Optical detection system

Abstract

The optical detection system comprises an optical device and an image processing device. The optical device comprises a light source and an image sensor. The light source is configured to generate a first laser beam and direct the first laser beam through the illuminator to an incident laser beam towards the wafer, thereby correspondingly generating a reflected laser beam. The image sensor is configured to capture the reflected laser beam through the objective lens, so that the reflected laser beam becomes a second laser beam, and correspondingly generates an image of the wafer. The image processing device is configured to generate a detection result according to the image. The wavelength of the incident laser beam is less than 120 nm.

Description

Technical Field

[0001] The present invention relates to a wafer inspection system, and in particular to an optical inspection system utilizing extreme ultraviolet (XUV) to inspect semiconductor devices. Background Art

[0002] In the modern semiconductor industry, wafer inspection is crucial in the manufacturing process. However, due to shrinking feature sizes and complex patterns, defects are becoming smaller and more complex, making defect detection more challenging. Typically, smaller defects (e.g., those below 20nm) require finer resolution for detection, but finer resolution requires detection light with shorter wavelengths. This short-wavelength light typically results in weaker signal strength, which in turn reduces detection sensitivity and makes it difficult to achieve the desired acquisition rate and interference rate. However, inspection using short-wavelength light (e.g., extreme ultraviolet (XUV)) is highly sensitive to thickness and material properties. For wafers with multilayer structures, variations in material or thickness (under short-wavelength light) can result in varying inspection quality (e.g., resolution). Furthermore, even with short-wavelength light, the low material contrast between materials used on the wafer can prevent some defects from being detected and reduce the signal-to-noise ratio (SNR) during inspection. Furthermore, generating short-wavelength light presents challenges such as conversion efficiency (or light intensity), bandwidth, polarization, spectral purity, resolution, and contamination management. These challenges are crucial to the light source performance of semiconductor inspection equipment.

[0003] On the other hand, as detection sensitivity increases, throughput decreases. Therefore, a new solution is needed to optimize the performance of wafer optical inspection systems. Summary of the Invention

[0004] One aspect of the present invention provides an optical inspection system. The optical inspection system includes an optical device and an image processing device. The optical device includes a light source and an image sensor. The light source is configured to generate a first laser beam and direct the first laser beam toward a wafer via an illuminator as an incident laser beam, thereby generating a corresponding reflected laser beam. The image sensor is configured to capture the reflected laser beam through an objective lens, converting it into a second laser beam, and correspondingly generate an image of the wafer. The image processing device is configured to generate a detection result based on the image. The wavelength of the incident laser beam is less than 120 nm.

[0005] Another aspect of the present invention provides an optical inspection system. The optical inspection system includes an optical device and an image processing device. The optical device includes a light source and an image sensor. The light source is configured to generate a first laser beam and guide the first laser beam through an illuminator as an incident laser beam toward a wafer, thereby generating a reflected laser beam accordingly. The light source includes a first wavelength conversion channel, a second wavelength conversion channel, and an XUV generator. The first wavelength conversion channel is configured to generate a first UV laser. The second wavelength conversion channel is configured to generate a second UV laser. The XUV generator is configured to generate the first laser beam based on the first UV laser or the second UV laser. The image sensor is configured to capture the reflected laser beam through an objective lens, making it a second laser beam, and generate an image of the wafer accordingly. The image processing device is configured to generate a detection result based on the image. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] When with Figure 1 The various aspects of the present invention are best understood from the following detailed description when read together. It should be noted that, in accordance with standard industry practice, various features are not drawn to scale. In fact, the dimensions of various features may be arbitrarily increased or decreased for clarity.

[0007] Figure 1 is a schematic diagram of an optical detection system according to some embodiments of the present invention;

[0008] Figure 2A is a schematic diagram of an optical device according to some embodiments of the present invention;

[0009] Figure 2B is a schematic diagram of a light source according to another embodiment of the present invention

[0010] Figure 3 is a schematic diagram of an optical device according to some embodiments of the present invention;

[0011] Figure 4 is a schematic diagram of a speckle detector according to some embodiments of the present invention;

[0012] Figure 5A is a schematic diagram of a homogenizer according to some embodiments of the present invention;

[0013] Figure 5B is a schematic diagram of laser transmission between a homogenizer and a concentrator according to some embodiments of the present invention;

[0014] Figure 6A is a schematic diagram of a homogenizer according to another embodiment of the present invention;

[0015] Figure 6B is a cross-sectional view of a homogenizer according to another embodiment of the present invention;

[0016] Figure 6Cis a cross-sectional view of a homogenizer according to another embodiment of the present invention;

[0017] Figure 6D is a cross-sectional view of a homogenizer according to another embodiment of the present invention;

[0018] Figure 7 is a schematic diagram of a homogenizer according to various embodiments of the present invention;

[0019] Figure 8 is a schematic diagram of a repeater according to some embodiments of the present invention;

[0020] Figure 9 is a schematic diagram of an objective lens according to some embodiments of the present invention;

[0021] Figure 10A and Figure 10B is a schematic diagram of the field of view of an illuminator, an objective lens, and an image sensor according to some embodiments of the present invention;

[0022] Figure 11 is a schematic diagram of an image processing device according to some embodiments of the present invention;

[0023] Figure 12 is a schematic diagram of a detection platform according to some embodiments of the present invention;

[0024] Figure 13 is a schematic diagram of a host according to some embodiments of the present invention;

[0025] Figure 14 is a schematic diagram of motion control during the operation phase according to some embodiments of the present invention;

[0026] Figure 15 is a schematic diagram of synchronization of illumination, image sensing, and wafer position according to some embodiments of the present invention. DETAILED DESCRIPTION

[0027] The following disclosure provides many different embodiments or examples for implementing different features of the provided subject matter. To simplify the present invention, specific examples of components and configurations are described below. Of course, these are merely examples and are not intended to be limiting. For example, in the following description, a first feature is formed to cover a second feature or a first feature is formed on top of a second feature, and the following may include embodiments in which the first and second features are formed to be in direct contact, and may also include embodiments in which additional features may be formed between the first and second features so that the first and second features are not in direct contact. In addition, the present invention may repeat the figure numbers and / or letters in the various examples. This repetition is for simplicity and clarity and does not, in itself, define the relationship between the various embodiments and / or configurations discussed.

[0028] Furthermore, for ease of description, spatially relative terms such as "below," "beneath," "lower," "above," and "upper" may be used herein to describe the relationship of one element or feature to another element or feature as illustrated in the figures. Spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. The device may be otherwise oriented (rotated 90 degrees or at other orientations), and the spatially relative descriptors used herein interpreted accordingly.

[0029] Although the numerical ranges and parameters setting forth the broad scope of the present invention are approximate, the numerical values ​​set forth in the specific examples are reported as precisely as possible. However, any numerical value inherently contains certain errors, necessarily due to the normal variation found in the corresponding testing measurements. Furthermore, as used herein, the terms "substantially," "approximately," or "about" generally refer to values ​​or ranges within a range that would be considered acceptable by one of ordinary skill in the art. Alternatively, the terms "substantially," "approximately," or "about" mean values ​​within an acceptable standard error of the mean, as considered by one of ordinary skill in the art. As one of ordinary skill in the art will appreciate, acceptable standard errors may vary from technology to technology. Except in the working examples, or unless expressly indicated otherwise, all numerical ranges, amounts, values, and percentages, such as amounts of materials, durations, temperatures, operating conditions, and quantitative ratios disclosed herein, should in all instances be construed as modified by the terms "substantially," "approximately," or "about." Therefore, unless otherwise indicated, the numerical parameters set forth in this Summary of the Invention and the appended claims are approximate and are subject to change as necessary. At the very least, each numerical parameter should at least be construed in light of the number of reported significant digits and by applying ordinary rounding techniques. Ranges can be expressed herein as from one endpoint to another endpoint or between two endpoints. Unless otherwise stated, all ranges disclosed herein include the endpoints.

[0030] Figure 1 FIG2 is a schematic diagram of an optical inspection system 10 according to some embodiments of the present invention. The optical inspection system 10 is used to inspect semiconductor components or wafers W to determine whether the semiconductor components or wafers W meet specific requirements.

[0031] In some embodiments, the optical inspection system 10 is configured to detect defects on a wafer W. Furthermore, the optical inspection system 10 is configured to determine whether the size, shape, type, density, and location of the detected defects are within tolerances and meet the specifications of the wafer W. In some embodiments, the optical inspection system 10 is capable of detecting defects with a size less than 100 nm. In some embodiments, the optical inspection system 10 is configured to detect defects with a size of approximately 10 nm to 20 nm.

[0032] In some embodiments, the optical inspection system 10 is used to inspect the wafer W after a lithography process. For example, the optical inspection system 10 inspects the photoresist and oxide patterns on the wafer W after the photoresist layer is developed. In some embodiments, the optical inspection system 10 is used to inspect the wafer W after an etching process. For example, the optical inspection system 10 inspects the hard mask on the wafer W after the mask layer is patterned. In some embodiments, the optical inspection system 10 is used to inspect the wafer W after a metal deposition process. For example, the optical inspection system 10 inspects the M0, M1, M2, M3, and / or M4 layers of the wafer W after metal deposition.

[0033] The optical inspection system 10 includes an optical device 100 , an image processing device 200 , a wafer stage 300 , a motion control device 400 , a position measurement module 500 , a synchronization device 600 , a host 700 , a spatial position measurement device SPM, a Z sensor ZS, and an optical microscope OM.

[0034] The optical device 100 is used to magnify the image of the wafer W to generate the image M0 . The optical device 100 generates a laser beam L1 with predetermined characteristics toward the region of interest of the wafer W and receives a reflected / scattered laser beam L2 from the region of interest to capture the image M0 .

[0035] The image processing device 200 is used to process the captured image M0 and perform defect detection and classification to generate a detection result DR.

[0036] The wafer platform 300 is used to carry and move the wafer W. In some embodiments, the wafer platform 300 is also referred to as an XYZ platform, that is, the wafer platform 300 can move along the X-axis, the Y-axis, the Z-axis, or a combination thereof.

[0037] During the inspection process, the wafer W is controlled to move along a predetermined inspection path P0 so that the laser beam L1 can be aligned with the same point and scan the wafer W along the predetermined inspection path P0. Specifically, the motion control device 400 is used to control the wafer stage 300 so that the movement path Pm of the wafer W is aligned with the predetermined inspection path P0.

[0038] However, in the real world, mechanical devices may have various vibrations during operation. Vibrations may be caused by motor operation, particles on the track, or any other suboptimal conditions. Similarly, the present optical inspection system 10 will also generate vibrations, thereby introducing an offset between the moving path Pm and the predetermined inspection path P0, wherein the offset may cause noise in the detection result. Therefore, the spatial position measurement device SPM is used to obtain the position Pxy of the wafer W during the inspection process. The Z sensor ZS is used to obtain the relative position Pz of the wafer W during the inspection. The position Pxy and the position Pz together represent the position of the wafer W defined in the platform coordinate system. The motion control device 400 controls the wafer platform 300 to move the wafer W according to the platform coordinate system.

[0039] The position measurement module 500 is used to obtain the position Pxy and position Pz of the wafer W to determine whether the wafer W deviates from the predetermined inspection path P0. When the deviation between the movement path Pm of the wafer W and the predetermined inspection path P0 exceeds a predetermined threshold, the position measurement module 500 generates a calibration signal Sc to the motion control device 400, which adjusts the movement path Pm of the wafer W based on the calibration signal Sc. In addition, the synchronization device 600, which is used to synchronize the timing of illumination, movement of the wafer W, and sensing, operates according to an image coordinate system that is different from the stage coordinate system. Furthermore, the position measurement module 500 is further used to convert the position Pxy and position Pz into the position PI of the wafer W defined in the image coordinate system and transmit the position PI to the synchronization device 600 for subsequent operation.

[0040] Because image M0 is generated by optical components, the depth of focus (DOF) must be considered. Position measurement module 500 is used to determine the difference between the image plane and wafer W based on position Pz obtained by Z sensor ZS. When the difference between the image plane and wafer W exceeds the tolerance, position measurement module 500 notifies motion control device 400 using calibration signal Sc. Motion control device 400 adjusts position Pz of wafer W in response to calibration signal Sc. Specifically, motion control device 400 controls wafer stage 300 to raise or lower wafer W.

[0041] The synchronization device 600 is used to synchronize the timing of the laser beam L1 generated by the light source 110 of the optical device 100 with the timing of the laser beam L2 captured by the image sensor 140 of the optical device 100 based on the position PI. Based on the position PI, the synchronization device 600 determines when the region of interest (ROI) of the wafer W reaches a desired position along the predetermined inspection path P0 and triggers the optical device 100 to generate the laser beam L1 to illuminate the region of interest. The synchronization device 600 further controls the optical device 100 to capture the laser beam L2 at the corresponding timing. As a result, the optical device 100 can capture an image M0 of the region of interest on the wafer W at the desired time and under the desired illumination.

[0042] The host computer 700 is used to set a predetermined inspection path P0, control the loading angle LA of the wafer W to control the orientation of the wafer W relative to the predetermined inspection path P0, determine the calibration algorithm CA for the image processing device 200, and determine the movement speed V of the wafer W for the motion control device 400 and the synchronization device 600 to synchronize the movement of the wafer W with the generation of the image M0. In some embodiments, the movement speed V ranges from 0.1 mm / s to 100 mm / s.

[0043] The host 700 is also used to set recipes according to different testing conditions.

[0044] The optical microscope OM is used to inspect the wafer W for optimal orientation. In some embodiments, the optical microscope OM is used to inspect whether the grooves on the wafer W are aligned in a specific direction. If the grooves on the wafer W are not aligned in the specific direction, the host 700 can adjust the predetermined inspection path P0 or loading angle LA based on the current orientation of the wafer W.

[0045] In some embodiments, optical inspection system 10 is capable of inspecting 20% ​​of the wafer area within one hour using a 10nm pixel size. This requires an image data channel processing data rate of at least 40G pixels / second. Synchronizer 600 has an image pixel size of 10nm to 50nm and a laser beam L1 pulse rate of up to 2MHz. Therefore, synchronizer 600 should be able to provide a pulse period of 0.5μs (calibrated synchronization timing) to light source 110 and image sensor 140.

[0046] In some embodiments, the position data refresh rate of the spatial position measurement device SPM is greater than 5 MHz, and the delay of the data link from the spatial position measurement device SPM to the synchronization device 600 is less than 0.3 μs.

[0047] Figure 2A 1 is a schematic diagram of an optical device 100 according to some embodiments of the present invention. Detection is performed in a bright field, so the optical device 100 adopts an off-axis design.

[0048] like Figure 2A As shown, optical device 100 includes a light source 110, an illuminator 120, an objective lens 130, and an image sensor 140. Light source 110 is used to generate a laser beam L0 and transmit it to illuminator 120. Illuminator 120 is used to convert laser beam L0 into laser beam L1 and project laser beam L1 onto a region of interest on wafer W. Objective lens 130 is used to receive and magnify the image constructed by laser beam L2 and transmit it to image sensor 140. Image sensor 140 is used to capture the image magnified by objective lens 130 to generate image M0.

[0049] In some embodiments, the wavelength of laser beam L0 is less than 120 nm. In some embodiments, laser beam L0 is a pulsed laser beam, such as a narrowband XUV laser beam, having a wavelength in the range of about 50 to 120 nm. In some embodiments, laser beam L0 has a repetition rate of about 200 Hz to about 2 MHz. In other embodiments, laser beam L0 is a continuous wave (CW) laser beam.

[0050] To generate the laser beam L0 having a wavelength of about 50 to 120 nm, the light source 110 converts infrared (IR) laser light into XUV laser light through several generation steps. To achieve sufficient defect detection sensitivity, the light source 110 is used to generate a brightness greater than 100 W / nm / srad / mm 2 Laser beam L0. Furthermore, the point stability of laser beam L0 is less than 1 μrad. Light source 110 is a low-pollution, debris-free source. Light source 110 is stable and has a long service life. In some embodiments, the sensitivity to detecting different types of defects is related to the polarization of laser beam L0. Therefore, light source 110 is configured to generate laser beam L0 with an arbitrary linear polarization direction to improve overall sensitivity to various defects. In other words, light source 110 can control the polarization of laser beam L0 to impart an arbitrary linear polarization direction to laser beam L0.

[0051] The light source 110 includes a pump laser 111 , a center wavelength selector 112 , a first harmonic generator 113 , and a second harmonic generator 114 .

[0052] The pump laser 111 is used to generate a source laser beam B0 to the central wavelength selector 112. In some embodiments, the spectrum of the source laser beam B0 may be from 400 nm to 1100 nm, for example, 1030 nm. The central wavelength selector 112 is used to select a wavelength range in the visible light or IR from the source laser beam B0 to generate a narrowband laser beam B1. In one embodiment, the central wavelength selector 112 is a bandpass filter that blocks a portion of the source laser beam B0 that exceeds the passband of the central wavelength selector 112 and allows the remaining portion of the source laser beam B0 within the passband of the wavelength selector 112 to pass. In another embodiment, the central wavelength selector 112 is a Raman shifter that shifts the spectrum of the source laser beam B0. In some embodiments, the wavelength selected by the central wavelength selector 112 is adjustable. Therefore, when the central wavelength selector 112 is used, a desired output wavelength (i.e., the wavelength of the laser beam L0) can be implemented. The first harmonic generator 113 is configured to convert the narrowband laser beam B1 into a deep ultraviolet (DUV) laser beam B2. The second harmonic generator 114 is configured to convert the DUV laser beam B2 into an XUV laser beam, namely, laser beam L0. In some embodiments, the wavelength of laser beam L0 is shorter than the wavelength of DUV laser beam B2. In some implementations, the wavelength of laser beam L0 is less than 120 nm.

[0053] The first harmonic generator 113 and the second harmonic generator 114 use nonlinear harmonic generation to increase the frequency of the narrowband laser beam B1 and the deep ultraviolet laser beam B2. In some embodiments, the first harmonic generator 113 converts the narrowband laser beam B1 in a solid-state material, meaning that the first harmonic generator 113 performs a solid-state nonlinear process. In some embodiments, the solid-state nonlinear process is second-order nonlinear generation. In some embodiments, the second harmonic generator 114 converts the deep ultraviolet laser beam B2 in an inert gas environment, meaning that the second harmonic generator 114 performs a gaseous nonlinear process. In some embodiments, the gaseous nonlinear process is third-order nonlinear generation. In other embodiments, the second harmonic generator 114 converts the deep ultraviolet laser beam B2 in a mixed inert gas environment, where the concentration of each inert gas is controlled by a gas mixer, a pressure controller, and the main unit 700. Due to the inert gas environment, the second harmonic generator 114 generates relatively clean, debris-free waves. After these two conversion stages, the wavelength of the laser beam L0 can reach a range of approximately 50 nm to 120 nm. In some embodiments, the bandwidth of laser beam L0 is less than 0.5% of the central wavelength of laser beam L0. It should be noted that the present invention is not limited to two-order harmonic generation. In various embodiments, using more than one order of harmonic generation to generate laser beam L0 is within the scope of the present invention. For example, first harmonic generator 113 may include at least two nonlinear harmonic generation steps in a solid-state material to convert narrowband laser beam B1 into DUV laser beam B2. In some embodiments, first harmonic generator 113 may perform second-order nonlinear harmonic generation, while second harmonic generator 114 may perform third-order nonlinear harmonic generation.

[0054] Figure 2B is a schematic diagram of a light source 110a according to another embodiment of the present invention. In some embodiments, the light source 110a may replace Figure 2A The light source 110 shown in FIG. 1 includes a laser unit 301 , a wave retarder 302 , a spectrum shaper 303 , a first wavelength conversion channel 310 , a second wavelength conversion channel 320 , an XUV generator 330 , a recipe controller 341 , a gas mixer 342 , and a gas controller 343 .

[0055] Laser unit 301 is configured to generate IR laser light C1 through wave retarder 302 to spectrum shaper 303. Wave retarder 302 is configured to adjust the polarization of IR laser light C1 and generate IR laser light C2 accordingly. In some embodiments, wave retarder 302 is configured to allow a portion of IR laser light C1 with a corresponding polarization to be transmitted to spectrum shaper 303. Spectrum shaper 303 is configured to filter the spectrum of IR laser light C2, thereby altering the spectrum of IR laser light C2 and generating IR laser light C3 accordingly. In some embodiments, spectrum shaper 303 is the same as central wavelength selector 112 of light source 110. IR laser light C3 is transmitted to first wavelength conversion channel 310 and second wavelength conversion channel 320.

[0056] The first wavelength conversion channel 310 includes a nonlinear (NOP) unit 311, a wavelength separator 312, a NOP unit 313, and a wavelength separator 314. The NOP unit 311 is used to convert the IR laser C3 into visible laser V1 and IR laser C4, and transmit them to the wavelength separator 312. Figure 2B Although visible laser light V1 and IR laser light C4 are shown as two separate arrows, they are actually mixed into a single beam and transmitted to wavelength separator 312. Wavelength separator 312 is used to separate visible laser light V1 from IR laser light C4 and transmit only visible laser light V1 to NOP unit 313. NOP unit 313 is configured to convert visible laser light V1 into visible laser light V2 and ultraviolet (UV) laser light E1, and transmit them to wavelength separator 314. Wavelength separator 314 is configured to separate visible laser light V2 from UV laser light E1 and transmit only UV laser light E1 to XUV generator 330.

[0057] The second wavelength conversion channel 320 includes an NOP unit 321, an NOP unit 323, and a wavelength separator 324. The NOP unit 321 is configured to convert the IR laser light C3 into visible laser light V3 and IR laser light C5, and transmit the converted light to the NOP unit 323. The NOP unit 323 is configured to convert the visible laser light V3 and IR laser light C5 into visible laser light V4, IR laser light C6, and UV laser light E2 through nonlinear harmonic generation, and transmit the converted light to the wavelength separator 324. The wavelength separator 324 is configured to separate the visible laser light V4, IR laser light C6, and UV laser light E2, and transmit only the UV laser light E2 to the XUV generator 330.

[0058] In some embodiments, NOP unit 311, NOP unit 313, NOP unit 321, and NOP unit 323 operate in a solid state material. In some embodiments, NOP unit 311, NOP unit 313, NOP unit 321, and NOP unit 323 are used to generate second-order nonlinear harmonics of received laser light.

[0059] Figure 2B The number of wavelength conversion channels shown is for illustration purposes only and the present invention is not limited thereto. In various embodiments, the light source 110a includes more than two wavelength conversion channels.

[0060] The XUV generator 330 includes a polarization switch 331, an NOP unit 332, a polarization switch 333, an NOP unit 334, a wavelength separator 335, and a wavelength separator 336. In some embodiments, the XUV generator 330 has a high-pressure region RG1 and a low XUV absorption region RG2. In some embodiments, the pressure of the high-pressure region RG1 is between about 3 Torr and about 2000 Torr. The polarization switch 331, the NOP unit 332, the polarization switch 333, and the NOP unit 334 are disposed in the high-pressure region RG1, and the wavelength separator 335 and the wavelength separator 336 are disposed in the low XUV absorption region RG2. In some embodiments, the optical device 100 and the wafer W are disposed in a low XUV absorption region such as the low XUV absorption region RG2. In some embodiments, the pressure of the low XUV absorption region RG2 is lower than that of the high-pressure region RG1, wherein the low XUV absorption region RG2 is evacuated by a suitable device (e.g., a mechanical pump and / or a turbo pump) so that the pressure of the low XUV absorption region RG2 is between about 10 -3 In some embodiments, the low XUV absorption region RG2 contains a low XUV absorption gas (e.g., helium, with a pressure of about 500 to about 800 Torr). In other embodiments, the low XUV absorption region RG2 contains other non-solid substances with higher ionization energy.

[0061] Polarization switch 331 is configured to receive UV laser light E1 and rotate the polarization of UV laser light E1, thereby generating UV laser light E3 for NOP unit 332. NOP unit 332 is configured to convert UV laser light E3 into UV laser light E5 and XUV laser light X1 through nonlinear harmonic generation. Polarization switch 333 is configured to receive UV laser light E2 and rotate the polarization of UV laser light E2, thereby generating UV laser light E4 for NOP unit 334. NOP unit 334 is configured to convert UV laser light E4 into UV laser light E6 and XUV laser light X2 through nonlinear harmonic generation. In some embodiments, NOP units 332 and 334 operate in a gas. In some embodiments, NOP units 332 and 334 operate in an inert gas. In some embodiments, NOP units 332 and 334 are configured to generate third-order nonlinear harmonics.

[0062] The wavelength separator 335 is used to separate the UV laser E5 and the XUV laser X1 and transmit only the XUV laser X1 to the illuminator 120. Similarly, the wavelength separator 336 is used to separate the UV laser E6 and the XUV laser X2 and transmit only the XUV laser X2 to the illuminator 120.

[0063] In some embodiments, the spectrum of UV laser E1, UV laser E2, UV laser E3, UV laser E4, UV laser E5, and UV laser E6 extends into the deep ultraviolet (DUV) range. In some embodiments, the central wavelengths of UV laser E1, UV laser E2, UV laser E3, UV laser E4, UV laser E5, and UV laser E6 are within the DUV range. In some embodiments, the central wavelengths of UV laser E1, UV laser E2, UV laser E3, UV laser E4, UV laser E5, and UV laser E6 are between approximately 200 nm and approximately 280 nm.

[0064] UV laser E1 and UV laser E2 are different, and XUV laser X1 and XUV laser X2 are also different. Specifically, the spectrum and central wavelength of UV laser E1 differ from those of UV laser E2, while the spectrum and central wavelength of XUV laser X differ from those of XUV laser X2. The first wavelength conversion channel 310 and the second wavelength conversion channel 320 operate independently. In other words, when the first wavelength conversion channel 310 is operating, the second wavelength conversion channel 320 is disabled, and vice versa. When the first wavelength conversion channel 310 is disabled, the polarization switch 331, the NOP unit 332, and the wavelength separator 335 are disabled. When the second wavelength conversion channel 320 is disabled, the polarization switch 333, the NOP unit 334, and the wavelength separator 336 are disabled. Therefore, the XUV generator 330 generates either the XUV laser X1 or the XUV laser X2 as the incident laser beam L0.

[0065] In some conventional technologies, when the object being measured changes, the detection light may be changed accordingly to adapt to the characteristics of the new object being measured. In this case, the light source must be replaced because traditional light sources can only emit light beams with a fixed spectrum and a fixed central wavelength. Compared with the light source 110a of the present invention, the light source 110a can generate more than one incident laser beam L0 (that is, XUV laser X1 and XUV laser X2). Therefore, when the light source 110a is applied, the optical detection system 10 can generate incident light L0 of different spectra onto different wafers W without changing the light source 100a.

[0066] The recipe controller 341 is configured to generate control signals SC1, SC2, and SC3 according to a predetermined recipe. In some embodiments, the predetermined recipe is provided by the host 700. The control signal SC1 is transmitted to the spectrum shaper 303, which selects a desired wavelength based on the control signal SC1.

[0067] The control signal SC2 is transmitted to the XUV generator 330, and the XUV generator 330 operates according to SC2. In some embodiments, the polarization switches 331 and 333 select the desired polarization according to the control signal SC2.

[0068] The gas mixer 342 is used to provide source gas GS1 to the recipe controller 341. In some embodiments, source gas SG1 comprises a single material, such as argon or neon. In other embodiments, source gas SG1 comprises several materials, such as a mixed inert gas. The gas controller 343 is used to purge source gas GS1 into the high-pressure region RG1 of the XUV generator 330 based on a control signal SC3 to maintain the pressure of the high-pressure region RG1 at a desired level. As the source gas GS1 is purged, its pressure may change. When the pressure of the source gas GS1 changes, the gas is represented by the reference numeral GS2.

[0069] In some embodiments, in response to wafers W of different materials, the light source 110 can change the center wavelength of the incident laser beam L0 by controlling the gas used for nonlinear harmonic generation performed by the second harmonic generator 114 and / or the wavelength selected by the center wavelength selector 112. Similarly, the light source 110a can change the center wavelength of the incident laser beam L0 by controlling the gas GS2 and / or the activated channel.

[0070] Figure 3 is a schematic diagram of an optical device 100 according to some embodiments of the present invention. The illuminator 120 is used to collect the laser beam L0 from the light source 110 (or the light source 110a) and convert the laser beam L0 into the laser beam L1. In some embodiments, the illuminator 120 directs the laser beam L1 toward the wafer W at an angle θ of 20 to 45 degrees. The numerical aperture (NA) of the illuminator 120 corresponds to the illuminated area on the wafer W and the field of view (FOV) of the image sensor 140. In some embodiments, the NA of the illuminator 120 and the objective lens 130 ranges from 0.2 to 0.5. In some embodiments, the NA of the illuminator 120 and the objective lens 130 is 0.3. In this embodiment, the illuminator 120 is a reflective system, that is, the elements in the system can be reflective optical elements.

[0071] The illuminator 120 includes a collector 121, a speckle remover 122, a collimator 123, a homogenizer 124, a condenser 115, and a relay 116. It should be noted that most optical elements are not transparent to XUV, so the optical device 100 of the present invention is a reflective device. Figure 3 The dotted lines shown in represent the laser beam L0 being reflected in each element within the illuminator 120. However, Figure 3 The dashed lines shown in FIG are for illustration purposes only, and the present invention is not limited thereto. In various embodiments, depending on the configuration of the illuminator 120 , the light beam in the illuminator 120 may have different shapes or directions.

[0072] The collector 121 is used to collect the laser beam L0 and project it onto the speckle detector 122 . The collector 121 controls the spot size of the laser beam L0 and converts the laser beam L0 into a laser beam L01 projected onto the speckle detector 122 , thereby controlling the etendue of the illuminator 120 .

[0073] Laser beams L0 and L01 are coherent light. In some embodiments, coherent light may cause strong speckle, severely affecting the detection result DR. To mitigate the effects of speckle, a speckle remover 122 is used to reduce the coherence of laser beam L01. In some embodiments, laser beams L0 and L01 are Gaussian beams with an NA of approximately 0.004. The speckle remover 122 scatters and reflects laser beam L01 into laser beam L02 with an NA in the range of 0.015 to 0.025. In other words, the coherence of laser beam L02 is lower than that of either laser beam L0 or L01.

[0074] The collimator 123 is used to collimate the laser beam L02 into a collimated laser beam L03 and transmit the collimated laser beam L03 to the homogenizer 124 .

[0075] The homogenizer 124 is used to homogenize the light intensity of the collimated laser beam L03 and shape the collimated laser beam L03 into the laser beam L04. The condenser 115 is used to condense the laser beam L04 into the laser beam L05. The laser beam L05 is condensed on the intermediate focal plane FP.

[0076] The relay 116 is used to adjust the NA and relay the laser beam L05 into the laser beam L1 to match the imaging field of view of the image sensor 140. The laser beam L1 is transmitted with an NA of 0.2 to 0.5. In some embodiments, the laser beam L1 is transmitted with an NA of approximately 0.3.

[0077] Figure 4Figure 1 is a schematic diagram of a speckle pattern detector 122 according to some embodiments of the present invention. The speckle pattern detector 122 includes a deformable mirror DM. The deformable mirror DM includes a membrane MB, whose random vibrations vary over time. Due to the random vibrations, the membrane MB exhibits random ripples. When the laser beam L01 encounters the random scattering conditions on the membrane MB, the phase of the laser beam L01 also experiences abrupt changes. Consequently, after being scattered by the speckle pattern detector 122, the coherence of the laser beam L01 is reduced.

[0078] In addition, the speckle detector 122 is used to adjust the spot size of the laser beam L01. Specifically, no matter what the spot size of the laser beam L01 on the deformable mirror DM is, the spot size of the laser beam L02 is substantially the same.

[0079] In other embodiments, the deformable mirror DM does not include the membrane MB, but instead includes a piezoelectric element (not shown). Applying a control voltage to the piezoelectric element can provide shear stress, causing the deformable mirror DM to vibrate at high speed. Consequently, the laser beam L01 experiences random scattering due to the high-speed vibration, thereby reducing the coherence of the laser beam L01.

[0080] Figure 5A 1 is a schematic diagram of a homogenizer 124 according to some embodiments of the present invention. The homogenizer 124 is implemented by an array of cylindrical concave mirrors. Figure 5A As shown, the homogenizer 124 includes cylindrical reflectors CL, each extending along a longitudinal direction D1 and arranged along a width direction D2. Each cylindrical reflector CL has a concave surface for receiving the collimated laser beam L03. In some embodiments, the cylindrical reflectors CL are identical. In other words, the pitch PH, sag SG, and radius of curvature RC of each cylindrical reflector CL are identical. In some embodiments, the peak value of the surface irregularities of the cylindrical reflectors CL is less than 150 nm.

[0081] Figure 5B FIG2 is a schematic diagram of laser transmission between homogenizer 124 and condenser 115 according to some embodiments of the present invention. In some embodiments, image sensor 140 is a time delay integration (TDI) sensor, and therefore the sensing area (or FOV) of image sensor 140 is rectangular. In this case, image sensor 140 may exhibit better performance if it receives a laser beam having an intensity distribution uniformly distributed within the rectangle or along the major dimension of the rectangle. However, the intensity distributions of laser beams L0, L01, L02, and collimated laser beam L03 are concentric. Therefore, homogenizer 124 is used to adjust the intensity distribution of collimated laser beam L03 from concentric circles to a rectangular shape.

[0082] Specifically, each cylindrical reflector CL scatters the collimated laser beam L03 into multiple beam segments L04n, each of which is substantially identical. Because the cylindrical reflectors CL are arranged along width direction D2, these beam segments L04n are also arranged along width direction D2 and do not overlap, thereby forming laser beam L04. Consequently, laser beam L04 can have a nearly rectangular intensity distribution.

[0083] It should be noted that the number of cylindrical reflectors CL (e.g. Figure 5A and Figure 5B The five cylindrical reflectors CL shown are for illustration purposes only. The present invention is not limited thereto. In various embodiments, the homogenizer 124 includes more cylindrical reflectors CL.

[0084] Figure 6A is a schematic diagram of a homogenizer 124 according to another embodiment of the present invention. The homogenizer 124 can be implemented using microelectromechanical system (MEMS) mirrors ML. The MEMS mirrors ML are arranged in a two-dimensional array. In some embodiments, each MEMS mirror ML is a hexagonal shape when viewed from above. The MEMS mirrors ML can be controlled to have different positions and tilt angles to form a randomly corrugated surface on the homogenizer 124.

[0085] Figure 6B FIG2 is a cross-sectional view of a homogenizer 124 according to another embodiment of the present invention. The homogenizer 124 includes a plurality of control units 114C. The control units 114C are used to control the position and tilt angle of the MEMS mirror ML.

[0086] Control unit 114C includes frame 1141, connector 1142, deformable element 1143, electrode 1144, and MEMS mirror ML. Electrode 1144 is provided at the bottom of frame 1141. Deformable element 1143 is provided in frame 1141 and connected to MEMS mirror ML via connector 1142.

[0087] In some embodiments, when a control voltage is applied to electrode 1144, deformable element 1143 is deformed in response to the electric field generated by electrode 1144. When the center portion of deformable element 1143 deforms and bends toward electrode 1144, connector 1142 pulls MEMS mirror ML down from level LV1 to level LV2, as shown in FIG. Figure 6C shown.

[0088] In some embodiments, when the center portion of the deformable element 1143 of one control unit 114C is deformed and bent toward the electrode 1144, while the deformable element 1143 of the adjacent control unit 114C is not deformed, a portion of the MEME mirror ML is lowered to level LV2, while another portion of the MEME mirror ML remains at level LV1. In this case, the MEMS mirror ML is tilted, as shown in FIG. Figure 6D shown.

[0089] In some embodiments, the maximum offset between level LV1 and level LV2 is about 3.5 μm. In some embodiments, the maximum tilt angle of the MEMS mirror ML is about 8 mrad.

[0090] Figure 7 Schematic diagram of a homogenizer 124 according to various embodiments of the present invention. The homogenizer 124 includes a plurality of first corrugated structures RP1 and a plurality of second corrugated structures RP2. Each of the first corrugated structures RP1 and the second corrugated structures RP2 extends along a longitudinal direction D1 and is arranged along a width direction D2. The first corrugated structures RP1 and the second corrugated structures RP2 are different. Specifically, the first corrugated structures RP1 and the second corrugated structures RP2 have different radii of curvature.

[0091] In some embodiments, when the collimated laser beam L03 is projected onto the homogenizer 124, the homogenizer 124 bounces back and forth along the width direction D2. In this way, the homogenizer 124 can transform the laser beam L03 from concentric circles to rectangles and unify the intensity distribution between the rectangles.

[0092] In some embodiments, the radius of curvature and the pitch angle of the corrugated structure are correlated with the longitudinal dimension LD1 of the beam shape within the field of view of the image sensor 140. For example, the larger the applied radius of curvature, the shorter the resulting longitudinal dimension LD1. For example, when the radius of curvature is equal to 4, 8, 16, and 32 mm, the longitudinal dimension LD1 is approximately 830, 410, 180, and 80 μm, respectively. In some embodiments, the homogenizer 124 can be moved along the width direction D2 to allow the collimated laser beam L03 to be impinged on the first corrugated structure RP1 or the second corrugated structure RP2, thereby adjusting the longitudinal dimension LD1 of the beam shape accordingly. In some embodiments, the homogenizer 124 can use both the first corrugated structure RP1 and the second corrugated structure RP2 to reflect the collimated laser beam L03.

[0093] In some embodiments, the homogenizer 124 does not change the beam size of the collimated laser beam L03. In some embodiments, the NA of the collimated laser beam L03 is substantially equal to the NA of the laser beam L04.

[0094] Figure 8 FIG2 is a schematic diagram of a repeater 116 according to some embodiments of the present invention. The repeater 116 is used to control the tilt angle of the focal plane and the size of the illumination field of the laser beam L1.

[0095] The repeater 116 includes a first repeater element 116a, a second repeater element 116b, and a fold mirror 116c. The repeater 116 guides the laser beam L05 through the first repeater element 116a, the fold mirror 116c, and the second repeater element 116b to become the laser beam L1.

[0096] The first repeater element 116a and the second repeater element 116b are both aspheric mirrors. In some embodiments, the first element 116a and the second repeater element 116b are off-axis parabolic (OAP) mirrors. By using aspheric mirrors, aberrations can be reduced.

[0097] Because the radius of curvature of the repeater element is related to the tilt angle of the focal plane, the radius of curvature of the first repeater element 116a and the second repeater element 116b can be designed to achieve a desired tilt angle. For example, in some embodiments, after laser beam L05 is reflected by the first repeater element 116a, the tilt angle of the focal plane is approximately 45 degrees. When the ratio of the radius of curvature of the first repeater element 116a to the radius of curvature of the second repeater element 116b is equal to 2:1, the tilt angle of the focal plane of laser beam L1 reflected by the second repeater element 116b can be adjusted to approximately 0 degrees.

[0098] Figure 9 FIG2 is a schematic diagram of objective lens 130 according to some embodiments of the present invention. Objective lens 130 is used to magnify the image formed by laser beam L2. Specifically, objective lens 130 provides a magnification of 100 to 600 for the image formed by laser beam L2. Objective lens 130 comprises two stages. The first stage comprises mirrors 131 and 132, and the second stage comprises mirrors 133 and 134.

[0099] In some embodiments, the magnification of the first stage is between 20 and 80, and the magnification of the second stage is between 5 and 25. In some embodiments, the magnification of the first stage and the second stage is constant. In other embodiments, the magnification of the first stage or the second stage is constant, and the other is adjustable.

[0100] Mirrors 131 and 132 are part of a Schwarzschild objective. Specifically, mirror 131 corresponds to the concave reflector of the Schwarzschild objective, and mirror 132 corresponds to the convex reflector of the Schwarzschild objective. Mirrors 131 and 132 are cut out of the concave and convex reflectors of the Schwarzschild objective based on the chief ray angle (CRA) of laser beam L2. For example, when the beam's CRA is between 20 and 45 degrees, only a portion of the concave reflector of the Schwarzschild objective is illuminated by the beam, and this illuminated portion of the concave reflector can be cut out to form mirror 131.

[0101] As described above, the laser beam L1 is directed toward the wafer W at an angle θ, and thus the laser beam L2 is also reflected at the same angle θ toward the objective lens 130. In this case, the CRA of the laser beam L2 is equal to the angle θ.

[0102] In some embodiments, the reflectors 131 and 132 are aspherical reflectors, while the reflectors 133 and 134 are spherical reflectors or plane reflectors. In other embodiments, the reflectors 133 and 134 are aspherical reflectors.

[0103] like Figure 9 As shown, there is an intermediate focus FS between the reflector 132 and the reflector 133 , and the reflector 133 and the reflector 134 can be aligned according to the intermediate focus FS.

[0104] Figure 10A and Figure 10B FIG. 1 is a schematic diagram of the field of view of the illuminator 120 , the objective lens 130 , and the image sensor 140 according to some embodiments of the present invention.

[0105] exist Figure 10A In the illustrated embodiment, the image sensor 140 may be implemented as a linesensor of a TDI sensor, and the field of view (FOVs) on the wafer W is rectangular. Therefore, the field of view (FOVi) of the illuminator 120 is designed to be rectangular and capable of covering the entire field of view (FOV) of the image sensor 140. Furthermore, the field of view (FOVo) of the objective lens 130 should cover the entire field of view (FOVi) of the illuminator 120.

[0106] In some embodiments, the longitudinal dimension LD1 of the field of view FOVi of the illuminator 120 is approximately 100 μm, while the longitudinal dimension LD2 of the field of view FOVs of the image sensor 140 is approximately 98 μm. In some embodiments, the width dimension LD3 of the field of view FOVi of the illuminator 120 is approximately 6 μm, while the width dimension LD4 of the field of view FOVs of the image sensor 140 is approximately 2.56 μm. The field of view FOVo of the objective lens 130 is circular, with a diameter greater than 100 μm.

[0107] exist Figure 10B In the illustrated embodiment, the image sensor 140 may be implemented as an area sensor, and the field of view (FOVs) of the sensor 140 on the wafer W is square. Therefore, the field of view (FOVi) of the illuminator 120 is designed to be square and capable of covering the entire field of view (FOV) of the sensor 140. Furthermore, the field of view (FOVo) of the objective lens 130 should cover the entire field of view (FOVi) of the illuminator 120.

[0108] In some embodiments, the diagonal FOVs of the sensor 140 is approximately 100 μm. The diameter of the FOV of the illuminator 120 is larger than the diagonal FOVs of the image sensor 140 , and the diameter of the FOVo of the objective lens 130 is larger than the diameter of the FOV of the illuminator 120 .

[0109] In some embodiments, the optical resolution of optical inspection system 10 is equal to 0.5*λ / NA, where NA is the numerical aperture of illuminator 120 and objective lens 130, and λ is the wavelength of laser beam L1. The pixel resolution of image sensor 140 is at least two times smaller than the optical resolution of optical inspection system 10. In some embodiments, the optical resolution of optical inspection system 10 is seven times the pixel resolution of image sensor 140.

[0110] In some embodiments, after the image sensor 140 captures an image of the wafer W, the image processing device 200 performs defect detection and classification based on the captured image, thereby providing a detection result DR for the wafer W. In some embodiments, the image processing device 200 may be a cluster of high-performance computing (HPC) servers that support machine learning models for defect detection and classification for HPC. Figure 11 FIG2 is a schematic diagram of an image processing apparatus 200 according to some embodiments of the present invention. The image processing apparatus 200 includes an image assembler 205, an image denoising unit 210, an image alignment unit 215, an image segmentation unit 220, a defect detection unit 225, a post-segmentation unit 230, a defect classification unit 235, an image quality monitor 240, a design database 245, a process information database 250, and a defect database 255.

[0111] Image assembler 205 receives image M0, position PI, and calibration algorithm CA. In some embodiments, each image M0 can be considered a small segment of wafer W. Image assembler 205 executes calibration algorithm CA to assemble these segments into a predefined image M1 based on position PI for subsequent processing. In some embodiments, the assembling operations include concatenation, overlaying, and header addition.

[0112] The image noise reduction unit 210 is configured to execute a noise reduction algorithm to reduce noise in the image M1 to improve detection accuracy. In some embodiments, the image noise reduction unit 210 may support multiple noise reduction algorithms and may employ at least one specific noise reduction algorithm based on the selection of the detection algorithm.

[0113] Image alignment unit 215 is configured to execute one of multiple alignment algorithms to align pixels and sub-pixels of image sensor 140. In some embodiments, image alignment unit 215 is further configured to receive a selection signal Ss from image quality monitor 240, where selection signal Ss indicates which alignment algorithm should be used. In other words, image quality monitor 240 is configured to determine the appropriate alignment algorithm for image alignment unit 215 to perform real-time alignment. In some embodiments, design database 245 may also provide information DB1 to image alignment unit 215 to improve pixel and sub-pixel alignment performance.

[0114] In some embodiments, the pixels of image sensor 140 are first aligned to ensure an alignment resolution of approximately one pixel dimension, and then the sub-pixels of image sensor 140 are aligned to improve the alignment resolution to a dimension less than one pixel. After the sub-pixels of image sensor 140 are aligned, the alignment resolution can be approximately 0.5, 0.1, or 0.01 times the dimension of one pixel. In some embodiments, both defects and circuit patterns are smaller than the dimension of one pixel, so the alignment resolution should be at least smaller than the dimension of one pixel; otherwise, defects cannot be distinguished from circuit patterns.

[0115] In some embodiments, information DB1 provides a die-to-die (D2D) algorithm to the image alignment unit 215 to improve pixel and sub-pixel alignment performance. In some embodiments, the D2D algorithm enables the image alignment unit 215 to align similar regions (e.g., pixels with the same pattern) with each other.

[0116] In some embodiments, information DB1 provides a die-to-database (D2DB) algorithm to the image alignment unit 215 to improve pixel and sub-pixel alignment performance. In some embodiments, the D2DB algorithm enables the image alignment unit 215 to align the image with a golden image (i.e., a desired image without defects).

[0117] The image segmentation unit 220 is used to execute a first segmentation algorithm to divide the image M1 into multiple segments M2. In some embodiments, the first segmentation algorithm includes smoothing and enhancement to reduce the noise level. The first segmentation algorithm is executed to isolate or highlight potential defective areas from the background or normal areas of the wafer image. That is, similar images (such as images of areas with similar patterns) are placed in the same group to improve operation speed and accuracy. In this case, images of suspicious / defective areas can be separated from the remaining images that may not have defects. Therefore, subsequent defect detection is easier to identify and analyze defects from the images that have passed the first segmentation algorithm. Therefore, the image segmentation unit 220 divides the image M1 into several segments M2 so that different algorithms can be used to check each segment M2. In some embodiments, the design database 245 can also provide information DB2 to the image segmentation unit 220 to improve the performance of image segmentation.

[0118] The defect detection unit 225 is configured to detect whether a defect exists in each segment M2 or what type of defect exists, and to generate an initial result M3 corresponding to each segment M2. The defect detection unit 225 is configured to receive information DB3 from the design database 245. In some embodiments, the information DB3 includes a D2D algorithm, a die-to-cell (D2C) algorithm, or a die-to-AI (D2AI) algorithm. The D2D, D2C, or D2AI algorithm provides functions such as pattern search, pattern browsing, pattern-centric management, and / or pattern-centric machine learning. The defect detection unit 225 applies the D2D, D2C, or D2AI algorithm to detect and determine defects.

[0119] Post-segmentation unit 230 receives information DB4 from design database 245, where information DB4 includes a second segmentation algorithm. Post-segmentation unit 230 is configured to execute the second segmentation algorithm to segment each initial result M3 into a plurality of segments M4. Post-segmentation unit 230 categorizes initial results M3 according to defect type. In some embodiments, post-segmentation unit 230 is also configured to detect the presence of defects detected by defect detection unit 225. In other words, post-segmentation unit 230 is capable of performing false defect filtering.

[0120] The defect classification unit 235 is configured to execute a classification algorithm to classify the defects in the segment M4 into different categories and generate a detection result DR accordingly. In some embodiments, the defect classification unit 235 classifies the defects according to their shape, size, and / or area.

[0121] In some embodiments, the aforementioned algorithms (denoising algorithm, matching algorithm, first segmentation algorithm, second segmentation algorithm, D2D algorithm, D2C algorithm, and D2AI algorithm) are traditional algorithms, machine learning (ML) algorithms, or a combination thereof.

[0122] Image quality monitor 240 is used to monitor the quality of image M1. In some embodiments, image quality monitor 240 monitors the alignment between the captured image and a reference wafer, the alignment between the captured image and a database (i.e., a golden image), defocus / focus, illumination intensity uniformity, objective lens aberration, and / or other calibration parameters of image M1, and feeds quality information QI back to host 700.

[0123] In some embodiments, the data in design library 245 is structured in GDSII or Oasis format. Design library 245 supports automatic recipe generation by combining process information library 250 with wafer information. In some embodiments, design library 245 helps determine the polarization of laser beam L1 based on defect type, as defects may be sensitive to laser beam L1 with a certain polarization. In some embodiments, design library 245 provides collective pupil configurations based on defect type, thereby reducing image diffraction.

[0124] The process information database 250 is used to collect and construct process-related data for the inspection. In some embodiments, the process-related data includes the technology node, process steps, equipment, materials, and optical properties of the optical device 100. The process information database 250 is also used to provide the required parameters when executing the denoising algorithm, alignment algorithm, first segmentation algorithm, second segmentation algorithm, D2D algorithm, D2DB algorithm, D2C algorithm, and D2AI algorithm, and transmit the required parameters along with information DB1 to DB4 to the corresponding units. In some embodiments, the process information database 250 is also used to transmit automatic inspection condition settings to the host computer 700, such as wavelength, aperture, and polarization settings.

[0125] The defect database 255 is used to store defect information, such as detection results DR. The defect database 255 can provide stored information for data analysis and incremental machine learning model adjustment.

[0126] In some embodiments, the design database 245 , the process information database 250 , and the defect database 255 provide an application programming interface (API) for searching, searching, and aggregating at any aggregation level (eg, layer, die, wafer, or lot level).

[0127] Figure 12FIG. 2 is a schematic diagram of a detection platform 20 according to some embodiments of the present invention. In some embodiments, the optical detection system 10 is implemented in the detection platform 20.

[0128] The operation of the optical inspection system 10 includes an initialization phase and an operational phase. During the initialization phase, the inspection platform 20 cooperates with the host computer 700 to calibrate the platform coordinate system and the image coordinate system. During the operational phase, the inspection platform 20 cooperates with the synchronization device 600 and the host computer 700 to synchronize the movement of the wafer W, the timing of generating the laser beam L1, and the timing of capturing the laser beam L2.

[0129] The detection platform 20 includes a frame 21, a top chamber 22, a bottom chamber 23, a shift-in / shift-out (SISO) plate 24, and a vibration isolator 25. The top chamber 22 is mounted on the frame 21 via the vibration isolator 25. The SISO plate 24 is mounted on the lower portion of the frame 21. The bottom chamber 23 can be assembled with the top chamber 22. Figure 12 As shown, the entire top chamber 22 and the bottom chamber 23 are raised and lowered by vibration isolators 25. The vibration isolators 25 are used to reduce vibrations transmitted through the frame 21 from the external environment.

[0130] In some embodiments, the optical device 100, optical microscope OM, and Z sensor (height sensor) ZS are mounted on the top chamber 22, and the wafer stage 300 for supporting the wafer W is disposed in the bottom chamber 23. In some embodiments, the spatial position measurement device SPM is mounted on the frame 21 or the bottom chamber 23. The top chamber 22 and the bottom chamber 23 enclose a space for wafer inspection. In one embodiment, the spatial position measurement device SPM includes an interferometer.

[0131] The operation of optical inspection system 10 also includes a maintenance phase. During this phase, bottom chamber 23 is disassembled from top chamber 22, and SISO board 24 is used to remove bottom chamber 23 from inspection platform 20. Because top chamber 22 and bottom chamber 23 are detachable from each other, maintenance on bottom chamber 23 and other portions of inspection platform 20 is easily performed.

[0132] Figure 13 is a schematic diagram of a host 700 according to some embodiments of the present invention.

[0133] The optical inspection system 10 utilizes a stage coordinate system and an image coordinate system. The movement of the wafer W is defined by the stage coordinate system, while the movement of the image (e.g., image M0) is defined by the image coordinate system. Initially, the host computer 700 executes a system calibration algorithm (SCA) to establish a transformation between the stage coordinate system and the image coordinate system. Once the transformation is established, when a position or movement in the stage coordinate system is obtained, the corresponding position or movement in the image coordinate system is determined by executing the transformation. The host computer 700 also executes the system calibration algorithm (SCA) to align the region of interest on the wafer W with the focus of the laser beam L1.

[0134] The host computer 700 is configured to execute a recipe setting function RSF to receive external instructions. In some embodiments, the user can set a recipe to the optical inspection system 10 through the recipe setting function RSF. In some embodiments, the user can view the wafer W through the optical microscope OM and determine the areas of the wafer W to be inspected. The user can then notify the host computer 700 through the recipe setting function RSF to include these areas in the predetermined inspection path P0.

[0135] In addition, the host 700 is configured to execute a system alignment algorithm SAA to correct the position of the wafer W to compensate for the offset accumulated during the wafer loading process.

[0136] The host computer 700 is also configured to execute the inspection planning and control function (IPCF) to determine the predetermined inspection path P0. Specifically, the host computer 700 receives data input via the recipe setting function (RSF), executes the system alignment algorithm (SAA), obtains alignment results, and accordingly determines the predetermined inspection path P0. The predetermined inspection path P0 is then transmitted to the image processing device 200, the motion control device 400, and the synchronization device 600.

[0137] Figure 14 is a schematic diagram of motion control during the operation phase according to some embodiments of the present invention. The motion control device 400 is configured to generate a target point based on a predetermined inspection path P0 and a movement speed V. The target point represents a location within a region of interest on wafer W. At the start of the operation phase, the synchronization device 600 sends a trigger signal ST to the motion control device 400 to notify the motion control device 400 to begin moving wafer W along the predetermined inspection path P0. In some embodiments, the trigger signal ST can be generated by the motion control device 400. In this embodiment, the motion control device 400 initiates the operation phase and notifies the synchronization device 600 via the trigger signal ST.

[0138] As the wafer W moves, the spatial position measurement device SPM and the Z sensor ZS continuously sense the position Pxy and position Pz, respectively. In some embodiments, the position Pxy and position Pz may include errors due to certain non-ideal factors. The position measurement module 500 is used to execute a compensation algorithm to correct the position Pxy and position Pz. The corrected position Pxy' and the corrected position Pz' are transmitted to the motion control device 400. It should be noted that the corrected position Pxy' and the corrected position Pz' are the instantaneous positions of the wafer W. The motion control device 400 can compare the instantaneous position with the target point (that is, the position on the predetermined detection path P0) to adjust the movement of the wafer W. For example, when the offset between the instantaneous position and the target point exceeds a threshold, the motion control device 400 executes a control algorithm to generate a calibration signal SC4 to the wafer platform 300. Then, the wafer platform 300 generates a force in response to the calibration signal Sca to change the position, velocity and / or acceleration of the wafer W so that the offset remains less than the threshold.

[0139] Figure 15 FIG6 is a schematic diagram illustrating synchronization of illumination, image sensing, and wafer position according to some embodiments of the present invention. Synchronization device 600 is configured to control the timing of illumination and sensing based on the position of wafer W, such that the movement of regions of interest (denoted by P1 through P5) on wafer W is synchronized with the illumination period PDi of laser beam L1 and the detection period PDd of image sensor 140.

[0140] like Figure 15 As shown, when determining the time it takes to capture images of positions P1 through P5, the illumination period PDi coincides with the time it takes for positions P1 through P5 to arrive at the locations to be illuminated. Therefore, positions P1 through P5 can be illuminated. Furthermore, the detection period PDd coincides with the illumination period PD, ensuring that image sensor 140 can receive the laser beams reflected from positions P1 through P5.

[0141] In some embodiments, the image sensor 140 is triggered by a trigger pulse TP. In some embodiments, the trigger pulse TP is externally emitted.

[0142] The features of several embodiments of the present invention are summarized above so that those skilled in the art can better understand the various aspects of the present invention. Those skilled in the art will recognize that they can easily use the present invention as a basis for designing or modifying other processes and structures to implement the same purposes and / or achieve the same advantages as the embodiments described herein. Those skilled in the art will also recognize that such equivalent interpretations do not depart from the spirit and scope of the present invention, and that they can make various changes, substitutions, and replacements herein without departing from the spirit and scope of the present invention.

[0143] Explanation of symbols

[0144] 10: Optical detection system

[0145] 20: Detection Platform

[0146] 21: Framework

[0147] 22: Top cavity

[0148] 23: Bottom cavity

[0149] 24: Move in and out of the board

[0150] 25: Vibration isolator

[0151] 100: Optical device

[0152] 110: Light Source

[0153] 110a: Light source

[0154] 111: Pump Laser

[0155] 112: Central wavelength selector

[0156] 113: First Harmonic Generator

[0157] 114: Second Harmonic Generator

[0158] 114C: Control unit

[0159] 115: Concentrator

[0160] 116: Repeater

[0161] 116a: First repeater element

[0162] 116b: Second repeater element

[0163] 116c: Folding Mirror

[0164] 120: Illuminator

[0165] 121: Collector

[0166] 122: Speckle Detector

[0167] 123: Collimator

[0168] 124: Homogenizer

[0169] 130: Objective lens

[0170] 131: Reflector

[0171] 132: Reflector

[0172] 133: Reflector

[0173] 134: Reflector

[0174] 140: Image sensor

[0175] 200: Image processing device

[0176] 205: Image assembler

[0177] 210: Image noise reduction unit

[0178] 215: Image alignment unit

[0179] 220: Image Segmentation Unit

[0180] 225: Defect Detection Unit

[0181] 230: Post-division unit

[0182] 235: Defect Classification Unit

[0183] 240: Image Quality Monitor

[0184] 245: Design Database

[0185] 250: Process Information Database

[0186] 255: Defect Database

[0187] 300: Wafer platform

[0188] 301: Laser unit

[0189] 302: Wave Delay

[0190] 303: Spectrum Shaper

[0191] 310: First wavelength conversion channel

[0192] 311: Nonlinear Unit

[0193] 312: Wavelength separator

[0194] 313: Nonlinear Unit

[0195] 314: Wavelength separator

[0196] 321: Nonlinear Unit

[0197] 323: Nonlinear Unit

[0198] 324: Wavelength separator

[0199] 330: XUV generator

[0200] 331: Polarization switch

[0201] 332: Nonlinear Unit

[0202] 333: Polarization switch

[0203] 334: Nonlinear Unit

[0204] 335: Wavelength separator

[0205] 336: Wavelength Separator

[0206] 341: Recipe Controller

[0207] 342: Gas Mixer

[0208] 343: Gas Controller

[0209] 400: Motion control device

[0210] 500: Position measurement module

[0211] 600: Synchronization device

[0212] 700: Host

[0213] 1141: Framework

[0214] 1142: Connector

[0215] 1143: Deformable Element

[0216] 1144: Electrode

[0217] B0: source laser beam

[0218] B1: Narrowband laser beam

[0219] B2: Deep ultraviolet laser beam

[0220] C1: Infrared laser

[0221] C2: Infrared laser

[0222] C3: Infrared laser

[0223] C4: Infrared laser

[0224] C5: Infrared laser

[0225] C6: Infrared laser

[0226] CA: Calibration Algorithm

[0227] CL: cylindrical mirror

[0228] D1: Vertical

[0229] D2: width direction

[0230] DB1: Information

[0231] DB2: Information

[0232] DB3: Information

[0233] DB4: Information

[0234] DM: Deformable Mirror

[0235] DR: Detection results

[0236] E1: UV laser

[0237] E2: UV laser

[0238] E3: Ultraviolet laser

[0239] E4: Ultraviolet laser

[0240] E5: Ultraviolet laser

[0241] E6: Ultraviolet laser

[0242] FOVi: Field of View

[0243] FOVo: Field of View

[0244] FOVs: Field of View

[0245] FP: Intermediate focal plane

[0246] GS1: Source Gas

[0247] GS2: Gas

[0248] IPCF: Inspection Planning and Control Function

[0249] L0: Laser beam

[0250] L01: Laser beam

[0251] L02: Laser beam

[0252] L03: Laser beam

[0253] L04: Laser beam

[0254] L04n: Bundle section

[0255] L05: Laser Beam

[0256] L1: Laser beam

[0257] L2: Laser beam

[0258] L3: Laser beam

[0259] LA: Loading Angle

[0260] LD1: longitudinal dimension

[0261] LD2: Longitudinal dimension

[0262] LD3: Latitude

[0263] LD4: Latitude

[0264] LV1: Level

[0265] LV2: Level

[0266] M0: Image

[0267] M1: Imaging

[0268] M2: Segment

[0269] M3: Initial Results

[0270] M4: Segment

[0271] MB: membrane

[0272] ML: MEMS Mirrors

[0273] OM: Optical microscope

[0274] P0: Predetermined detection path

[0275] P1: Position

[0276] P2: Position

[0277] P3: Position

[0278] P4: Position

[0279] P5: Location

[0280] PDd: detection period

[0281] PDi: Lighting period

[0282] PH: Pitch

[0283] PI: Position

[0284] Pm: moving path

[0285] Pxy: Position

[0286] Pxy': Position

[0287] Pz: Position

[0288] Pz': Position

[0289] QI: Quality Information

[0290] RC: radius of curvature

[0291] RG1: High-voltage area

[0292] RG2: Low extreme ultraviolet absorption region

[0293] RP1: First corrugated structure

[0294] RP2: Second corrugated structure

[0295] RSF: Recipe Setting Function

[0296] SAA: System Alignment Algorithm

[0297] Sc: calibration signal

[0298] SC1: Control signal

[0299] SC2: Control signal

[0300] SC3: Control signal

[0301] SC4: Calibration signal

[0302] SCA: System Calibration Algorithm

[0303] SG: slack

[0304] SG1: Source gas

[0305] SG2: Gas

[0306] SPM: Spatial Position Measurement Device

[0307] Ss: Select signal

[0308] ST: trigger signal

[0309] TP: Trigger Pulse

[0310] V: Movement Speed

[0311] V1: Visible light laser

[0312] V2: Visible light laser

[0313] V3: Visible light laser

[0314] V4: Visible light laser

[0315] W: wafer

[0316] X1: Extreme ultraviolet laser

[0317] X2: Extreme ultraviolet laser

[0318] ZS: Z sensor

[0319] θ: angle

Claims

1. An optical detection system comprising: An optical device comprising: a light source for generating a first laser beam by performing a first nonlinear harmonic generation and a second nonlinear harmonic generation and directing the first laser beam through an illuminator to become an incident laser beam toward a wafer, thereby correspondingly generating a reflected laser beam; and an image sensor for capturing the reflected laser beam passing through an objective lens to become a second laser beam and correspondingly generating an image of the wafer; and an image processing device for generating a detection result according to the image, wherein a wavelength of the incident laser beam is less than 120 nm, wherein the wavelength of the first laser beam is adaptively adjustable, The first nonlinear harmonic generation is performed in a solid-state material, and the second nonlinear harmonic generation is performed in a gas.

2. The optical detection system of claim 1, wherein the incident laser beam is a pulsed laser, and the wavelength of the incident laser beam ranges from 50 nm to 120 nm.

3. The optical inspection system of claim 1 , wherein the light source comprises: a laser pump for generating a source laser beam; a central wavelength selector for selecting a wavelength range from the source laser beam to generate a narrowband laser beam; as well as A first harmonic generator is used to perform the second nonlinear harmonic generation to generate the first laser beam according to the narrowband laser beam.

4. The optical inspection system of claim 3, wherein the light source further comprises: a second harmonic generator for performing the first nonlinear harmonic generation to convert the narrowband laser beam into a DUV laser beam, The second nonlinear harmonic generation is a third-order nonlinear generation, and the first nonlinear harmonic generation is a second-order nonlinear generation.

5. The optical detection system as claimed in claim 3, wherein the spectrum range of the source laser beam is from 400 nm to 1100 nm, and the wavelength range selected by the central wavelength selector is visible light or infrared light.

6. The optical detection system as claimed in claim 3, wherein the first harmonic generator operates in an inert gas.

7. The optical detection system of claim 1 , wherein the optical device further comprises: the luminaire; as well as The objective lens, The illuminator comprises: a collector for collecting and projecting the first laser beam into a first beam; a speckle detector for receiving the first beam and reducing a coherence of the first beam to form a second beam; a collimator for collimating the second beam into a third beam; a homogenizer for making the light intensity of the third beam uniform and shaping the third beam into a fourth beam; a concentrator for concentrating the fourth beam into a fifth beam; and A relay is used for adjusting a numerical aperture of the fifth beam and relaying the fifth beam to become the first laser beam.

8. The optical detection system of claim 7, wherein the illuminator and the objective lens are reflective optical devices.

9. The optical inspection system of claim 7, wherein the objective lens comprises: a first curved reflector; a second curved reflector; a first reflecting mirror; as well as a second reflecting mirror, The reflected laser beam is guided to the image sensor via the first curved reflector, the second curved reflector, the first reflector, and the second reflector in sequence. The first curved reflector is a reflector formed by cutting the concave mirror of the Schwarzschild objective lens, the second curved reflector is a reflector formed by cutting the convex mirror of the Schwarzschild objective lens, and the first curved reflector and the second curved reflector together constitute a part of the Schwarzschild objective lens. 10 . The optical inspection system as claimed in claim 9 , wherein the first curved reflector and the second curved reflector are aspherical reflectors.

11. The optical inspection system of claim 9, wherein an intermediate focus is formed between the second curved reflector and the first reflector.

12. The optical inspection system of claim 1, wherein the image sensor is implemented by a time delay integration sensor.

13. The optical inspection system of claim 1, further comprising: a wafer platform for carrying the wafer; a motion control device for controlling the wafer platform to move the wafer along a predetermined inspection path; as well as A spatial position measuring device is used to obtain a position of the wafer in a platform coordinate system.

14. The optical inspection system of claim 13, further comprising: A position measurement module is used to determine whether the wafer deviates from the predetermined detection path according to the position of the wafer.

15. The optical inspection system of claim 13, further comprising: A host is used to set the predetermined detection path.

16. The optical inspection system of claim 1, further comprising: A synchronization device is used to synchronize the timing of the incident laser beam generated by the light source with the timing of the image sensor capturing the reflected laser beam.

17. An optical detection system comprising: An optical device comprising: A light source is configured to generate a first laser beam by performing multiple nonlinear harmonic generation operations and direct the first laser beam through an illuminator to become an incident laser beam toward a wafer, thereby correspondingly generating a reflected laser beam, wherein the light source comprises: A first wavelength conversion channel for generating a first UV laser: a second wavelength conversion channel for generating a second UV laser; and an XUV generator, configured to generate the first laser beam according to the first UV laser or the second UV laser; an image sensor for capturing the reflected laser beam through an objective lens as a second laser beam and correspondingly generating an image of the wafer; and An image processing device is used to generate a detection result according to the image.

18. The optical inspection system of claim 17, wherein the light source further comprises: a laser unit for generating a first IR laser; a wave retarder for adjusting a polarization of the first IR laser to form a second IR laser; and a spectrum shaper for filtering a spectrum of the second IR laser and generating a third IR laser accordingly, The third IR laser is emitted to the first wavelength conversion channel and the second wavelength conversion channel, and the first wavelength conversion channel and the second wavelength conversion channel respectively generate the first UV laser and the second UV laser according to the third IR laser.

19. The optical detection system of claim 18, wherein the first wavelength conversion channel comprises a first NOP unit and a second NOP unit, wherein the first NOP unit is configured to convert the third IR laser light into a first visible light laser light and a fourth IR laser light, and the second NOP unit is configured to generate a third UV laser light based on the first visible light laser light. wherein the first UV laser is generated according to the third UV laser, The first NOP unit and the second NOP unit operate in solid state materials.

20. The optical inspection system of claim 17, wherein the XUV generator comprises: a third NOP unit for generating an XUV laser and a fourth UV laser according to the first UV laser; as well as a wavelength separator for separating the XUV laser from the fourth UV laser to generate the first laser beam, The third NOP unit operates in an inert gas.

21. The optical inspection system of claim 17, wherein the optical device further comprises: The luminaire comprises: a collector for collecting and projecting the first laser beam into a first beam; a speckle detector for receiving the first beam and reducing the coherence of the first beam to form a second beam; a collimator for collimating the second beam into a third beam; a homogenizer for making the light intensity of the third beam uniform and shaping the third beam into a fourth beam; a concentrator for concentrating the fourth beam into a fifth beam; and a relay for adjusting a numerical aperture of the fifth beam and relaying the fifth beam to the first laser beam; and The objective lens comprises: a first curved reflector; a second curved reflector; a first reflecting mirror; and a second reflecting mirror, The reflected laser beam is guided to the image sensor via the first curved reflector, the second curved reflector, the first reflector, and the second reflector in sequence. The first curved reflector is a reflector formed by cutting the concave mirror of the Schwarzschild objective lens, the second curved reflector is a reflector formed by cutting the convex mirror of the Schwarzschild objective lens, and the first curved reflector and the second curved reflector together constitute a part of the Schwarzschild objective lens.

Images