herein, but should be tied to the maximum technical scope that corresponds to the principles and novel features disclosed herein.

[0118] With regard to virtually all use of plural and / or singular forms in this specification, those skilled in the art will be able to substitute plural forms for singular forms and / or singular forms for plural forms as appropriate to the context and / or use. For clarity, various singular / plural substitutions will not be explicitly described here.

[0119] All methods described herein may include storing the results of one or more steps of the method embodiments in memory. These results may include any of the results described herein and may be stored in any manner known in the art. The memory may include any memory described herein or any other suitable storage medium known in the art. After the results are stored, they may be accessible in memory and used by any embodiment of the method or system described herein, and may also be formatted for user display and used by another software module, method, or system. Furthermore, these results may be stored “permanently,” “semi-permanently,” “temporarily,” or for a certain period of time. For example, the memory may be random-access memory (RAM), but these results may not necessarily persist in memory indefinitely.

[0120] Furthermore, each of the embodiments of the methods described above may incorporate any other step(s) of any other method(s) described herein. In addition, each of the embodiments of the methods described above may be performed using any of the systems described herein.

[0121] The subject matter described herein may, in some cases, represent different components that are contained within or related to other components. It should be understood that the architectures described in this manner are merely typical, and in fact, many other architectures can be implemented to achieve the same functionality. Conceptually, any configuration consisting of components that achieve the same functionality is effectively “associated” in such a way that the desired functionality is achieved. Therefore, any two components combined herein to achieve a particular functionality can be considered “associated” with each other, regardless of the architecture or intervening components, in such a way that the desired functionality is achieved. Similarly, any two such associated components can be considered “connected” or “joined” with each other in order to achieve the desired functionality, and any two components that can be associated in this way can be considered “joinable” with each other in order to achieve the desired functionality. Specific examples of operably joinable components include, non-limitingly, physically interlocking components and / or physically interacting components, as well as / or wirelessly interacting components and / or wirelessly interacting components, and / or logically interacting components and / or logically interactable components.

[0122] Furthermore, it should be understood that the present invention is defined by the appended claims. As will be understood by those skilled in the art, generally, the terms used herein, and in particular in the appended claims (e.g., the body of the appended claims), are used as “open” terms (for example, the term “including” should be interpreted as “including but not limited to,” the term “having” should be interpreted as “having at least,” and the term “includes” should be interpreted as “includes but not limited to,” etc.). Furthermore, as will be understood by those skilled in the art, if a particular number of items are to be enumerated in the claims, such enumeration will be explicitly stated in the claims; if such enumeration is not made, such enumeration will not be stated in the claims. For example, for ease of understanding, the appended claims may use the introductory phrases “at least one” and “one or more” to enumerate items in the claims. However, the use of such phrases should not be interpreted as limiting any particular claim containing such items enumerated in a claim to embodiments containing only one of the enumerated items, even if the same claim contains the introductory phrase "one or more" or "at least one" and an indefinite article such as "a" or "an" (e.g., "a" and / or "an" should be interpreted as meaning "at least one" or "one or more"). The same applies to the use of the definite article used to enumerate items in a claim.Furthermore, even when a specific number of items are explicitly enumerated in a claim, a person skilled in the art will understand that such an enumeration should be interpreted as meaning a number greater than or equal to the number enumerated (for example, the enumeration of "two recitations" without any other modifiers means at least two items, or two or more items). In addition, in examples where a conventional phrase similar to "at least one of A, B, and C, or the like" is used, such a sequence of words is generally used in a sense that a person skilled in the art would understand (for example, "a system having at least one of A, B, and C" includes, but is not limited to, a system having only A, only B, only C, both A and B, both A and C, both B and C, and / or all of A, B, and C). In examples where a conventional phrase similar to "at least one of A, B, or C, or the like" is used, such a sequence of words is generally used in a sense that a person skilled in the art would understand the conventional phrase to mean (for example, "a system having at least one of A, B, or C" includes, but is not limited to, a system having only A, only B, only C, both A and B, both A and C, both B and C, and / or all of A, B and C). Furthermore, a person skilled in the art should understand that almost all separate words and / or phrases indicating that there are two or more options can be considered to include one of these items, either item, or both, regardless of whether they are included in the detailed description, claims, or drawings. For example, the phrase "A or B" would be understood to include the possibility of "A," "B," or "A and B."

[0123] We believe that many of the advantages of this disclosure and its associated benefits can be understood from the above description. It will also be clear that various modifications can be made to the form, structure, and arrangement of the components without deviating from the disclosed subject matter or compromising any of its main advantages. The described forms are for illustrative purposes only, and the intent of the claims below is to encompass and include such modifications. Furthermore, it should be understood that the present invention is defined by the attached claims.

Claims

1. It is a system, The system includes a processing unit that is communicatively connected to the detector array of an optical wafer characterization system. The aforementioned processing unit is From the detector array, one or more target images of the target location on the wafer are acquired. A noise reduction filter is applied to at least one or more of the above target images. The system is configured to determine one or more difference images from one or more reference images and one or more difference images after denoising at least one or more of the aforementioned target images, A system capable of detecting one or more wafer defects in the one or more difference images.

2. The processing unit is configured to acquire one or more reference images from the detector array. The system according to claim 1, wherein the one or more reference images are of reference locations near the target location on the wafer.

3. The system according to claim 1, wherein the processing unit is configured to determine one or more difference images from the one or more reference images and the one or more target images by subtracting one or more reference images from the one or more target images.

4. The system according to claim 1, wherein the noise reduction filter is a spatial frequency filter and is configured to remove a portion of the noise in at least one or more target images acquired by the detector array, and is configured to suppress changes in the noise-free image.

5. The spatial frequency filter is a Fast Fourier Transform (FFT) filter, The system according to claim 4, wherein the at least one or more target images are subjected to a Fourier transform by the FFT filter, multiplied by a function of spatial frequency, and then subjected to an inverse Fourier transform.

6. The spatial frequency filter is a finite impulse response (FIR) filter, The system according to claim 4, wherein the at least one or more target images are convolved by the FIR filter.

7. The processing unit applies a noise reduction filter to the one or more reference images. The system according to claim 1, configured to determine one or more difference images from the one or more reference images and the one or more target images after noise reduction of the one or more reference images and the one or more target images.

8. The system according to claim 1, wherein the processing unit includes a field-programmable gate array (FPGA), an application-specific integrated circuit, an image processing device, or a central processing unit.

9. The system according to claim 1, wherein the optical wafer characterization system is a broadband plasma (BBP) tool.

10. The processing unit is configured to determine one or more control signals based on one or more wafer defects detected in the difference image. The system according to claim 1, wherein the one or more control signals are configured to perform one or more adjustments on at least one of the one or more optical wafer characterization components of the optical wafer characterization system and the one or more semiconductor manufacturing components of the semiconductor manufacturing system.

11. The processing unit is connected to the controller in a communication manner, The controller includes one or more processors and memory, The memory is configured to store program instructions. The one or more processors are configured to execute the program instructions, The program instructions are sent to the one or more processors: The processing unit receives one or more of the difference images. Based on the one or more wafer defects detected in the one or more difference images, one or more control signals are determined. The system according to claim 1, wherein the one or more control signals are configured to perform one or more adjustments on at least one of the one or more optical wafer characterization components of the optical wafer characterization system and the one or more semiconductor manufacturing components of the semiconductor manufacturing system.

12. It is a method, The processing unit acquires one or more target images of target locations on the wafer from the detector array of the optical wafer characterization system, The processing unit applies a noise reduction filter to at least one or more of the target images, The processing unit includes, after denoising at least one or more target images, determining one or more reference images and one or more difference images from the one or more target images, A method for detecting one or more wafer defects in the one or more difference images.

13. The processing unit further includes acquiring the one or more reference images from the detector array, The method according to claim 12, wherein the one or more reference images are of reference locations near the target location on the wafer.

14. The method according to claim 12, further comprising determining one or more difference images from the one or more reference images and the one or more target images by subtracting the one or more reference images from the one or more target images using the processing unit.

15. The method according to claim 12, wherein the noise reduction filter is a spatial frequency filter and is configured to remove a portion of the noise in at least one or more target images acquired by the detector array, and is configured to suppress changes in the noise-free image.

16. The spatial frequency filter is a Fast Fourier Transform (FFT) filter, The method according to claim 15, wherein the at least one or more target images are Fourier transformed by the FFT filter, multiplied by a function of spatial frequency, and inverse Fourier transformed.

17. The spatial frequency filter is a finite impulse response (FIR) filter, The method according to claim 15, wherein the at least one or more target images are convolved by the FIR filter.

18. The processing unit applies a noise reduction filter to the one or more reference images, The method according to claim 12, further comprising: using the processing unit to remove noise from the one or more reference images and the one or more target images, and then determining one or more difference images from the one or more reference images and the one or more target images.

19. The method according to claim 12, wherein the processing unit includes a field-programmable gate array (FPGA), an application-specific integrated circuit, an image processing device, or a central processing unit.

20. The method according to claim 12, wherein the optical wafer characterization system is a broadband plasma (BBP) tool.

21. The processing unit further includes determining one or more control signals based on one or more wafer defects detected in the difference image. The method according to claim 12, wherein the one or more control signals are configured to perform one or more adjustments on at least one of the one or more optical wafer characterization components of the optical wafer characterization system and the one or more semiconductor manufacturing components of the semiconductor manufacturing system.

22. The processing unit is connected to the controller in a communication manner, The method is, The controller receives the one or more difference images from the processing unit, The controller further includes determining one or more control signals based on one or more wafer defects detected in the one or more difference images. The method according to claim 12, wherein the one or more control signals are configured to perform one or more adjustments on at least one of the one or more optical wafer characterization components of the optical wafer characterization system and the one or more semiconductor manufacturing components of the semiconductor manufacturing system.

23. An optical wafer characterization system, A light source configured to generate and emit light, A lighting arm having one or more lighting optical systems configured to irradiate a wafer placed on a stage assembly with light received from the light source, A focusing arm having one or more focusing optical systems, A detector array having one or more sensors, wherein the one or more focusing optical systems are configured to image the wafer onto the detector array, The system comprises a processing unit that is communicatively connected to the detector array, The aforementioned processing unit is From the detector array, one or more target images of target locations on the wafer are acquired. A noise reduction filter is applied to at least one or more of the above target images. The system is configured to determine one or more difference images from one or more reference images and one or more difference images after denoising at least one or more of the aforementioned target images, An optical wafer characteristic evaluation system capable of detecting one or more wafer defects in the one or more difference images.

24. The processing unit is configured to acquire one or more reference images from the detector array. The optical wafer characteristic evaluation system according to claim 23, wherein the one or more reference images are of reference locations on the wafer near the target location.

25. The optical wafer characteristic evaluation system according to claim 23, wherein the processing unit is configured to determine one or more difference images from the one or more reference images and the one or more target images by subtracting one or more reference images from the one or more target images.

26. The optical wafer characteristic evaluation system according to claim 23, wherein the noise reduction filter is a spatial frequency filter and is configured to remove a portion of the noise in at least one or more target images acquired by the detector array, and is configured to suppress changes in the noise-free image.

27. The spatial frequency filter is a Fast Fourier Transform (FFT) filter, The optical wafer characteristic evaluation system according to claim 26, wherein the at least one or more target images are Fourier transformed by the FFT filter, multiplied by a function of spatial frequency, and inverse Fourier transformed.

28. The spatial frequency filter is a finite impulse response (FIR) filter, The optical wafer characteristic evaluation system according to claim 26, wherein the at least one or more target images are convolved by the FIR filter.

29. The processing unit applies a noise reduction filter to the one or more reference images. The optical wafer characteristic evaluation system according to claim 23, configured to determine one or more difference images from the one or more reference images and the one or more target images after noise reduction of the one or more reference images and the one or more target images.

30. The optical wafer characterization system according to claim 23, wherein the processing unit includes a field-programmable gate array (FPGA), an application-specific integrated circuit, an image processing device, or a central processing unit.

31. The optical wafer characterization system according to claim 23, wherein the optical wafer characterization system is a broadband plasma (BBP) tool.

32. The processing unit is configured to determine one or more control signals based on one or more wafer defects detected in the difference image. The optical wafer characterization system according to claim 23, wherein the one or more control signals are configured to perform one or more adjustments on at least one of the one or more optical wafer characterization components of the optical wafer characterization system and the one or more semiconductor manufacturing components of the semiconductor manufacturing system.

33. The optical wafer characteristic evaluation system is The processing unit further comprises a controller that is communicatively connected to the processing unit, The controller includes one or more processors and memory, The memory is configured to store program instructions. The one or more processors are configured to execute the program instructions, The program instructions are sent to the one or more processors: The processing unit receives one or more of the difference images. Based on the one or more wafer defects detected in the one or more difference images, one or more control signals are determined. The optical wafer characterization system according to claim 62, wherein the one or more control signals are configured to perform one or more adjustments on at least one of the one or more optical wafer characterization components of the optical wafer characterization system and the one or more semiconductor manufacturing components of the semiconductor manufacturing system.

Images