A synchronizable imaging astigmatism differential confocal alignment system and method

Abstract

This invention relates to the field of precision alignment technology, specifically to an astigmatic differential confocal alignment system and method capable of synchronous imaging. The astigmatic differential confocal alignment system includes: an astigmatic differential confocal optical path system and an imaging optical path system sharing the same objective lens; the astigmatic differential confocal optical path system includes a laser, a collimating and expanding beam system, a polarizing beam splitter cube, a quarter-wave plate, a long-pass filter, a short-pass dichroic mirror, an objective lens, a non-polarizing beam splitter cube, a plano-convex cylindrical lens, and two four-quadrant photodetectors connected in sequence; the imaging optical path system includes a visible light point source, a beam splitter, a short-pass dichroic mirror, an objective lens, a sleeve lens, and a camera connected in sequence; the two optical path systems are coaxial and synchronized. When the defocus error signal FES crosses zero, the camera synchronously acquires a clear morphological image of the surface of the sample under test, achieving synchronization of alignment and imaging.

Description

Technical Field

[0001] This invention relates to the field of precision alignment technology, and specifically to an astigmatic differential confocal alignment system and method capable of synchronous imaging. Background Technology

[0002] Precision alignment technology is a core component of modern precision manufacturing and micro / nano fabrication, especially in wafer fabrication, mask alignment, and defect detection of third-generation semiconductor materials. The accuracy and efficiency of alignment directly determine the performance and yield of devices. Among various alignment methods, optical confocal technology has become mainstream due to its high lateral resolution and tomographic capabilities. Specifically, astigmatic differential confocal technology generates astigmatism by introducing cylindrical mirrors and uses the differential signal (FES) caused by the change in the spot shape of two photodetectors before and after focus to determine the defocus state. Due to its advantages such as high sensitivity, large measurement range, and relatively simple structure, it has gradually been widely used in wafer topography measurement and autofocus systems. This technology, through precise locking of the FES zero point, can achieve high-precision positioning of the measured surface, becoming one of the important means of high-precision displacement detection in this field.

[0003] The shortcomings and deficiencies of existing technologies are as follows: First, they are limited in functionality and lack in-situ synchronous observation capabilities. Traditional astigmatic differential confocal technology is essentially a precision ranging or focusing method based on light intensity response. Its core lies in determining defocus displacement by analyzing the magnitude and sign of the FES signal. However, during the focusing process, the system cannot simultaneously acquire a two-dimensional topographic image of the measured surface while searching for the focal plane and zeroing the FES. The operator or host computer must wait for focusing to complete before switching to or starting another imaging system for observation. This "focus first, then image" serial working mode is not only cumbersome, but the physical separation between the focusing reference point and the imaging observation point also easily introduces coordinate system transformation errors, making it difficult to guarantee that the "aligned point" is indeed the "target point." Second, the single light intensity signal is easily affected by sample characteristics and lacks multi-dimensional criteria support. Existing astigmatic differential confocal technology relies solely on the light intensity difference signal in the FES dimension for defocus determination. In practical engineering applications, especially when processing transparent or semi-transparent wafers such as SiC, the FES signal is easily affected by various factors. For example, reflected light from the back of the wafer, scattering from defects within the material, or thin film structures of different materials on the surface can all be superimposed on the reflected light field, causing distortion or zero-point drift in the FES curve. Lacking a synchronous visual image originating from the same objective optical path as a reference, the system struggles to distinguish whether changes in FES stem from genuine surface defocus or spurious signals caused by stray light within the material. This one-dimensional criterion mechanism challenges the system's robustness and alignment confidence when dealing with complex samples. Finally, the separate setup of the two functional modules results in low system integration and poor synchronization stability: To meet the dual requirements of high-precision focusing and visual observation, a common approach in existing technologies is to combine an astigmatic differential confocal module with a separate machine vision imaging module. These two modules typically have their own independent light sources, detectors, and optical path structures, with only rough mechanical alignment in spatial position. This discrete design leads to a large overall system size, a bulky optical path structure, and requires complex mechanical calibration mechanisms to ensure the relative positional relationship between the optical axes of the two modules. Even with meticulous initial calibration, when two modules share different objectives or optical paths, it is difficult to ensure that their respective focusing reference planes are strictly consistent. During long-term operation or when the ambient temperature fluctuates, the mechanical relative positions of the two modules are prone to drift, resulting in misalignment between the "focus measurement point" and the "visual observation point," making it difficult to achieve truly stable and synchronized "what you see is what you measure." Summary of the Invention

[0004] The purpose of this invention is to provide an astigmatic differential confocal alignment system capable of synchronous imaging, in order to solve the problems of signal detection and image observation being separated, system being discrete, and optical axis being prone to drift in the prior art.

[0005] To address the aforementioned problems, the technical solution of the astigmatic differential confocal alignment system capable of synchronous imaging according to the present invention is as follows: A synchronous imaging astigmatic differential confocal alignment system includes: an astigmatic differential confocal optical path system and an imaging optical path system that share the same short-pass dichroic mirror, objective lens and sample under test; The astigmatic differential confocal optical path system includes a laser, a collimating and expanding system, a polarizing beam splitter cube, a zero-order quarter-wave plate, a long-pass filter, a short-pass dichroic mirror, an objective lens, and a non-polarizing beam splitter cube, a plano-convex cylindrical lens, and two four-quadrant photodetectors, which are connected in sequence to the original optical path of the objective lens after the polarizing beam splitter cube. The imaging optical path system includes a visible light point source, a 50 / 50 flat beam splitter, a short-pass dichroic mirror, an objective lens, and a sleeve lens and a CMOS camera that are sequentially connected to the objective lens via the original return optical path after the 50 / 50 flat beam splitter. The astigmatic differential confocal detection optical path system is coaxial and synchronized with the imaging optical path system. The output signals of the two four-quadrant photodetectors are used to calculate the defocus error signal FES. When FES crosses zero, it is determined that the surface of the sample under test is located on the focal plane of the objective lens. At the same time, the CMOS camera synchronously acquires a clear topographic image of the surface of the sample under test, realizing the synchronization of alignment and imaging.

[0006] Optionally, the optical path corresponding to the astigmatic differential confocal optical path system specifically includes: The laser beam output from the laser is collimated and expanded by a beam-splitting system. After the S-beam is filtered by a polarization-splitting cubic filter, the P-beam is transmitted to a zero-order quarter-wave plate. At this point, the linearly polarized light is converted to circularly polarized light. Then, it is reflected by a long-pass filter and a short-pass dichroic mirror and enters the objective lens. After being focused by the objective lens and reflected from the surface of the sample being measured, it returns along the original path. It is then reflected again by a short-pass dichroic mirror, a long-pass filter, and a zero-order quarter-wave plate. The circularly polarized light is converted to linearly polarized light, and the P-beam is converted to the S-beam. After being reflected by a polarization-splitting cubic filter and split into two beams of equal energy by a 50 / 50 non-polarization-splitting cubic filter, the beams are focused by plano-convex cylindrical lenses and irradiated onto two four-quadrant photodetectors. The defocusing error signal FES is calculated based on the output signals of the two four-quadrant photodetectors.

[0007] Optionally, the optical path corresponding to the imaging optical path system specifically includes: Visible light emitted from a visible light point source is split by a 50 / 50 flat beam splitter. The transmitted light is physically blocked, and the reflected light enters a short-pass dichroic mirror. The low-wavelength visible light is transmitted into the objective lens, focused by the objective lens onto the surface of the sample under test, reflected, and then returns along the original path. After passing through the short-pass dichroic mirror and the 50 / 50 flat beam splitter in sequence, it is focused by the sleeve lens onto the CMOS camera.

[0008] Optionally, the defocus error signal FES is calculated as follows: FES=(A1+A3-A2-A4) / (A1+A2+A3+A4)-(B1+B3-B2-B4) / (B1+B2+B3+B4) Among them, the output quantities of four-quadrant photodetector A are A1, A2, A3, and A4, and the output quantities of four-quadrant photodetector B are B1, B2, B3, and B4.

[0009] Optionally, in the astigmatic differential confocal optical path system, an aperture is provided between the collimating beam expander system and the polarization beam splitter cube, between the polarization beam splitter cube and the 50 / 50 non-polarization beam splitter cube, and between the short-pass dichroic mirror and the objective lens to block stray light with a spot diameter greater than 14 mm.

[0010] Optionally, the astigmatic differential confocal alignment system capable of synchronous imaging further includes an astigmatic differential confocal optical path adjustment system, comprising: Using the clear image acquired by the CMOS camera as a reference, the positive focus position of the sample surface under test is calibrated; then the output signals of the two diagonal quadrants in the four-quadrant detector are adjusted to make the output signal values ​​of the two diagonal quadrants in each group equal.

[0011] Optionally, the synchronous imaging astigmatic differential confocal alignment system further includes a spot morphology determination system for a four-quadrant photodetector, comprising: When the surface of the sample to be tested is located at the focal plane of the objective lens, the spot morphology on the two four-quadrant photodetectors is the same. When the surface of the sample to be tested is in a defocused state, the spot morphology on the two four-quadrant photodetectors changes in opposite directions, tending to be linear and circular, respectively.

[0012] Optionally, the two four-quadrant photodetectors are respectively located 2 mm in front of and behind the focal plane of the plano-convex cylindrical lens, and rotated 45 degrees along the optical axis.

[0013] Optionally, the long-pass filter has a cutoff wavelength of 600nm and a transmittance of >90% for light with wavelength ≥600nm, and is used to block visible light with wavelength <600nm; the short-pass dichroic mirror divides the light beam into transmitted light and reflected light according to the wavelength range, wherein the transmitted band and the reflected band are separated by a cutoff wavelength of 567nm, and the transmittance of >90% for visible light with wavelength <567nm and the reflectance of >90% for light with wavelength >567nm.

[0014] The technical solution of the astigmatic differential confocal alignment method for synchronous imaging proposed in this invention is as follows: A synchronous imaging astigmatic differential confocal alignment method includes the following steps: Construct an astigmatic differential confocal detection optical path and a visible light imaging optical path that share the same objective lens; The astigmatic differential confocal detection optical path uses a laser emitted from a laser source. The laser beam is sequentially reflected by a collimating and expanding system, a polarizing beam splitter cube, a zero-order quarter-wave plate, a long-pass filter, and a short-pass dichroic mirror before entering the objective lens. After being focused by the objective lens and reflected from the surface of the sample, the beam returns along the same path and is again split into two beams of equal energy by a short-pass dichroic mirror, a long-pass filter, a zero-order quarter-wave plate, a polarizing beam splitter cube, and a 50 / 50 unpolarized beam splitter cube. These beams are then focused by plano-convex cylindrical lenses and irradiated by two four-quadrant photodetectors. The defocusing error signal FES is calculated based on the output signals of the two four-quadrant photodetectors. Simultaneously, visible light emitted from a visible light point source enters the objective lens through a 50 / 50 flat beam splitter and a short-pass dichroic mirror. After being focused by the objective lens and reflected from the surface of the sample, the beam returns along the same path, sequentially passing through a short-pass dichroic mirror and a 50 / 50 flat beam splitter before being focused by a sleeve lens onto a CMOS camera. By coaxially synchronizing the astigmatic differential confocal probe optical path and the imaging optical path, while the FES signal crosses zero to determine that the surface of the sample under test is located on the focal plane of the objective lens, the CMOS camera simultaneously acquires a clear morphological image of the surface of the sample under test, thus achieving real-time unification of alignment and imaging.

[0015] Compared with the prior art, this application has the following beneficial effects: The astigmatic differential confocal alignment system and method provided in this application integrates independent astigmatic differential confocal detection and visible light imaging optical paths behind the same objective lens. Specifically, it integrates an astigmatic differential confocal detection optical path for high-precision displacement detection and a visible light imaging optical path for morphology observation. This allows the system to perform high-precision defocus search based on the FES curve while simultaneously observing the actual morphology of the sample surface in real time through a CMOS camera. By establishing an inherent optical correspondence between the FES zero point and the CMOS clear imaging plane, during stage scanning, when the FES curve crosses zero, the system automatically determines that the current objective lens focal plane coincides with the sample surface. At this point, the image acquired by the CMOS is the clearest surface image. In other words, when the FES signal crosses zero, the image captured by the CMOS corresponds to the clearest focus state, achieving spatiotemporal unification of "signal alignment" and "visual alignment," i.e., physical synchronization of the defocus error signal FES returning to zero and the clear image capture time. This design completely eliminates the sequential waiting time between focusing and imaging, allowing operators to intuitively and in real-time monitor the entire alignment process. It overcomes the functional separation between signal detection and visual observation, achieving a real-time visual closed loop for the alignment process and significantly improving operational efficiency and ease of use. Simultaneously, this application introduces a dual-optical-path independent parallel architecture based on the same objective lens, enabling the simultaneous acquisition and mutual verification of multi-dimensional information. This invention combines a laser-based astigmatic differential confocal detection system with a visible light point source imaging system, using a short-pass dichroic mirror for optical beam combining. This allows two beams with different wavelengths and functions to share the same detection objective lens, focusing on the same sample surface before returning to their respective detectors. This ensures high-sensitivity detection of the FES signal and enables visual observation of the alignment process. This parallel architecture allows the system to simultaneously acquire two dimensions of information: a high-sensitivity FES signal reflecting axial defocus and a two-dimensional optical image reflecting surface morphology. During the alignment process of transparent or semi-transparent materials such as SiC wafers, when the FES signal is distorted due to back reflection or internal defect interference, the operator can simultaneously refer to the CMOS image for comprehensive judgment. This effectively avoids misfocusing that may be caused by a single light intensity signal, significantly improving the confidence of the alignment results and the system's anti-interference capability. Furthermore, this invention designs two optical paths into a single astigmatic differential confocal alignment system capable of synchronous imaging. Both share a single objective lens, and the optical paths are combined and split using optical elements, rather than relying on mechanical calibration to maintain relative position. This optical path fusion design fundamentally eliminates the relative drift between the two optical paths, ensuring the long-term stable and reliable correspondence between the FES numerical information and the CMOS observation field of view.Meanwhile, the integrated design significantly reduces system size and assembly and debugging difficulty, laying a solid foundation for the miniaturization and engineering application of the equipment. It achieves a high degree of system architecture integration and fundamentally guarantees optical axis stability, solving the drift problem of discrete modules. Finally, while realizing synchronous imaging, this invention fully retains the core performance indicators of astigmatic differential confocal technology. Experimental verification shows that the system achieves nanometer-level high sensitivity in axial displacement detection, meeting the stringent requirements of sub-micron precision in SiC wafer dicing, alignment, and other processes. Simultaneously, its effective working range reaches the centimeter level, covering the entire process from coarse positioning to fine alignment without frequent switching of detection modes. The organic combination of high sensitivity and large working range, coupled with the visual guidance capability of synchronous imaging, makes this invention not only suitable for precision measurements in laboratory environments but also directly supports high-efficiency, high-reliability production applications in industrial settings such as semiconductor packaging and micro / nano fabrication. It possesses the dual advantages of high-sensitivity detection and wide-range operation, providing a powerful tool for precision manufacturing.

[0016] This invention effectively solves the problems of signal detection and image observation being separated, system being discrete, and optical axis being prone to drift in the prior art. While maintaining nanometer-level high sensitivity and millimeter-level large range, it significantly improves the alignment efficiency, ease of operation, and system robustness in precision manufacturing. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the astigmatic differential confocal alignment system capable of synchronous imaging according to this application. Figure 2 This is a schematic diagram showing the placement of the four-quadrant photodetector in the astigmatic differential confocal alignment system capable of synchronous imaging according to this application. Figure 3 The defocus error signal FES variation curve is calculated from the output signals of the two four-quadrant photodetectors in the synchronous imaging astigmatic differential confocal alignment system of this application. Figure 4 This is a CMOS observation image taken at the zero crossing of the FES in the synchronous imaging astigmatic differential confocal alignment system of this application.

[0018] In the diagram, 1. Laser, 2. Collimation and beam expansion system, 3. Aperture, 4. Polarizing beam splitter cube, 5. Zero-order quarter-wave plate, 6. Long-pass filter, 7. Short-pass dichroic mirror, 8. Objective lens, 9. Wafer, 10. 50 / 50 unpolarized beam splitter cube, 11. Plano-convex cylindrical lens, 12. Four-quadrant photodetector, 13. Visible light point source, 14. 50 / 50 planar beam splitter, 15. Sleeve lens, 16. CMOS camera. Detailed Implementation

[0019] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0020] The following description, with reference to the accompanying drawings, describes an astigmatic differential confocal alignment system and method capable of synchronous imaging according to embodiments of this application.

[0021] The following is combined Figure 1 This application provides a detailed description of the astigmatic differential confocal alignment system capable of synchronous imaging. Figure 1 This is a schematic diagram of the astigmatic differential confocal alignment system capable of synchronous imaging according to an embodiment of this application.

[0022] In this embodiment, as Figure 1 As shown, the astigmatic differential confocal alignment system capable of synchronous imaging includes an astigmatic differential confocal optical path system and an imaging optical path system that share the same short-pass dichroic mirror 7, objective lens 8, and the sample under test. The astigmatic differential confocal optical path system includes, in sequence, a laser 1, a collimating and beam expanding system 2, a polarizing beam splitter 4, a zero-order quarter-wave plate 5, a long-pass filter 6, a short-pass dichroic mirror 7, and an objective lens 8, as well as a 50 / 50 non-polarizing beam splitter 10, a plano-convex cylindrical lens 11, and two four-quadrant photodetectors 12, which are sequentially connected to the original return optical path of the objective lens 8 after the polarizing beam splitter 4. Here, the sample under test is a wafer 9, whose main materials include silicon, silicon carbide, sapphire, and glass, which interact with the focused laser to achieve micro-nano processing. Laser 1: The generated laser has a wavelength of 642nm, a divergence angle (full angle) of 1mrad, an output power ≥20mW, a linear polarization state, a polarization extinction ratio ≥500:1, and an output spot size of 1mm, forming a fundamental mode Gaussian beam. Two four-quadrant photodetectors 12 are respectively positioned 2mm in front of and behind the focal face of the plano-convex cylindrical lens 11, and rotated 45 degrees along the optical axis.

[0023] Specifically, the optical path corresponding to the astigmatic differential confocal optical path system includes: the beam output from laser 1 passes through collimation and beam expansion system 2, then through polarization beam splitter cube 4 to filter S-beams, and then transmits P-beams to zero-order quarter-wave plate 5. At this time, the linearly polarized light is converted into circularly polarized light, and then reflected by long-pass filter 6 and short-pass dichroic mirror 7 into objective lens 8. After being focused by objective lens 8 and reflected by the surface of the sample under test, it returns along the original path, and passes through short-pass dichroic mirror 7, long-pass filter 6, and zero-order quarter-wave plate 5 again. The circularly polarized light is converted into linearly polarized light, and the P-beam is converted into S-beams. After being reflected by polarization beam splitter cube 4 and split into two beams of equal energy by 50 / 50 non-polarization beam splitter cube 10, they are focused by plano-convex cylindrical lens 11 and irradiated by two four-quadrant photodetectors 12. The defocus error signal FES is calculated based on the output signals of the two four-quadrant photodetectors 12.

[0024] The imaging optical path system includes a visible light point source 13, a 50 / 50 flat beam splitter 14, a short-pass dichroic mirror 7, and an objective lens 8, all connected sequentially in a optical path. A sleeve lens 15 and a CMOS camera 16 are also connected sequentially after the 50 / 50 flat beam splitter 14 via the original optical path return path of the objective lens 8. Here, the visible light point source 13 outputs white light with a uniformity of illumination ≥95%, a color temperature of 6500~7000K, an illuminance of 1700~2300Klux, and a power consumption of 5.5W. The 50 / 50 flat beam splitter 14 has a dielectric beam-splitting film coated on its surface, used in the 400-700nm range. The dielectric film performs best at a 45° incident angle, achieving a beam splitting ratio of 50:50.

[0025] Specifically, the optical path of the imaging optical path system includes: visible light emitted from the visible light point source 13 is split by a 50 / 50 flat beam splitter, the transmitted light is physically blocked, the reflected light enters the short-pass dichroic mirror 7, the low-wavelength visible light is transmitted into the objective lens 8, and after being focused by the objective lens 8 onto the surface of the sample being measured and reflected, it returns along the original path, and passes sequentially through the short-pass dichroic mirror 7 and the 50 / 50 flat beam splitter 14, and is then focused by the sleeve lens 15 onto the CMOS camera 16.

[0026] The astigmatic differential confocal detection optical path system and the imaging optical path system are coaxial and synchronized. The output signals of the two four-quadrant photodetectors 12 are used to calculate the defocus error signal FES. When FES crosses zero, it is determined that the surface of the sample under test is located on the focal plane of the objective lens 8. At the same time, the CMOS camera 16 synchronously acquires a clear topographic image of the surface of the sample under test, realizing the synchronization of alignment and imaging.

[0027] Specifically, the defocus error signal FES is calculated as follows: FES=(A1+A3-A2-A4) / (A1+A2+A3+A4)-(B1+B3-B2-B4) / (B1+B2+B3+B4) Among them, the output quantities of four-quadrant photodetector A in the two four-quadrant photodetectors 12 are A1, A2, A3, A4, and the output quantities of four-quadrant photodetector B are B1, B2, B3, B4.

[0028] In a specific embodiment of this application, the long-pass filter 6 effectively isolates specific regions in the spectrum, with a cutoff wavelength of 600nm and a transmittance of >90% for light with wavelengths ≥600nm, used to block visible light with wavelengths <600nm. The short-pass dichroic mirror 7 separates the light beam into transmitted and reflected light according to the wavelength range. The transmitted and reflected wavelengths are separated by a cutoff wavelength of 567nm, with a transmittance of >90% for visible light with wavelengths <567nm and a reflectance of >90% for light with wavelengths >567nm. Here, in the astigmatic differential confocal optical path system, the long-pass filter 6 does not block the laser; therefore, the laser passes directly through the long-pass filter 6 without change. The long-pass filter 6 is mainly used in imaging optical path systems. Since the wavelength range of the 642nm laser 1 overlaps with that of visible light, the short-pass dichroic mirror 7 is used to transmit visible light with wavelengths less than 567nm for imaging, while reflecting and blocking visible light with wavelengths greater than 567nm. The 642nm wavelength laser is only used for the final astigmatic differential confocal test. Since the short-pass dichroic mirror 7 has a reflectivity of only slightly over 90%, a small amount of visible light with wavelengths greater than 567nm is transmitted into the objective lens 8. After reflection from the wafer surface 9, this light enters the astigmatic differential confocal optical path system along with the 642nm wavelength laser. Therefore, a long-pass filter 6 is needed to intercept visible light with wavelengths less than 600nm to prevent it from affecting the output values ​​of the four-quadrant photodetector 12.

[0029] In another specific embodiment of this application, in the astigmatic differential confocal probe optical path system, the collimation and beam expansion system 2 can achieve collimation and beam expansion of the incident beam, with a beam magnification of 2x to 8x and a transmittance >97%. The polarization beam splitter cube 4 (PBS) has a dielectric layer coated on the inclined surface between the two right-angle prisms that make up this cube. This coating reflects s-light and transmits p-light, achieving beam splitting effects for beams with different polarization states. The zero-order quarter-wave plate 5 is composed of two multi-order quartz waveplates, capable of generating a λ / 4 optical path difference, realizing the mutual conversion between linearly polarized light and circularly polarized light. The objective lens 8 precisely focuses the processed beam inside or on the surface of the wafer 9; the numerical aperture NA = 0.65, and the magnification is 50X. The 50 / 50 unpolarized beam splitter cube 10 (BS) is composed of two prisms, one of which has a dielectric beam splitting film coated on its inclined surface, enabling beam splitting with a 50:50 splitting ratio. Plano-convex cylindrical lens 11: Focuses the light beam in only one direction, forming a straight spot at the focal plane. Its focal length is 25mm, operating wavelength is 400nm ~ 700nm, and dimensions are H×L = 10mm × 12mm. Four-quadrant photodetector 12: A surface-mount, four-segment detector capable of outputting four current signals. It has an applicable wavelength range of 320nm ~ 1100nm, a photosensitivity of 0.72A / W, a maximum dark current of 2000pA, a cutoff frequency of 25MHz, and a junction capacitance of 10pF.

[0030] In the imaging optical path system, the sleeve lens 15 is designed to be used in conjunction with the infinity correction objective lens 8 to build an infinity correction optical system. Both sides of all lenses in the sleeve lens 15 are coated with broadband anti-reflection coatings with wavelength ranges of 350-750 nm and 700-1100 nm. The CMOS camera 16 has an effective pixel count of 3072×2048, a pixel size of 2.4μm × 2.4μm, and uses a monochrome sensor.

[0031] In one possible implementation of this application embodiment, in the astigmatic differential confocal optical path system, an aperture stop 3 is provided between the collimating beam expander system 2 and the polarization beam splitter 4, the reflection of the polarization beam splitter 4 and the 50 / 50 unpolarized beam splitter 10, the reflection of the short-pass dichroic mirror 7 and the objective lens 8, to block stray light with a spot diameter greater than 14 mm. Multiple aperture stops 3 are placed in the entire synchronous imaging astigmatic differential confocal optical system, avoiding the influence of stray light in the system on the numerical values ​​of the four-quadrant photodetector 12, and effectively reducing testing errors.

[0032] In another possible implementation of this application embodiment, the astigmatic differential confocal alignment system capable of synchronous imaging further includes an astigmatic differential confocal optical path debugging system, including: using the clear image acquired by the CMOS camera 16 as a reference, completing the positive focus position calibration of the surface of the sample under test; subsequently, adjusting the output signals of the two sets of diagonal quadrants in the four-quadrant photodetectors 12 respectively, so that the output signal values ​​of the two diagonal quadrants in each set are equal. During the astigmatic differential confocal optical path debugging process, the positive focus position of the wafer 9 surface should first be calibrated according to the clear image presented by the CMOS camera 16, and the output signal values ​​in quadrants 1 and 3, and quadrants 2 and 4 of the two four-quadrant photodetectors 12 should be adjusted to be equal, otherwise the FES curve will be shifted, but the trend of the curve will not be changed.

[0033] In another possible implementation of this application embodiment, the astigmatic differential confocal alignment system capable of synchronous imaging further includes a spot morphology judgment system for the four-quadrant photodetector 12, comprising: when the surface of the sample under test is located at the focal plane of the objective lens 8, the spot morphology on the two four-quadrant photodetectors 12 is the same; when the surface of the sample under test is in a defocused state, the spot morphology on the two four-quadrant photodetectors 12 changes in opposite directions, respectively approaching a linear shape and a circle. Specifically, the focal length of the plano-convex cylindrical lens 11 is 25mm, and the two four-quadrant photodetectors 12 need to be placed 2mm in front of and behind the focal plane, respectively. Assuming that the surface of the wafer 9 is at the positive focal position of the objective lens 8, the beam returned by the objective lens 8 must be a collimated parallel beam, and at this time, the irradiated spot morphology on the two four-quadrant photodetectors 12 is the same elliptical shape. When the surface of wafer 9 is in a defocused state, the spot morphology on the two quadrant photodetectors 12 will inevitably show opposite changes, with one being closer to a linear shape and the other closer to a circle. The resulting change in the FES value is used to determine the direction and distance of defocus.

[0034] Using the aforementioned synchronous imaging astigmatic differential confocal alignment system, the present invention also provides a synchronous imaging astigmatic differential confocal alignment method. This synchronous imaging astigmatic differential confocal alignment method includes the following steps: Construct an astigmatic differential confocal detection optical path and a visible light imaging optical path that share the same objective lens 8; The astigmatic differential confocal detection optical path uses a laser emitted from laser 1, which is sequentially reflected by collimation and beam expansion system 2, polarization beam splitter 4, zero-order quarter-wave plate 5, long-pass filter 6, and short-pass dichroic mirror 7 before entering objective lens 8. After being focused by objective lens 8 onto the surface of the sample under test, the laser returns along the original path and is again split into two beams of equal energy by short-pass dichroic mirror 7, long-pass filter 6, zero-order quarter-wave plate 5, polarization beam splitter 4, and 50 / 50 non-polarization beam splitter 10. These beams are then focused by plano-convex cylindrical lens 11 and irradiated by two four-quadrant photodetectors 12. The defocusing error signal FES is calculated based on the output signals of the two four-quadrant photodetectors 12. Simultaneously, visible light emitted from visible light point source 13 is reflected by a 50 / 50 non-polarization beam splitter 10. The flat beam splitter 14 and the short-pass dichroic mirror 7 enter the objective lens 8. After being focused by the objective lens 8 and reflected by the surface of the sample under test, the beam splitter returns along the original path and passes through the short-pass dichroic mirror 7 and the 50 / 50 flat beam splitter 14 in sequence. Finally, it is focused by the sleeve lens 15 onto the CMOS camera 16. By coaxially synchronizing the astigmatic differential confocal probe optical path and the imaging optical path, while the FES signal zero-crossing determines that the surface of the sample under test is located on the focal plane of objective lens 8, the CMOS camera 16 simultaneously acquires a clear topographic image of the surface of the sample under test, thus achieving real-time unification of alignment and imaging.

[0035] Based on the aforementioned astigmatic differential confocal alignment system capable of simultaneous imaging, the optical path in the astigmatic differential confocal alignment method comprises two parts: an astigmatic differential confocal optical path and an imaging optical path. The specific technical solution is as follows: First, a detailed explanation of the astigmatic differential confocal optical path. The linearly polarized Gaussian beam output from the 642nm laser 1, after passing through the collimation and beam expanding system 2, becomes a collimated parallel beam with a spot diameter of 4mm. After passing through the polarization beam splitter 4, the S-beam is reflected and blocked, while the P-beam is transmitted to the zero-order quarter-wave plate 5. At this point, the linearly polarized light becomes circularly polarized light. Then, after passing through the long-pass filter 6, it is reflected by the short-pass dichroic mirror 7 and enters the objective lens 8. The circularly polarized light transmitted through the objective lens 8 is reflected by the surface of the wafer 9, returns along the original path, is reflected again by the short-pass dichroic mirror 7, enters the long-pass filter 6, and then passes through the zero-order quarter-wave plate 5 again. At this point, the circularly polarized light becomes linearly polarized light again, and the P-beam becomes the S-beam. Subsequently, the S-beam is reflected by polarized beam splitter cube 4 and passes through 50 / 50 unpolarized beam splitter cube 10, where it is split into two beams of equal energy. These beams are then incident on a plano-convex cylindrical lens 11 with a focal length of 25 mm, and subsequently irradiate two four-quadrant photodetectors 12 located 2 mm in front of and behind the focal length of the plano-convex cylindrical lens 11, respectively. The spatial arrangement of the two four-quadrant photodetectors 12 is illustrated in the diagram below. Figure 2As shown, the surfaces of the two four-quadrant photodetectors 12 are rotated 45 degrees along the optical axis, so that the elliptical light spots symmetrically irradiate the four quadrants. Assuming that the outputs of the two four-quadrant photodetectors A and B are A1~A4 and B1~B4 respectively, the defocus error signal FES of the astigmatic differential confocal system is FES=(A1+A3-A2-A4) / (A1+A2+A3+A4)-(B1+B3-B2-B4) / (B1+B2+B3+B4). When the FES change curve passes through the zero point, that is, the surface of the wafer 9 being processed is located at the focal plane of the objective lens 8.

[0036] Second, a detailed description of the imaging optical path. Visible light emitted from the visible light point source 13 is split into two beams of equal energy by the 50 / 50 flat beam splitter 14. One beam is transmitted and blocked, while the other is reflected and enters the short-pass dichroic mirror 7. Subsequently, visible light with wavelengths higher than 567nm is reflected and blocked, while visible light with wavelengths lower than 567nm is transmitted and enters the objective lens 8. After passing through the objective lens 8, the visible light is reflected by the surface of the wafer 9 and returns along the original path. It then passes sequentially through the short-pass dichroic mirror 7 and the 50 / 50 flat beam splitter 14, and is focused by the sleeve lens 15 onto the CMOS camera 16, where a clear image of the wafer 9 surface is displayed by the computer.

[0037] Since the same objective lens 8 is used, the distance information between the objective lens 8 and the surface of the wafer 9 can be synchronously fed back to both the defocus error signal FES detection and CMOS imaging, realizing the physical synchronization of the two signals and providing convenience for real-time observation for precision measurement personnel.

[0038] The astigmatic differential confocal alignment system and method capable of synchronous imaging of the present invention were tested experimentally. The mounting positions of the optical components and... Figure 1 Completely identical. The wafer 9 is moved in a uniform linear motion using an electric displacement platform, its trajectory moving from away from objective lens 8 to near objective lens 8, covering the focal plane of objective lens 8. During this process, the eight output current signals of the two four-quadrant photodetectors 12 are continuously acquired, converted into voltage signals, and the FES value of the astigmatic differential confocal system is calculated synchronously, and the relevant variation curves are plotted as follows: Figure 3 As shown. By Figure 3It can be seen that when moving away from objective lens 8, the FES value remains stable. As the wafer moves closer to objective lens 8, the FES value gradually decreases, then rises sharply, reaches its maximum value, and then gradually decreases again, tending towards a stable value. This trend is due to the change in spot morphology caused by the change in defocus distance. When the FES values ​​are positive and negative, it indicates that the surface of wafer 9 is in a negative or positive defocus state, respectively. It is particularly important to note that this conclusion is related to the order in which the signals of the two quadrant photodetectors 12 are subtracted, which can be agreed upon as needed. When the FES signal rises sharply, the change in spot morphology is most significant because the wafer 9 has moved past the focal plane of objective lens 8. At this time, the spot morphology on the two quadrant photodetectors 12 changes in opposite directions. When the FES signal crosses zero, it proves that the spot morphology of the two quadrant photodetectors 12 is the same, and the surface of the wafer 9 being measured is located at the focal plane of objective lens 8, which is the positive focus position. At this time, the image viewed by the CMOS camera 16 is as follows. Figure 4 As shown, by Figure 4 As can be seen, the surface morphology of wafer 9 can be clearly presented, achieving the spatiotemporal unity of "signal alignment" and "visual alignment". The FES signal tends to a stable value at the beginning or end of the motion. This is because the defocus distance is large, the two four-quadrant photodetectors 12 have the same astigmatic characteristics, and the similar spot morphology leads to a small change in the FES value.

[0039] Traditional astigmatic differential confocal technology only outputs an FES electrical signal, leaving the operator in a "blind" state while searching for the focal plane. The operator must wait for focusing to complete before observing the image again, resulting in low efficiency and susceptibility to coordinate deviations. The astigmatic differential confocal alignment system and method of this invention, capable of simultaneous imaging, possesses nanometer-level displacement sensitivity and centimeter-level effective working range, showing broad application prospects in precision manufacturing and micro / nano fabrication. For transparent or semi-transparent materials like SiC, the "signal-image" dual criterion provided by this invention significantly improves the confidence level of the alignment results. Simultaneously, the integrated design of coaxial fusion of dual optical paths ensures the optical axis stability of the system during long-term operation in industrial settings, reducing the frequency and cost of maintenance and calibration. This invention is not only a theoretical innovation but also a technical solution that has been experimentally verified and possesses direct engineering transfer value. It can provide core alignment and observation methods for equipment development in fields such as semiconductor packaging, advanced displays, and precision measurement, and is expected to play a significant role in high-end optical measurement and intelligent manufacturing.

[0040] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. The scope of patent protection of the present invention shall be determined by the claims. Similarly, any equivalent structural changes made based on the description and drawings of the present invention shall also be included within the scope of protection of the present invention.

Claims

1. A synchronous imaging astigmatic differential confocal alignment system, characterized in that, include: They share the same short-pass dichroic mirror (7), objective lens (8), astigmatic differential confocal optical path system, and imaging optical path system of the sample under test; The astigmatic differential confocal optical path system includes a laser (1), a collimating beam expander (2), a polarizing beam splitter (4), a zero-order quarter-wave plate (5), a long-pass filter (6), a short-pass dichroic mirror (7), an objective lens (8), and a 50 / 50 non-polarizing beam splitter (10), a plano-convex cylindrical lens (11), and two quadrant photodetectors (12) connected in sequence to the original return optical path of the objective lens (8) after the polarizing beam splitter (4). The imaging optical path system includes a visible light point source (13), a 50 / 50 flat beam splitter (14), a short-pass dichroic mirror (7), an objective lens (8) connected in sequence, and a sleeve lens (15) and a CMOS camera (16) connected in sequence to the 50 / 50 flat beam splitter (14) via the original optical path return of the objective lens (8). The astigmatic differential confocal detection optical path system and the imaging optical path system are coaxial and synchronized. The output signals of the two four-quadrant photodetectors (12) are used to calculate the defocus error signal FES. When FES crosses zero, it is determined that the surface of the sample under test is located on the focal plane of the objective lens (8). At the same time, the CMOS camera (16) synchronously acquires a clear image of the surface of the sample under test, thus achieving synchronization of alignment and imaging.

2. The astigmatic differential confocal alignment system capable of synchronous imaging according to claim 1, characterized in that, The optical path corresponding to the astigmatic differential confocal optical path system specifically includes: The beam output from the laser (1) is collimated and expanded by the beam-expanding system (2), and then filtered by the polarization beam-splitter (4) to transmit the P beam to the zero-order quarter-wave plate (5). At this time, the linearly polarized light is converted into circularly polarized light. It is then reflected by the long-pass filter (6) and the short-pass dichroic mirror (7) and enters the objective lens (8). After being focused by the objective lens (8) and reflected by the surface of the sample under test, it returns along the original path and passes through the short-pass dichroic mirror (7), the long-pass filter (6), and the zero-order quarter-wave plate (5) again. The circularly polarized light is converted into linearly polarized light, and the P beam is converted into the S beam. After being reflected by the polarization beam-splitter (4) and split into two beams of equal energy by the 50 / 50 non-polarization beam-splitter (10), they are focused by the plano-convex cylindrical lens (11) and irradiated by two four-quadrant photodetectors (12). The defocus error signal FES is calculated based on the output signals of the two four-quadrant photodetectors (12).

3. The astigmatic differential confocal alignment system capable of synchronous imaging according to claim 1, characterized in that, The optical path corresponding to the imaging optical path system specifically includes: Visible light emitted from the visible light point source (13) is split by a 50 / 50 flat beam splitter. The transmitted light is physically blocked, and the reflected light enters a short-pass dichroic mirror (7). The low-wavelength visible light is transmitted into the objective lens (8), and after being focused by the objective lens (8) onto the surface of the sample being tested and reflected, it returns along the original path. After passing through the short-pass dichroic mirror (7) and the 50 / 50 flat beam splitter (14) in sequence, it is focused by the sleeve lens (15) onto the CMOS camera (16).

4. The astigmatic differential confocal alignment system capable of synchronous imaging according to claim 1, characterized in that, The defocus error signal FES is calculated as follows: FES=(A1+A3-A2-A4) / (A1+A2+A3+A4)-(B1+B3-B2-B4) / (B1+B2+B3+B4) Among them, the output quantities of the four-quadrant photodetector A in the two four-quadrant photodetectors (12) are A1, A2, A3, A4, and the output quantities of the four-quadrant photodetector B are B1, B2, B3, B4.

5. The astigmatic differential confocal alignment system capable of synchronous imaging according to claim 1, characterized in that, In the astigmatic differential confocal optical path system, an aperture stop (3) is provided between the collimation beam expansion system (2) and the polarization beam splitter (4), the polarization beam splitter (4) and the 50 / 50 non-polarization beam splitter (10), the short-pass dichroic mirror (7) and the objective lens (8) to block stray light with a spot diameter greater than 14 mm.

6. The astigmatic differential confocal alignment system capable of synchronous imaging according to claim 1, characterized in that, The synchronous imaging astigmatic differential confocal alignment system also includes an astigmatic differential confocal optical path adjustment system, comprising: Using the clear image acquired by the CMOS camera (16) as a reference, the positive focus position of the sample surface under test is calibrated; then the output signals of the two diagonal quadrants in the four-quadrant photodetector (12) are adjusted respectively so that the output signal values ​​of the two diagonal quadrants in each group are equal.

7. The astigmatic differential confocal alignment system capable of synchronous imaging according to claim 1, characterized in that, The synchronous imaging astigmatic differential confocal alignment system also includes a spot morphology determination system for the four-quadrant photodetector (12), comprising: When the surface of the sample to be tested is located on the focal plane of the objective lens (8), the spot morphology on the two quadrant photodetectors (12) is the same; When the surface of the sample to be tested is in a defocused state, the spot morphology on the two four-quadrant photodetectors (12) changes in opposite directions, tending to be linear and circular respectively.

8. The astigmatic differential confocal alignment method for synchronous imaging according to claim 1, characterized in that, The two quadrant photodetectors (12) are respectively located 2 mm in front of and behind the focal plane of the plano-convex cylindrical lens (11), and rotated 45 degrees along the optical axis.

9. The astigmatic differential confocal alignment system capable of synchronous imaging according to claim 3, characterized in that, The long-pass filter (6) has a cutoff wavelength of 600nm and a transmittance of >90% for light with wavelength ≥600nm, and is used to block visible light with wavelength <600nm; the short-pass dichroic mirror (7) divides the light beam into transmitted light and reflected light according to the wavelength range, wherein the transmitted band and the reflected band are separated by a cutoff wavelength of 567nm, and the transmittance of >90% for visible light with wavelength <567nm and the reflectance of >90% for light with wavelength >567nm.

10. A method for synchronous imaging with astigmatic differential confocal alignment, characterized in that, The astigmatic differential confocal alignment system capable of synchronous imaging according to any one of claims 1-9 includes the following steps: Construct an astigmatic differential confocal detection optical path and a visible light imaging optical path that share the same objective lens (8); The astigmatic differential confocal detection optical path uses a laser emitted by a laser (1), which is reflected sequentially through a collimation and beam expansion system (2), a polarization beam splitter (4), a zero-order quarter-wave plate (5), a long-pass filter (6), and a short-pass dichroic mirror (7) into an objective lens (8). After being focused by the objective lens (8) onto the surface of the sample under test and reflected, the laser returns along the original path and is split again by a short-pass dichroic mirror (7), a long-pass filter (6), a zero-order quarter-wave plate (5), a polarization beam splitter (4), and a 50 / 50 unpolarized beam splitter (10) into two beams of equal energy. These beams are then focused by plano-convex cylindrical lenses (11) and irradiated onto two quadrant photodetectors (12). The defocus error signal FES is calculated based on the output signals of the two quadrant photodetectors (12). At the same time, visible light emitted by a visible light point source (13) is reflected by a 50 / 50 unpolarized dichroic mirror. The flat beam splitter (14) and short-pass dichroic mirror (7) enter the objective lens (8), and after being focused by the objective lens (8) onto the surface of the sample to be tested and reflected, they return along the original path, passing through the short-pass dichroic mirror (7) and the 50 / 50 flat beam splitter (14) in sequence, and are then focused by the sleeve lens (15) onto the CMOS camera (16). By coaxially synchronizing the astigmatic differential confocal detection optical path and the imaging optical path, while the FES signal crosses zero to determine that the surface of the sample under test is located on the focal plane of the objective lens (8), the CMOS camera (16) simultaneously acquires a clear morphological image of the surface of the sample under test, thereby achieving real-time unification of alignment and imaging.

Images