Defect inspection device

Abstract

The present invention provides a defect inspection device comprising: a stage portion that holds and moves a sample; an illumination unit that irradiates the sample with light emitted from a light source and forms a beam spot for scanning the sample; a detection unit that detects scattered light from the beam spot and photoelectrically converts a scattered light amount; and a signal processing unit that processes the photoelectrically converted electric signal. The detection unit includes a first scattered light detection optical system and a second scattered light detection optical system that respectively detect scattered light from a first illumination region and a second illumination region in the beam spot. The exposure time of an optical sensor included in the first scattered light detection optical system is longer than the exposure time of an optical sensor included in the second scattered light detection optical system.

Description

Defect Inspection Equipment

[0001] The present invention relates to a defect inspection apparatus that inspects the surface of a sample and outputs the position, type, size, etc. of a defect.

[0002] In manufacturing lines for semiconductor substrates, thin-film substrates, and other devices, surface defects are inspected to maintain and improve product yield. As circuit patterns on semiconductor substrates become increasingly finer, the size of defects that affect product yield also becomes smaller, necessitating improved sensitivity in inspection equipment. Defect inspection using optical inspection equipment detects scattered light from defects, and the amount of scattered light is proportional to approximately the sixth power of the defect size. As defect size decreases, the amount of detectable reflected light decreases dramatically. Therefore, for high-speed detection of small defects, improving detector sensitivity and illumination power density is essential to detect small amounts of light. However, increasing detector sensitivity and illumination power density can saturate the detection signal for relatively large defects, making it impossible to properly measure the amount of scattered light from defects. In other words, improving illumination and detection systems for detecting small defects lowers the upper limit of the defect size at which scattered light can be measured without saturation. Generally, larger defects have a greater impact on product quality. Saturation of the detection signal makes it difficult to estimate defect size and classify defect types. Therefore, defect inspection requires detecting both minute defects and relatively large defects without saturating their signals. Furthermore, irradiating foreign particles with illumination can heat them and cause them to explode (particle ablation, see Patent Document 1). Increasing illumination power is effective for detecting minute defects, but this increases the amount of heat, making foreign particles on the sample surface more likely to explode. It is also known that larger foreign particles are more likely to explode. Since explosions of foreign particles contaminate the sample surface, some method is required to detect foreign particles of a size that may explode before high-power-density illumination is applied, and to reduce the power density irradiated on the foreign particles. Patent Document 2 uses an optical modulation element to reduce the illumination power for large-diameter foreign particles detected in advance.

[0003] In Patent Document 3, when the amount of scattered light is greater than the linear region of the detector and photoelectric conversion becomes nonlinear, both minute defects and relatively large defects are detected by correcting the nonlinearity.

[0004] JP-T-2023-544498A JP-A-2018-71970A WO 2019 / 159334

[0005] To detect smaller defects, a longer integration time for the detector output is advantageous. This is because the longer the integration time and the fewer readout times, the more the effects of readout noise and other factors are suppressed. However, as the integration time becomes longer, the detection signal becomes more likely to saturate, lowering the upper limit of the defect size for which the amount of scattered light can be measured without saturation.

[0006] An object of the present invention is to provide a defect inspection apparatus capable of detecting defects ranging from minute defects to relatively large defects with high sensitivity.

[0007] In order to achieve the above-mentioned object, the present invention provides a defect inspection device comprising a stage unit that holds and moves a sample, an illumination unit that irradiates light emitted from a light source onto the sample and forms a beam spot that scans the sample, a detection unit that detects scattered light from the beam spot and photoelectrically converts the amount of scattered light, and a signal processing unit that processes the photoelectrically converted electrical signal, wherein the detection unit includes a first scattered light detection optical system and a second scattered light detection optical system that detect scattered light from a first illumination area and a second illumination area in the beam spot, respectively, and the exposure time of an optical sensor included in the first scattered light detection optical system is longer than the exposure time of an optical sensor included in the second scattered light detection optical system.

[0008] According to the present invention, defects ranging from minute to relatively large can be detected with high sensitivity.

[0009] FIG. 1 is a schematic diagram showing the overall configuration of a defect inspection apparatus according to the present invention; FIG. 2 is a diagram showing a method for scanning an illumination spot on a sample surface; FIG. 3 is a diagram showing the arrangement and detection direction of an oblique detection unit (first scattered light detection system) according to the present invention as viewed from above; FIG. 4 is a diagram showing the arrangement and detection direction of an oblique detection unit (first scattered light detection system) according to the present invention as viewed from the side; FIG. 5 is a schematic diagram showing an example of the configuration of an oblique detection unit (first scattered light detection system) provided in a first embodiment of the present invention; FIG. 6 is a schematic diagram showing an example of the positional relationship between an illumination spot and a sensor for an oblique detection unit (first scattered light detection system) provided in a first embodiment of the present invention; FIG. 7 is a schematic diagram showing an example of the configuration of a vertical detection unit (second scattered light detection system) provided in a first embodiment of the present invention; FIG. 8 is a diagram showing an illumination spot and a measurement area on a sample surface according to a first embodiment of the present invention; FIG. 9 is a diagram showing an illumination spot and a measurement area on a sample surface according to a second embodiment of the present invention; FIG. 10 is a diagram showing an illumination spot and a measurement area on a sample surface according to a third embodiment of the present invention; and FIG. 11 is a diagram showing an illumination spot and a measurement area on a sample surface according to a fifth embodiment of the present invention. FIG. 10 is an example of a detection signal of an oblique detection unit (first scattered light detection system) provided in the defect inspection apparatus according to the present invention. FIG. 11 is a diagram showing a position change of an illumination spot due to a displacement in the normal direction of a sample surface. FIG. 12 is a diagram showing an example of a change over time in the positional relationship between a defect and an illumination spot and an example of light intensity adjustment in a second embodiment of the present invention. FIG. 13 is a diagram showing an example of a change over time in the positional relationship between a defect and an illumination spot and an example of light intensity adjustment in a third embodiment of the present invention. FIG. 14 is a diagram showing an example of a configuration of a vertical detection unit (second scattered light detection system) in a third embodiment of the present invention. FIG. 15 is a diagram showing an example of a mirror included in the vertical detection unit (second scattered light detection system) in the third embodiment of the present invention. FIG. 16 is a diagram showing the range of defect sizes measurable by the oblique detection unit (first scattered light detection system) and the leading area detection system and trailing area detection system of the vertical detection unit (second scattered light detection system) in the third embodiment of the present invention. FIG. 17 is a diagram showing an example of a configuration of a vertical detection unit (second scattered light detection system) in a fourth embodiment of the present invention. FIG. 18 is a diagram showing an example of an aperture included in the vertical detection unit (second scattered light detection system) in the fourth embodiment of the present invention. FIG. 13 is a diagram showing an example of the configuration of a vertical detection unit (second scattered light detection system) in a sixth embodiment of the present invention.

[0010] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0011] 1 is a schematic diagram showing the overall configuration of a defect inspection apparatus according to a first embodiment of the present invention. The defect inspection apparatus shown in the figure includes an illumination unit 101, a stage unit 201 on which a sample can be placed, a sample surface displacement detection system 210, a detection unit 301, a signal processing unit 401, a control unit 501, an input unit 601, and a display unit 701. A typical example of a sample 1 to be inspected by this defect inspection apparatus is a disk-shaped semiconductor silicon wafer with no pattern formed on it and a flat surface.

[0012] The signal processing unit 401 and the control unit 501 are computers, and are configured to include a CPU, an FPGA, a timer, and the like, in addition to ROM, RAM, and other memories. The signal processing unit 401 performs defect inspection and the like based on a detection signal obtained by integrating the output of the detection unit 301 for a time width of a set integration time (exposure time) for any coordinate on the sample surface. As an example, the signal processing unit 401 is assumed to be configured as a single computer that forms a unit with the main body of the defect inspection apparatus (mechanical parts that perform inspection, such as the illumination unit 101, the stage unit 201, and the detection unit 301). However, it may also be configured as multiple computers connected via a network. In this case, a server may be used as one of the multiple computers, and the server may also be included as a component of the defect inspection apparatus and the signal processing unit 401. For example, a configuration may be adopted in which a computer attached to the main body of the defect inspection apparatus acquires defect detection signals from the main body, processes the detection signals as necessary, and transmits them to a server, where processing such as defect classification is performed by the server.

[0013] The explosive foreign object detection unit 42, defect determination unit 43, and other functions of the signal processing unit 401 may be virtually realized by software, or may be realized by hardware such as electronic circuits. The signal processing unit 401 and control unit 501 may also be configured as a single computer. While FIG. 1 illustrates the explosive foreign object detection unit 42 described in the second embodiment, the explosive foreign object detection unit 42 is not necessarily required in this embodiment.

[0014] The illumination unit 101 irradiates the sample surface with light emitted from a light source to form an illumination spot 20 (beam spot) that scans the sample surface. This illumination unit 101 appropriately includes a laser light source 1011, an attenuator 1012, a beam expander 1013, an electro-optical element 1014, a polarization voltage control unit 1015, a polarizer 1016, and an illumination intensity distribution control unit 1017. The laser light beam emitted from the laser light source 1011 is adjusted to a desired beam intensity by the attenuator 1012, and adjusted to a desired beam diameter by the beam expander 1013, and is then irradiated onto the surface of the sample 1 (hereinafter referred to as the sample surface).

[0015] A Pockels cell is typically used as the electro-optical element 1014. The electro-optical element 1014 controls the polarization direction of the light input from the beam expander 1013 in response to a control voltage input from a polarization voltage control unit 1015. The polarizer 1016 separates the incident light according to its polarization direction. The combination of the electro-optical element 1014 and the polarizer 1016 controls the amount of light incident on the sample surface. The polarization voltage control unit 1015 is an example of a light amount adjustment unit that adjusts the amount of light irradiated from the laser light source 1011 onto the sample surface in response to a control signal input from the signal processing unit 401 (or a control signal input from the control unit 501 in response to the signal from the signal processing unit 401).

[0016] The optimal polarization direction and incident angle of the incident light change depending on the size of the defect to be detected and the magnitude of scattered light from minute irregularities on the sample surface that cause noise. Therefore, although not shown in Figure 1, the illumination unit 101 may be appropriately equipped with an incident light control unit that controls the polarization direction of the light incident on the sample surface or the incident angle on the sample surface.

[0017] As the laser light source 1011, in order to detect minute defects near the surface of the sample, a laser light source that emits a short wavelength (355 nm or less) ultraviolet or vacuum ultraviolet laser beam, which has a wavelength that does not easily penetrate into the interior of the sample, and has a high output of 2 W or more is used. The diameter of the emitted beam is about 1 mm. In order to detect defects inside the sample, a laser light source that emits a visible or infrared laser beam, which has a wavelength that easily penetrates into the interior of the sample, is used.

[0018] The beam expander 1013 has two or more lens groups and functions to expand the diameter of the incident parallel beam of light. For example, a Galilean beam expander equipped with a combination of concave and convex lenses is used. The beam expander 1013 is mounted on a translation stage with two or more axes, and its position can be adjusted so that its center coincides with a predetermined beam position. The beam expander 1013 is also provided with a function for adjusting the tilt angle of the entire beam expander 1013 so that the optical axis of the beam expander 1013 coincides with the optical axis of a predetermined beam. The expansion rate of the beam diameter can be controlled by adjusting the spacing between the lenses (zoom mechanism).

[0019] The illumination intensity distribution control unit 1017 includes optical elements such as an aspherical lens, a diffractive optical element, a cylindrical lens array, and a light pipe, and controls the light intensity distribution of the illumination on the sample surface.

[0020] When performing oblique incidence illumination, variations in the height of the sample surface cause variations in the position of the illumination intensity distribution and disturbances in the illumination intensity distribution due to defocusing. To prevent this, the height variation of the sample surface is measured during inspection using the sample surface displacement detection system 210, and if the height deviates from the ideal state, the deviation in the illumination intensity distribution is corrected by adjusting the height using the illumination intensity distribution control unit 1017 or the Z axis of the stage unit 201. Although not shown in Figure 1, the illumination unit 101 may include an illumination light path switching unit 1021 (Figure 7) and a normal incidence optical system 102 (Figure 7) in addition to the optical system that causes the illumination light to be obliquely incident on the sample surface, so that the illumination light can also be incident from a direction perpendicular to the sample surface.

[0021] Next, using FIG. 2 , we will explain the illumination spot 20 formed on the sample surface by the illumination unit 101 and the method by which the illumination spot 20 scans the sample surface. Assume that the sample 1 is a circular semiconductor silicon wafer. The illumination spot 20 has an illumination intensity distribution that is elongated in one direction. In this embodiment, the illumination intensity distribution is elongated in the direction indicated by the arrow R. The illumination spot 20 typically has an intensity profile with a two-dimensional Gaussian distribution. The stage unit 201 holds and moves the sample 1 and includes a translation stage, a rotation stage, and a Z-stage (none of which are shown) for adjusting the height of the sample surface. During inspection, the rotation stage rotates the sample in the circumferential direction, causing the illumination spot 20 to rotate on the sample surface in the direction indicated by the arrow θ. Simultaneously, the linear stage drives the sample linearly, causing the illumination spot 20 to move linearly on the sample surface in the direction indicated by the arrow R. Here, the distance traveled in the R direction by the linear stage while the illumination spot 20 completes one revolution in the θ direction by the rotation stage is less than the longitudinal length of the illumination spot 20 in the R direction. By combining this rotational movement and linear movement, the stage unit 201 moves the sample relative to the illumination spot 20, and the illumination spot 20 traces a spiral trajectory to scan the entire surface of the sample.

[0022] 2, the stage unit 201 may be provided with a linear stage having a movement axis (Y-axis here) that horizontally intersects with the movement axis (X-axis here) of the linear stage. In this case, the entire sample surface can be scanned along a trajectory in which the illumination spot 20 repeatedly moves in the X direction and the −X direction while being sequentially shifted in the Y-axis direction.

[0023] The configuration of the detection unit 301 will be described using FIGS. 1 and 3 to 7. The detection unit 301 detects scattered light from the illumination spot 20 and photoelectrically converts the amount of scattered light. The detection unit 301 includes a first scattered light detection optical system that detects scattered light from region 202 ( FIG. 8A ), which is the first illuminated region in the illumination spot 20, and a second scattered light detection optical system that detects scattered light from region 203 ( FIG. 8A ), which is the second illuminated region in the illumination spot 20. In this embodiment, the first scattered light detection optical system is an imaging optical system and is an oblique detection unit 32 that detects light scattered from the illumination spot 20 in a direction oblique to the normal to the sample surface. The second scattered light detection system is a focusing optical system and is a vertical detection unit 303 that detects light scattered from the illumination spot 20 in the normal direction to the sample surface. The oblique detection unit 32 includes multiple oblique detection optical systems to detect scattered light emanating from the illumination spot 20 in multiple directions.

[0024] The arrangement of the oblique detection optical system relative to the illumination spot 20 will be described using FIGS. 3 and 4. As shown in FIG. 3, in this embodiment, the oblique detection unit 32 includes multiple (four in this example) oblique detection optical systems 302a-302d. FIG. 3 shows a plan view of the arrangement of the oblique detection optical systems 302a-302d. The angle between the direction of travel of the oblique incident illumination and the detection direction of each of the oblique detection optical systems 302a-302d, as viewed perpendicular to the sample surface, is defined as the detection azimuth angle. Each detection direction of the oblique detection optical systems 302a-302d is the direction in which the optical axis (center line of each detection aperture) of each of the oblique detection optical systems 302a-302d extends. The oblique detection unit 32 includes the oblique detection optical system 302a located in front of the illumination incident direction, the oblique detection optical system 302b located behind the oblique detection optical system 302a, and the oblique detection optical systems 302c and 302d located symmetrically to the oblique detection optical system 302a-302b with respect to the illumination incident plane. For example, the oblique detection optical system 302a is installed so that the detection azimuth angle is 0 degrees or more and 90 degrees or less, and the oblique detection optical system 302b is installed so that the detection azimuth angle is 90 degrees or more and 180 degrees or less. The illumination incident plane is a plane that includes the optical axis of the illumination unit 101 and is perpendicular to the sample surface at the illumination spot 20.

[0025] Figure 4 shows a side view of the arrangement of the oblique detection optical systems 302a and 302b. The oblique detection optical systems 302c and 302d are omitted from Figure 4. The angle formed by each detection direction of the oblique detection optical systems 302a-302d with respect to the normal to the sample surface is defined as the detection zenith angle. The number and positions of the oblique detection optical systems may be changed as appropriate, without being limited to the examples shown in Figures 3 and 4.

[0026] FIG. 5 shows an example of a specific configuration diagram of the oblique detection optical systems 302a-302d. Each of the oblique detection optical systems 302a-302d has an optical axis 3029. The R and θ directions shown in FIG. 5 correspond to the R and θ directions in FIG. 2. In the oblique detection optical systems 302a-302d, scattered light generated from the illumination spot 20 is collected by the objective lens 3021, and the polarization direction is controlled by the polarization control filter 3022. The polarization control filter 3022 can be, for example, a half-wave plate whose rotation angle can be controlled by a driving mechanism such as a motor. To efficiently detect scattered light, the detection NA of the objective lens 3021 is preferably 0.3 or greater. The objective lens 3021, which is positioned close to the sample surface, has its bottom notched as necessary to prevent interference with the sample surface. The scattered light, whose polarization direction has been controlled by the polarization control filter 3022, enters the imaging lens 3023, and an intermediate image of the illumination spot 20 is formed at the position of the aperture 3024. The aperture 3024 has an opening diameter set so as to pass only light from a region of the image of the illumination spot 20 that is detected by a sensor (optical sensor) 3028. The aperture 3024 passes only the central portion of the illumination intensity distribution, where the light intensity is strong in the R direction, and blocks the regions at the beam edges where the light intensity is weak. Furthermore, the opening diameter of the aperture 3024 in the θ direction is approximately the same size as the image of the illumination spot 20 formed on the aperture 3024, thereby suppressing disturbances such as air scattering that occur when the illumination passes through air. The intermediate image formed at the position of the aperture 3024 is re-focused by the focusing lens 3025 and enters the polarizing beam splitter 3026, where it is dispersed according to the polarization direction controlled by the polarization control filter 3022. The light beams separated by the polarizing beam splitter 3026 enter imaging lenses 3027-1 and 3027-2, respectively, and are focused on the light receiving surfaces of sensors 3028-1 and 3028-2, which are photoelectric conversion units, to form an image of the illumination spot 20 on each light receiving surface. Cylindrical lenses can also be used as the imaging lenses 3027-1 and 3027-2 to form an image in only one direction.

[0027] FIG. 6 shows the relationship between the sensor 3028 and the illumination spot 20. The optical path branching by the polarizing beam splitter 3026 is omitted from FIG. 6. The linear illumination spot 20 is focused onto multiple sensors 3028 by the oblique detection optical systems 302a-302d as described above, and optical images 21 of the illumination spot 20 are formed on the light-receiving surfaces of the multiple sensors 3028. The sensors 3028 are line sensors or time-delay-integration (TDI) sensors. Examples of sensor types that can be used include a complementary metal-oxide-semiconductor (CMOS) sensor, a charge-coupled device (CCD) sensor, a silicon photomultiplier (SiPM) sensor, a photomultiplier tube (PMT) sensor, and an avalanche photodiode (APD) sensor. Typically, a line sensor is positioned so that the normal direction of its light-receiving surface coincides with the optical axis of the detection optical system. However, in this embodiment, the array-shaped light-receiving units 30281 (light-receiving surfaces) of the sensor 3028 are inclined with respect to the optical axis 3029 of each of the oblique detection optical systems 302a-302d. The light-receiving units 30281 are positioned so that the array direction is parallel to the longitudinal direction of the optical image 21. Although not shown in FIG. 5 , a pair of cylindrical lenses 30210 and 30211 are positioned between the objective lens 3021 and the imaging lens 3023. These cylindrical lenses 30210 and 30211 form a cylindrical beam expander, which reduces the spread of the optical image 21 formed by the illumination spot 20 in the short-side direction γ to be smaller than the spread of the optical image 22 in the long-side direction σ.

[0028] The sensor 3028 captures the optical image 21 and outputs it as an electrical signal. This sensor 3028 includes a light-receiving unit 30281 and an anti-reflection coating at a position one-dimensionally conjugate with the illumination spot 20 irradiated on the surface of the sample 1. The light-receiving unit 30281 is conjugate with the sample surface in the longitudinal direction σ, but not in the transverse direction γ. However, because the transverse direction γ corresponds to the transverse direction of the illumination spot 20, the image height of the optical image 21 is low and almost no defocus occurs. Therefore, by increasing the imaging magnification of the optical image 21 in the transverse direction γ, it is possible to reduce the variation in the angle of incidence on the light-receiving unit 30281.

[0029] The configuration of the vertical detection unit 303 and its relationship with the vertical incidence optical system 102 will be described using Figure 7. In the vertical detection unit 303, scattered light generated from the illumination spot 20 in a direction of a small detection zenith angle (for example, the normal direction) is collected by a collecting relay lens system 3031 including a plurality of lenses. The intermediate image of the illumination spot 20 formed by the collecting relay lens system 3031 is enlarged or reduced at a desired magnification by a magnification adjustment lens system 3032, and is focused at the position of an aperture 3033 to form an intermediate image 23. This intermediate image 23 is collected again by a collecting lens 3034 and guided to a sensor (optical sensor) 3035, which is a photoelectric conversion unit of the vertical detection unit.

[0030] Although not shown in FIG. 1 , the illumination unit 101 may include a normal incidence optical system 102 that directs light in the normal direction to the sample surface, as shown in FIG. 7 . In this embodiment, the illumination unit 101 has an illumination optical path switching unit 1021 that switches the optical path at a predetermined position, allowing for switching between oblique incidence illumination and normal illumination. The illumination optical path switching unit 1021 is composed of multiple mirrors and other components disposed in the illumination optical path. As shown in FIG. 7 , the normal detection unit 303 and the normal incidence optical system 102 share some optical elements, forming an illumination spot 20 on the sample surface in the same manner as with oblique incidence illumination. The illumination relay system of the normal incidence optical system 102 includes a lens 1022 and a mirror 1023. The lens of the condensing relay lens system 3031 closest to the sample surface is shared with the normal detection unit 303. The illumination light passing through the lens 1022 is reflected by the mirror 1023 and irradiated onto the sample surface along the same optical axis as the condensing relay lens system 3031. The mirror 1023 is movable and is inserted into the optical path only during vertical illumination.

[0031] 8A , the relationship between the region 202 on the sample surface measured by the oblique detection unit 32 and the region 203 on the sample surface measured by the perpendicular detection unit 303 in this embodiment will be described. These regions 202 and 203 can be set, for example, by the size and shape of the openings of the apertures 3024 and 3033. The region 202 measured by the oblique detection unit 32 is obtained by projecting the light receiving unit 30281 of the sensor 3028 included in the oblique detection unit 32 onto the sample surface. The region 203 measured by the perpendicular detection unit 303 is obtained by projecting the outline of the light receiving surface of the sensor 3035 included in the perpendicular detection unit 303 onto the sample surface. The region 202 measured by the oblique detection unit 32 is the center of the illumination spot 20 in the longitudinal direction R, and is determined by the aperture 3024 included in the oblique detection unit 32.

[0032] In this embodiment, the area 203 measured by the vertical detection unit 303 is determined by an aperture 3033 included in the vertical detection unit 303 so as to include the entire area 202 measured by the oblique detection unit 32. Therefore, the vertical detection unit 303 detects scattered light from the area 203 including the area 202 measured by the oblique detection unit 32. As shown in FIG. 8A , the aperture 3033 included in the vertical detection unit 303 passes only the central part of the illumination spot 20, where the illumination intensity distribution has a strong light amount, in the R direction, and blocks the areas at both ends of the beam where the light amount is weak. Furthermore, the opening diameter of the aperture 3033 in the θ direction is approximately the same size as the image of the illumination spot 20 formed on the aperture 3033, thereby suppressing disturbances such as air scattering that occur when illumination passes through air.

[0033] For small defects on the sample surface, the amount of scattered light scattered from the illumination spot 20 toward the oblique detection optical systems 302a-302d, i.e., obliquely, is greater than the amount of scattered light scattered toward the vertical detection unit 303. Therefore, the detection sensitivity of the oblique detection unit 32 is higher than that of the vertical detection unit 303. To further improve sensitivity to small defects, the integration time of the output of each sensor 3028 included in the oblique detection unit 32 is set longer than the integration time of the output of the sensor 3035 of the vertical detection unit 303. This makes it possible to suppress a decrease in S / N ratio due to readout noise, etc. On the other hand, extending the integration time increases the amount of light integrated per exposure, making it easier for the detection signal from a relatively large defect to saturate. Saturation of the detection signal makes it difficult to estimate the size and type of defect. The size and type of defect are important for estimating the impact of the defect on the final semiconductor device product.

[0034] To solve this problem, in this embodiment, the following difference is provided between the sensor 3035 of the vertical detection unit 303 and the sensor 3028 of the oblique detection unit 32.

[0035] First, in this embodiment, a laser with a higher power than that typically used for inspecting the sample 1 is used to detect scattered light from minute defects. When increasing the laser output, a pulsed laser is more advantageous than a continuous-wave laser due to the nonlinearity of wavelength conversion elements. Therefore, in this embodiment, a QCW (Quasi Continuous Wave) pulsed laser with an oscillation frequency on the order of several MHz to several tens of MHz is used as the laser light source 1011. The control unit 501 inputs a laser oscillation synchronization signal to the sensor 3035 of the vertical detection unit 303. This causes the sensor 3035 to perform sampling synchronized with the input signal and A / D conversion. Synchronization with laser oscillation reduces the influence of noise such as dark current. For example, when the rotation speed of the rotary stage of the stage unit 201 is a constant angular velocity, the movement speed of the field of view on the sample surface varies between the inner and outer peripheries of the sample surface. Because the emission time per pulse is typically on the order of picoseconds, the movement of the field of view during the emission time of one pulse can be ignored. Therefore, even if the rotation speed of the rotary stage is a constant angular velocity, the influence of changes in the movement speed of the field of view on the sample surface can be suppressed by synchronizing the sampling of the sensor 3035 with the oscillation timing of the pulsed laser. In contrast, the sensor 3028 of the oblique detection unit 32 places emphasis on detecting minute amounts of light, and is set to sample multiple pulses in one integration, that is, in one integration of the signal for one coordinate position.

[0036] In this embodiment, the integration time of the output of the sensor 3035 of the vertical detection unit 303 is set shorter than the integration time of the output of the sensor 3028 of the oblique detection unit 32. That is, the exposure time of the sensor 3028 included in the oblique detection unit 32, which is the first scattered light detection optical system, is longer than the exposure time of the sensor 3035 included in the vertical detection unit 303, which is the second scattered light detection optical system. By setting the integration time of the output of the vertical detection unit 303 shorter than the integration time of the output of the oblique detection unit 32, the defect size that can be detected by the vertical detection unit 303 without saturating the detection signal can be larger than that of the oblique detection optical system 302. As with the sensor 3028, the type of the sensor 3035 of the vertical detection unit 303 can be a CMOS sensor, a CCD sensor, a SiPM sensor, a PMT sensor, or an APD sensor. However, when sampling is synchronized with an oscillation frequency of several tens of MHz, it is difficult for a CMOS sensor or a CCD sensor to move a minute charge in accordance with the oscillation frequency.

[0037] The signal processing unit 401 includes a function (defect determination unit 43 shown in FIG. 1) for determining the size of a defect from the detection signals of the sensor 3035 and the sensor 3028. The defect determination unit 43 calculates defect information by appropriately integrating the detection signals of the vertical detection unit 303 and the oblique detection unit 32. For a relatively small defect for which the detection signal of the sensor 3028 does not saturate in all of the oblique detection optical systems 302a-302d, the detection signal of the vertical detection unit 303 is small, so the defect information is calculated using only the signal of the oblique detection unit 32. For a relatively large defect for which the detection signal of the sensor 3028 saturates in at least one of the oblique detection optical systems 302a-302d, the defect information is calculated using the detection signals of both the vertical detection unit 303 and the oblique detection unit 32.

[0038] The estimation of the defect size of a relatively large defect will now be described in detail. Defect size estimation is generally performed by comparing the ratio of the power density of light irradiated on the defect to the measured amount of scattered light with a sensitivity curve obtained by calibration using standard particles or simulation. However, because the vertical detection unit 303 is a light-collecting detection system, although the amount of scattered light can be determined based on the detection signal of the vertical detection unit 303, the spatial position information of the defect cannot be determined. Because the illumination power density distribution is a Gaussian distribution rather than a uniform distribution, information on the defect's position is required to determine the power density irradiated on the defect.

[0039] An example of the detection signal of the oblique detection unit 32 and the estimation of defect position information will be described using FIG. 9 . The horizontal axis in FIG. 9 represents the R direction, and the horizontal axis value of the detection signal (plot) of each pixel of the line sensor 3028 corresponds to the position on the sample surface. The vertical axis represents the amount of scattered light measured at each pixel. In the example of FIG. 9 , due to the large defect size, the pixels detecting scattered light from near the defect position are saturated, and the amount of scattered light cannot be directly obtained from the detection signal of the oblique detection unit 32. However, the defect position can be estimated, for example, from the pixel whose signal is saturated, as its center of gravity. In the example of FIG. 9 , three pixels near the defect position are saturated. The center of these three pixels is considered to coincide with the actual defect position with an error on the order of the pixel size in the R direction. To estimate the position information more accurately, for example, using signals from surrounding pixels that detect light but are not saturated can be expected to improve the accuracy of defect position estimation.

[0040] As described above, the vertical detection unit 303 has a short sensor integration time, so the detection signal is less likely to saturate, and it is therefore possible to measure the amount of scattered light even for relatively large defects. However, since the defect position cannot be determined from the signal from the vertical detection unit 303, it is difficult to accurately calculate the size. However, as described above, the defect position can be estimated from the detection signal from the oblique detection unit 32. This makes it possible to determine the power density distribution of the illumination light obtained in advance by optical simulation, or the power density distribution of the illumination light actually measured using a power density distribution measurement optical system (not shown), as will be described later.

[0041] By combining the detection signals of the vertical detection unit 303 and the oblique detection unit 32 in this manner, the defect position, the amount of scattered light, and the illumination power density irradiated onto the defect can be measured, and the signal processing unit 401 can accurately calculate the size of the defect.

[0042] As described above, the illumination power density distribution on the sample surface is used to calculate the illumination power density irradiated to the defect. Because the sample rotates at high speed during inspection, the position of the sample surface can fluctuate in an oscillatory manner on the order of microns in the normal direction of the sample surface. This fluctuation can cause illumination blurring and a change in the illumination power density distribution. Furthermore, when using oblique incidence illumination, as shown in FIG. 10 , the sample surface is displaced in the normal direction, causing the illumination spot 20 to displace in the R direction, which also changes the illumination power density distribution. These factors cause errors in the estimation of the power density irradiated to the defect, resulting in a large error in the calculation of the defect size. To solve this problem, the sample surface displacement in the normal direction of the sample surface during inspection is measured using the sample surface displacement detection system 210. The sample surface displacement measured by the sample surface displacement detection system 210 is input to the defect determination unit 43. The defect determination unit 43 uses pre-stored data (e.g., table data) regarding the change in illumination power density distribution for each position in the normal direction of the sample surface to calculate the power density irradiated to the defect based on the position in the normal direction of the sample surface. This makes it possible to reduce measurement errors of defect size due to displacement in the normal direction of the sample surface.

[0043] According to the present invention, the vertical detection unit 303, which is the second scattered light detection optical system, is set to have a shorter exposure time than the oblique detection unit 32, which is the first scattered light detection optical system, and therefore the detection signal is less likely to become saturated than the oblique detection unit 32, and the scattered light amount can be measured even for relatively large defects. In this way, by using multiple detection optical systems with different exposure times, the dynamic range is widened, and defects ranging from minute to relatively large can be detected with good sensitivity.

[0044] In this embodiment, the integral time of the output of the sensor 3035 included in the vertical detection unit 303 is set to be relatively short, but the sensor whose integral time is relatively short does not necessarily have to be the sensor 3035. For example, the integral time of the sensor 3028 may be set to be shorter for some (at least one) of the multiple oblique detection optical systems 302a-302d of the oblique detection unit 32 than for the other detection optical systems. In other words, some of the oblique detection optical systems can replace the vertical detection unit 303.

[0045] Second Embodiment In this embodiment, a case will be described in which an area 203 (second illumination area) on the sample surface measured by the perpendicular detection unit 303 includes an area located in the R-positive direction relative to an area 202 (first illumination area) measured by the oblique detection unit 32. It is sufficient that at least a portion of the area 203 is located in the R-positive direction relative to the area 202, and the areas 202 and 203 may partially overlap. In other words, the area 203 (second illumination area) includes a preceding area that scans the sample temporally prior to the area 202 (first illumination area).

[0046] The aperture 3033 included in the vertical detection unit 303 passes through the region on the R-positive side in FIG. 2 of the low-intensity portion of the beam end in the intermediate image 23 that is blocked by the aperture 3024 in the oblique detection unit 32. The relationship between the regions 202 and 203 in this embodiment is illustrated in FIG. 8B . As described in FIG. 2 , the illumination spot 20 scans the entire sample surface in a spiral pattern from the center of the sample outward. In other words, the region 203 is a region where weak light is irradiated at certain coordinates on the sample surface prior to the region 202. Each coordinate on the sample surface is first irradiated with weak light when crossing the region 203, and is irradiated with stronger light when crossing the region 202 on the next or subsequent revolutions. In this embodiment, the regions 202 and 203 do not overlap even partially.

[0047] This makes it possible to detect defects such as foreign particles that have the potential to explode (hereinafter sometimes abbreviated as explosive foreign particles) before they are irradiated with high-power density illumination, thereby preventing explosions. This detection of explosive foreign particles must be performed in real time during defect inspection, in which the sample 1 rotates several tens of times per second, and therefore must be processed at high speed. In this regard, the vertical detection unit 303 is a light-collecting detection system, and therefore outputs a single signal, making it suitable for high-speed processing.

[0048] The signal processing unit 401 includes an explosive foreign object detection unit 42 that detects explosive foreign objects from the detection signal of the sensor 3035 of the vertical detection unit 303. The explosive foreign object detection unit 42 extracts explosive foreign objects by comparing the detection signal acquired by the sensor 3035 with a predetermined threshold value after passing it through a high-pass filter. The explosive foreign object detection unit 42 estimates, for example, a foreign object detected by the vertical detection unit 303 whose defect exceeds a predetermined threshold value as an explosive foreign object in the form of a particle having a size equal to or larger than a predetermined value. When the vertical detection unit 303 detects an explosive foreign object in the form of a particle having a size (light amount) equal to or larger than a predetermined threshold value in the region 203, the light intensity adjustment unit (polarized voltage control unit 1015) adjusts the power density of the illumination light irradiated on the explosive foreign object in the region 202 as described below. In this case, the signal processing unit 401 can also be configured to change the threshold value based on the measurement results of the sample surface displacement input from the sample surface displacement detection system 210. By changing the threshold value in accordance with the displacement of the sample surface, it is possible to deal with changes in the illumination power density distribution due to displacement in the normal direction of the sample surface.

[0049] FIG. 11A illustrates a case where an explosive particle is detected by the explosive particle detection unit 42. FIG. 11A (1) shows an example of the change in the illumination power density irradiated onto a defect during inspection. The horizontal axis represents the R direction, and the longitudinal illumination power density distribution (Gaussian) is plotted. Furthermore, the defect positions at times t1, t2, and t3 relative to the illumination power density distribution are indicated by dotted lines. The radial movement pitch 204 is the distance the illumination spot 20 moves on the R stage while the rotating stage rotates the sample once, and varies depending on the inspection mode. A defect signal at coordinate X within region 203 detected by the vertical detection unit 303 at time t1 is input to the explosive particle detection unit 42, which determines whether the defect is an explosive particle. If an explosive particle is detected, the control unit 501 stores the defect position (coordinate X) and controls the light intensity when coordinate X is detected in region 202. FIG. 11A (2) shows an example of the temporal change in the power density irradiated onto the sample surface by the polarization voltage control unit 1015 and the electro-optical element 1014. If the signal at coordinate X detected in region 203 at time t1 is determined to be an explosive foreign particle, the control unit 501 controls the polarization voltage control unit 1015 and the electro-optical element 1014 to reduce the illumination power density at the timing (times t2 and t3) when the illumination spot 20 irradiates the defect position (coordinate X) on the next or subsequent rotations. The illumination power density is reduced to a predetermined value to prevent explosion based on the defect size estimated from the detection signal at time t1. The illumination power density of the light irradiated onto the explosive foreign particle may be a constant value, but it can also be determined based on the signal intensity detected in region 203 using, for example, data (table data) that predefines the relationship between the size of the explosive foreign particle (signal intensity detected in region 203) and the illumination power density.

[0050] The size estimation of explosive particles in this embodiment will now be described. As described in the first embodiment, the scattered light amount can be obtained from the detection signal of the vertical detection unit 303. In the example of FIG. 11A(1), the power density P(t1) irradiated to the explosive particle at time t1 can be calculated as follows using the position information of the explosive particle estimated from the saturated detection signal output from the oblique detection unit 32 at time t (t1<t): P(t1)=Pr(x(t)-Rp*N) where Pr is the illumination power density distribution in the R direction, x(t) is the position of the explosive particle estimated from the saturated detection signal measured by the oblique detection unit 32 at time t, Rp is the radial movement pitch 204, and N is the number of rotations (number of revolutions) of the sample 1 by the θ stage between time t1 and time t. Pr may be an illumination power density distribution calculated by optical simulation or may be an illumination power density distribution measured using a spot monitor optical system (not shown). Pr is a power density distribution that takes into account the amount of light source power reduction required to prevent foreign matter from exploding, as input from the explosive foreign matter detection unit 42 or the control unit 501. Furthermore, Pr may be an illumination power density distribution that takes into account changes due to specimen surface displacement at time t1, based on the measurement results of the specimen surface displacement detection system 210.

[0051] The other configurations and functions of this embodiment are similar to those of the previously described embodiment.

[0052] In this embodiment, the same effects as those of the previously described embodiments can be obtained. In addition, explosive foreign particles can be detected in region 202, which scans the sample surface in advance of region 202. Furthermore, by reducing the illumination power density irradiating the detected explosive foreign particle, it is possible to effectively suppress the explosion of the defect.

[0053] Third Embodiment In this embodiment, a case will be described in which the vertical detection unit 303 includes multiple detection systems (leading-region detection system 312 and trailing-region detection system 311) that each detect a different region on the sample surface. The relationship between the regions 202 and 203 in this embodiment is illustrated in FIG. 8C . As described in FIG. 2 , the illumination spot 20 scans the entire sample surface in a spiral pattern from the center of the sample outward. That is, the region 203 arrives at a certain coordinate on the sample surface before the region 202. However, unlike the second embodiment, in this embodiment, a portion of the region 203 overlaps the region 202. The portion of the region 203 that does not overlap with the region 202 is the leading region 2031, and the portion of the region 203 that overlaps with the region 202 is the trailing region 2032. The trailing region 2032 is a region that scans the sample temporally following the leading region 2031. In this embodiment, the following region 2032 includes the entire region 202 , but a part of the region 202 (for example, the end in the −R direction) may be outside the following region 2032 .

[0054] The configuration of the vertical detection unit 303 in this embodiment will be described using Figure 12. In the vertical detection unit 303 in this embodiment, an aperture 3041 is provided at the position of the intermediate image of the illumination spot 20 formed by the condensing relay lens system 3031. The aperture 3041 blocks light from areas of the illumination spot 20 that are not to be measured in order to prevent stray light. In this embodiment, an optical path branching mirror 3040 is provided at the position where the aperture 3033 is provided in the first and second embodiments, and the optical path branching mirror 3040 branches the optical path into the following area detection system 311 and the leading area detection system 312.

[0055] The optical path branching mirror 3040 will be described using FIG. 13 . The optical path branching mirror 3040 is placed at the position of the intermediate image 23 and includes a light-shielding portion 30401, a mirror portion 30402, and a transmission portion 30403. The transmission portion 30403 may be made of a material with high transmittance at the light source wavelength, or may simply be configured by cutting out the relevant portion so that light reaching this region passes directly through (i.e., an opening without a structure). The mirror portion 30402 reflects light from the trailing region 2032 ( FIG. 8C ) of the region 203 and guides it to the trailing region detection system 311. The transmission portion 30403 transmits light from the leading region 2031 ( FIG. 9C ), which is located on the R-positive side within the region 203, and guides it to the leading region detection system 312. The light-shielding portion 30401 blocks light from the beam end in the R-negative direction. The light-shielding portion 30401 has the effect of preventing unnecessary reduction in laser power when detecting explosive foreign matter of this embodiment.

[0056] Returning to FIG. 12 , the vertical detection unit 303 will be further described. Scattered light from the leading region 2031 passes through the optical path branching mirror 3040 and reaches the leading region detection system 312. The intermediate image 23 that passes through the optical path branching mirror 3040 is condensed by a condensing lens 3034. In this embodiment, a SiPM sensor is used for the sensor (optical sensor) 3042, which is the photoelectric conversion unit of the leading region detection system 312. The SiPM sensor can achieve a high current amplification factor, but has a dead region on the light-receiving surface, which is characterized by the fact that photons entering the dead region are not detected. In this embodiment, the intermediate image 23 condensed by the condensing lens 3034 is guided to the sensor 3042 by a microlens array 3036, reducing the number of photons entering the dead region and improving detection efficiency.

[0057] On the other hand, scattered light from the trailing region 2032 is reflected by an optical path branching mirror 3040 and guided to a trailing region detection system 311. The intermediate image 23 reflected by the optical path branching mirror 3040 is collected by a condenser lens 3037 and detected by a sensor (optical sensor) 3039, which is a photoelectric conversion unit of the trailing region detection system 311. An ND filter 3038, which transmits light incident on the sensor 3039, is configured to switch between multiple ND filters with different optical densities, so that the photon collection efficiency of the trailing region detection system 311 can be adjusted within a predetermined range.

[0058] Next, the calculation of defect size in this embodiment will be described. In this embodiment, the defect size is calculated by integrating signals from three detection systems: the leading region detection system 312, the following region detection system 311, and the oblique detection unit 32. The leading region 2031 is an area at the edge of the illumination spot 20 where the power density is low, and therefore the amount of scattered light is also small. In other words, the leading region detection system 312 can detect the largest range of defect sizes. Both the following region detection system 311 and the oblique detection unit 32 detect scattered light from the region 202 where the illumination power density is high, but due to the difference in integration time described above, the oblique detection unit 32 can detect smaller defect sizes.

[0059] The minimum defect size measurable by the leading region detection system 312 depends on the power density distribution within the leading region 2031. From the perspective of inspection throughput, it is desirable to set the region 202 as large as possible to expand the range that can be inspected in one revolution. In this embodiment, the power density of the illumination spot 20 has a two-dimensional Gaussian distribution, and the region 202 is set to a range where the illumination power density is 1 / e2 or more of the peak value. In this case, the illumination power density within the leading region 2031 is at most 13.5% of the peak value within the region 202. Depending on the size of the radial movement pitch 204, even defects that would saturate the oblique detection unit 32 may not be detected by the leading region detection system 312. The photon collection efficiency of the trailing region detection system 311 is adjusted by the ND filter 3038 so that it can detect defects of a size that would saturate the oblique detection unit 32 but that are difficult to detect with the leading region detection system 312 due to the small amount of light. In other words, the photon collection efficiency of the trailing region detection system 311 is set lower than that of the leading region detection system 312.

[0060] FIG. 14 shows the relationship between the defect sizes measurable by the three detection systems, the leading area detection system 312, the trailing area detection system 311, and the oblique detection unit 32. The horizontal axis represents defect size, indicating the range of defect sizes within which each detection system can detect a sufficient amount of scattered light and the detection signal does not saturate. Measurement ranges 1301, 1302, and 1303 represent the ranges of defect sizes measurable by the oblique detection unit 32, the trailing area detection system 311, and the leading area detection system 312, respectively. For defects of a size that saturates the oblique detection unit 32, the defect size is calculated in the same manner as in the first and second embodiments. That is, the defect determination unit 43 determines the amount of scattered light from the detection signal of the trailing area detection system 311 or the leading area detection system 312. The defect determination unit 43 also calculates the defect position from the saturated detection signal of the oblique detection unit 32 and calculates the illumination power density irradiated on the defect from the illumination profile. The defect size is calculated from the illumination power density and scattered light amount calculated in this manner.

[0061] Detection of explosive foreign matter in this embodiment will be described with reference to FIGS. 11B and 14 . In FIG. 14 , defect size 1304 is the lower limit of the defect size that may explode when the highest power density is applied within the illumination spot 20. In other words, defects larger than the defect size 1304 are explosive foreign matter, and adjustment of the illumination power is required. In this embodiment, the illumination power density in the leading region 2031 is small, and it is difficult for the leading region detection system 312 to detect defects of a size close to the defect size 1304, as shown in FIG. 14 . Therefore, two-stage explosive foreign matter detection is performed, also using the trailing region detection system 311.

[0062] For defects of a size that can be detected by the leading region detection system 312, explosive foreign object detection is performed in the same manner as in the second embodiment described in FIG. 11A . FIG. 11B shows the process for explosive foreign objects that are difficult to detect by the leading region detection system 312 and are larger than the defect size 1304. In FIG. 11B (1), assume that at time t1, an explosive foreign object is located within the leading region 2031, but is too small to be detected by the leading region detection system 312. At time t2, which corresponds to the next rotation, the explosive foreign object enters the trailing region 2032 and region 202. The explosive foreign object described here is small for an explosive foreign object. Therefore, the illumination region where the explosive foreign object is located at time t2 is a low region within the trailing region 2032 and region 202 where the power density has not yet reached its peak, and therefore the explosive foreign object does not explode at time t2. However, at time t3, which corresponds to the next rotation, a high power density is irradiated onto the explosive foreign object, which may result in the explosive foreign object exploding.

[0063] In this embodiment, an explosive foreign object that was not detected by the leading area detection system 312 at time t1 is detected by the trailing area detection system 311 at time t2. While the detection signal in the oblique detection unit 32 becomes saturated, the trailing area detection system 311 does not become saturated and is therefore able to detect explosive foreign objects. The detection signal from the trailing area detection system 311 is input to the explosive foreign object detection unit 42, which determines whether or not the object is an explosive foreign object. If it is determined to be an explosive foreign object, the control unit 501 stores the position of the explosive foreign object and controls the amount of light irradiated on the explosive foreign object from the next rotation onward.

[0064] 11B(2), the control unit 501 controls the polarization voltage control unit 1015 and the electro-optical element 1014 to reduce the illumination power density at the timing when the illumination spot 20 is irradiated onto the stored position of the explosive foreign particle from the next revolution (time t3) onwards. The illumination power density is reduced to a preset value according to the size of the explosive foreign particle estimated from the detection signal at time t2 so as to prevent the explosive foreign particle from exploding.

[0065] That is, in this embodiment, when the leading region detection system 312 detects a particle having a size equal to or greater than the first threshold in the leading region 2031, the polarization voltage control unit 1015 adjusts the power density of the illumination light irradiated onto the explosive foreign matter, which is a particle having a size equal to or greater than the first threshold, in the trailing region 2032 and region 202 in accordance with a control signal from the signal processing unit 401 input via the control unit 501, in the same manner as in the second embodiment. In addition, when the trailing region detection system 311 detects an explosive foreign matter, which is a particle having a size equal to or greater than the second threshold (a particle smaller than the particle having a size equal to or greater than the first threshold), in the trailing region 2032, the polarization voltage control unit 1015 adjusts the power density of the illumination light irradiated onto the particle having a size equal to or greater than the second threshold in region 202 in accordance with a control signal from the signal processing unit 401 input via the control unit 501. The power density of the illumination light irradiating the explosive foreign particle detected by the leading region detection system 312 and the power density of the illumination light irradiating the explosive foreign particle detected by the trailing region detection system 311 can be set to the same value if they are constant, or, for example, the latter can be set higher than the former. The power density of the illumination light irradiating the explosive foreign particle may also be varied depending on the size (light intensity) of the explosive foreign particle. The signal processing unit 401 may also be configured to change at least one of the first and second thresholds based on the measurement results of the sample surface displacement input from the sample surface displacement detection system 210. Changing these thresholds in response to the displacement of the sample surface can accommodate changes in the illumination power density distribution due to displacement in the normal direction of the sample surface.

[0066] Furthermore, in this embodiment, if the position of the explosive foreign object moves in the negative R direction in region 202 after time t3, the light-shielding portion 30401 of the optical path branching mirror 3040 prevents the scattered light from the explosive foreign object from being detected by the following region detection system 311, and laser power reduction control is not executed when the explosive foreign object crosses region 202 during subsequent revolutions. This prevents unnecessary reductions in laser power and reduces sensitivity to minute foreign objects.

[0067] In this embodiment, the other configurations and functions are the same as those of the previously described embodiments, and the same effects as those of the previously described embodiments can be obtained. Additionally, according to this embodiment, explosive foreign particles that are difficult to detect in the leading region 2031 can be detected in the trailing region 2032, thereby improving the effectiveness of suppressing the explosion of explosive foreign particles. In this regard, by setting the photon collection efficiency of the trailing region detection system 311 lower than that of the leading region detection system 312, defects of a size that would saturate the oblique detection unit 32 but are difficult to detect in the leading region detection system 312 due to the small amount of light emitted, can be more easily detected by the trailing region detection system 311. The effect of the light-shielding portion 30401 of the optical path branching mirror 3040 prevents an unnecessary decrease in laser power and prevents a decrease in sensitivity to minute foreign particles.

[0068] Fourth Embodiment In this embodiment, the configurations of the vertical incidence optical system 102 and the vertical detection unit 303 are partially different. An example of the configuration of the vertical detection unit 303 and the vertical incidence optical system 102 in this embodiment is shown in Fig. 15. In Fig. 15, elements that are the same as or correspond to those in Fig. 7 or Fig. 12 are given the same reference numerals as in Fig. 7 or Fig. 12, and descriptions thereof will be omitted.

[0069] The optical axes of the normal incidence optical system 102 and normal detection unit 303 shown in Figure 15 are branched by a half mirror 1024. When the sample 1 is inspected using normal incidence light with this configuration, the illumination light specularly reflected from the sample surface is collected by the light collecting relay lens system 3031, and the specularly reflected light becomes noise, hindering the detection of minute scattered light from defects. Therefore, in the configuration example of Figure 15, the normal detection unit 303 additionally has a spatial filter (light intensity adjustment aperture) 3043 installed at the pupil plane position to block the specularly reflected light.

[0070] The spatial filter 3043 will be described using FIG. 16 . In FIG. 16 , X and Y are Cartesian coordinate systems on the pupil plane, corresponding to the θ and R directions in FIG. 2 . Because the illumination light is narrowed into a thin line on the sample surface with the θ direction as its short side, the specularly reflected light of the illumination spreads widely in the X direction, which corresponds to the θ direction, on the pupil plane. The light-shielding portion 30431 of the spatial filter 3043 is a rod that shields a rectangular region that is long in the X direction on the pupil plane. The light-shielding portion 30431 is configured to be switchable to rods of different thicknesses, allowing the width of the light-shielding portion 30431 in the Y direction to be changed. This makes it possible to adjust the photon collection efficiency of the vertical detection portion 303 during vertical illumination.

[0071] In this embodiment, other configurations, functions, and effects are the same as those of the previously described embodiment.

[0072] Fifth Embodiment In this embodiment, a case will be described in which the illumination spot 20 in the second embodiment is a multi-spot, and the regions 202 and 203 are different illumination spot regions. For example, an optical element such as a beam splitter is incorporated into the illumination unit 101, and multiple illumination spots can be formed by forming multiple parallel beams from the light emitted from the laser light source 1011. The power density of the multiple illumination spots can be differentiated, for example, by the spectral ratio of the beam splitter. The method of forming the multiple illumination spots is not limited to this example, and appropriate design modifications are possible, such as using multiple light sources.

[0073] FIG. 8D shows the relationship between the regions 202 and 203 and the illumination spots in this embodiment. In this embodiment, the illumination unit 101 irradiates the sample surface with not only the main illumination spot 20a, which has the highest power density, but also the sub-illumination spot 20b, which has a relatively lower power density. Region 202 includes the main illumination spot 20a, and region 203 includes the sub-illumination spot 20b. In the second embodiment ( FIG. 8B ), region 203 is at the beam end of the illumination spot 20, so the scattered light intensity is low, which can reduce the detection sensitivity for relatively small explosive foreign particles. In this embodiment, a sub-illumination spot 20b is formed for region 203 separately from the main illumination spot 20a. Therefore, by appropriately setting the area power density of the sub-illumination spot 20b, the detection sensitivity for defects in region 203 can be improved.

[0074] In this embodiment, other configurations, functions, and effects are the same as those of the previously described embodiment.

[0075] Sixth Embodiment In this embodiment, a case will be described in which the vertical detection unit 303 in the third embodiment includes image formation detection instead of light collection detection. An example of the vertical detection unit 303 in this embodiment is shown in FIG. 17 . In the vertical detection unit 303 according to this embodiment, the intermediate image 23 is formed on the sensor 3039 of the following region detection system 311 by the light collection lens 3037 and the image formation lens 3044. The sensor 3039 of the following region detection system 311 is a line sensor, like the sensor 3028 of the oblique detection unit 32. The sensor 3042 of the leading region detection system 312 is the same as in the previously described embodiments.

[0076] In this embodiment, defect position information is obtained from the detection signal of the trailing region detection system 311 of the vertical detection unit 303, so when calculating the size of the defect at which the diagonal detection unit 32 becomes saturated, there is no need to combine this with the position information obtained from the detection signal of the diagonal detection unit 32, thereby simplifying the calculation process for the defect size.

[0077] Furthermore, compared to the third embodiment, unnecessary power and sensitivity reductions are suppressed. This will be explained using FIG. 11C . In the example of FIG. 11C (1), the trailing region detection system 311 again measures approximately the same amount of light at time t4 as the amount of light measured at time t2. In this case, if the trailing region detection system 311 is using focused light detection, the explosive foreign matter detection unit 42 cannot acquire defect position information, and the location of the detected defect within the region 202 cannot be determined. Therefore, it is difficult to determine whether the power density irradiated to the defect will increase or decrease relatively in the next rotation from the amount of scattered light measured at time t4. Therefore, when the third embodiment is applied to the case of FIG. 11C , a process is performed to reduce the light intensity at time t5 based on the measured light intensity at time t4. However, in reality, the power density of the illumination irradiated to the defect decreases at time t5 compared to time t4, so adjusting the light intensity at time t5 is unnecessary and may actually reduce the decrease in detection sensitivity for small defects.

[0078] In contrast, in this embodiment, by using an imaging detection system as the trailing area detection system 311, it is possible to use defect position information to determine whether or not light intensity adjustment is required in the explosive foreign object detection unit 42. In this case, it is possible to grasp in real time that the power density of the illumination irradiated at time t5 on the defect detected at time t4 in the example of Figure 11C will decrease, thereby preventing unnecessary power and sensitivity reductions such as those described above.

[0079] In this embodiment, other configurations, functions, and effects are the same as those of the previously described embodiment.

[0080] <Modifications> The present invention is not limited to the above-described embodiments and may include various modifications. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and are not necessarily limited to those including all of the described configurations. It is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is also possible to add, delete, or replace part of the configuration of each embodiment with other configurations.

[0081] Some or all of the above-described configurations, functions, processes, processing means, etc. may be implemented by hardware such as an integrated circuit. The above-described configurations, functions, etc. may also be implemented by software, with a processor interpreting and executing a program that implements each function. Information such as programs, tables, and files that implement each function can be stored in various storage media. Examples of various storage media include recording devices such as memory, hard disks, and solid-state drives (SSDs), as well as flash memory cards and digital versatile disks (DVDs).

[0082] In each embodiment, the signal input / output lines are those that are considered necessary for explanation, and do not necessarily show all of the signal input / output lines in the product. In reality, it can be considered that almost all components are interconnected.

[0083] 20... illumination spot (beam spot), 20a... main illumination spot (beam spot), 20b... secondary illumination spot (beam spot), 32... oblique detection unit (first scattered light detection optical system), 101... illumination unit, 201... stage unit, 202... region (first illumination region), 203... region (second illumination region), 210... specimen surface displacement detection system, 301... detection unit, 303... vertical detection unit (second scattered light detection optical system), 311... trailing region detection system, 312... leading region detection system, 401... signal processing unit, 1011... laser light source (light source), 1015... polarization voltage control unit (light intensity adjustment unit), 2031... leading region, 2032... trailing region, 3028... sensor (optical sensor), 3035... sensor (optical sensor), 3039... sensor (optical sensor), 3042... sensor (optical sensor), 3043... spatial filter

Claims

1. A defect inspection device comprising: a stage unit that holds and moves a sample; an illumination unit that irradiates light emitted from a light source onto said sample and forms a beam spot that scans said sample; a detection unit that detects scattered light from said beam spot and photoelectrically converts the amount of scattered light; and a signal processing unit that processes the photoelectrically converted electrical signal, wherein said detection unit includes a first scattered light detection optical system and a second scattered light detection optical system that detect scattered light from a first illumination area and a second illumination area in said beam spot, respectively, and the exposure time of an optical sensor included in said first scattered light detection optical system is longer than the exposure time of an optical sensor included in said second scattered light detection optical system.

2. A defect inspection apparatus according to claim 1, wherein the second illumination area includes a preceding area that scans the sample temporally prior to the first illumination area.

3. A defect inspection device according to claim 2, characterized in that the second scattered light detection optical system includes a leading region detection system that detects scattered light from the leading region, and a trailing region detection system that detects scattered light from a trailing region that follows the leading region in time and scans the sample.

4. A defect inspection device according to claim 3, wherein the photon collection efficiency of said trailing area detection system is set lower than the photon collection efficiency of said leading area detection system.

5. The defect inspection device according to claim 3, further comprising a light amount adjusting unit for adjusting the amount of light from said light source.

6. A defect inspection device according to claim 5, wherein the light intensity adjusting unit adjusts the power density of the first illumination area that is irradiated onto particles having a size equal to or larger than a threshold value when the second scattered light detection optical system detects particles having a size equal to or larger than a threshold value in the second illumination area.

7. A defect inspection device according to claim 5, wherein the light intensity adjustment unit adjusts the power density of the trailing region and the first illumination region that are irradiated onto particles that are larger than a first threshold when the leading region detection system detects a particle in the leading region that is larger than a first threshold, and adjusts the power density of the first illumination region that is irradiated onto particles that are larger than a second threshold when the trailing region detection system detects a particle in the trailing region that is larger than a second threshold.

8. A defect inspection apparatus according to claim 6, wherein the beam spot is a multi-spot, and the first illumination area and the second illumination area are areas of different beam spots.

9. A defect inspection device according to claim 1, wherein the first scattered light detection optical system is disposed at an angle to the sample surface, and the light receiving surface of the optical sensor included in the first scattered light detection optical system is inclined with respect to the optical axis so as to be conjugate with the sample surface.

10. The defect inspection device according to claim 1, wherein the light source is a pulsed laser.

11. A defect inspection apparatus according to claim 10, wherein the optical sensor included in the second scattered light detection optical system performs sampling in synchronization with the oscillation timing of the pulsed laser.

12. A defect inspection apparatus according to claim 1, wherein the first scattered light detection optical system is an imaging optical system, and the second scattered light detection optical system is a focusing optical system.

13. A defect inspection device according to claim 1, characterized in that the signal processing unit calculates the size of the defect using defect position information obtained from the first scattered light detection optical system and a scattered light quantity signal obtained from the second scattered light detection optical system.

14. A defect inspection apparatus according to claim 1, wherein the second scattered light detection optical system includes a spatial filter that blocks illumination light that is specularly reflected from the sample surface.

15. The defect inspection device of claim 6, further comprising a sample surface displacement detection system for measuring displacement of the sample surface, wherein the signal processing unit changes the threshold value based on the measurement results of the sample surface displacement input from the sample surface displacement detection system.

Images