Method for controlling a stage in an optical detection system and optical detection system

Abstract

The application provides a control method of a worktable in an optical detection system and the optical detection system. The control method of the worktable in the optical detection system comprises the following steps: collecting a real-time interference fringe image formed by single-path diffraction light on a surface of a sample to be detected; reducing the real-time interference fringe image from a two-dimensional image to a one-dimensional real-time signal; performing phase correlation calculation on the one-dimensional real-time signal and a one-dimensional reference signal obtained in advance to obtain a phase change amount of the one-dimensional real-time signal relative to the one-dimensional reference signal; and converting the phase change amount into an axial displacement amount of the worktable, and controlling the worktable to move along a lifting direction according to the axial displacement amount to realize real-time focusing and tracking. According to the scheme, the reference signal enhances the contrast and stability of the interference fringe, and provides reliable support for the phase correlation calculation; in combination with sub-pixel level phase analysis and displacement conversion, the worktable realizes nanoscale real-time focusing and tracking, and the detection precision and stability are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of semiconductor technology, and in particular to a control method for a stage in an optical inspection system and an optical inspection system. Background Technology

[0002] As the feature sizes of semiconductor integrated circuits continue to shrink to the nanometer and even sub-nanometer level in accordance with Moore's Law, high-end manufacturing has placed higher demands on the accuracy and stability of inspection and measurement technologies in the process. Against this backdrop, cutting-edge precision inspection equipment such as Electron Beam Inspection (EBI) and Optical Critical Dimension (OCD) have become indispensable key tools for ensuring production line yield and process iteration.

[0003] The primary prerequisite for the proper functioning of such equipment is that the observed wafer surface must be extremely stable and precisely controlled at the instrument's optimal focal plane. Since the depth of focus in advanced manufacturing processes has shrunk to the nanometer scale, even minute defocusing can lead to blurred images, reduced edge contrast, and deteriorated signal-to-noise ratio. This makes it difficult to identify minute defects and causes significant deviations or even distortions in critical dimension measurements. Ultimately, these measurement errors propagate through layers, directly affecting the judgment of process windows, the accurate assessment of production yield, and even the feedback control of subsequent process parameters, thus having an immeasurable impact on the performance and output of the final product. Therefore, achieving and maintaining nanometer-level focal plane control has become a key bottleneck restricting the development of advanced manufacturing technologies.

[0004] Currently, common technical solutions for autofocusing on Z-axis tables mainly include the following: The first type is the intensity / contrast-based focusing method. This method analyzes the sharpness of the acquired image (e.g., calculating gray-level gradient, variance, Laplacian operator, etc. as evaluation functions) and searches for the Z-axis position with the largest function value as the focal plane. The inherent drawbacks of this method are: First, the evaluation function curve is relatively flat near the focal point, making it difficult to accurately locate the peak point, resulting in limited focusing accuracy and difficulty in consistently achieving nanometer-level precision; second, this method heavily relies on the overall brightness of the image and is prone to misjudgments due to differences in the reflectivity of the sample surface material (e.g., the different reflectivities of silicon and glass) and fluctuations in the illumination source, resulting in poor robustness; finally, it requires multiple scans to search for the peak, leading to slow algorithm convergence and failing to meet the real-time requirements of high-speed online detection.

[0005] The second method is differential astigmatism: This method introduces optical elements such as cylindrical mirrors into the optical path to convert the defocusing amount into a change in the shape of the light spot, and the differential signal is detected by a four-quadrant detector. Although it is faster, its disadvantages are: the optical system is complex and difficult to calibrate; the measurement accuracy is easily affected by the laser source mode itself, the collimation of the optical path, and the performance drift of the detector; its linear measurement range is narrow, and its ability to track large-scale defocusing or surface undulations is insufficient.

[0006] The third method is traditional interferometry. Although nanometer-level precision can theoretically be achieved using the principles of optical interference, traditional interferometer solutions are typically complex, costly, and extremely sensitive to environmental factors (vibration, temperature). When processing wafer surfaces with complex morphologies or multiple materials, the interference signal is prone to degradation, the signal-to-noise ratio decreases, and the anti-interference capability is insufficient. Furthermore, traditional time-domain phase calculation algorithms are computationally intensive, and their response speed is difficult to meet the demands of modern high-speed production lines.

[0007] In summary, existing technologies cannot simultaneously meet the development needs of high-end manufacturing in terms of the three core performance indicators: response speed, computational accuracy, and anti-interference capability. In particular, when facing application scenarios involving wafers of different materials, complex working conditions, and high-speed scanning, existing solutions cannot combine high speed, high precision, and high robustness, making it difficult to achieve stable and reliable nanometer-level real-time focusing and tracking. Summary of the Invention

[0008] One objective of this invention is to solve the problems of limited accuracy and susceptibility to interference from sample surface characteristics in traditional optical detection focusing methods.

[0009] A further objective of this invention is to solve the technical problem of being unable to accurately and in real-time focus and track the axial position of the stage during sample testing by relying on single-path diffraction interference phase analysis and sub-pixel-level solution.

[0010] Specifically, the present invention provides a method for controlling a stage in an optical inspection system, comprising: acquiring a real-time interference fringe image formed on the surface of a sample under test by a single-path diffracted light; reducing the real-time interference fringe image from a two-dimensional image to a one-dimensional real-time signal; performing phase correlation calculation between the one-dimensional real-time signal and a pre-acquired one-dimensional reference signal to obtain the phase change of the one-dimensional real-time signal relative to the one-dimensional reference signal; and converting the phase change into an axial displacement of the stage, and controlling the stage to move along the lifting direction based on the axial displacement to achieve real-time focusing and tracking.

[0011] Optionally, the step of pre-acquiring a one-dimensional reference signal includes: adjusting the surface of the sample to be tested to the optimal focal plane, acquiring a reference interference fringe image formed by a single-path diffraction beam on the surface of the sample; projecting the reference interference fringe image along the direction of maximum phase gradient of the interference fringes to generate an initial reference signal; and performing peak detection and sub-pixel-level peak and valley localization on the initial reference signal, and using adaptive contrast enhancement to obtain an enhanced one-dimensional reference signal.

[0012] Optionally, after the step of reducing the real-time interference fringe image from a two-dimensional image to a one-dimensional real-time signal, the method further includes: performing background correction and denoising processing on the one-dimensional real-time signal, smoothing high-frequency speckle noise and eliminating low-frequency background drift through local window polynomial fitting, and preserving the peak shape characteristics of the interference fringes.

[0013] Optionally, after the step of performing background correction and denoising on the one-dimensional real-time signal, the method further includes: applying a window function to the one-dimensional real-time signal after background correction and denoising to suppress spectral leakage caused by truncation during the fast Fourier transform of the one-dimensional real-time signal.

[0014] Optionally, the step of performing phase correlation calculation between the one-dimensional real-time signal and the pre-acquired one-dimensional reference signal includes: performing fast Fourier transform on the one-dimensional reference signal and the one-dimensional real-time signal respectively, calculating the cross power spectrum, and then performing inverse Fourier transform on the cross power spectrum to obtain the phase correlation function.

[0015] Optionally, after obtaining the phase correlation function, the method further includes: performing parabolic fitting on the main peak of the phase correlation function to obtain a one-dimensional signal shift, which is used to determine the phase change.

[0016] Optionally, the step of converting the phase change into the axial displacement of the worktable includes: converting the phase change into the axial displacement according to a preset phase-displacement mathematical model, wherein the phase-displacement mathematical model is established based on the optical parameters of the optical detection system and is used to characterize the correspondence between the phase change and the axial displacement.

[0017] Optionally, before acquiring the real-time interference fringe image formed by a single-path diffraction light on the surface of the sample under test, the method further includes: loading a corresponding preset algorithm parameter set according to the type of the sample under test to adapt to surfaces with different reflectivities. The preset algorithm parameter set includes window function type and filtering parameters.

[0018] Optionally, the projection direction used to reduce the real-time interference fringe image from a two-dimensional image to a one-dimensional real-time signal is consistent with the projection direction used to generate the one-dimensional reference signal.

[0019] According to another aspect of the present invention, an optical inspection system is also provided, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the control method of the stage in any of the above-described optical inspection systems.

[0020] The stage control method in the optical detection system of the present invention significantly improves the uniformity and contrast of the interference fringe signal by performing peak detection, sub-pixel level positioning and adaptive contrast enhancement processing on the reference signal, providing a more stable template for phase correlation calculation and effectively reducing interference caused by noise and signal attenuation in the edge region.

[0021] Furthermore, the stage control method in the optical detection system of the present invention, relying on high-precision phase correlation calculation and sub-pixel level fitting, can quickly and accurately resolve the phase change of the real-time interference fringes, and convert it into the axial displacement of the stage through a phase-displacement mathematical model, thereby realizing high-precision real-time focusing and tracking of the sample surface during optical detection, and significantly improving focusing accuracy and detection stability.

[0022] The above and other objects, advantages and features of the present invention will become more apparent to those skilled in the art from the following detailed description of specific embodiments of the invention in conjunction with the accompanying drawings. Attached Figure Description

[0023] The following sections will describe some specific embodiments of the invention in detail by way of example and not limitation, with reference to the accompanying drawings. The same reference numerals in the drawings denote the same or similar parts or portions. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings: Figure 1 This is a schematic diagram of a control method for a worktable in an optical inspection system according to an embodiment of the present invention; Figure 2 This is a comparison diagram showing the effect of reference frame signal enhancement processing in the stage control method of an optical inspection system according to an embodiment of the present invention; Figure 3 This is an overall block diagram of an optical detection system according to an embodiment of the present invention; Figure 4 This is a schematic diagram of a computer program product according to an embodiment of the present invention; Figure 5 This is a schematic diagram of a computer-readable storage medium according to an embodiment of the present invention; and Figure 6 This is a schematic diagram of an optical detection system according to an embodiment of the present invention. Detailed Implementation

[0024] This embodiment provides a method for controlling the stage in an optical inspection system. By enhancing the reference signal, the contrast and stability of the interference fringes are effectively improved, providing a reliable basis for high-precision phase correlation calculation. Relying on sub-pixel-level phase resolution and displacement conversion, high-precision real-time focusing and tracking of the sample surface by the stage in optical inspection is realized, significantly improving the detection accuracy and stability. Figure 1 This is a schematic diagram of a stage control method in an optical inspection system according to an embodiment of the present invention. Figure 1 As shown, the stage control method in the optical inspection system of this embodiment generally includes: Step S102: Acquire a real-time interference fringe image formed on the surface of the sample under test by a single-path diffraction beam; Step S104: Reduce the real-time interference fringe image from a two-dimensional image to a one-dimensional real-time signal; Step S106: Perform phase correlation calculation between the one-dimensional real-time signal and the pre-acquired one-dimensional reference signal to obtain the phase change of the one-dimensional real-time signal relative to the one-dimensional reference signal. In step S108, the phase change is converted into the axial displacement of the worktable, and the worktable is controlled to move along the lifting direction based on the axial displacement to achieve real-time focusing and tracking.

[0025] In the above steps, the worktable is a Z-axis worktable, meaning the Z-axis direction is the lifting direction. The sample to be tested can be a wafer. In a specific embodiment, before step S102, which involves acquiring a real-time interference fringe image formed by single-path diffraction light on the surface of the sample to be tested, the method may further include: loading a corresponding preset algorithm parameter set according to the type of the sample to be tested to adapt to surfaces with different reflectivities. The preset algorithm parameter set includes window function type and filtering parameters.

[0026] This is the system initialization and parameter loading process. In one specific embodiment, the system starts and executes an initialization self-test process. The user can select the appropriate mode according to the current wafer type (such as silicon or glass), and the system loads a preset set of algorithm parameters (such as window function type, filtering parameters, etc.) to adapt to different reflective surfaces and ensure optimal algorithm performance.

[0027] The step of pre-acquiring the one-dimensional reference signal in step S106 may specifically include: adjusting the surface of the sample to be tested to the optimal focal plane, acquiring a reference interference fringe image formed on the surface of the sample by a single-path diffraction beam; projecting the reference interference fringe image along the direction of maximum phase gradient of the interference fringes to generate an initial reference signal; and performing peak detection and sub-pixel-level peak-valley localization on the initial reference signal, and using adaptive contrast enhancement to obtain the enhanced one-dimensional reference signal. It should be noted that the projection direction used in step S104 to reduce the real-time interference fringe image from a two-dimensional image to a one-dimensional real-time signal is consistent with the projection direction used to generate the one-dimensional reference signal.

[0028] The first step involves the acquisition of a reference frame and the generation of a one-dimensional reference signal, where a frame refers to the interference fringe image captured by the camera in a single acquisition. In one specific embodiment, the Z-axis stage is controlled to precisely adjust the wafer surface to the optimal focal plane. In this state, the camera acquires a high signal-to-noise ratio, high-contrast reference interference fringe image, which is stored as I_ref(x,y). Along the dimension perpendicular to the direction of the interference fringes (i.e., the direction of maximum phase gradient), the pixel grayscale values ​​of the two-dimensional reference interference fringe image are projected to generate a one-dimensional initial reference signal S_ref(n). This signal completely preserves the periodic phase information of the interference fringes.

[0029] Then, reference frame signal enhancement can be performed. Specifically, peak detection is performed on the initial reference signal S_ref(n) to identify the peak / valley regions of each interference fringe. The information obtained at this point is pixel-level positional information, and its sub-pixel-level peak / valley positions and peak values ​​are calculated through polynomial fitting. Based on this information, an adaptive contrast enhancement method is used to improve the intensity of the target signal, resulting in a more ideal one-dimensional reference signal S_ref_enhanced(n).

[0030] Figure 2 This is a comparison diagram showing the effect of reference frame signal enhancement processing in the stage control method of an optical inspection system according to an embodiment of the present invention. Figure 2As shown, the reference frame signal enhancement in this embodiment leverages the human eye's sensitivity to high-frequency signals, effectively increasing the strength of the effective signal embedded in low-frequency background noise through a local enhancement strategy. Unlike traditional global enhancement methods (such as linear contrast stretching and histogram equalization), this embodiment avoids the noise amplification problem caused by uniform processing. The core innovation of this embodiment lies in using pre-acquired peak / valley information to adaptively identify key signal regions and implement differentiated enhancement based on this: significant enhancement is performed in effective signal regions such as peaks and troughs, while the enhancement intensity is maintained or suppressed in background regions. The method of this embodiment can significantly improve the sharpness of interference fringe edges and the clarity of periodic features, providing a high-quality, high signal-to-noise ratio input signal for subsequent phase correlation calculations, thereby comprehensively improving the system's measurement accuracy and anti-interference capability.

[0031] In one specific embodiment, after reducing the real-time interference fringe image from a two-dimensional image to a one-dimensional real-time signal in step S104, the method may further include: performing background correction and denoising processing on the one-dimensional real-time signal, smoothing high-frequency speckle noise and eliminating low-frequency background drift through local window polynomial fitting, while preserving the peak shape characteristics of the interference fringes. Furthermore, after the step of performing background correction and denoising processing on the one-dimensional real-time signal, the method may further include: applying a window function to the one-dimensional real-time signal after background correction and denoising processing to suppress spectral leakage caused by truncation during the fast Fourier transform of the one-dimensional real-time signal.

[0032] This describes the real-time frame acquisition and signal processing process. After the optical detection system enters the scanning detection phase, the optical imaging module continuously acquires real-time interference fringe images I_curr(x,y) at a high frame rate. The acquisition process is hardware-triggered and synchronized to ensure strict correspondence with the scanning position. For each frame I_curr(x,y), the same one-dimensional signal transformation operation as the one-dimensional reference signal acquisition described earlier is performed to generate a one-dimensional real-time signal S_curr(n).

[0033] To ensure computational accuracy and robustness, background correction and windowing are performed on the one-dimensional real-time signal S_curr(n). This generates S_preprocessed_curr(n). In one specific embodiment, the Savitzky-Golay filter can be used to process the original one-dimensional real-time signal S_curr(n) for background correction and denoising. This filter, through polynomial fitting within a local window, effectively smooths high-frequency speckle noise, eliminates low-frequency background drift, and preserves the sharp peak characteristics of the interference fringe signal to the maximum extent, resulting in S1_curr(n) = SG(S_curr(n)).

[0034] In one specific embodiment, regarding windowing, a window function (such as a Hamming window) W(n) can be applied to the background-corrected signal S1(n) to obtain S_preprocessed_curr(n) = S1_curr(n)•W(n). This step aims to reduce spectral leakage caused by signal truncation during the Fast Fourier Transform (FFT), making subsequent frequency domain analysis results more accurate.

[0035] Step S106, which involves calculating the phase correlation between the one-dimensional real-time signal and the pre-acquired one-dimensional reference signal, may specifically include: performing Fast Fourier Transform on both the one-dimensional reference signal and the one-dimensional real-time signal to calculate the cross-power spectrum, and then performing an Inverse Fourier Transform on the cross-power spectrum to obtain the phase correlation function. Furthermore, after obtaining the phase correlation function, the step may further include: performing parabolic fitting on the main peak of the phase correlation function to obtain the one-dimensional signal shift, which is used to determine the phase change.

[0036] This describes the phase correlation calculation and sub-pixel-level peak localization process. In a specific embodiment, the following operations are performed on the preprocessed one-dimensional reference signal S_ref_enhanced(n) and the one-dimensional real-time signal S_preprocessed_curr(n): Fast Fourier Transform is performed respectively to obtain their frequency domain representations F_ref(f) and F_curr(f). The cross-power spectrum (CPS) of the two signals is calculated: CPS(f) = [F_curr(f)] F_ref'(f)] / |F_curr(f) F_ref'(f)|. Where F_ref'(f) is the complex conjugate of F_ref(f).

[0037] The inverse Fourier transform (IFFT) of the cross-power spectrum CPS(f) yields the phase correlation function P(n), which, ideally, is a pulse peak in the form of a Dirac delta function. The position of the dominant peak in the phase correlation function P(n) is then located. The horizontal displacement Δn of this peak represents the one-dimensional signal translation of the real-time frame relative to the reference frame. To achieve nanometer-level precision, subpixel-level interpolation is required to locate the dominant peak. This subpixel-level precise one-dimensional signal translation Δn_subpixel is obtained through parabolic fitting.

[0038] In a specific embodiment, the step of converting the phase change amount into the axial displacement amount of the worktable in step S108 may specifically include: converting the phase change amount into the axial displacement amount according to a preset phase-displacement mathematical model. The phase-displacement mathematical model is established based on the optical parameters of the optical detection system and is used to characterize the correspondence between the phase change amount and the axial displacement amount.

[0039] This is the process of phase to Z-displacement conversion. In a specific embodiment, the one-dimensional signal translation Δn_subpixel directly corresponds to the phase change Δφ of the interference fringes. The phase-displacement mathematical model ΔZ=k is established based on the specific optical parameters of the optical detection system (such as light wavelength λ, interference angle θ, etc.). Δφ (where k is a proportionality coefficient) converts the phase change Δφ into a precise axial displacement ΔZ. This axial displacement ΔZ is then used as a control signal and sent to the Z-axis stage driver to move the stage a corresponding distance, compensating for defocusing and thus achieving nanometer-level real-time focusing and tracking.

[0040] It should be noted that the two-dimensional image correlation localization algorithm is based on classic computer vision image registration techniques. Its core assumption is that when the Z-axis stage defocuses, the acquired single-path diffraction interference fringe image will undergo a slight translation or deformation on the two-dimensional plane. In some other embodiments, the defocus amount can be deduced by calculating the relative displacement between the real-time acquired interferogram and the reference interferogram pre-stored at the optimal focal plane position. However, this method is computationally intensive and sensitive to image deformation, making it difficult to compete with the one-dimensional phase correlation scheme of this embodiment in the demanding nanometer-scale real-time focusing scenarios requiring high speed, high precision, and high robustness. This embodiment's scheme, through dimensionality reduction and innovative preprocessing, achieves a comprehensive performance improvement over this approach.

[0041] Specifically, the objectives of the stage control method in the optical inspection system of this embodiment include the following: Improved response speed: Through innovative algorithm design, the traditional two-dimensional image processing problem is reduced to one-dimensional signal processing, and a high-efficiency frequency domain phase correlation calculation core is adopted to greatly reduce the amount of computation. The aim is to achieve high-speed real-time closed-loop control with millisecond-level latency to meet the requirements of high-speed scanning detection.

[0042] Improve computational accuracy: By introducing a frequency domain phase correlation algorithm, it becomes insensitive to light intensity fluctuations, fundamentally overcoming the shortcomings of intensity-based methods; and by combining sophisticated signal preprocessing techniques (such as Savitzky-Golay filtering) and sub-pixel-level peak positioning algorithms, it aims to achieve higher positioning accuracy and meet the detection requirements of the most advanced processes.

[0043] Enhanced anti-interference capability: By designing a combined signal preprocessing workflow that includes background noise suppression and window function spectral leakage suppression, the system aims to effectively resist signal interference caused by speckle, uneven illumination, electronic noise, etc., and ensure that the system can maintain extremely high measurement stability and reliability when facing wafers of different materials and complex working conditions.

[0044] Improve system adaptability: By establishing a configurable parameter set, the system can quickly call the preset optimal processing parameters according to different detection objects (such as silicon and glass), simplifying the operation process and improving the versatility and automation level of the equipment.

[0045] Ultimately, the overall objective of the stage control method in the optical inspection system of this embodiment is to provide a Z-axis stage control scheme with superior comprehensive performance, providing core technical support for achieving nanometer-level real-time, high-precision, and high-stability autofocus and focal plane tracking.

[0046] To achieve the above objectives, the core of the stage control method in the optical detection system of this embodiment lies in: innovatively applying the phase correlation algorithm to the sequential image processing of single-path diffraction interference fringes, efficiently compressing two-dimensional image information into one-dimensional signals, and after a series of preprocessing steps, calculating the sub-pixel-level phase change of the fringes in real time in the frequency domain, and finally converting it into a high-precision Z-axis stage control signal.

[0047] It should be emphasized that the core of the stage control method in the optical inspection system of this embodiment lies in building a vision-based nanoscale real-time focusing system. Its key technologies embody multi-level innovation, covering three aspects: algorithm principle, signal processing flow, and system architecture.

[0048] This paper presents a nanoscale displacement sensing technology based on one-dimensional projection and phase correlation. It innovatively reduces the two-dimensional interferometric image processing problem to a one-dimensional signal processing problem, and employs a frequency-domain phase correlation algorithm as the core displacement calculation engine. The classic phase correlation algorithm, originally developed for two-dimensional image registration, is creatively applied to the processing of one-dimensional interferometric signals, achieving high-speed, high-precision, and interference-resistant measurement of minute phase changes.

[0049] Regarding the combined adaptive preprocessing technique for one-dimensional interferometric signals, this system designs a dedicated preprocessing flow tailored to the characteristics of one-dimensional interferometric signals, greatly improving signal quality and providing optimal input for the phase correlation algorithm, which is key to achieving high robustness. The combined application of Savitzky-Golay filtering and windowing processing is a preprocessing scheme specifically designed for the characteristics of one-dimensional interferometric signals, solving the two core interference sources affecting accuracy: noise and spectral leakage.

[0050] Regarding the regional adaptive contrast enhancement technique based on signal features, this method further utilizes prior information of the interferometric signal for local enhancement on the basis of preprocessing, thereby improving the signal-to-noise ratio. It achieves a leap from "global enhancement" to "local adaptive enhancement," avoiding the drawbacks of traditional methods (such as histogram equalization) that blindly amplify global noise, and specifically enhances the useful interferometric signal.

[0051] The parameterized configurable system adaptation technology decouples the core parameters of the algorithm from specific application scenarios, and enables rapid system adaptation through software configuration, thereby improving versatility and ease of use.

[0052] Furthermore, the stage control method in the optical inspection system of this embodiment achieves significant results by innovatively applying the phase correlation algorithm to the one-dimensional signal processing of single-path diffraction interference fringes: First, by compressing two-dimensional image information into a one-dimensional signal and performing efficient frequency domain Fourier transform operations, a high-speed response with millisecond-level delay is achieved, which can meet the stringent requirements for real-time closed-loop control in high-speed scanning inspection; Second, based on frequency domain phase information processing, the system becomes insensitive to light intensity fluctuations, and combined with sub-pixel-level interpolation positioning technology, it breaks through the accuracy limitations of traditional intensity methods and achieves ultra-high measurement accuracy at the nanometer level; Third, by adopting a combined preprocessing scheme that combines Savitzky-Golay filtering with windowing processing, speckle noise and spectral leakage are effectively suppressed, enabling the system to exhibit excellent anti-interference capabilities and stability when facing wafers of different materials and complex working conditions; Finally, by using a preset parameter set, one-click switching between different processing modes is achieved, significantly improving the system's adaptability to different process materials and ease of operation. This enables the control method of the worktable in the optical inspection system of this embodiment to achieve a breakthrough improvement in three aspects: response speed, calculation accuracy and anti-interference capability, providing a reliable technical means to solve the problem of nanoscale real-time focusing and tracking in the field of high-precision manufacturing.

[0053] In summary, the stage control method in the optical inspection system of this embodiment has the following significant advantages: Fast response speed: It reduces two-dimensional image processing to one-dimensional signal processing and combines it with efficient algorithms such as FFT / IFFT, which greatly reduces the amount of computation and meets the real-time control requirements of high speed and high frame rate.

[0054] High computational accuracy: It adopts a phase correlation algorithm in the frequency domain, which is not sensitive to light intensity fluctuations, and through sub-pixel level peak positioning technology, it ultimately achieves nanometer-level displacement measurement accuracy.

[0055] Strong anti-interference capability: The innovative preprocessing process (Savitzky-Golay filtering and windowing) effectively suppresses speckle noise and spectral leakage, enabling the system to maintain extremely high stability and robustness under complex operating conditions.

[0056] Good adaptability: By pre-loading parameter sets for different materials, it can quickly adapt to wafer surfaces with different reflective properties, such as silicon and glass.

[0057] This embodiment also provides a computer program product, a computer-readable storage medium, and an optical inspection system. Figure 3 This is an overall block diagram of an optical detection system 400 according to an embodiment of the present invention. Figure 4 This is a schematic diagram of a computer program product 500 according to an embodiment of the present invention. Figure 5 This is a schematic diagram of a computer-readable storage medium 300 according to an embodiment of the present invention. Figure 6 This is a schematic diagram of an optical detection system 400 according to an embodiment of the present invention.

[0058] like Figure 3 As shown, the core components of the optical inspection system 400 include: an optical imaging module 210, a signal processing module 220, and a motion control module 230. The optical imaging module 210 generates and acquires interference fringe patterns formed by single-path diffracted light on the surface of the wafer under test, and includes a light source, a beam splitter, and a high dynamic range scientific-grade CCD or CMOS camera. The signal processing module 220 may include an image acquisition card and a processor 410, used to execute the core algorithm flow described in the above embodiments. The motion control module 230 may include a precision Z-axis stage and its driver, used to perform focusing and tracking actions.

[0059] Computer program product 500 includes computer program 310, which, when executed by processor 410, implements the stage control method in any of the aforementioned optical inspection systems. Computer-readable storage medium 300 stores the aforementioned computer program 310, which, when executed by processor 410, implements the stage control method in any of the aforementioned optical inspection systems. Optical inspection system 400 may include memory 420, processor 410, and computer program 310 stored in memory 420 and running on processor 410, and when processor 410 executes computer program 310, it implements the stage control method in any of the aforementioned optical inspection systems.

[0060] The computer program 310 used to perform the operations of the present invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, status setting data, integrated circuit configuration data, or source code or object code written in any combination of one or more programming languages ​​and procedural programming languages.

[0061] Computer program 310 may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0062] In some embodiments, in order to perform aspects of the present invention, electronic circuits including, for example, programmable logic circuits, field-programmable gate arrays (FPGAs) or programmable logic arrays (PLAs) can execute computer-readable program instructions to personalize the electronic circuits by utilizing state information of computer-readable program instructions.

[0063] For the purposes of this embodiment, computer program product 500 is a related product containing computer program 310. For the purposes of this embodiment, computer-readable storage medium 300 is a tangible device capable of holding and storing computer program 310, and can be any device capable of containing, storing, communicating, propagating or transmitting computer program 310 for use by or in conjunction with an instruction execution system, apparatus or device.

[0064] More specific examples (a non-exhaustive list) of computer-readable storage media 300 include the following: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable optical disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, and any suitable combination of the foregoing.

[0065] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any machine-readable storage medium for use by, or in conjunction with, instruction execution systems, apparatuses or devices (such as computer-based systems, processor-based systems or other systems that can fetch and execute instructions from, or instruction execution systems, apparatuses or devices).

[0066] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system.

[0067] The optical inspection system 400 may include a computer device, such as a server, desktop computer, laptop computer, tablet computer, or smartphone. In some examples, the computer device may be a cloud computing node. The computer device can be described in the general context of computer system executable instructions (such as program modules) executed by the computer system. Typically, program modules may include routines, programs, object programs, components, logic, data structures, etc., that perform specific tasks or implement specific abstract data types. The computer device may be implemented in a distributed cloud computing environment where tasks are performed by remote processing devices linked through a communication network. In a distributed cloud computing environment, program modules may reside on local or remote computing system storage media, including storage devices.

[0068] The computer device may include a processor 410 adapted to execute stored instructions and a memory 420 that provides temporary storage space for the operation of the instructions during operation. The processor 410 may be a single-core processor, a multi-core processor, a computing cluster, or any other configuration. The memory 420 may include random access memory (RAM), read-only memory, flash memory, or any other suitable storage system.

[0069] The processor 410 can be connected via a system interconnect (e.g., PCI, PCI-Express, etc.) to an I / O interface (input / output interface) suitable for connecting a computer device to one or more I / O devices (input / output devices). I / O devices may include, for example, a keyboard and indicating devices, where indicating devices may include a touchpad or touchscreen, etc. I / O devices may be built into the computer device or may be external devices connected to the computing device.

[0070] The processor 410 can also be linked via a system interconnect to a display interface suitable for connecting computer devices to a display device. The display device may include a display screen that is a built-in component of the computer device. The display device may also include a computer monitor, television, or projector, etc., externally connected to the computer device. Furthermore, a network interface controller (NIC) may be adapted to connect computer devices to a network via a system interconnect. In some embodiments, the NIC may use any suitable interface or protocol (such as an Internet Minicomputer System Interface) to transmit data. The network may be a cellular network, a radio network, a wide area network (WAN), a local area network (LAN), or the Internet, etc. Remote devices can connect to the computing device via the network.

[0071] Therefore, those skilled in the art should recognize that although numerous exemplary embodiments of the present invention have been shown and described in detail herein, many other variations or modifications conforming to the principles of the present invention can be directly determined or derived from the disclosure of the present invention without departing from the spirit and scope of the invention. Thus, the scope of the present invention should be understood and construed as covering all such other variations or modifications.

Claims

1. A method for controlling a stage in an optical inspection system, comprising: Real-time interference fringe images formed on the surface of the sample under test by a single-path diffraction beam are acquired. The real-time interference fringe image is reduced from a two-dimensional image to a one-dimensional real-time signal; The phase correlation between the one-dimensional real-time signal and the pre-acquired one-dimensional reference signal is calculated to obtain the phase change of the one-dimensional real-time signal relative to the one-dimensional reference signal. as well as The phase change is converted into the axial displacement of the worktable, and the worktable is controlled to move along the lifting direction based on the axial displacement to achieve real-time focusing and tracking.

2. The method according to claim 1, wherein the step of pre-acquiring the one-dimensional reference signal comprises: The surface of the sample under test is adjusted to the optimal focal plane, and a reference interference fringe image formed on the surface of the sample under test by the single-path diffraction light is acquired. The reference interference fringe image is projected along the direction of maximum phase gradient of the interference fringes to generate an initial reference signal; as well as Peak detection and subpixel-level peak-valley localization are performed on the initial reference signal, and adaptive contrast enhancement is used to obtain the enhanced one-dimensional reference signal.

3. The method according to claim 1, wherein after the step of reducing the real-time interference fringe image from a two-dimensional image to a one-dimensional real-time signal, it further comprises: The one-dimensional real-time signal is subjected to background correction and denoising processing. High-frequency speckle noise is smoothed by local window polynomial fitting, low-frequency background drift is eliminated, and the peak shape characteristics of the interference fringes are preserved.

4. The method according to claim 3, wherein after the step of performing background correction and denoising processing on the one-dimensional real-time signal, it further comprises: A window function is applied to the one-dimensional real-time signal after background correction and denoising to suppress spectral leakage caused by truncation during the fast Fourier transform of the one-dimensional real-time signal.

5. The method according to claim 1, wherein the step of performing phase correlation calculation between the one-dimensional real-time signal and the pre-acquired one-dimensional reference signal includes: Fast Fourier transform is performed on the one-dimensional reference signal and the one-dimensional real-time signal respectively to calculate the cross-power spectrum, and then inverse Fourier transform is performed on the cross-power spectrum to obtain the phase correlation function.

6. The method of claim 5, further comprising, after obtaining the phase correlation function: Parabolic fitting is performed on the main peak value of the phase correlation function to obtain a one-dimensional signal shift, which is used to determine the phase change.

7. The method of claim 1, wherein the step of converting the phase change into the axial displacement of the worktable comprises: The phase change is converted into the axial displacement according to a preset phase-displacement mathematical model. The phase-displacement mathematical model is established based on the optical parameters of the optical detection system and is used to characterize the correspondence between the phase change and the axial displacement.

8. The method according to claim 1, further comprising, before the step of acquiring a real-time interference fringe image formed by a single-path diffracted light on the surface of the sample under test: The corresponding preset algorithm parameter set is loaded according to the type of the sample to be tested to adapt to surfaces with different reflectivities. The preset algorithm parameter set includes window function type and filtering parameters.

9. The method according to claim 1, wherein, The projection direction used to reduce the real-time interference fringe image from the two-dimensional image to the one-dimensional real-time signal is consistent with the projection direction used to generate the one-dimensional reference signal.

10. An optical inspection system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the control method for the stage in the optical inspection system according to any one of claims 1 to 9.

Images