A wafer surface defect detection system and defect detection method based on polarized light

Abstract

This invention provides a wafer surface defect detection system and method based on polarized light. The system includes: an illumination optical path; a collection optical path, including a phase modulator, a polarization modulator, a polarization beam splitter, and a differential detector arranged sequentially; the differential detector is used to receive... P Light and S The system outputs differential signals from two optical paths. A signal processing module acquires calibration parameters for the phase modulator and polarization modulator, which are parameters that ensure the differential signal corresponding to the surface region of the calibrated wafer is zero. It also fixes the phase modulator and polarization modulator to the corresponding calibration parameters, sends a scanning strategy for scanning the test area of ​​the wafer under test, generates an image based on the differential signal obtained from scanning the wafer, and detects defects on the wafer surface based on the image. The material of the test area is the same as the material of the surface region of the calibrated wafer. This invention improves the detection sensitivity of minute defects on the wafer surface.

Description

Technical Field

[0001] This invention relates to the field of defect detection technology, and in particular to a wafer surface defect detection system and method based on polarized light. Background Technology

[0002] Related technologies provide a wafer defect detection system that acquires images of the wafer surface using a CCD or CMOS detector and detects defects in the images based on defect detection methods. However, wafer surfaces can contain minute defects, such as shallow scratches, tiny particles, and photoresist residue. Existing technologies struggle to identify these minute defects in the images, resulting in insufficient detection sensitivity and a low detection rate for minute defects on the wafer surface. Summary of the Invention

[0003] This invention provides a wafer surface defect detection system and method based on polarized light, in order to improve the detection sensitivity and detection rate of minute defects on the wafer surface.

[0004] In a first aspect, embodiments of the present invention provide a wafer surface defect detection system based on polarized light, comprising: The illumination optical path is used to generate polarized light that illuminates the wafer surface at an oblique incidence. An optical path for collecting reflected light beams from the wafer surface includes a phase modulator, a polarization modulator, a polarization beam splitter, and a differential detector arranged sequentially; the phase modulator is used to modulate the reflected light beam... P Light and S The phase difference of light forms a first modulated beam, and the polarization modulator is used to adjust the polarization direction of the first modulated beam to form a second modulated beam. The second modulated beam is split by the polarization beam splitter. P Light and S The differential detector is used to receive light. P Light and S The two optical paths output differential signals. The signal processing module is used to obtain the calibration parameters of the phase modulator and the polarization modulator, wherein the calibration parameters are parameters that make the differential signal corresponding to the wafer surface region of the calibration wafer zero; The signal processing module is further configured to fix the phase modulator and the polarization modulator to the corresponding calibration parameters, send a scanning strategy for scanning the test area of ​​the wafer under test, generate an image based on the differential signal obtained by scanning the wafer under test, and detect defects on the surface of the wafer under test based on the image, wherein the material of the test area is the same as the material of the wafer surface area of ​​the calibration wafer.

[0005] Secondly, embodiments of the present invention provide a method for detecting defects on the surface of a wafer, which detects defects on the surface of the wafer under test based on the defect detection system as described in the first aspect, and includes the following steps: S1: Using the calibration wafer, obtain the calibration parameters of the phase modulator and the polarization modulator; S2: Fix the phase modulator and the polarization modulator to the corresponding calibration parameters, and send a scanning strategy for scanning the test area of ​​the wafer under test; S3: Generate an image based on the differential signal obtained from scanning the wafer under test, and locate the defects on the surface of the wafer under test based on the abnormal areas in the image.

[0006] This invention proposes a wafer surface defect detection system based on polarized light. By calibrating the wafer, the reflected light beam from the material (i.e., background material) of the test area on the wafer surface is polarized and modulated to zero the differential signal at non-defect locations. A differential detector collects data from scanning the wafer, and a signal processing module generates an image characterizing the wafer surface state based on this data. Defects on the wafer surface can be located by performing defect identification on this image. Specifically, when the incident light beam scans a defect (such as a foreign object, scratch, or material anomaly) in the test area of ​​the wafer, the polarization state of the reflected light beam changes due to the difference in optical properties between the defect and the background material. This causes the differential detector to output a non-zero signal, thus highlighting the defect with high contrast. This defect detection system improves the detection sensitivity, enabling high-sensitivity localization of minute defects on the wafer surface and increasing the defect detection rate.

[0007] This invention also proposes a wafer surface defect detection method, which detects defects on the surface of the wafer under test based on the defect detection system as described in the first aspect, and thus has the beneficial effects described above. Attached Figure Description

[0008] Figure 1 A schematic diagram of a wafer surface defect detection system based on polarized light provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of a light beam illuminating the wafer surface in an embodiment of the present invention; Figure 3 This is a schematic diagram of butterfly frequency domain filters with different light transmission ratios in an embodiment of the present invention; Figure 4 This is a schematic diagram of ring frequency domain filters with different light transmission ratios in an embodiment of the present invention; Figure 5 This is a schematic diagram of a strip frequency domain filter with different shading positions in an embodiment of the present invention; Figure 6A schematic diagram of another wafer surface defect detection system based on polarized light provided in an embodiment of the present invention; Figure 7 This is a schematic diagram of the polarization state of the reflected light beam in an embodiment of the present invention; Figure 8 This is a flowchart of a wafer surface defect detection method provided in an embodiment of the present invention; Figure 9 This is a flowchart of another wafer surface defect detection method provided in an embodiment of the present invention; Figure 10 This is a schematic diagram of the polarization state of a beam modulated by a first half-wave plate in an embodiment of the present invention; Figure 11 This is a schematic diagram of the polarization state of the light beam modulated by the liquid crystal phase delayer in an embodiment of the present invention; Figure 12 This is a schematic diagram of the polarization state of a beam modulated by a first half-wave plate in another embodiment of the present invention; Figure 13 This is a schematic diagram of the wafer scanning process in an embodiment of the present invention; Figure 14 This is a schematic diagram of the signal architecture of the defect detection system in an embodiment of the present invention; Figure 15 This is a schematic diagram of image reconstruction in an embodiment of the present invention; Figure 16 This is a partial schematic diagram of residual photoresist on the surface of a patternless wafer in an embodiment of the present invention; Figure 17 This is a simulated image of the wafer surface in an embodiment of the present invention.

[0009] The components include: 1. Light source; 2. Beam modulator; 3. Signal generator; 4. Beam splitter; 5. Photodetector; 6. Third mirror; 7. Polarization modulation module; 8. Second half-wave plate; 9. Quarter-wave plate; 10. Second mirror; 11. Converging lens; 12. Wafer; 13. Collimating lens; 14. First mirror; 15. Aperture stop; 16. Relay mirror group; 17. Frequency domain filter; 18. Phase modulator; 19. Polarization modulator; 20. Polarization beam splitter; 21. Fiber optic coupler; 22. Fiber optic cable; 23. Differential detector; 24. Lock-in amplifier; 25. Analog-to-digital converter; 26. Host computer; 27. Frequency domain filtering component. Detailed Implementation

[0010] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0011] In the description of this application, unless otherwise specified or stated, the term "plural" means at least two; unless otherwise specified or stated, the terms "joining," "attaching," "installing," "connecting," and "linking" should be interpreted broadly, for example, they can be fixed connections or movable connections; they can be non-detachable connections or detachable connections, and non-detachable connections can be integral connections or welded connections; they can be mechanical connections or electrical connections; they can be internal communication between two components or the interaction between two components; they can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0012] In this embodiment of the invention, the background material refers to the material of the user's region of interest (i.e., the region to be measured) on the surface of the wafer being tested. Wafers can be divided into two categories: unpatterned wafers and patterned wafers. For unpatterned wafers, the wafer surface can be the surface of a substrate (such as a Si substrate) or the surface of a film layer (such as a SiO2 film layer) located above the substrate; that is, the wafer surface layer can be either a substrate or a film layer. For patterned wafers, many layers are formed on the substrate during the wafer manufacturing process. The material of each layer can be Si, SiO2, SiN, W, or Cu, etc. These layers are divided into patterned layers and unpatterned layers (unpatterned layers can be understood as film layers of the wafer). The wafer surface layer can be either a patterned layer or an unpatterned layer. When the surface layer of a patterned wafer is an unpatterned layer, part or all of the area in the unpatterned layer is the user's region of interest, and the material of this layer is the background material when scanning that layer.

[0013] In this embodiment of the invention, after the light beam illuminates the background material, the reflected light beam undergoes polarization modulation through a phase modulator and a polarization modulator (such as a half-wave plate) before reaching the differential detector. The phase modulator and the half-wave plate can make the differential signal output by the differential detector zero, thus realizing the zeroing of the signal based on the background material in this embodiment of the invention. Moreover, when the light beam illuminates the same material, as long as the state of the material does not change, the differential signal output by the differential detector is always zero. For defects on the unpatterned layer, such as scratches, damage, or small particles of other materials (different from the unpatterned layer), when the light beam illuminates these areas, the reflected light beam will carry information of other polarization directions. The reflected light beam passes through the phase modulator and the polarization modulator before reaching the differential detector, and the differential signal output by the differential detector is no longer zero, thus detecting the defects. The defects can represent abnormal morphology or material abnormalities. The states of phase modulators and polarization modulators (such as half-wave plates) differ depending on the unpatterned layer. This means that, for different wafer surface layer materials, the material of the region of interest within the wafer surface layer can be selectively used as the background material, and the differential signal output by the corresponding differential detector can be set to 0, thereby improving the applicability of the defect detection system. The wafer surface layer material can be metallic, dielectric, or compound materials, etc.

[0014] When the surface layer of a patterned wafer is patterned, defects can be not only those mentioned above, but also pattern defects (such as photoresist residue defects). Users select the material to be zeroed based on the area of ​​interest. For example, if a layer is mostly Si, and the pattern is composed of SiO2, the SiO2 material can be filled into the grooves of the Si layer. To observe defects in non-patterned areas, the differential signal output by the differential detector corresponding to the reflected beam from Si should be set to 0. In this case, the background material is Si, and the signal from defects contrasts strongly with the signal from Si. Similarly, to observe defects in patterned areas, i.e., the quality of the SiO2 pattern, the differential signal output by the differential detector corresponding to the reflected beam from SiO2 should be set to 0. In this case, the background material is SiO2, and the signal from defects contrasts strongly with the signal from SiO2.

[0015] Figure 1 This is a schematic diagram of a wafer surface defect detection system based on polarized light, provided in an embodiment of the present invention. (Refer to...) Figure 1The defect detection system includes an illumination optical path, a collection optical path, and a signal processing module. The illumination optical path generates polarized light, which is incident obliquely onto the surface of wafer 12. Wafer 12 may include a calibration wafer for calibration or a test wafer for detecting defects. The calibration wafer is determined based on the material of the region of interest (i.e., the background material) of the test wafer, such that the surface of the calibration wafer includes this background material. The collection optical path collects the reflected light beam from the surface of wafer 12. The collection optical path includes a phase modulator 18, a polarization modulator 19, a polarization beam splitter 20, and a differential detector 23 arranged sequentially. The phase modulator 18 modulates the reflected light beam... P Light and S The phase difference of light forms a first modulated beam, and polarization modulator 19 is used to adjust the polarization direction of the first modulated beam to form a second modulated beam. The second modulated beam is split by polarization beam splitter 20. P Light and S The optical differential detector 23 is used to receive light. P Light and S The optical signals from both paths are output as differential signals. The signal processing module is used to obtain the calibration parameters of the phase modulator 18 and the polarization modulator 19. The calibration parameters are those that make the differential signal corresponding to the wafer surface region of the calibration wafer zero. The differential signal may include a differential voltage signal or a differential current signal.

[0016] in, P The polarization direction of light is parallel to the plane of incidence, and its electric field vibrates within the plane of incidence. S The polarization direction of the light is perpendicular to the plane of incidence, and the direction of its electric field vibration is perpendicular to the plane of incidence. P Light and S Light depends on how the beam of light illuminates the surface. Once the plane of incidence is changed, S Light and P The direction of the light also changes accordingly.

[0017] The signal processing module is also used to fix the phase modulator 18 and polarization modulator 19 to corresponding calibration parameters, send a scanning strategy for scanning the test area of ​​the wafer under test, generate an image based on the differential signal obtained from scanning the wafer under test, and detect defects on the surface of the wafer under test based on the image. The wafer under test can also be referred to as the wafer to be tested, and the wafer surface region (or wafer surface layer) of the wafer under test includes the test area. The material of the test area is the same as the material of the wafer surface region of the calibration wafer.

[0018] In this embodiment of the invention, the scanning strategy includes scanning path planning (such as serpentine scanning or raster scanning), a sequence of scanning positions, and a preset image size (the image size is represented by the number of pixels in the image, such as M×N). The scanning strategy is used to achieve ordered scanning of different scanning positions in the area to be tested of the wafer under test. The content regarding the construction of the scanning strategy is prior art and will not be elaborated here. In one embodiment, the defect detection system further includes a wafer displacement system, which includes a displacement stage and a displacement stage controller. The scanning strategy is sent to the displacement stage controller, and the displacement stage controller controls the displacement stage to move the wafer under test according to the scanning strategy to achieve scanning. In another embodiment, the illumination optical path further includes a scanning galvanometer and a galvanometer controller. The scanning strategy is sent to the galvanometer controller, and the galvanometer controller controls the scanning galvanometer to move the wafer under test according to the scanning strategy to achieve scanning.

[0019] This invention proposes a wafer surface defect detection system based on polarized light. By calibrating the wafer, the reflected light beam from the material (i.e., background material) of the test area on the wafer surface is polarized and modulated to zero the differential signal at non-defect locations. A differential detector 23 collects data from scanning the wafer, and a signal processing module generates an image characterizing the wafer surface state based on this data. Defects on the wafer surface can be located by performing defect identification on this image. Specifically, when the incident light beam scans a defect (such as a foreign object, scratch, or material anomaly) in the test area of ​​the wafer, the polarization state of the reflected light beam changes due to the difference in optical properties between the defect and the background material. This causes the differential detector 23 to output a non-zero signal, thus highlighting the defect with high contrast. This defect detection system improves the detection sensitivity, enabling high-sensitivity localization of minute defects on the wafer surface and increasing the defect detection rate.

[0020] Optionally, the phase modulator 18 includes a liquid crystal phase retarder or a photoelastic modulator; and / or, the polarization modulator 19 includes a first half-wave plate. The phase modulator 18 is used to modulate the reflected beam. P Light and S The phase difference of light modulates the elliptically polarized reflected beam into a linearly polarized beam or a first modulated beam with elliptically polarization in a different state. A first half-wave plate is used to adjust the polarization direction of the first modulated beam, for example, by rotating the direction of the major axis of the linearly polarized or elliptically polarized light, to form a second modulated beam.

[0021] Exemplarily, the optical collection path also includes a polarizing beam splitter 20, a fiber coupler 21, and an optical fiber 22. The polarizing beam splitter 20 is also called a PBS, and the differential detector 23 is also called a differential photodetector. The optical fiber 22 includes a multimode fiber. The polarizing beam splitter 20 splits the second modulated beam into... P Light and SThe light beam consists of two parts, which are coupled into two optical fibers 22 through corresponding fiber couplers 21 and then enter the differential detector 23. Figure 1 As shown, there can be two fiber optic couplers 21. The use of fiber optic couplers 21 and optical fibers 22 can increase the stability of the input signal of the differential detector 23. The differential signal of the differential detector 23 is either a voltage signal or a current signal. In this embodiment, the differential signal is a voltage signal as an example. The two optical fibers 22 transmit the following signals respectively: P Light and S The greater the energy difference between the two polarization components of light, the greater the voltage value of the voltage signal output by the differential detector 23. This voltage value can be positive or negative. When the second modulation beam is linearly polarized light at a 45° angle, after being split by the polarizing beam splitter 20, the two optical fibers 22 transmit their respective polarization components... P Light and S Since the two polarization components of light have the same energy, the differential signal of the differential detector 23 is set to 0.

[0022] Because the material on the surface of wafer 12 is determined according to requirements, and different materials have different surface optical properties, when an unmodulated reflected light beam directly enters the differential detector 23, there will be a certain signal output, which will affect the detection sensitivity of the entire defect detection system. Therefore, this embodiment of the invention proposes a modulation method. By using phase modulator 18 and polarization modulator 19 in combination, the defect detection system can control the energy difference of the light beam reaching the differential detector 23. When the incident light beam (the illumination light beam provided by the illumination optical path that illuminates the wafer 12) irradiates the material on the surface of wafer 12, taking a film layer as an example, the differential signal of the differential detector 23 is 0. However, when the incident light beam irradiates a defect on the film layer, there is other material at the defect location. The optical properties of this material are different from those of the film layer material. The modulation of the reflected light beam at this location is targeted at the film layer material. Therefore, the energy of the light beam reaching the differential detector 23 changes, and the differential detector 23 outputs a non-zero voltage signal. Because the differential signal of the differential detector 23 corresponding to the film material is 0, the defect detection system is in a metastable state. Any slight disturbance will change the differential signal of the differential detector 23. For non-film materials, the intensity of the two beams of light reaching the differential detector 23 will increase while the intensity decreases. Therefore, the differential detector 23 will output a double change, further increasing the difference between the two beams. This results in the differential detector 23 outputting a very clear differential signal, thus enabling the detection of defects on the surface of wafer 12. Furthermore, for scratch defects, the intensity of the two beams of light reaching the differential detector 23 will also increase while the intensity decreases, similarly causing the differential detector 23 to output a very clear differential signal. Therefore, the differential signal of the differential detector 23 is particularly clear. This modulation method greatly improves the signal contrast, thereby increasing the detection sensitivity of the defect detection system.

[0023] Optionally, the optical path also includes a frequency domain filter component 27, which is located before the phase modulator 18 and on the propagation path of the reflected beam. The frequency domain filter component 27 is located in the optical path between the wafer 12 and the phase modulator 18. The frequency domain filter component 27 is used to perform frequency domain filtering on the reflected beam propagating to the phase modulator 18, thereby improving the beam quality of the reflected beam propagating to the phase modulator 18.

[0024] The frequency domain filtering component 27 includes a frequency domain filter 17 located on the conjugate Fourier surface of the optical path. The frequency domain filter 17 selectively transmits specific spatial frequency components of the reflected beam on the conjugate Fourier surface, specifically transmitting the beam at certain positions on the conjugate Fourier surface while blocking the beam at other positions. For example, the frequency domain filter 17 can be used to block diffraction components caused by periodic patterns on the surface of wafer 12.

[0025] Optionally, depending on the scattering characteristics of the particles or other defects, the frequency domain filter 17 placed on the conjugate Fourier surface of the collecting optical path may include at least one of a butterfly frequency domain filter, a ring frequency domain filter, or a strip frequency domain filter. Figure 3 This is a schematic diagram of butterfly frequency domain filters with different transmittance ratios in an embodiment of the present invention. Figure 3 The diagram illustrates three butterfly-shaped frequency domain filters with different light transmission ratios. Figure 4 This is a schematic diagram of ring frequency domain filters with different transmittance ratios in an embodiment of the present invention. Figure 4 The diagram illustrates three different ring frequency domain filters with varying diameter ratios (transmission ratios). Figure 5 This is a schematic diagram of a strip frequency domain filter with different shading positions in an embodiment of the present invention. Figure 5 The diagram illustrates three different strip frequency domain filters with varying shading positions. (Reference) Figures 3-5 Black represents the area that is blocked, and white represents the area that is transparent.

[0026] After the reflected beam is expanded by the collimating lens 13, it carries information about the measured position, resulting in a non-uniform spatial energy distribution on the conjugate Fourier surface. Some locations have a high proportion of additional polarization components, while others have a low proportion. For example, the transmission ratio of the frequency domain filter can be configured to allow beams with a high proportion of additional polarization components to pass through, further improving detection sensitivity. The additional polarization component refers to the polarization element in the reflected beam that changes relative to the original incident beam or the polarization state relative to the background material due to surface characteristics when the incident beam irradiates the surface of wafer 12. The frequency domain filter 17 further filters and purifies these beams carrying defect information from the physical optical path.

[0027] For example, the transmittance ratio of the frequency domain filter can be configured so that the frequency domain filter 17 selectively transmits beams at certain locations within the reflected beam. By blocking high-power-density regions, the signal contrast entering the differential detector 23 is enhanced, further increasing the detection sensitivity of the defect detection system. The high-power-density regions correspond to specularly reflected beams, while beams carrying defect information are contained within non-spectrally reflected beams, such as scattered light caused by defects. By blocking the high-power-density regions, the intensity of the background light entering the differential detector 23 is reduced, thereby increasing the power ratio of the beam signal light to the background light and improving signal contrast.

[0028] Optionally, the light-collecting path also includes an aperture stop 15 and a relay lens group 16. The aperture stop 15 is located on the Fourier surface of the light-collecting path, which is also the pupil surface of the light-collecting path. The aperture stop 15 and the frequency domain filter 17 are conjugate with respect to the relay lens group 16. The aperture stop 15 is located at the back focal plane of the collimating lens 13, which is also the pupil surface position of the light-collecting path. For example, a variable aperture stop with a variable aperture is placed at this position as the aperture stop 15, which can select the diameter of the transmitted beam as needed and block peripheral stray beams, thereby further improving the signal sensitivity. The reflected beam after passing through the pupil surface forms a 1:1 conjugate image at the frequency domain filter 17 after being processed by the relay lens group 16. That is, the pupil surface of the light-collecting path forms a conjugate Fourier surface (a new Fourier surface) at the frequency domain filter 17 after being processed by the relay lens group 16.

[0029] Optionally, the calibration parameters of phase modulator 18 and / or polarization modulator 19 differ for different background materials under test. For example, the states of phase modulator 18 and polarization modulator 19 differ for different unpatterned layers (i.e., film layers). For one unpatterned layer, calibration parameters suitable for phase modulator 18 and polarization modulator 19 are obtained; for another unpatterned layer, calibration parameters for other phase modulator 18 and polarization modulator 19 are obtained. For the material of the surface layer of wafer 12, the material of the wafer surface layer can be specifically used as the background material, and the differential signal of the corresponding differential detector 23 can be set to 0, thereby improving the applicability of the defect detection system.

[0030] Exemplarily, the light collection path also includes a collimating lens 13 and a first reflecting mirror 14. The reflected beam enters the collimating lens 13 and becomes parallel light again after passing through it. The parallel light carrying information about the illuminated area of ​​the wafer 12 enters the aperture stop 15, and then enters the relay mirror group 16, the frequency domain filter 17, and the phase modulator 18. As an optional configuration, the first reflecting mirror 14 can be placed in the optical path between the collimating lens 13 and the aperture stop 15 to control the propagation direction of the reflected beam, for example, changing the beam propagating in an oblique direction to propagating in a horizontal direction.

[0031] Aperture stop 15, relay lens group 16 and frequency domain filter 17 are optional optical elements.

[0032] In this embodiment of the invention, the reflected beam after frequency domain filtering enters the phase modulator 18, and the polarization angle of the reflected beam entering the phase modulator 18 at this time is as follows: Figure 7 As shown, Figure 7 This is a schematic diagram of the polarization state of the reflected light beam in an embodiment of the present invention. Figure 7 The horizontal axis in the figure represents the normalized value. PThe optical amplitude is represented by the normalized S-axis. Depending on the optical properties of the wafer 12 surface material, elliptically polarized light will be generated at different angles. Figure 7 An example is shown of elliptically polarized light.

[0033] In this embodiment of the invention, the phase modulator 18 is aligned during installation, and the modulation direction of the phase modulator 18 is aligned with the incident beam. P The light directions are the same. The modulation direction of phase modulator 18 refers to the fast axis direction of phase modulator 18. The function of phase modulator 18 is to provide [modulation / modulation] for the reflected beam. P Light and S The phase difference of the light allows it to remain linearly polarized even after reflection from wafer 12, effectively changing the polarization state of the reflected beam from elliptically polarized to linearly polarized. This linearly polarized beam then passes through polarization modulator 19, modulating its polarization direction to 45°.

[0034] Optionally, the illumination optical path includes a light source 1, a beam modulator 2, and a polarization modulation module 7 arranged sequentially. The light source 1 may include a laser that provides linearly polarized light, or it may include a light source 1 (such as a white light source) for providing unpolarized light and a polarizer, which converts the unpolarized light into linearly polarized light. The light source 1 emits a linearly polarized first illumination beam; the beam modulator 2 modulates the presence or absence of the first illumination beam to form a second illumination beam; the polarization modulation module 7 adjusts the polarization state of the second illumination beam to obtain circularly polarized light or linearly polarized light with a polarization state different from that of the first illumination beam.

[0035] For example, in this embodiment of the invention, the light source 1 is an ultraviolet pulsed laser, which can emit a collimated single-wavelength ultraviolet beam. The laser beam is linearly polarized and horizontal, thus being horizontally linearly polarized light, with a repetition frequency of 80MHz or 120Hz. The beam modulator 2 is an electro-optic modulator (EOM) or an acousto-optic modulator, which can modulate the presence or absence of the beam at an extremely high modulation frequency (generated by the signal generator 3, the frequency of which is preset as needed). For example, in conjunction with the signal generator 3 to generate a modulation frequency of 30MHz, the passing linearly polarized laser exhibits a modulation frequency of two pulses passing and two pulses disappearing. The vibration direction of the electric field of the horizontally linearly polarized light is parallel to the tabletop, i.e., parallel to the plane where wafer 12 is located. The vibration direction of the electric field of the vertically linearly polarized light is perpendicular to the tabletop, i.e., perpendicular to the plane where wafer 12 is located. The horizontally and vertically linearly polarized light are descriptions of the vibration direction of the electric field of the light wave in a fixed coordinate system. S Light and P Light is a description of the direction of the electric field vibration of a light wave in a relative coordinate system relative to the plane of incidence.

[0036] Optionally, the polarization modulation module 7 includes a second half-wave plate 8 and a quarter-wave plate 9. The second half-wave plate 8 is located in the optical path between the beam modulator 2 and the quarter-wave plate 9, or the quarter-wave plate 9 is located between the beam modulator 2 and the half-wave plate 8. Alternatively, the polarization modulation module 7 includes an optical rotator, such as a Faraday rotator or a polarization-adjusting fiber. In this embodiment of the invention, the polarization modulation module 7 includes a second half-wave plate 8 and a quarter-wave plate 9 as an example. As a polarization device in a defect detection system, the polarization modulation module 7 can form a standard circularly polarized beam or linearly polarized beams at different angles by adjusting the angles of the second half-wave plate 8 and the quarter-wave plate 9. The polarization state of the polarization-modulated beam can be... S polarization, P Different polarization states, such as circular polarization, result in different detection sensitivities. To obtain optimal sensitivity, in this embodiment of the invention, circular polarization is used as the polarization direction of the light beam illuminating wafer 12. For example, the optical axis angles of the second half-wave plate 8 and the quarter-wave plate 9 are both 45° to the polarization direction of the illumination beam incident on the polarization modulation module 7. The optical axis directions of the second half-wave plate 8 and the quarter-wave plate 9 are parallel. The second half-wave plate 8 converts horizontally linearly polarized light into vertically linearly polarized light, and the quarter-wave plate 9 converts vertically linearly polarized light into circularly polarized light.

[0037] For example, the polarization modulation module 7, as a polarization device in a defect detection system, can modulate natural light generated by a white light source into light with a specific polarization state. It can also modulate linearly polarized light generated by the light source into linearly polarized light with higher contrast or light with other polarization states. For instance, after modulation by the polarization modulation module 7, the ratio of the brightness of the transmission axis to the brightness of the absorption axis of the linearly polarized light can reach 3000:1 to 10000:1.

[0038] Exemplarily, the illumination optical path also includes a converging mirror (such as a converging lens 11 or an off-axis parabolic mirror) and a second reflecting mirror 10. The beam is focused onto the surface of the wafer 12 by the converging lens 11, with an incident angle of approximately 70° (optionally between 65° and 80°), and the size of the converged spot is approximately 5 μm. In some scenarios, the second reflecting mirror 10 can be replaced with a Dammann grating to achieve the purpose of converging multiple beams onto the surface of the wafer 12, further improving the throughput of the defect detection system.

[0039] Figure 2 This is a schematic diagram of the light beam illuminating the wafer surface in an embodiment of the present invention, as shown below. Figure 2As shown, the incident light beam illuminating the wafer surface is a converging beam. When the beam is incident obliquely on the wafer surface, for example, at a large angle of incidence, depending on the state of the wafer surface, the beam will change from the incident polarization state to a complex polarization state of reflection, because the optical properties of the wafer surface introduce polarization components in other directions into the incident beam. For example, the wafer surface is either flat or a patterned wafer that has undergone CMP polishing, and the incident beam is reflected into the collecting optical path.

[0040] Exemplarily, the illumination optical path also includes a beam splitter 4 and a photodetector 5. The beam splitter 4 is located in the propagation path of the second illumination beam. The beam splitter 4 is used to partially reflect and partially transmit the second illumination beam. The beam splitter 4 can be used to make the energy of the reflected beam less than or equal to the energy of the transmitted beam, for example, to reflect less than 10% (e.g., about 1%) of the energy in the second illumination beam into the photodetector 5 (PD). The signal received by the photodetector 5 is modulated light. The photodetector 5 generates a corresponding electrical signal based on the modulated periodic light signal, which serves as the reference signal for the lock-in amplifier 24 in the signal processing module. This reference signal is connected to the reference terminal ref of the lock-in amplifier 24. Using the directly detected light as the corresponding electrical signal as the reference signal of the lock-in amplifier 24 improves the phase stability of the reference signal of the lock-in amplifier 24, resulting in a higher effective gain.

[0041] Figure 6 This is a schematic diagram of another wafer surface defect detection system based on polarized light provided in an embodiment of the present invention, with reference to... Figure 6 There is no need to set up a beam splitter 4 and a photodetector 5 in the illumination optical path. The electrical signal output by the signal generator 3 is directly used as the reference signal of the lock-in amplifier 24. The output terminal of the signal generator 3 is connected to the reference terminal ref of the lock-in amplifier 24.

[0042] Exemplarily, the illumination optical path further includes a third reflector 6, located in the optical path between the beam modulator 2 and the polarization modulation module 7, for controlling the propagation direction of the second illumination beam. For example, changing a beam propagating in a vertical direction to propagate in a horizontal direction. A second reflector 10, located in the optical path between the polarization modulation module 7 and the converging lens 11, is also used to control the propagation direction of the second illumination beam. For example, changing a beam propagating in a horizontal direction to propagate in an oblique direction. The third reflector 6, the second reflector 10, and the first reflector 14 are optional optical elements.

[0043] Optionally, the signal processing module includes a lock-in amplifier 24, an analog-to-digital converter 25, and a host computer 26. The analog-to-digital converter 25 is abbreviated as ADC. The reference terminal ref of the lock-in amplifier 24 receives a reference signal, and the signal terminal sig (i.e., the signal input terminal) of the lock-in amplifier 24 is connected to the output terminal of the differential detector 23. The signal terminal sig of the lock-in amplifier 24 receives the differential signal output by the differential detector 23. The input terminal of the analog-to-digital converter (ADC) 25 is connected to the output terminal of the lock-in amplifier 24, and the output terminal of the ADC 25 is connected to the input terminal of the host computer 26. The ADC 25 acquires the amplitude of the DC signal at the output terminal of the lock-in amplifier 24 based on a trigger signal. The trigger signal can be a trigger signal issued by the displacement stage controller or a trigger signal issued by the galvanometer controller. The ADC 25 is used to synchronously acquire the amplitude of the DC signal based on the trigger signal. The trigger signal includes multiple trigger pulses. The synchronous acquisition of the DC signal amplitude by the ADC 25 specifically includes: the ADC 25 responds to one trigger pulse received from the trigger signal to synchronously acquire one amplitude, such that the number of trigger pulses is the same as the number of amplitudes acquired by the ADC 25. The host computer 26, based on the correspondence between the trigger pulse timing in the trigger signal and the preset image size (e.g., M×N) in the scanning strategy, splits and reassembles the one-dimensional amplitude to form a two-dimensional image. The reference signal of the lock-in amplifier 24 originates from the electrical signal corresponding to the modulated beam signal in the illumination optical path. This electrical signal can be the electrical signal generated by the photodetector 5 or the electrical signal output by the signal generator.

[0044] The defect detection system also includes a wafer displacement system (not shown in the figure). The wafer displacement system includes a displacement stage and a displacement stage controller. The displacement stage is used to drive the wafer 12 to reciprocate. During the scanning process, the displacement stage controller sends a trigger signal to the analog-to-digital converter 25 according to the moving position of the displacement stage. The trigger signal sent by the displacement stage controller is used as a control signal to control the analog-to-digital converter 25 to synchronously acquire the amplitude of the DC signal output by the lock-in amplifier 24. The analog-to-digital converter 25 sends the acquired data (i.e., the amplitude of the DC signal) to the host computer 26. The host computer 26 generates an image based on the received amplitude. The data collected by the host computer 26 (including amplitude values ​​acquired at multiple different times) is a one-dimensional data in time. Each trigger pulse corresponds to an amplitude. The host computer 26 splits and reassembles the one-dimensional amplitude according to the image size preset in the scanning strategy to form a two-dimensional image based on the correspondence between the trigger pulse time (i.e., the moment of the trigger pulse) and the amplitude in the trigger signal. Taking the image size as M×N as an example, the first N amplitude values ​​are used as the pixel values ​​of one row, and the next N amplitude values ​​are used as the pixel values ​​of another row, and so on, to obtain an image with M rows and N columns of pixel values. The image includes M×N pixel values, and each pixel value is the amplitude at a scanning position, which is used to represent the surface reflection at that scanning position. By identifying the high bright spots in the image, defects on the wafer surface can be located.

[0045] Figure 8 This is a flowchart of a wafer surface defect detection method provided in an embodiment of the present invention, with reference to... Figures 1-8 The defect detection method includes the following steps: S1: Use the calibration wafer to obtain the calibration parameters of the phase modulator and polarization modulator.

[0046] The wafer 12 used in the calibration process is called the calibration wafer. The surface layer (i.e., the surface area) of the calibration wafer includes background material. The calibration wafer can be the same wafer 12 as the wafer under test, but it is not limited to this. For example, if the surface layer of the wafer under test is a SiO2 film layer and the material of the area of ​​interest to the user is SiO2, then the surface layer of the calibration wafer also includes SiO2 material. The calibration wafer includes a substrate and a SiO2 film layer located on the substrate. In this embodiment of the invention, the material of the surface layer of the calibration wafer is usually a single material to avoid the influence of other patterns. The calibration wafer has the characteristics of single material, no pattern influence, and few dust particles.

[0047] S2: Fix the phase modulator and polarization modulator to the corresponding calibration parameters, and send a scanning strategy for scanning the test area of ​​the wafer under test.

[0048] S3: Generate an image based on the differential signal obtained from scanning the wafer under test, and locate the defects on the surface of the wafer under test based on the abnormal areas in the image.

[0049] The signal processing module executes various steps in the embodiments of the present invention, such as the host computer 26 acquiring the calibration parameters of the phase modulator 18 and the polarization modulator 19.

[0050] The defect detection method provided in this embodiment uses the defect detection system described in the above embodiment. The defect detection system's workflow includes a background material signal calibration process and a scanning imaging process. Before actual scanning, signal calibration is performed on the background material of the wafer surface under test using a calibration wafer, ensuring that the differential signal output by the differential detector is zero. The purpose of the calibration process is to obtain the calibration parameters of the phase modulator 18 and polarization modulator 19 when the differential signal output by the differential detector is zero. After calibration, the calibration parameters of the phase modulator and polarization modulator are obtained. Then, the host computer sends the calibration parameters to the phase modulator and polarization modulator respectively, fixing them to the corresponding calibration parameters. At this time, the defect detection system is in the optimal sensitivity state for the background material. When the beam scans onto the defect on the wafer under test, the polarization state of the reflected beam changes due to the difference in optical properties between the defect and the background material, causing the differential detector 23 to output a non-zero signal, thereby highlighting the defect with high contrast. This defect detection system improves the detection sensitivity for defects, enabling high-sensitivity localization of minute defects on the surface of wafer 12 and improving the defect detection rate.

[0051] Figure 9 This is a flowchart of another wafer surface defect detection method provided in an embodiment of the present invention, with reference to... Figures 1-9 The defect detection method includes the following steps: S201: Pre-aligned phase modulator.

[0052] For example, in the pre-alignment phase modulator 18, the signal processing module (specifically, the host computer 26 within the signal processing module) sends a control signal to align the fast axis of the phase modulator 18 with the optical axis of the rear polarizing beam splitter 20, and the modulation direction of the phase modulator 18 is... P Light direction or S Optical direction. The modulation direction of the phase modulator 18 is either the fast axis or the slow axis direction of the phase modulator 18. In this embodiment, the modulation direction is the fast axis direction of the phase modulator 18. The optical axis direction of the polarizing beam splitter 20 is the transmission axis direction of the polarizing beam splitter 20, that is, the optical direction of the polarizing beam splitter 20. P Light direction. In this embodiment of the invention, the polarization analysis and modulation of the entire defect detection system are performed within a defined reference frame, which is determined by... P Light direction and S The coordinate system defined by the direction of light.

[0053] S202: Move the calibration wafer to the position where the illumination spot converges, and adjust the height of the calibration wafer to place it at the position where the illumination spot converges the least.

[0054] The illumination beam illuminating wafer 12 is the incident beam, which is a converging beam. The incident beam forms an illumination spot on wafer 12. The displacement stage controller controls the height of the displacement stage according to the signal sent by the host computer. When the height of the calibration wafer is adjusted by the displacement stage, the size of the illumination spot on the wafer will change. The smallest illumination spot is found among all the illumination spots corresponding to multiple different heights. At this time, the position of the displacement stage is the position required for the calibration wafer and the wafer under test. Subsequently, the calibration wafer is calibrated at this position to obtain the calibration parameters of the phase modulator and the polarization modulator, and the wafer under test is scanned and imaged at this position.

[0055] S203: Control the polarization modulator to rotate at a preset angle, find the second angle that maximizes the amplitude of the differential signal, and fix it.

[0056] For example, step S203 includes: the host computer 26 in the signal processing module sends a control signal to control the rotary motor to drive the polarization modulator 19 (e.g., the first half-wave plate) to rotate, scanning a range greater than or equal to 180° with a step size of 0.01°. The host computer 26 records the differential signals of the differential detector 23 at different angles. The step size ranges from [0.01, 1], and the unit is degrees. It can be 0.05° or 0.1°, etc. The step size affects the inspection sensitivity of defect detection. The smaller the step size, the better. The user can set the step size as needed. During the rotation of the first half-wave plate, the differential signal of the differential detector 23 will exhibit a strong-weak-strong sinusoidal periodic change.

[0057] Step S203 further includes: the host computer 26 in the signal processing module obtains the angle of the first half-wave plate when the amplitude (such as the voltage value) of the differential signal output by the differential detector 23 is at its maximum by fitting the angle of the first half-wave plate and the amplitude (taking the voltage value as an example) of the differential signal output by the differential detector 23, and controls the rotating motor of the first half-wave plate to rotate to this angle and fix it. At this time, the polarization state of the beam modulated by the first half-wave plate is as follows: Figure 10 As shown. The purpose of this step is to modulate the reflected beams from the surface of wafer 12 (which is the calibration wafer at this time) to the same polarization. The detection sensitivity is highest under this polarization condition because the amplitude of the output voltage is proportional to the detection sensitivity. The higher the amplitude of the output voltage, the higher the signal-to-noise ratio and the higher the detection sensitivity.

[0058] This step does not directly search for the zero point, but rather establishes a highly sensitive reference state by finding the point of maximum signal. SLight and P Light has high energy, which reduces the ratio of disturbance to base intensity, thereby improving the signal-to-noise ratio and detection sensitivity.

[0059] S204: Obtain multiple different phase delay values ​​of the phase modulator. The phase delay values ​​are values ​​within the range of the first phase value to the second phase value. Use a calibration wafer and a differential detector to obtain the differential signal corresponding to each phase delay value, and find the phase delay value that minimizes the amplitude of the differential signal to serve as the calibration parameter of the phase modulator.

[0060] The difference between the first phase value and the second phase value is greater than or equal to π / 2, for example, greater than or equal to π.

[0061] In this embodiment of the invention, step S204 includes: the host computer 26 in the signal processing module is used to acquire multiple different phase delay values. These phase delay values ​​are, for example, phase delay values ​​obtained by adjusting the phase modulator 18 from 0 to π with a preset step size, such as 0.01. The host computer 26 is also used to acquire the differential signal of the differential detector 23 under the above-mentioned different phase delay values. The value range of the step size of the phase delay value is [0.01, 1], and the unit is radians. It can be 0.05 or 0.1 radians or other step sizes. The step size affects the inspection sensitivity of defect detection. The smaller the step size, the better. The user can set the step size as needed. In addition, the range of the phase delay value can also be 0 to π / 2 or π / 2 to π, etc., as long as the difference between the two endpoints (the left endpoint is the first phase value and the right endpoint is the second phase value) is greater than or equal to π / 2. To ensure that the scanning range is large enough, the scanning range of 0 to π is used in this embodiment of the invention. Similarly, as the phase delay gradually increases, the differential signal of the differential detector 23 will exhibit a strong-weak-strong sinusoidal periodic change.

[0062] In this embodiment of the invention, step S204 further includes: the host computer 26 in the signal processing module obtains the phase delay amount when the amplitude of the differential signal output by the differential detector 23 is at its minimum through fitting. That is, the host computer 26 in the signal processing module records the calibration parameters of the liquid crystal phase delay at this time, which are used as calibration parameters for the liquid crystal phase delay. The modulation result of the liquid crystal phase delay at this position is to modulate the reflected elliptically polarized beam into a linearly polarized beam, and the polarization state of the linearly polarized beam at this time is as follows: Figure 11 As shown.

[0063] The host computer 26 in the signal processing module controls the phase delay of the phase modulator 18 (e.g., a liquid crystal phase delayer) to precisely adjust the reflected beam from the background material from an elliptically polarized state to a linearly polarized state.

[0064] S205: Then control the polarization modulator to rotate at a preset angle to find the first angle that makes the amplitude of the differential signal zero, and use it as the calibration parameter of the polarization modulator.

[0065] For example, step S205 includes: the host computer 26 in the signal processing module again controls the rotation angle of the polarization modulator 19 (e.g., the first half-wave plate) to scan a range greater than or equal to 180° in steps of 0.01°, and records the differential signals of the differential detector 23 at different angles. The step size ranges from [0.01, 1], and the unit is degrees. It can be 0.05° or 0.1°, etc. The step size affects the inspection sensitivity of defect detection. The smaller the step size, the better. The user can set the step size as needed. During the rotation, the differential detector 23 outputs a strong-weak-strong sinusoidal periodic change.

[0066] In this embodiment of the invention, step S205 further includes: the host computer 26 in the signal processing module obtains the waveplate angle, i.e., the first angle, when the amplitude of the differential signal output by the differential detector 23 is at its minimum through fitting, and uses it as a calibration parameter for the polarization modulator. At this time, the polarization angle of the beam incident on the polarization beam splitter 20 should be modulated to 45 degrees, that is, the polarization direction of the beam is relative to... P The angle between the beams is 45 degrees, and the differential signal of the differential detector 23 is 0. That is, the host computer 26 in the signal processing module records the calibration parameters of the first half-wave plate at this time as the calibration parameters of the first half-wave plate. Reaching the state of highest detection sensitivity, the polarization direction of the beam entering the polarizing beam splitter 20 is as follows... Figure 12 As shown. Figures 10-12 The horizontal and vertical axes in the middle are Figure 7 The same applies, so I won't repeat it here.

[0067] The core function of the first half-wave plate is to rotate the vibration direction of incident linearly polarized light relative to its fast axis by an angle twice the included angle. This rotation effect does not require the incident light beam to be in a specific direction. As long as it is linearly polarized light, the first half-wave plate can rotate it to a new, predictable direction according to its own fast axis direction. The rotation capability of the first half-wave plate is universal, thus automatically adapting to the current background material.

[0068] In steps S203-S205 of this embodiment of the invention, calibration parameters of the phase modulator and polarization modulator can be obtained at a preset location on the wafer surface region of the calibration wafer. These calibration parameters can be parameters obtained from a location on the wafer surface region of the calibration wafer, but are not limited to this. They can also be parameters obtained through mathematical statistics from candidate calibration parameters obtained from multiple preset locations on the wafer surface region of the calibration wafer. Taking the latter approach as an example, the defect detection method includes step S206.

[0069] S206: At multiple different locations on the surface of the calibration wafer, candidate calibration parameters for the phase modulator and polarization modulator are obtained respectively. The average calibration parameter of all candidate calibration parameters, or the average calibration parameter calculated after removing outliers from all candidate calibration parameters, is used as the calibration parameter for the phase modulator and polarization modulator.

[0070] In this embodiment of the invention, step S206 includes: the host computer 26 in the signal processing module controls the displacement stage to move the calibration wafer to other positions, repeating steps S201-S205. The number of other positions is one or more, usually multiple (i.e., at least two). Each time a new position is reached, the calibration parameters obtained by repeating steps S201-S205 once are called candidate calibration parameters. The host computer 26 in the signal processing module obtains the candidate calibration parameters (e.g., phase delay and waveplate angle at all different positions) of the phase modulator 18 and polarization modulator 19 at all different positions. The host computer 26 in the signal processing module performs statistics on the recorded data. The host computer 26 can calculate the average calibration parameter of all candidate calibration parameters as the calibration parameters of the phase modulator and polarization modulator; or, the host computer 26 can use the average calibration parameter calculated after removing outliers from all candidate calibration parameters as the calibration parameters of the phase modulator and polarization modulator. In this embodiment, taking the removal of outliers as an example, if a certain data point reaches or exceeds a preset multiple (the preset multiple is greater than or equal to two), such as twice the data standard deviation, the point can be considered an outlier. After removing the data outliers, the host computer 26 in the signal processing module averages the phase delay and waveplate angle obtained from testing at multiple locations. The calculated average angle and average delay are the calibration parameters for the background material.

[0071] After steps S201-S206 are completed, the defect detection system has obtained the average angle and average delay for the calibrated wafer, which can be considered as setting the system to the highest sensitivity state for the background material of the wafer under test. Subsequently, the states of each modulation device in the light collection path are maintained, i.e., the calibration parameters of phase modulator 18 and polarization modulator 19 are kept constant, and scanning imaging of the wafer under test begins.

[0072] In other embodiments, step S206 can be omitted, and calibration can be performed based on one location on the wafer surface region of the calibration wafer to obtain the calibration parameters of the phase modulator and polarization modulator, instead of calibrating based on multiple locations on the wafer surface region of the calibration wafer.

[0073] Steps S203-S206 can be understood as a specific description of the content in step S1, and step S1 includes steps S203-S206.

[0074] S207: Fix the phase modulator and polarization modulator to the corresponding calibration parameters, and send a scanning strategy for scanning the test area of ​​the wafer under test; generate an image based on the differential signal obtained from scanning the wafer under test, and locate the defects on the surface of the wafer under test based on the abnormal areas in the image.

[0075] In this embodiment of the invention, the polarization modulator includes a first half-wave plate and a rotary motor. The host computer 29 can send calibration parameters to the rotary motor to fix the first half-wave plate to the corresponding calibration parameters.

[0076] In this embodiment of the invention, generating an image based on the differential signal obtained from scanning the wafer under test includes: S301: Obtain the trigger signal, which includes multiple trigger pulses; S302: Acquire the differential signal according to the trigger pulse time to obtain the amplitude corresponding to the trigger pulse time; S303: The amplitude is split and recombined according to the image size preset in the scanning strategy to form a two-dimensional image.

[0077] The trigger signal, the amplitude corresponding to the trigger pulse time, and the image size have been described in the previous text and will not be repeated here. The execution subject of step S301 can be understood as analog-to-digital converter 25, the execution subject of step S302 can be understood as analog-to-digital converter 25, and the execution subject of step S302 can be understood as host computer 26. In step S302, the differential signal is the output signal of the differential detector obtained by scanning at the scanning position of the wafer under test according to the scanning strategy.

[0078] If the wafer under test is not the same wafer as the calibration wafer, the calibration wafer is removed from the wafer and then the wafer under test is mounted onto the displacement stage. The wafer is then fixed in place by a chuck on the displacement stage (such as a mechanical chuck or a vacuum chuck). The chuck can be used to fix either the calibration wafer or the wafer under test, which will not be elaborated further here.

[0079] For the wafer under test, the control stage performs step scanning. After scanning all positions, the host computer 26 can synchronize with the amplitude of the differential signal from the differential detector 23 and the trigger pulse time of the trigger signal to obtain an image. When there are bright spots (i.e., abnormal areas) in the image, it indicates that the image contains defects, and that the wafer under test contains defects. In this embodiment, the area to be tested on the wafer contains defects, hence the bright spots in the image.

[0080] The host computer 26 in the signal processing module locates defects based on the positions of high-brightness areas in the image. Defect location based on high-brightness areas is a current technology. Defect identification methods can be used to locate defects. This can be achieved by comparing the image with a standard image (the image of the calibrated wafer, without high-brightness areas) to determine the defect location, or by comparing the image with a preset threshold. Areas in the image above the threshold are identified as defect locations. The host computer 26 in the signal processing module performs defect detection on the image to determine the defect location. When calibrating the wafer as the wafer under test, an image of a defect-free area of ​​the wafer under test can be selected as the standard image, and images of other areas of the wafer under test can be compared with the standard image.

[0081] The following describes the working process of the defect detection system. The defective wafer being inspected is called the test wafer. The wafer used in the above calibration process can be a calibration wafer with the same surface material as the wafer to be scanned, based on the actual working state. As described above, the calibration wafer can be the same wafer as the wafer under test or a different wafer. It is only necessary to ensure that the surface material of the calibration wafer in the calibration process is the same as the surface material of the wafer under test in the subsequent actual working process. If the two are the same wafer, the host computer 26 in the signal processing module in step S207 does not need to control the wafer loading mechanism to load the wafer again. If the two are different wafers, the host computer 26 in the signal processing module needs to control the wafer loading mechanism to load the wafer under test onto the displacement stage. Taking the latter as an example, the host computer 26 in the signal processing module controls the wafer under test to be aligned by the front-end device (EFEM) in the defect detection system and then sent onto the displacement stage. The displacement stage can vacuum adsorb the wafer under test and drive it to move to achieve scanning. The scanning method of the displacement stage can be snake scanning or raster scanning. In this embodiment of the invention, snake scanning is used as an example. At this time, the displacement stage drives the wafer to perform snake reciprocating motion. The illumination beam is focused onto wafer 12 (which is now the wafer being measured) after passing through converging lens 11. x After a row of scan positions in the direction has been completely scanned, y Step a certain distance in the direction, then repeat the process for the next row. x The scanning proceeds along a single line of scan positions in a specific direction until all scan positions are completed. These scan positions are pre-planned locations within the test area of ​​the wafer under test. The number of scan positions is multiple and depends on the specific needs; for example, the number of scan positions should ideally cover the entire surface of wafer 12. The displacement stage is capable of... x direction and y Directional movement, x direction and y The plane formed by the direction is usually a horizontal plane. Figure 1 The meaning shown x direction,y direction and z The directions are perpendicular to each other.

[0082] Unlike traditional imaging methods that directly output images, the images in this embodiment are analog images generated by signal reconstruction. The amplitude of the DC signal output by the lock-in amplifier 24 corresponds to different scanning positions of the wafer under test in time. The two are correlated through the trigger pulse timing of the trigger signal, with each trigger pulse timing corresponding to a scanning position and an amplitude. For example, the displacement stage moves the wafer under test... x The reciprocating motion in the direction, after each scan x After the first line of scan position in the direction, in y Make a tiny stepping motion in the direction, so as to... y Stepping a certain distance in the direction. The pixel size of the final output image corresponding to the actual physical space can be preset in the scanning strategy. The pixel size includes... x Pixel size in direction and y Pixel size in direction, x Pixel size in direction and y The pixel size in each direction is usually equal. y The tiny step distance in the direction corresponds to the actual physical size of this pixel. This is performed on the translation stage. x During the directional scanning process, the displacement stage controller immediately outputs a high-level signal for each distance in the actual physical space corresponding to the size of the pixel moved by the displacement stage. This signal is a trigger signal and is connected to the analog-to-digital converter 25.

[0083] Taking the scanning process above as an example: First, set the area to be scanned on the surface of the wafer to be measured as a rectangle. Figure 13 This rectangle is shown in the diagram, with coordinates ranging from X. min ~X max Y min ~Y max The output image has M×N pixels, where M and N are both greater than or equal to 3. The scanning strategy also includes step sizes in the x and y directions. The step size is used to move the incident beam from one scanning position to another. The step size can be understood as the movement step size of the displacement stage. In the x direction, X... min ~X max The distance is L x The step size in the x-direction is L. x / M, in the y direction, Y min ~Y max The distance is L y The step size in the y-direction is L. y / N. X min ~X maxThis distance is divided into x0, x1, x2, x3...x M There are M+1 coordinate points, each of which can be understood as a scanning position. The displacement stage starts moving from position x0. When it reaches the next position (currently x1), the displacement stage controller sends a trigger signal, and so on, until it reaches position x1. M When the position is reached, M trigger signals are emitted, meaning that the M+1 scan positions in a row correspond to the M trigger signals. At this point, one row of scanning is completed, and the y-axis steps by one step, moving the y-coordinate from y0 to y1. During this time, the displacement stage controller does not output trigger signals. After reaching position y1, the displacement stage moves in the opposite direction along the x-axis, from x... M The position moves to x0, and M trigger signals are issued again. After all position points have been scanned, the displacement stage controller outputs a total of M×N signals, corresponding to M×N pixel data in the final image. It should be noted that although the scanning range is X... min ~X max However, since this embodiment of the invention uses a serpentine scan, not every point in the x-direction is recorded. The data is divided into M steps, X... min ~X max The coordinates of the point within the range are x0 (that is, X). min x1, x2...x M (that is, X) max When scanning odd-numbered rows (1st, 3rd, 5th, etc.), x0 is the starting point of that row. No trigger signal is output at this point, and no data is recorded at the position corresponding to x0. The actual recorded positions for the M pixels are x1 to x2. M When scanning even-numbered rows (2nd, 4th, 6th...), x M This is the starting point of the line; no trigger signal is output at this point. M The corresponding positions are not recorded; the actual positions recorded for M pixels are x. M-1 ~x0. It should be noted that this embodiment uses a serpentine scanning path as an example to describe how the host computer 26 obtains the amplitude corresponding to the trigger pulse moment, but it is only an example and is not limited thereto.

[0084] In this embodiment of the invention, the laser remains on during the scanning process of the sample driven by the displacement stage. The signal generator 3 generates a 30MHz square wave signal and connects it to the beam modulator 2 (e.g., an electro-optic modulator, EOM), applying it to the EOM via a power amplifier. The power amplifier is located between the signal generator 3 and the beam modulator 2, and the illumination optical path includes this power amplifier. In this embodiment of the invention, the differential detector 23 continuously outputs a differential signal to the lock-in amplifier 24. The lock-in amplifier 24 continuously outputs a DC signal by comparing the reference signal of the photodetector 5 with the actual received differential signal from the differential detector 23. The amplitude of the DC signal is related to the amplitude of the differential signal output by the differential detector 23 and is transmitted to the analog-to-digital converter 25 via a signal line. The trigger terminal of the analog-to-digital converter 25 is connected to the trigger signal issued by the displacement stage controller. When the analog-to-digital converter 25 receives a high-level trigger signal from the displacement stage controller, it begins recording the signal output by the lock-in amplifier 24 and transmitting it to the host computer 26. The signal architecture of this embodiment of the invention can be referred to... Figure 14 Because the high-level trigger signal of the displacement stage controller is only emitted after the displacement stage has moved by the size corresponding to the pixel size in the actual physical space, the trigger signal can be regarded as a continuous pulse signal in time. The trigger signal includes multiple trigger pulses. For the analog-to-digital converter 25, data is recorded only when a trigger pulse occurs. The analog-to-digital converter 25 records data in response to the trigger pulse, and the data obtained by the host computer 26 is discontinuous. Each segment of the signal output by the lock-in amplifier 24 represents the surface information of the wafer under test at that position. The output segment of the lock-in amplifier 24 means that the input of the lock-in amplifier 24 is continuous, and the output signal of the lock-in amplifier 24 is a numerical value that represents the continuous input over a period of time. The lock-in amplifier 24 is used to take the average of the product of the continuous input values ​​over a period of time and the reference signal as the output value of the lock-in amplifier 24. This is an inherent function of the lock-in amplifier. After the stage moves with the wafer under test and is illuminated by the laser spot across all scanning positions, the scanning process is complete. At this point, the analog-to-digital converter 25 receives M×N trigger pulses, where M×N represents the number of pixels on the wafer surface divided according to the actual physical space corresponding to the pixel size. Simultaneously, the analog-to-digital converter 25 also transmits M×N data points to the host computer 26. After processing, the host computer 26 obtains M×N surface information data representing different positions. Because the stage moves in a sequential order over time, arranging the M×N data points according to their actual spatial positions allows for the reconstruction of the two-dimensional surface information of the wafer under test, i.e., an image, which can be called an analog image. A schematic diagram of image reconstruction can be found in [reference needed]. Figure 15 .

[0085] The above describes a method where a displacement stage moves the wafer while the beam remains stationary to complete the scanning of the wafer under test. To improve scanning speed, this embodiment of the invention can achieve scanning by keeping the displacement stage stationary while moving the incident beam through a scanning galvanometer. The illumination optical path in the defect detection system also includes a scanning galvanometer and a galvanometer controller. The scanning galvanometer is located on the output optical path of the polarization modulation module 7 and is used to control the propagation direction of the second illumination beam, which serves as the incident beam. The galvanometer controller sends a trigger signal to the analog-to-digital converter 25 based on the angle of the scanning galvanometer, so that the analog-to-digital converter 25 synchronously acquires the amplitude of the DC signal output by the lock-in amplifier 24. Please refer to the relevant description above; it will not be repeated here. In this scenario, Figure 1 The second reflecting mirror 10 can be replaced with a scanning galvanometer and used in conjunction with a galvanometer controller, but it is not limited to this. The scanning galvanometer can also be located between the second reflecting mirror 10 and the converging lens 11.

[0086] For example, unlike the above, the stage controller does not issue a trigger signal; instead, the galvanometer controller issues the trigger signal. During scanning, the galvanometer controller issues a trigger signal to the analog-to-digital converter 25 based on the angle of the scanning galvanometer. The output of the analog-to-digital converter 25 is connected to the input of the host computer 26. The analog-to-digital converter 25 obtains the amplitude of the DC signal at the output of the lock-in amplifier 24 based on the trigger signal. The host computer 26, based on the correspondence between the trigger pulse timing and amplitude in the trigger signal, splits and reassembles the one-dimensional amplitude according to the preset image size in the scanning strategy to form a two-dimensional image. The trigger signal includes multiple trigger pulses. The galvanometer controller controls the scanning galvanometer to change an angle and issues a trigger pulse. At this time, the position of the light spot illuminating the wafer is also... x The direction moves from x0 to x1. Similarly, each trigger pulse corresponds to a scanning position and an amplitude. After the galvanometer controller controls the scanning galvanometer to rotate by M angles, the spot position moves exactly from x0 to x1. M The galvanometer controller outputs M trigger pulses, at which point the displacement stage steps in the y-direction. The difference between this and the displacement stage scanning method is that in the displacement stage scanning process, the beam remains stationary while the wafer under test moves, whereas in the scanning galvanometer method, the beam moves while the wafer under test remains stationary.

[0087] When a light beam strikes the surface of a wafer, the intensity of the reflected beam is related to the incident angle θ and the refractive index n of the wafer surface material. The intensity of the reflected beam is determined by the following relationship, and the formula for reflectivity is given in the prior art, for example: When the material is metal or silicon, its refractive index can be expressed as: , k Let be the imaginary part of the refractive index, representing the extinction coefficient. According to current technology, the reflectivity of the material to the incident light beam at this point is: R s for S Reflectivity of light r s for S The reflectance coefficient of light, Angle of incidence for P Reflectivity of light for P The reflectance coefficient of light. And because of the reflectance coefficient... r s and r p Both contain imaginary parts, indicating that the phase of the beam changes after the reflection process, and the changes in the two beams are different.

[0088] When the material is SiO2 or other low-absorbing materials, according to existing technology, the reflectivity of the material to the incident light beam is determined by the following formula: In the above formula, Let be the refractive index of the incident medium. Let be the refractive index of the transmission medium. Angle of incidence The transmission angle is denoted as . It can be observed that, regardless of whether it is a metallic or dielectric material, the surface reflection is related to the polarization direction of the incident light beam. Therefore, in this embodiment of the invention, circularly polarized light is used to illuminate the surface of the wafer under test, which can be decomposed into . P Light and S Light has two components, and both components have the same intensity. After reflection from the wafer surface, P Light and S The intensity of the light changes, and the magnitude of the change varies, causing the polarization direction of the emitted beam to become elliptically polarized. Depending on the material properties, the major axis direction and ellipticity of the elliptically polarized light also differ. This beam can have its major axis direction adjusted by a polarization modulator 19 (e.g., a first half-wave plate). When the major axis direction of the elliptically polarized light aligns with the polarization beam splitter 20... P When the polarization directions are the same, the two components split by the polarization beam splitter 20 have the greatest intensity difference, and at this time, the differential detector 23 outputs the strongest (i.e., maximum) signal; when the polarization direction of the linearly polarized light is the same as that of the polarization beam splitter 20, the polarization beam splitter 20 outputs the strongest (i.e., maximum) signal. PWhen the direction is 45°, the two components split by the polarizing beam splitter 20 have the same intensity, and the output of the differential detector 23 is 0. Therefore, in this embodiment of the invention, the background material can be calibrated by adjusting the angle of the first half-wave plate, so that the output of the differential detector 23 is 0. First, the angle of the first half-wave plate is set so that the output of the differential detector 23 is 0 when it is targeting a certain material (such as SiO2 or Si). When there is a material change at other locations on the surface of the wafer being measured, such as changes in the height of the wafer surface pattern or changes in the refractive index due to changes in the doping concentration on the wafer surface, the differential detector 23 will immediately output a non-zero signal. This situation is very sensitive to changes in refractive index, and theoretically, it can detect refractive index differences of up to 10. -5 The following are minute refractive index changes. This highly sensitive polarization detection method has significant advantages for detecting doping concentration in wafer manufacturing processes and for detecting minute pattern defects in high-node processes. When the two input components (i.e., P Light and S When the intensity of the light is the same, the output of the differential detector 23 is 0; when the intensities of the two input components entering the differential detector 23 are different, the intensity of the differential signal output by the differential detector 23 is related to the energy difference between the two input components.

[0089] Example 1: A scenario for detecting photoresist residue on wafers. During wafer fabrication, multiple processes are involved in transferring patterns from a photomask to the wafer surface. Photolithography is the method chosen by the vast majority of wafer fabrication manufacturers. It involves developing a pattern from the photomask onto a photoresist surface, altering the chemical properties of the photoresist. This photoresist is then removed using solvents or plasma ablation. However, photoresist residue often remains during processing due to changes in its properties or varying surface adhesion, posing a potential hazard for subsequent manufacturing processes. This invention provides highly sensitive detection of such defects.

[0090] In one embodiment, the example illustrates a scenario where a layer of photoresist approximately 5 nm thick remains on the Si etched surface, and the photoresist has a size of approximately 10 nm. Figure 16 and Figure 17 As shown. This situation is already quite demanding in actual manufacturing processes. Due to the extremely thin photoresist, neither bright-field nor dark-field detection designs can image it, and existing equipment can hardly detect this type of photoresist residue. In this embodiment of the invention, to address this detection scenario, an additional beam expander can be added to the illumination optical path to adjust the size of the laser beam, making the spot size illuminating the surface of the wafer under test approximately 1 μm. The beam expander is located between the light source 1 and the beam modulator 2. When the beam irradiates the dielectric surface and is reflected, its... P andS The reflection coefficients of the two components follow the following relationship: in, , For Si surfaces For photoresist Therefore, when a light beam illuminates a Si surface free of residue, , The reflectivity Rs = 0.791 and Rp = 0.268 can be calculated; the phase delay of the reflected beam is: φ s =-176.381°, φ p =21.340°. The emitted beam at this point is elliptically polarized with an ellipticity of 1.8696; the major axis of the ellipse is perpendicular to... S The included angle between the polarization directions is 14.3113°, and the polarization state of the emitted beam is as follows: Figure 7 As shown. At this point, the first half-wave plate can be placed at 29.65°, which will change the major axis direction of this elliptically polarized light to 45 degrees. After being split by the polarizing beam splitter 20, the two components have the same intensity, and the output of the differential detector 23 is 0. When the beam illuminates a surface with residual Si, , The reflectivity of the reflected light beam Rs =0.2817, Rp =0.0081; At this point, the emitted beam is elliptically polarized, with the major axis of the ellipse aligned with... S The polarization directions are the same, but the ellipticity is 5.8728. Under the influence of the first half-wave plate, the major axis of the ellipse of the beam passing through the first half-wave plate becomes 59.3°. The energy ratio of the two beams after being split by the polarizing beam splitter 20 is approximately 0.6. When the power of the incident beam is 0.1W, if the detected range consists entirely of residual photoresist, the power values ​​detected by the differential detector 23 are 0.0461W and 0.0274W, respectively. The differential detector 23 can be selected from existing models. For example, a certain existing differential detector at the lowest gain (10 3 The output voltage is 3V, and the background voltage is 0V, which is sufficient to distinguish the residual photoresist. Considering that the photoresist size is 10 nm and the spot size is 1 μm, approximately 1 / 10000 of the collected beam carries information about the photoresist. Therefore, when the spot covers the area with residual photoresist, the amplitude of its output differential signal is approximately 0.3 mV. This corresponds to the noise equivalent power (NEP) of the differential detector 23 (NEP = 123 pW / Hz). 1 / 2The signal still maintains a relatively high signal-to-noise ratio. This signal is received by the lock-in amplifier 24. The reference signal is a beam signal that has not passed through the wafer surface, generated by the high-speed photodetector 5. Therefore, when the reference signal enters the reference terminal (ref), the lock-in amplifier 24 can easily identify minute signal changes through phase-locking and further amplify the output DC signal. Calculating based on a gain of only 1dB, the signal output by the lock-in amplifier 24 can increase the original 0.3mV signal to 3mV. This signal enters the analog-to-digital converter 25 and is acquired by the host computer 26.

[0091] At this point, the simulated image of the wafer surface is as follows: Figure 17 As shown. Therefore, for residual photoresist on silicon, even if the thickness is 5 nm and the size is 10 nm, this system can still highlight the residue by adjusting the angle of the phase modulator 18 (e.g., a liquid crystal phase retarder) and the first half-wave plate to set the differential signal corresponding to the background material to 0, thereby achieving high-sensitivity detection of residual photoresist. Furthermore, even higher sensitivity can be achieved by increasing the beam power density and increasing the gain of the differential detector 23.

[0092] Example 2: High-sensitivity detection of patterned defects on patterned wafer surfaces Wafer manufacturing involves multi-layered structures, most of which have patterns. In another embodiment, an example scenario involves a SiO2 film layer covered with a metal layer (TaO) with a specific pattern. During actual manufacturing, patterns may appear in unwanted locations, or the patterns may be distorted. Using the defect detection system of this invention, the signal is first zeroed out for the SiO2 material, and then a scan is performed on the planned scanning positions across the entire wafer. This results in a high-intensity signal at the TaO location, which, after pattern reconstruction, forms the actual TaO pattern. For example, the size and shape of the aperture stop 15 and the frequency domain filter 17, placed on two Fourier surfaces, can be adjusted to increase the signal amplitude of the TaO material. By comparing the image of the actual pattern with the image of the calibrated wafer, the location of the pattern defect can be found.

[0093] As can be seen from the two examples above, the embodiments of the present invention can achieve high-sensitivity detection of minute defects in the wafer manufacturing process.

[0094] It should be understood that all the embodiments described above can be combined with each other without conflict, and for any part not described in detail in a certain embodiment, please refer to the relevant description in other embodiments.

[0095] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, combinations, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A wafer surface defect detection system based on polarized light, characterized in that, include: The illumination optical path is used to generate polarized light that illuminates the wafer surface at an oblique incidence. The optical path for collecting reflected light beams from the wafer surface includes a phase modulator, a polarization modulator, a polarization beam splitter, and a differential detector arranged sequentially. The phase modulator is used to modulate the reflected beam. P Light and S The phase difference of light forms a first modulated beam, and the polarization modulator is used to adjust the polarization direction of the first modulated beam to form a second modulated beam. The second modulated beam is split by the polarization beam splitter. P Light and S The differential detector is used to receive light. P Light and S The two optical paths output differential signals. The signal processing module is used to obtain the calibration parameters of the phase modulator and the polarization modulator, wherein the calibration parameters are parameters that make the differential signal corresponding to the wafer surface region of the calibration wafer zero; The signal processing module is further configured to fix the phase modulator and the polarization modulator to the corresponding calibration parameters, send a scanning strategy for scanning the test area of ​​the wafer under test, generate an image based on the differential signal obtained by scanning the wafer under test, and detect defects on the surface of the wafer under test based on the image, wherein the material of the test area is the same as the material of the wafer surface area of ​​the calibration wafer.

2. The defect detection system according to claim 1, characterized in that, The phase modulator includes a liquid crystal phase delayer or a photoelastic modulator; And / or, the polarization modulator includes a first half-wave plate.

3. The defect detection system according to claim 1, characterized in that, The illumination optical path includes a light source, a beam modulator, and a polarization modulation module arranged sequentially. The light source emits a linearly polarized first illumination beam; the beam modulator is used to modulate the presence or absence of the first illumination beam to form a second illumination beam; the polarization modulation module is used to adjust the polarization state of the second illumination beam to obtain circularly polarized light or linearly polarized light with a polarization state different from that of the first illumination beam.

4. The defect detection system according to claim 3, characterized in that, The polarization modulation module includes a second half-wave plate and a quarter-wave plate; Alternatively, the polarization modulation module may include an optical rotator.

5. The defect detection system according to claim 1, characterized in that, The signal processing module includes a lock-in amplifier, an analog-to-digital converter (ADC), and a host computer. The signal terminal of the lock-in amplifier is connected to the output terminal of the differential detector. The reference signal of the lock-in amplifier originates from the electrical signal corresponding to the modulated beam in the illumination optical path. The input terminal of the ADC is connected to the output terminal of the lock-in amplifier, and the output terminal of the ADC is connected to the input terminal of the host computer. The ADC obtains the amplitude of the DC signal at the output terminal of the lock-in amplifier according to the trigger signal. The host computer, based on the correspondence between the trigger pulse time in the trigger signal and the amplitude, splits and reassembles the one-dimensional amplitude according to the preset image size in the scanning strategy to form a two-dimensional image.

6. The defect detection system according to claim 5, characterized in that, It also includes a wafer displacement system, which includes a displacement stage and a displacement stage controller. During the scanning process, the displacement stage controller sends a trigger signal to the analog-to-digital converter according to the moving position of the displacement stage, so that the analog-to-digital converter can synchronously acquire the amplitude of the DC signal. Alternatively, the illumination optical path includes a light source, a beam modulator, and a polarization modulation module arranged sequentially. The light source emits a linearly polarized first illumination beam; the beam modulator modulates the presence or absence of the first illumination beam to form a second illumination beam; the polarization modulation module adjusts the polarization state of the second illumination beam to obtain circularly polarized light or linearly polarized light with a polarization state different from that of the first illumination beam; the defect detection system further includes a scanning galvanometer and a galvanometer controller, the scanning galvanometer being located on the output optical path of the polarization modulation module and used to control the propagation direction of the second illumination beam, the galvanometer controller sending a trigger signal to the analog-to-digital converter according to the angle of the scanning galvanometer, so that the analog-to-digital converter synchronously acquires the amplitude of the DC signal.

7. The defect detection system according to claim 1, characterized in that, The optical path for collecting light also includes a frequency domain filtering component, which is located before the phase modulator. The frequency domain filtering component includes a frequency domain filter located on the conjugate Fourier surface of the optical path for selectively transmitting specific spatial frequency components of the reflected light beam on the conjugate Fourier surface.

8. The defect detection system according to claim 7, characterized in that, The optical path for collecting light also includes an aperture stop and a relay mirror group. The aperture stop is located on the Fourier surface of the optical path for collecting light, and the aperture stop and the frequency domain filter are conjugate with respect to the relay mirror group.

9. A method for detecting defects on a wafer surface, characterized in that, The method is based on the defect detection system as described in any one of claims 1-8 to detect defects on the surface of the wafer under test, and includes the following steps: S1: Using the calibration wafer, obtain the calibration parameters of the phase modulator and the polarization modulator; S2: Fix the phase modulator and the polarization modulator to the corresponding calibration parameters, and send a scanning strategy for scanning the test area of ​​the wafer under test; S3: Generate an image based on the differential signal obtained from scanning the wafer under test, and locate the defects on the surface of the wafer under test based on the abnormal areas in the image.

10. The method according to claim 9, characterized in that, Step S1 includes: Multiple different phase delay values ​​of the phase modulator are obtained, the phase delay values ​​being values ​​within the range of a first phase value to a second phase value. The differential signal corresponding to each phase delay value is obtained using the calibration wafer and the differential detector. The phase delay value that minimizes the amplitude of the differential signal is found and used as the calibration parameter of the phase modulator. The difference between the first phase value and the second phase value is greater than or equal to π. Then, the polarization modulator is controlled to rotate at a preset angle to find the first angle at which the amplitude of the differential signal is zero, which is used as the calibration parameter of the polarization modulator.

11. The method according to claim 10, characterized in that, Before obtaining multiple different phase delay values ​​of the phase modulator, step S1 further includes: The polarization modulator is controlled to rotate at a preset angle to find and fix a second angle that maximizes the amplitude of the differential signal.

12. The method according to claim 9, characterized in that, Prior to step S1, the defect detection method further includes: Pre-align the phase modulator; Move the calibration wafer to the position where the illumination spot converges, and adjust the height of the calibration wafer to position it at the position where the illumination spot converges the least.

13. The method according to claim 9, characterized in that, The step of generating an image based on the differential signal obtained from scanning the wafer under test includes: Acquire a trigger signal, wherein the trigger signal includes multiple trigger pulses; The differential signal is acquired based on the trigger pulse time to obtain the amplitude corresponding to the trigger pulse time; The amplitude is split and recombined according to the preset image size in the scanning strategy to form a two-dimensional image.

Images