JP4387089B2 - Defect inspection apparatus and defect inspection method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention detects a foreign matter existing in a thin film substrate, a semiconductor substrate, a photomask or the like when manufacturing a semiconductor chip or a liquid crystal product, or a defect generated in a circuit pattern, and analyzes the detected foreign matter or defect to take a countermeasure. The present invention relates to an apparatus and a method for inspecting the occurrence state of foreign matters or defects in a device manufacturing process.
[0002]
[Prior art]
In a semiconductor manufacturing process, if foreign matter is present on a semiconductor substrate (wafer), it may cause defects such as wiring insulation failure or short circuit. Furthermore, with the miniaturization of semiconductor elements, finer foreign substances also cause failure of insulation of capacitors and destruction of gate oxide films. These foreign substances are various due to various causes such as those generated from the moving part of the transfer device, those generated from the human body, those generated in the processing apparatus by the process gas, those mixed in chemicals and materials, etc. It is mixed in the state of.
[0003]
Similarly, in the manufacturing process of the liquid crystal display element, if a foreign substance is mixed on the pattern or some defect occurs, the liquid crystal display element cannot be used as a display element. The situation is the same in the manufacturing process of the printed circuit board, and the inclusion of foreign matters causes a short circuit of the pattern and a defective connection.
[0004]
As one of the conventional techniques for detecting foreign matters on a semiconductor substrate of this type, as described in Japanese Patent Application Laid-Open No. 62-89336 (conventional technique 1), a semiconductor is irradiated with a laser to produce a semiconductor. By detecting scattered light from foreign matter generated when foreign matter is attached to the substrate and comparing it with the inspection result of the same type semiconductor substrate inspected immediately before, there is no false information due to the pattern, and high sensitivity and high reliability What enables a high degree of foreign matter and defect inspection is disclosed. Further, as disclosed in Japanese Patent Application Laid-Open No. 63-135848 (Prior Art 2), the scattering from the foreign matter generated when the semiconductor substrate is irradiated with a laser and the foreign matter adheres to the semiconductor substrate. There is known one that detects light and analyzes the detected foreign matter by an analysis technique such as laser photoluminescence or secondary X-ray analysis (XMR).
In addition, as a technique for inspecting the foreign matter, there is a method of irradiating the wafer with coherent light and removing light emitted from the repeated pattern on the wafer with a spatial filter to emphasize and detect the foreign matter or defect having no repeatability. It is disclosed. Further, the circuit pattern formed on the wafer is irradiated from a direction inclined 45 degrees with respect to the main straight line group of the circuit pattern, and the 0th-order diffracted light from the main straight line group enters the aperture of the objective lens. A foreign matter inspection apparatus which is not allowed to be used is known in Japanese Patent Laid-Open No. 1-117024 (Prior Art 3). This prior art 3 also describes shielding other straight line groups that are not the main straight line group with a spatial filter. Further, as prior art relating to a defect inspection apparatus and method for foreign matters, etc., Japanese Patent Laid-Open No. 1-250847 (prior art 4), Japanese Patent Laid-Open No. 6-258239 (prior art 5), and Japanese Patent Laid-Open No. 6-324003. (Prior Art 6), JP-A-8-210989 (Prior Art 7), JP-A-8-271437 (Prior Art 8), and JP-A 2000-105203 (Prior Art 9) are known. In particular, the prior art 9 describes that the detection pixel size is changed by switching the detection optical system. Further, JP-A-2001-60607 (Prior Art 10) and JP-A-2001-264264 (Prior Art 11) are disclosed as a foreign matter size measurement technique.
[0005]
[Problems to be solved by the invention]
However, in the prior arts 1 to 9, it has not been possible to easily detect fine foreign matters or defects on the substrate in which repeated patterns and non-repeated patterns are mixed with high sensitivity and high speed. That is, in the prior arts 1 to 9, there is a problem that the detection sensitivity (minimum detection foreign matter size) is low in a portion other than a repetitive portion such as a memory cell portion. Further, in the related arts 1 to 9, there is a problem that the detection sensitivity of a minute foreign matter or defect of a 0.1 μm level in a region where the pattern density is high is low. Moreover, in the said prior arts 1-9, there existed a subject that the detection sensitivity of the foreign material or defect which short-circuits between wiring, or the detection sensitivity of a thin-film-like foreign material is low. Moreover, in the said prior arts 10-11, there existed a subject that the measurement precision of a foreign material or a defect was low.
[0006]
The first object of the present invention is to solve the above-mentioned problems by rapidly and finely removing a minute foreign substance or defect of 0.1 μm level on a substrate to be inspected in which a repeated pattern and a non-repeated pattern are mixed. It is an object of the present invention to provide a defect inspection apparatus and method capable of inspecting with high accuracy.
[0007]
A second object of the present invention is to provide a defect inspection apparatus and method capable of inspecting foreign matter or defects with high sensitivity even in a region where the pattern density is high.
[0008]
A third object of the present invention is to provide a defect inspection apparatus and method capable of inspecting foreign matters or defects short-circuiting between wirings or thin-film foreign matters with high sensitivity.
[0009]
A fourth object of the present invention is to provide a defect inspection apparatus and method capable of classifying and displaying foreign matters or defects existing on a substrate to be inspected.
[0010]
A fifth object of the present invention is to provide a defect inspection apparatus and method capable of measuring the size of a foreign substance or a defect existing on a substrate to be inspected.
[0011]
[Means for Solving the Problems]
In order to achieve the above object, the present invention comprises an illumination optical system configured to be able to switch and illuminate an illumination light beam emitted from an illumination light source at different inclination angles with respect to the surface of a substrate to be inspected, and An objective lens that collects reflected and scattered light from the substrate to be inspected, a variable-magnification imaging optical system that can form images of reflected and scattered light collected by the objective lens at different imaging magnifications, and the variable magnification A variable magnification detection optical system having a photodetector that receives reflected scattered light imaged by the imaging optical system and converts it into an image signal, and an image signal obtained from the photodetector of the variable magnification detection optical system. A defect inspection apparatus comprising a signal processing system that converts a digital image signal and detects a defect based on the converted digital image signal.
[0012]
According to the present invention, in the variable magnification detection optical system of the defect inspection apparatus, the variable magnification imaging optical system is configured such that the imaging magnification is variable while the relative distance between the inspection target substrate and the photodetector is constant. It is characterized by doing. Further, the present invention is characterized in that, in the variable magnification detection optical system of the defect inspection apparatus, the variable magnification imaging optical system is configured such that the imaging magnification is variable while the size of the Fourier transform image is constant. .
[0013]
According to the present invention, in the signal processing system of the defect inspection apparatus, the digital image signal is merged with neighboring pixels, and a defect is detected based on the merged image signal. In the signal processing system of the defect inspection apparatus according to the present invention, the defect detection apparatus further includes classification means for classifying the detected defects by category. According to the present invention, the signal processing system of the defect inspection apparatus further comprises size measuring means for measuring the size of the detected defect.
[0014]
The present invention also provides an illumination optical system configured to irradiate an illumination light beam emitted from an illumination light source by switching between a high inclination angle and a low inclination angle with respect to the surface of the inspection target substrate, and the inspection target An objective lens that collects the reflected and scattered light from the substrate, an imaging optical system that forms an image of the reflected and scattered light collected by the objective lens, and the reflected and scattered light imaged by the imaging optical system A detection optical system having a photodetector for converting into a signal, and when the illumination optical system is illuminated at a high tilt angle and when the illumination optical system is illuminated at a low tilt angle, the detection optical system obtains from the photodetector An A / D converter that converts the image signal to be converted into a digital image signal, a defect detection processor that detects a defect based on the digital image signal converted by the A / D converter, and the defect detection processor Calculate feature values for defects About the defect in which the defect detected from the defect detection processing unit when illuminated with the collection amount calculating unit and the high inclination angle is identical to the defect detected from the defect detection processing unit when illuminated at the low inclination angle A defect inspection apparatus comprising: a signal processing system including an integrated processing unit that acquires a feature amount from the feature amount calculation unit and classifies a defect category based on the acquired feature amount of the defect It is.
[0015]
According to the present invention, in the integrated processing unit of the defect inspection apparatus, the feature amount of the defect is constituted by a detected light amount and a planar area.
[0016]
Further, the present invention irradiates the illumination light beam emitted from the illumination light source to the surface of the substrate to be inspected by the illumination optical system at the first inclination angle, and the reflected scattered light from the irradiated substrate to be inspected. Is focused by an objective lens and imaged by an imaging optical system, and the reflected and scattered light thus imaged is received by a photodetector and converted into a first image signal. The converted first image A signal is converted into a first digital image signal by an A / D converter, a defect is detected based on the converted first digital image signal, and a feature amount of the detected defect is converted into the first digital image signal. A first step of calculating based on the digital image signal; and a second tilt angle different from the first tilt angle with respect to the surface of the substrate to be inspected by the illumination optical system for the illumination light beam emitted from the illumination light source The reflected scattered light from the irradiated substrate to be inspected is Condensed by a lens and imaged by an imaging optical system, the imaged reflected and scattered light is received by a photodetector and converted into a second image signal, and the converted second image signal is converted into a second image signal. A second digital image signal is converted by an A / D converter, a defect is detected based on the converted second digital image signal, and a feature amount of the detected defect is converted to the second digital image signal. Calculated in the first step for the second step of calculating based on the signal and the defect in which the defect detected in the first step and the defect detected in the second step are identified. A defect inspection method comprising a classification step of classifying a defect category based on a feature amount and the feature amount calculated in the second step.
[0017]
Further, the present invention irradiates the illumination light beam emitted from the illumination light source to the surface of the substrate to be inspected by the illumination optical system at a desired inclination angle, and reflects the reflected scattered light from the irradiated substrate to be inspected. Condensed by an objective lens and imaged by an imaging optical system, the imaged reflected and scattered light is received by a photodetector and converted into an image signal, and the converted image signal is converted into an A / D converter. Is converted into a digital image signal, a defect is detected based on the converted digital image signal, a plurality of types of merge processing are performed on the digital image signal for the detected defect, and the plurality of merge processing performed A defect inspection method characterized by classifying defects based on a result.
[0018]
Further, the present invention irradiates the illumination light beam emitted from the illumination light source to the surface of the substrate to be inspected by the illumination optical system at a desired inclination angle, and reflects the reflected scattered light from the irradiated substrate to be inspected. Condensed by an objective lens and imaged by an imaging optical system, the imaged reflected and scattered light is received by a photodetector and converted into an image signal, and the converted image signal is converted into an A / D converter. To detect a defect based on the converted digital image signal, and classify the defect based on the detected light amount calculated from the digital image signal for the detected defect. This is a defect inspection method.
[0019]
As described above, according to the above configuration, a minute foreign matter or defect can be inspected at high speed and with high accuracy against a substrate to be inspected in which a repeated pattern and a non-repeated pattern are mixed. Moreover, according to the said structure, a foreign material or a defect can be test | inspected with high sensitivity also in the area | region where pattern density is high. Moreover, according to the said structure, the foreign material which short-circuits between wiring, a defect, or a thin film-like foreign material can be test | inspected with high sensitivity. Moreover, according to the said structure, the size of the foreign material or defect which exists on a to-be-inspected board | substrate can be measured. Moreover, according to the said structure, the foreign material or defect which exists on a to-be-inspected board | substrate can be classified and displayed.
[0020]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments according to the present invention will be described below with reference to the drawings.
[0021]
The defect inspection apparatus according to the present invention increases various types of defects such as foreign matter, pattern defects and micro scratches on a substrate to be inspected such as a wafer in various varieties and various manufacturing processes. It is intended to enable inspection with high sensitivity. To that end, as shown in FIG. 1, as a defect inspection apparatus according to the present invention, the irradiation angle α of the slit beam 201 illuminated by the illumination optical system 10 can be changed according to the object to be inspected. The magnification of the detection optical system 20 is made variable so that the detection pixel size matches the size of the detection defect.
[0022]
Furthermore, the defect inspection apparatus according to the present invention is, for example, a detection pixel that indicates the amount of scattered light obtained by detecting from a defect by low-angle illumination, the amount of scattered light obtained by detecting from a defect by high-angle illumination, and the spread of the defect. Various defects can be classified from the relationship with the number.
[0023]
Next, an embodiment of the defect inspection apparatus according to the present invention will be specifically described. In the following embodiments, a case where small / large foreign matters, pattern defects, micro scratches, and other defects on a semiconductor wafer are inspected will be described. However, the present invention is not limited to a semiconductor wafer, but a thin film substrate, photomask, TFT, PDP. The present invention can also be applied.
[0024]
By the way, the defect inspection apparatus according to the present invention, as shown in FIG. 1, has xyz stages 31, 32, 33 for placing and moving a substrate 1 to be inspected such as a wafer obtained from various varieties and various manufacturing processes. , 34 and the controller 35, and the light emitted from the laser light source 11 is passed through the concave lens 12, the convex lens 13, the conical spherical lens 14, and the mirror 15 from the oblique direction on the substrate 1 to be inspected. Illuminating optical system 10, objective lens 21, spatial filter 22, imaging lens 23, ND (Neutral Density) filter (adjusts the amount of light regardless of wavelength band) 24, optical filter 25 such as a polarizing plate, TDI It comprises a light detector 26 such as an image sensor, and detects reflected diffracted light (or scattered light) from an area illuminated by the illumination optical system 10. The variable rate detection optical system 20, the signal processing system 40 for detecting foreign matter based on the image signal detected by the photodetector 26, inspection conditions, etc. are set, and the illumination optical system 10 and the variable magnification detection optical system 20 are set. And an overall control unit 50 that controls the entire transport system 30 and the signal processing system 40. The overall control unit 50 includes an input / output unit 51 (including a keyboard and a network), a display unit 52, and a storage unit 53.
[0025]
The foreign matter inspection apparatus includes an automatic focus control system (not shown) so as to form an image of the surface of the wafer 1 on the light receiving surface of the photodetector 26.
[0026]
[Illumination optical system 10]
As the illumination optical system 10, as described in Japanese Patent Application Laid-Open No. 2000-105203, a beam expanding optical system including a concave lens 12 and a convex lens 13, and a conical curved lens 14 are used. As shown in FIG. 2, the slit beam 201 is placed on the sample mounting table 34 from one or more directions (three directions 820, 821, 822 in FIG. 2) via the mirror 15. It is configured to irradiate a wafer (substrate to be inspected) 1. At this time, illumination is performed such that the longitudinal direction of the slit beam 201 is directed to the chip arrangement direction. Note that the slit beam 201 is used as the illumination light in order to inspect the foreign matter at high speed. That is, as shown in FIG. 2, the slit beam 201 illuminated on the wafer 1 in which the chips 202 are arranged in the scanning direction of the x stage 31 and the scanning direction of the y stage 32 is scanned in the scanning direction y of the y stage 32. And a wide slit-shaped beam in the vertical direction x (scanning direction of the x stage 31). The slit beam 201 is illuminated so as to be parallel light in the x direction so that an image of the light source is formed in the y direction.
[0027]
By the way, the longitudinal direction of the slit beam 201 is directed to the chip arrangement direction with respect to the wafer 1 by keeping the pixel direction 203 of the photodetector 26 and the traveling direction of the y stage 32 parallel to each other. This is because the inter-chip comparison can be easily performed, and the position coordinates of the foreign matter can be easily calculated. As a result, the foreign matter can be inspected at high speed.
[0028]
In particular, in order to form the illumination of the slit beam 201 from the direction 821 or 822 so as to be directed to the arrangement direction of the chips 202 with respect to the wafer 1 and to be perpendicular to the scanning direction y of the y stage 32. The conical curved lens 14 is required. The conical curved lens 14 is a lens in which the focal length is different at the longitudinal position of the cylindrical lens and the focal length is linearly changed. With this configuration, as shown in FIG. 3, it is possible to illuminate with the slit beam 201 that is narrowed down in the y direction and collimated in the x direction even when illuminated obliquely (with both the angle α and the inclination in the direction φ).
Thereby, it is possible to realize illumination having parallel light in the x direction and having φ = 45 degrees. In particular, by making the slit beam 201 parallel to the x direction, a diffracted light pattern can be obtained from a circuit pattern in which the main straight line groups are directed in the x direction and the y direction, and can be shielded by the spatial filter 22. become.
[0029]
By the way, since the manufacturing method of the conical curved lens 14 is described in Unexamined-Japanese-Patent No. 2000-105203, description is abbreviate | omitted.
[0030]
Next, an embodiment in which the illumination angle α and the illumination direction φ of the illumination optical system 10 are changed according to the inspection target substrate 1 placed on the stage based on a command from the overall control unit 50 will be described. In other words, as far as the illumination angle α is concerned, the illumination angle α may be determined according to the type of foreign matter generated in the inspection object. For example, when mainly detecting foreign matter on the wafer surface, an angle close to the angle parallel to the wafer surface is good, and is about 1 to 5 degrees above the wafer surface, that is, α is about 1 to 5 degrees. It is good to illuminate. By setting the illumination angle α to an angle parallel to the wafer surface, the SN ratio of foreign matter on the outermost surface of the wafer is improved. Further, when it is desired to mainly detect pattern defects or to detect a foreign object having a low height, it is desirable to illuminate from a high angle. However, if the angle is increased, the amount of reflected diffracted light from the underlying circuit pattern increases and the S / N ratio decreases, so the illumination angle α is preferably set to about 45 to 55 degrees. In addition, in order to detect the above-described foreign matter and pattern defects on the wafer surface uniformly, it is preferable to set an intermediate angle between the above-described angles and set the illumination angle α to about 20 degrees. Furthermore, if there is a correspondence between the process to be inspected and the type of foreign object to be detected, for example, if it is known that the foreign object is mainly to be detected in the wiring process, or the defect is a low-height foreign object, Then, you may decide to illuminate from a high angle.
[0031]
Further, with regard to the illumination direction φ, when the wafer is Line & Space, for example, after being etched in the wiring process, it is desirable to select illumination from a direction parallel to the wiring direction of the circuit pattern. That is, it becomes easy to detect foreign matter between the wirings by matching the parallel direction of the illumination light and the direction of the wiring pattern. In addition, when the circuit pattern of the wafer is not a wiring pattern but a contact hole, a capacitor, or the like, since there is no specific directionality, it is desirable to illuminate the chip from a direction near 45 degrees.
[0032]
As the laser light source 11, it is preferable to use the second harmonic SHG of a high-power YAG laser and a wavelength of 532 nm considering that foreign substances can be inspected with high sensitivity and the maintenance cost is low, but it is not necessarily limited to 532 nm. The light source may be an ultraviolet laser, a far ultraviolet laser, a vacuum ultraviolet laser, an Ar laser, a nitrogen laser, a He-Cd laser, an excimer laser, a semiconductor laser, or the like. As an advantage when using each laser, if the laser wavelength is shortened, the resolution of the detected image is increased, so that highly sensitive inspection is possible. When the wavelength is about 0.34 μm, the NA of the objective lens 21 is about 0.4, and when the wavelength is about 0.17 μm, the NA of the objective lens 21 is about 0.2. Detection sensitivity can be improved by making a lot of light incident on the objective lens 21. In addition, when an inexpensive and high-power laser such as a semiconductor laser is used, the device can be manufactured at a low cost.
[0033]
Further, the illumination optical system 10 will be specifically described. First, a method of changing the illumination angle α can be realized by changing the angle of the mirror 15. In FIG. 1, the illumination position 701 is irradiated with laser illumination by the mirror 15. When changing the illumination angle α, the mirror 702 having a different angle from the mirror 15 may be replaced with the mirror 15, and the mirror 702 may be moved in the z direction to irradiate the illumination position 701 with laser light. At this time, since the distance from the convex lens 13 to the illumination position 701 changes, it is necessary to change the convex lens 13 to a convex lens having a different focal length. In this embodiment, an example in which mirrors having different angles are used has been described. However, a configuration in which the mirrors are rotated may be used.
[0034]
Next, a method for changing the illumination direction φ will be described. FIG. 2 is a plan view in the case where three illumination optical systems 10 are configured using one laser light source 11. The laser beam emitted from the laser light source 11 is branched into two optical paths by a branching optical element 801 such as a half mirror, and one is reflected by the mirror 802 and incident downward on the concave lens 12 by the mirror 804. An illumination beam from 820 can be obtained, while the other proceeds to a branching optical element 805 such as a half mirror. One of the beams branched by the branching optical element 805 is reflected by the mirror 806 and incident downward on the concave lens 12 by the mirror 807, whereby an illumination beam from the illumination direction 821 can be obtained, and the other is mirrored by the mirror 808. The illumination beam from the illumination direction 822 can be obtained by making it incident on the concave lens 12 toward. Here, when illuminating only from the illumination direction 820, it can be realized by switching from the branch optical element 801 to the mirror element 809. Further, in the case of illuminating from two directions of the illumination direction 821 and the illumination direction 822, it can be realized by retracting the branching optical element 801 from the optical path or switching to a normal optical element. Of the illuminations from the illumination direction 821 and the illumination direction 822, for example, illumination from only the illumination direction 822 can be realized by switching from the branch optical element 805 to the mirror element 810, and from the illumination direction 821. In the case of only illuminating, the branching optical element 805 may be withdrawn from the optical path.
[0035]
[Detection optical system 20]
The variable magnification detection optical system 20 will be described. The variable magnification detection optical 20 is configured to convert light reflected and diffracted from the inspection target substrate 1 such as a wafer into an objective lens 21, a spatial filter 22, an imaging lens (variable magnification imaging optical system) 23, an ND filter 24, and an optical filter. 25 is configured to be detected by a photodetector 26 such as a TDI image sensor.
[0036]
The spatial filter 22 shields the Fourier transform image by the reflected diffracted light from the repetitive pattern on the wafer 1, and is in the spatial frequency region of the objective lens 21, that is, in the Fourier transform imaging position (corresponding to the exit pupil). Placed. Note that the spatial filter 22 can be produced by printing out an image obtained by capturing a Fourier transform image, reversing the image in black and white, and widening the light shielding portion on a transparent sheet.
[0037]
Next, a method of changing the magnification by the variable magnification detection optical system 20 based on a command from the overall control unit 50 will be described. First, a method for changing the magnification with the imaging lens (variable magnification imaging optical system) 23 will be described. 1 and 4, the imaging lens (variable magnification imaging optical system) 23 includes movable lenses 1201 and 1202, a fixed lens 1203, and a moving mechanism 1204.
[0038]
4A shows a schematic position of the movable lenses 1201 and 1202 at a low magnification, and FIGS. 1 and 4B show a schematic position of the movable lenses 1201 and 1202 at a high magnification. The variable magnification detection optical system 20 is characterized in that the position of the objective lens 21 is not changed when the magnification is changed, the position of the spatial filter 22 is not changed, and the position of the fixed lens 1203 is not changed. That is, it is not necessary to change the position of the photodetector 26. That is, even if the magnification is changed, there is no need to change the relative position between the substrate 1 to be inspected and the photodetector 26. Therefore, the drive mechanism for changing the magnification is only the moving mechanism 1204, and the magnification is changed with a simple configuration. Furthermore, since the size of the Fourier transform plane does not change, there is an advantage that the spatial filter 22 need not be changed.
[0039]
An advantage of changing the magnification of the variable magnification detection optical system 20 is that high-speed and high-sensitivity inspection can be performed in accordance with the inspection object. That is, an object or region to be inspected whose circuit pattern is manufactured at a high density can be inspected with high sensitivity because an image signal with high resolution can be obtained by increasing the magnification. In addition, an inspection object or region in which circuit patterns are manufactured at a low density can be inspected at high speed with high sensitivity by reducing the magnification.
[0040]
As an operation at the time of changing the magnification, a method of moving the movable lenses 1201 and 1202 in the z direction using the moving mechanism 1204 is used. At this time, the magnification M of the variable magnification detection optical system 20 is obtained by changing the focal length 1205 of the objective lens 21 to f. ₁ , The focal length 1206 of the imaging lens 23 is f ₂ Then, it is computable by (1) Formula.
[0041]
M = f ₁ / F ₂ (1)
Therefore, in order to set the magnification variable detection optical system 20 to the magnification M, f ₁ Is a fixed value, so f ₂ Is (f ₁ / M).
[0042]
Next, details of the moving mechanism 1204 will be described with reference to FIG. FIG. 5 shows an embodiment of a mechanism for moving the movable lens 1201. FIG. 5A shows an embodiment in which the movable lens 1201 is moved to an arbitrary location, and FIG. 5B shows an embodiment in which the movable lens 1201 is positioned at a specific location. 5A includes a lens holding portion 4101 of the movable lens 1201, a ball screw 4102, and a motor 4103.
[0043]
As an operation, first, the overall control unit 50 determines the magnification used for the inspection according to the substrate 1 to be inspected placed on the stage. Here, as a policy for determining the magnification, as described above, when inspecting a wafer with a high density of circuit patterns, a high magnification is selected, and when inspecting a wafer with a low density of circuit patterns. Or, when inspecting at high speed, a low magnification may be selected. Next, since the setting position of the movable lens 1201 is determined when the magnification is determined, the ball screw 4102 is rotated using the motor 4103 to move the lens holding portion 1401 and the movable lens 1201. At this time, in order to accurately position, it is preferable to move in one direction in consideration of the backlash of the motor 4103.
[0044]
Further, when several magnifications are set, the movable lens 1201 may be positioned at a specific location, so the configuration shown in FIG. 5B may be used. FIG. 5B shows a configuration in which positioning sensors 4104, 4105, 4106, and 4107 are added to the configuration of FIG. The positioning sensor 4105 is an upper limit sensor in the z direction, and the positioning sensor 4107 is a lower limit sensor in the z direction. Further, the positioning sensor 4106 is attached to the set position of the movable lens 1201.
[0045]
As the operation, the ball screw 4102 is rotated by the motor 4103 and the movable lens 1201 is moved by the same way as in FIG. 5A, and the motor 4103 is stopped at a place where the positioning sensor 4104 and the positioning sensor 4106 correspond. Here, the positioning sensor may be an optical sensor or a sensor using magnetism. The mechanism described here can also be applied to the movable lens 1202.
[0046]
Further, in another embodiment of the variable magnification detection optical system 20, as shown in FIG. 6A, the magnification of the variable magnification detection optical system 20 is changed by replacing the lens unit 3301 with lens units 3305 and 3306. Is. The lens units 3305 and 3306 have the same outer shape as the lens unit 3301 and are configured with a different magnification from the lens unit 3301 by changing the positions of the relay lenses 3302 and 3303. That is, when changing the magnification of the variable magnification detection optical system 20, the lens unit 3301 is first removed, and the lens unit 3305 or the lens unit 3306 is installed instead. The advantage in the case of using this embodiment is that adjustment can be performed with each lens unit, and further, adjustment with respect to one type of magnification can be performed with one lens unit, so that the number of adjustment steps for the lens unit can be reduced. However, since it is necessary to install a plurality of lens units, it may be used when there is a sufficient installation space.
[0047]
Further, in another embodiment of the variable magnification detection optical system 20, as shown in FIG. 6B, the magnification of the variable magnification detection optical system 20 is changed by changing the magnification of the objective lens of the lens unit 4401. Is. Here, the objective lens 4403 is a lens having a magnification different from that of the objective lens 21. The imaging lens 4404 is a lens designed in correspondence with the objective lens 4403. By the way, when changing the magnification of the variable magnification detection optical system 20, the lens unit 4401 is first removed, and the lens unit 4402 may be installed instead. The advantage of using this embodiment is that the objective lens detection N.P. A. Therefore, if the magnification is increased, the resolution is improved and high-sensitivity inspection is possible.
[0048]
Next, the ND filter 24 is for adjusting the amount of light detected by the photodetector 26. The ND filter 24 is not necessarily required when the illumination optical system unit 10 can adjust the amount of irradiation light. However, by using the ND filter 24, the adjustment range of the detected light amount can be increased, and various inspections can be performed. The amount of light can be adjusted to be optimal for the target. For example, the laser light source 11 can adjust the output from 1 W to 100 W, and if a 100% transmission filter and 1% transmission filter are prepared as ND filters, the light amount can be adjusted from 10 mW to 100 W, and a wide range of light amounts can be adjusted. Can do.
[0049]
Next, the optical filter 25 is, for example, a polarizing element. The polarizing element shields the polarized component due to the reflected diffracted light generated from the edge of the circuit pattern when polarized illumination is performed by the illumination optical system unit 10, and transmits a part of the polarized component due to the reflected diffracted light generated from the foreign matter. Is not always necessary.
[0050]
Next, the photodetector 26 is an image sensor for receiving the reflected diffracted light collected by the imaging lens 25 and photoelectrically converting it, for example, a TV camera, a CCD linear sensor, a TDI sensor, or an anti-blooming TDI. It is a sensor or photomultiplier.
[0051]
Here, as a selection method of the light detector 26, a TV camera or a CCD linear sensor is preferable when an inexpensive inspection device is used, and when detecting weak light with high sensitivity, for example, about 0.1 μm. In the case of detecting the following extremely small foreign matter, a sensor or photomultiplier having a TDI (Time Delay Integration) function is preferable. When the dynamic range of light received by the photodetector 26 is large, that is, when light that saturates the sensor is incident, a sensor with an anti-blooming function is preferable.
[0052]
Further, when using a photomultiplier, as shown in FIG. 7, a sensor in which the photomultipliers are arranged in a one-dimensional direction may be used. In this case, since it can be used as a highly sensitive one-dimensional sensor, a highly sensitive inspection is possible. As a configuration at this time, as shown in FIG. 7A, a microlens 5002 is attached to the image forming lens 23 side of the photomultiplier 5001, and the reflected diffracted light collected by the image forming lens 23 is detected. Just do it. Here, the microlens 5002 has a function of condensing light in the same range as the photomultiplier surface onto the photomultiplier 5001. Further, as shown in FIG. 7B, an optical fiber 5004 may be attached via a jig 5003 installed downstream of the microlens 5002, and a photomultiplier 5001 may be attached to the output end of the optical fiber 5004. In this case, since the diameter of the optical fiber is smaller than the diameter of the photomultiplier, the sensor pitch can be made smaller than in FIG. 7A, so that a sensor with high resolution can be obtained.
[0053]
Next, the conveyance system 30 will be described. Stages 31 and 32 are stages for moving the sample mounting table 34 to the xy plane, and have a function of moving the entire surface of the inspected substrate 1 to the illumination area of the illumination optical system 10. The stage 33 is a z stage, and has a function of moving the sample setting table 34 in the optical axis direction (z direction) of the variable magnification detection optical system 20. The sample mounting table 34 has a function of holding the wafer 1 and rotating the substrate 1 to be inspected in the plane direction. The stage controller 35 has a function of controlling the stages 31, 32, 33 and the sample setting table 34.
[0054]
[Signal processing system 40]
Next, the contents of the signal processing system 40 will be described with reference to FIG. The signal processing system 40 includes an A / D converter 1301, a data storage unit 1302 that stores the A / D converted detected image signal f (i, j), and a threshold calculation that performs a threshold calculation process based on the detected image signal. The detection image signal 410 obtained from the processing unit 1303 and the data storage unit 1302 and the threshold image signal (Th (H), Th (Hm), Th (Lm), Th (L)) 420 obtained from the threshold calculation processing unit 1303 Foreign matter detection processing units 1304a to 1304n that perform foreign matter detection processing for each pixel merge based on the above, for example, the amount of scattered light obtained by detecting from a defect by low-angle illumination, obtained by detecting from a defect by high-angle illumination Based on the feature amount for each merge obtained from the feature amount calculation circuit 1310 for calculating the feature amount such as the scattered light amount and the number of detected pixels indicating the extent of the defect, and the feature amount calculation circuit 1310 Integration processing unit 1309 classifies the various defect to classify the defects such as small / large particle or pattern defects or micro-scratches on the semiconductor wafer, and a result display unit 1311. Each of the foreign object detection processing units 1304a to 1304n corresponds to each of 1 × 1, 3 × 3, 5 × 5,..., N × n merge operators, for example, pixel merge circuit units 1305a to 1305n, 1306a to 1306n, Foreign matter detection processing circuits 1307a to 1307n and inspection area processing units 1308a to 1308n are provided.
[0055]
In particular, the present invention is characterized by foreign object detection processing units 1304a to 1304n, a feature amount calculation circuit 1310, and an integrated processing unit 1309.
[0056]
Next, the operation will be described. First, the signal obtained by the photodetector 26 is digitized by the A / D converter 1301. The detected image signal f (i, j) 410 is stored in the data storage unit 1302 and sent to the threshold value calculation processing unit 1303. The threshold value calculation processing unit 1303 calculates a threshold image Th (i, j) 420 for foreign object detection, and the foreign object detection processing circuit 1307 is based on the signals processed by the pixel merge circuits 1305 and 1306 for each of various merge operators. To detect foreign matter. The detected foreign matter signal and the threshold image are processed by the inspection area processing unit 1308 according to the detection location. At the same time, signals obtained from the pixel merge circuits 1305a to 1305n, 1306a to 1306n, the foreign matter detection processing circuits 1307a to 1307n, and the inspection area processing units 1308a to 1308n of the foreign matter detection processing units 1304a to 1304n provided for the various merge operators. Based on the above, the feature amount calculation circuit 1309 calculates the feature amount (for example, the scattered light amount obtained by high angle illumination, the scattered light amount obtained by low angle illumination, the number of detected pixels of the defect, etc.) The feature amounts are integrated by the integration processing unit 1309 and the inspection result is displayed on the result display unit 1311.
[0057]
Details are described below. First, the A / D converter 1301 is a circuit having a function of converting an analog signal obtained by the photodetector 26 into a digital signal, and the number of conversion bits is preferably about 8 to 12 bits. This is because when the number of bits is small, the resolution of signal processing is low, so it is difficult to detect minute light. On the other hand, when the number of bits is large, the A / D converter is expensive and the device price is high. Because there is. Next, the data storage unit 1302 is a circuit for storing A / D converted digital signals. Next, the threshold value calculation processing unit 1303 will be described with reference to FIG.
[0058]
FIG. 9 is a diagram showing an embodiment of the threshold value calculation processing unit 1303 described in Japanese Patent Laid-Open No. 2000-105203. The threshold calculation processing unit 1303 generates a reference image signal g (i, j), a delay circuit unit 1401, and a difference signal ΔS (i, j) between the detected image signal f (i, j) and the reference image signal g (i, j). j) a difference circuit unit 1402, a maximum / minimum value removal unit 1403, a square value calculation unit 1404 that calculates a square value of the difference signal, a square value sum calculation unit 1405 that calculates the sum of the square values, and a difference signal The signal sum calculating unit 1406, the input signal number counting unit 1407, the standard deviation calculating unit 1408 for calculating the standard deviation, the average value calculating unit 1409 for calculating the average value, and the input counted by the counting unit 1407 A coefficient calculation unit 1410 that calculates a coefficient (n) based on the number of signals, a threshold calculation unit 1411 that calculates an image Th (i, j) 420 of thresholds (Th (H), Th (L)), and a data storage unit 1412 Consists of It has been.
[0059]
Next, the operation will be described. The detected image signal f (i, j) obtained by the A / D converter 1301 is input to the delay circuit unit (delay memory unit) 1401 and the difference circuit unit 1402, and the difference circuit unit 1402 uses the difference signal ΔS (i, j j). In the case of chip comparison, (i, j) is a pixel coordinate in the chip. Here, the delay circuit unit 1401 is a circuit that delays a signal by a repetition pitch of a circuit pattern on a substrate to be inspected such as a wafer to generate a reference image signal g (i, j). For example, in a semiconductor wafer, a plurality of chips having the same structure are manufactured on the wafer by the same process. Therefore, when a wafer is inspected by the foreign matter inspection apparatus of the present invention, a circuit pattern is repeated for each chip, and thus a similar circuit pattern signal is obtained for each chip. In this case, the delay circuit unit 1401 may be a circuit that delays signals for one chip on the wafer. However, it is not always necessary for one chip. Depending on the object to be inspected, a plurality of chips may be delayed for a wafer having repetitions in units of a plurality of chips. If there is a repetitive pattern, the repetitive pattern may be delayed by that amount. These signal delay amounts can be changed by a command from a CPU or the like.
[0060]
Next, the difference signal ΔS (i, j) processed by the difference circuit unit 1402 is stored in the data storage unit 1412 and the square value (ΔS) of the difference signal ΔS is passed through the maximum / minimum value removal unit 1403. ² ) For calculating the square value of the difference signal obtained from the square value calculation unit 1404 and the square value calculation unit 1404 (ΣΔS) ² ) For calculating the sum of square values 1405, a sum calculating unit 1406 for calculating the sum of difference signals ΔS (ΣΔS), and a counting unit 1407 for the number of input signals n. Here, the maximum / minimum value removal unit 1403 has a function of obtaining a maximum value and a minimum value from a plurality of input data and removing the maximum value and the minimum value from the input data group. The square value is a value obtained by squaring the input data (ΔS ² The signal sum is a value obtained by calculating the sum of input data (ΣΔS).
[0061]
Using the value calculated in the above processing, the standard deviation value calculation unit 1408 uses the standard deviation value (σ (ΔS) = √ (ΣΔS ² / N−ΣΔS / n)), and the average value calculation unit 1409 calculates the average value of the input data (μ (ΔS) = ΣΔS / n). Further, the coefficient calculation unit 1410 calculates a coefficient (magnification) k for setting a threshold corresponding to the number n of input data.
[0062]
The threshold value calculation unit 1411 calculates the image Th (i, j) 420 of the threshold values (Th (H), Th (L)) using the values described above. The threshold values (Th (H), Th (L)) are calculated by the average value calculation unit 1409, μ (ΔS), the standard deviation value calculation unit 1408, σ (ΔS), and the coefficient calculation unit 1410, Alternatively, if the set value is k and the verification coefficient is m (m is smaller than 1), the following equation (2) is used.
Th (H) = μ + k × σ or Th (L) = μ−k × σ or
Th (Hm) = m × (μ + k × σ) or Th (Lm) = m × (μ−k × σ) (2)
In this manner, the threshold image {detection threshold (Th (H), Th (L)), verification threshold (Th (Hm), Th (Lm))} calculated by the threshold calculation unit 1411} is stored in the data storage unit 1412. The threshold value image data may be output from the data storage unit 1412. The threshold image data may be changed for each region set by the inspection region processing units 1308a to 1308n. In short, the data storage unit 1412 can create and store a threshold map. Further, in order to reduce the detection sensitivity in a certain region, the threshold value in that region may be increased.
[0063]
Next, the pixel merging circuit units 1305 and 1306 for signals will be described with reference to FIG. The pixel merge circuit units 1305 a to 1305 n and 1306 a to 1306 n are configured by different merge operators 1504. The merge operator 1504 includes a detection image signal f (i, j) 410 obtained from the data storage unit 1302, a detection threshold image Th (H), a detection threshold image Th (L) obtained from the threshold calculation processing unit 1303, and verification. This is a function of combining each of the threshold image Th (Hm) and the difference threshold image signal 420 composed of the verification threshold image Th (Lm) in the range of n × n pixels. For example, an average value of n × n pixels is output. Circuit. Here, the pixel merge circuit units 1305a and 1306a are configured by, for example, a merge operator that merges 1 × 1 pixels, and the pixel merge circuit units 1305b and 1306b are configured by, for example, a merge operator that merges 3 × 3 pixels. The units 1305c and 1306c are composed of, for example, a merge operator that merges 5 × 5 pixels, and the pixel merge circuit units 1305n and 1306n are composed of, for example, a merge operator that merges n × n pixels. The merge operator that merges 1 × 1 pixels outputs the input signals 410 and 420 as they are.
[0064]
The threshold image signal is composed of four image signals (Th (H), Th (Hm), Th (Lm), Th (L)) as described above, and therefore, the four pixel merge circuit units 1306a to 1306n A merge operator Op is required. Therefore, from each of the pixel merge circuit units 1305a to 1305n, the detected image signals are merged by various merge operators 1504 and output as merge process detected image signals 431a to 431n. On the other hand, four threshold image signals (Th (H), Th (Hm), Th (Lm), Th (L)) are merged by various merge operators Op1 to Opn from the pixel merge circuit units 1306a to 1306n. The merge processing threshold image signals 441a (441a1 to 441a4) to 441n (441n1 to 441n4) are output. Note that the merge operators in the pixel merge circuit units 1306a to 1306n are the same.
[0065]
Here, the effect of merging pixels will be described. In the foreign matter inspection apparatus of the present invention, it is necessary to detect not only a minute foreign matter but also a large thin film-like foreign matter spread over a range of several μm without overlooking it. However, since the detection image signal from the thin film-like foreign material does not necessarily become large, the detection image signal in units of one pixel has a low SN ratio and may be overlooked. Therefore, when the average detection image signal level of one pixel is S and the average variation is σ / n, the detection is performed by cutting out in units of n × n pixels corresponding to the size of the thin film-like foreign material and performing a convolution operation. The image signal level is n ² × S, and the variation (N) is n × σ. Therefore, the SN ratio is n × S / σ. On the other hand, if an attempt is made to detect thin film-like foreign matters in units of one pixel, the detected image signal level is S and the variation is σ, so the SN ratio is S / σ. Accordingly, the SN ratio can be improved by n times by cutting out in units of n × n pixels corresponding to the size of the thin film foreign matter and performing the convolution operation.
[0066]
For a minute foreign substance of about one pixel unit, the detected image signal level detected in one pixel unit is S and the variation is σ, so the SN ratio is S / σ. If a minute foreign substance of about one pixel is cut out in units of n × n pixels and a convolution operation is performed, the detected image signal level is S / n. ² Since the variation is n × σ, the SN ratio is S / n ³ / Σ. Therefore, for a minute foreign substance of about one pixel unit, the signal of the pixel unit as it is can improve the SN ratio.
[0067]
In this embodiment, the example in which the merge range is a square (n × n pixels) has been described, but the merge range may be a rectangle (n × m pixels). In this case, the detection of a directional foreign substance or the detection pixel in the photodetector 26 is rectangular, but the signal processing is effective when it is desired to process with a square pixel.
[0068]
Further, the function of the merge operator described in the present embodiment has been described in the embodiment that outputs an average value of n × n pixels, but even if the maximum value, minimum value, or median value of n × n pixels is output. good. When the median is used, a stable signal can be obtained. Furthermore, the output value may be a value obtained by multiplying or dividing an average value of n × n pixels by a specific value.
[0069]
Next, FIG. 11 is a diagram showing an embodiment of the foreign object detection processing circuit 1306. FIG. 11 shows a pixel merge circuit unit 1305a and a pixel merge circuit unit 1306a that merge 1 × 1 pixels, and a pixel merge circuit unit 1305n and a pixel merge circuit unit 1306n that merge n × n pixels.
[0070]
Then, the foreign object detection processing circuits 1307a to 1307n correspond to the respective merge operators, and compare circuits 1601a to 1601n that compare the magnitudes of the merge processing difference signals 471a to 471n and the merge processing threshold signals 441a to 441n, and foreign object detection. It comprises detection location determination processing units 1602a to 1602n that specify locations. The comparison circuits 1601a to 1601n include delay memories 451a to 451n that repeatedly delay, for example, chips for the pixel-merged detection image signals obtained from the pixel merge circuits 1305a to 1305n, as shown in FIG. Difference processing circuits 461a to 461n for forming a difference signal between the pixel-merged detected image signals 431a to 431n and the pixel-merged reference image signals delayed by the delay memories 451a to 451n are provided. Therefore, the comparison circuits 1601a to 1601n are merge processing threshold images Th (H) (i, j) and Th (Hm) (i, j) obtained from the pixel merge circuits M1 to M4 of the pixel merge circuit units 1306a to 1306n. , Th (Lm) (i, j) and Th (L) (i, j), for example, merge processing difference detection signals 471a to 471n are obtained from the merge processing threshold image Th (i, j). If it is too large, it has a function of determining as a foreign object. In this embodiment, four types of threshold values are prepared, and foreign matter determination processing is performed by the comparison circuits 1601a to 1601n on the merge processing threshold images 1603, 1604, 1605, and 1606 for each merge operator.
[0071]
Next, the detection location determination processing units 1602a to 1602n will be described with reference to FIGS. The detection location determination process is a process for identifying a chip having a foreign object or a defect detected by signal processing. The idea of this processing is that detection thresholds (Th (H), Th (L)) for detecting foreign matter or defects and verification thresholds (Th (Hm), Th () that are smaller than the detection thresholds). Lm)) is used to identify the chip on which a foreign object or defect is detected.
[0072]
A processing method will be described. FIG. 12A shows a signal when a foreign substance 1704 is present in the central chip 1702 among the chips 1701, 1702, and 1703. FIG. A signal 1706 indicates a signal from the foreign substance 1704, and is a convex or concave signal. On the other hand, the signal 1705 and the signal 1707 are signals from a portion where no foreign matter exists, that is, from a normal pattern.
[0073]
FIG. 12B shows a signal subjected to differential processing in units of chips. Difference signals 1708 and 1709 represent difference signals of signals obtained by the chips 1701, 1702, and 1703. The difference signal 1708 is a difference signal (B−A) between the signal 1706 “B” of the chip 1702 and the signal 1705 “A” of the chip 1701, and the difference signal 1709 is the signal 1707 “C” of the chip 1703 and the signal of the chip 1702. The difference signal (C−B) from 1706 “B”. Further, the detection threshold Th (H) and the verification threshold Th (Hm) are thresholds for detecting a convex difference signal, and the detection threshold Th (L) and the verification threshold Th (Lm) are concave differences. This is a threshold for detection.
[0074]
In FIG. 12B, when the difference signal 1708 (B-A) is positive, if the value is larger than the threshold value Th (H) or Th (Hm), it is detected as a foreign object or a defect, and the difference signal 1709 is detected. When (C−B) is negative, the value is smaller than the threshold value Th (L) or Th (Lm) (when the signed value is small because both the difference value and the threshold value are negative values). As a value, the difference value is larger than the threshold value), and it is detected as a foreign object or a defect. When the foreign matter 1704 is present on the chip 1702, either or both of the differential signals 1708 (B-A) and 1709 (CB) are detected as foreign matter, and thus either of the differential signals 1708 and 1709 is detected. Even so, it must be determined that the chip 1702 has a foreign substance as the signal B.
[0075]
Therefore, when the foreign matter or defect signal 1706 is convex (the signal amount is larger than that of the normal part), the following equation (3) is established from the truth table shown in FIG. In the case of “true”, it may be considered that a foreign substance or a defect exists as the signal B in the chip 1702. 12C and 12C, the signal of the chip 1701 is “A”, the signal of the chip 1702 is “B”, and the signal of the chip 1703 is “C”.
[0076]
{(Difference signal of verification threshold Th (Hm) <(B−A)) AND (Difference signal of (C−B) <Verification threshold Th (Lm))} AND {(Detection threshold Th (H) <(B− A) differential signal) OR ((C−B) differential signal <detection threshold Th (L))} (3) Also, the foreign matter or defect signal 1706 is concave (the signal amount is smaller than the normal part). Inverted from the signal waveform shown in FIG. 12B, the signal 1708 becomes a negative signal, and the signal 1709 becomes a positive signal. Therefore, when the following expression (4) is “true”, it can be considered that a foreign substance or a defect exists as the signal B in the chip 1702.
{(Differential signal of verification threshold Th (Lm) <(B−A)) AND (Differential signal of (CB) <Verification threshold Th (Hm))} AND {(Detection threshold Th (L) <(B− (A) differential signal) OR ((C−B) differential signal <detection threshold Th (H))} (4) Next, when there is no chip next to the chip in which the foreign substance exists, that is, at the top of the wafer. A detection location determination process when a foreign object is detected by the outer peripheral chip will be described. That is, the detection location determination processing units 1602a to 1602n compare the adjacent chips {(BA), based on the control signal when going to the outermost peripheral chip obtained from the overall control unit (not shown). (C-B)} is switched to the comparison processing {(BA), (C-A)} between adjacent chips that will be described next. Here, the adjacent chip is a chip in the direction of chip comparison, and will be described with reference to FIGS. 13 and 14 depending on the direction in which the difference processing is performed during chip comparison.
[0077]
First, FIG. 13A shows a signal when a foreign object 1904 is present on the chip 1901 among the chips 1901, 1902, and 1903. A signal 1905 indicates a signal from the foreign material 1904 and is a convex signal. On the other hand, the signal 1906 and the signal 1907 are signals from a portion where no foreign matter exists, that is, from a normal pattern. Also, differential signals 1908 and 1909 shown in FIG. 13B represent differential signals of signals obtained by the chips 1901, 1902, and 1903. The difference signal 1908 is a difference signal (B−A) between the signal 1906 “B” and the signal 1905 “A”. The difference signal 1909 is a difference signal (C−A) between the signal 1907 “C” and the signal 1905 “A”.
[0078]
In this case, the foreign material 1904 is present on the chip 1901 in the case of the table shown in FIG. Therefore, when the expression (5) is “true”, it can be considered that the foreign substance 1904 exists as the signal A in the chip 1901. 13C and 13, the signal of the chip 1901 is “A”, the signal of the chip 1902 is “B”, and the signal of the chip 1903 is “C”.
[0079]
{((B−A) differential signal <verification threshold Th (Lm)) AND ((C−A) differential signal <verification threshold Th (Lm))} AND {((B−A) differential signal <detection Threshold Th (L)) OR (difference signal of (C−A) <detection threshold Th (L))} (5) Similarly, in FIG. 13A, the chip 1901 has a concave defect (the signal amount is normal). If there is a smaller foreign matter or defect), the formula (6) may be applied. {(Difference signal of verification threshold Th (Hm) <(BA)) AND (Difference signal of verification threshold Th (Hm) <(C−A))} AND {(Detection threshold Th (H) <(B− A) differential signal) OR (detection threshold Th (H) <(C−A) differential signal)} (6) Furthermore, FIG. 14A shows a foreign substance 2004 on the chip 2003 out of the chips 2001, 2002, and 2003. Is shown, and there is no chip on the right side of the chip 2003. A signal 2007 indicates a signal from the foreign substance 2004 and is a convex signal. On the other hand, the signal 2005 and the signal 2006 are signals from a portion where no foreign matter exists, that is, a normal pattern. Also, differential signals 2008 and 2009 shown in FIG. 14B represent differential signals of signals obtained by the chips 2001, 2002, and 2003. The difference signal 2008 is a difference signal (C-A) between the signal 2007 and the signal 2005, and the difference signal 2009 is a difference signal (C-B) between the signal 2007 and the signal 2006.
[0080]
In this case, the foreign substance 2004 is present on the chip 2003 in the case of the table shown in FIG. Accordingly, when the expression (7) is “true”, it can be considered that the foreign substance 2004 exists as the signal C in the chip 2003. 14C and 14, the signal of the chip 2001 is “A”, the signal of the chip 2002 is “B”, and the signal of the chip 2003 is “C”.
[0081]
{(Difference signal of verification threshold Th (Hm) <(C−A)) AND (Difference signal of verification threshold Th (Hm) <(C−B))} AND {(Detection threshold Th (H) <(C− A) differential signal) OR (detection threshold Th (H) <(C−B) differential signal)} (7) Similarly, in FIG. 14A, the chip 2003 has a concave defect (the signal amount is normal). If there is a smaller foreign matter or defect), the equation (8) may be applied.
{((C−A) differential signal <verification threshold Th (Lm)) AND ((C−B) differential signal <verification threshold Th (Lm))} AND {((C−A) differential signal <detection Threshold value Th (L)) OR ((C−B) difference signal <detection threshold value Th (L))} (8) As described above, the detection location determination processing units 1602a to 1602n correspond to various merge operators. Thus, it is possible to specify a chip in which a foreign object or a defect exists and calculate the position coordinates (i, j).
[0082]
Next, the inspection area processing units 1308a to 1308n will be described. The inspection area processing units 1308a to 1308n store data of areas (including areas in the chip) that do not need to be inspected with respect to the foreign substances or defect detection signals obtained by specifying the chips from the foreign substance detection processing circuits 1307a to 1307n. It is used when removing, changing the detection sensitivity for each region (including the region in the chip), or selecting the region to be inspected. For example, when the detection sensitivity may be low among the regions on the inspection target substrate 1, the inspection region processing units 1308a to 1308n may detect the threshold value of the corresponding region obtained from the threshold value calculation unit 1411 of the threshold value calculation processing unit 1303. May be set high, or a method of leaving only the foreign matter data of the region to be inspected based on the foreign matter coordinates from the foreign matter data output from the foreign matter detection processing circuits 1307a to 1307n may be used.
[0083]
Here, the region where the detection sensitivity may be low is, for example, a region where the circuit pattern density is low in the substrate 1 to be inspected. The advantage of lowering the detection sensitivity is to efficiently reduce the number of detections. That is, a high-sensitivity inspection apparatus may detect tens of thousands of foreign objects. At this time, what is really important is foreign matter in the region where the circuit pattern exists, and countermeasures against this important foreign matter are a shortcut to improving the yield of device manufacturing. However, when the entire area on the substrate 1 to be inspected is inspected with the same sensitivity, important foreign matters cannot be easily extracted because important foreign matters and non-important foreign matters are mixed. In view of this, the inspection area processing units 1308a to 1308n are efficiently important by reducing the detection sensitivity of areas that do not have much influence on yield, such as circuit patterns that do not exist, based on in-chip CAD information or threshold map information. Foreign matter can be extracted. However, the foreign matter extraction method is not limited to the method of changing the detection sensitivity, but the important foreign matter may be extracted based on the foreign matter classification described later, or the important foreign matter may be extracted based on the foreign matter size.
[0084]
The region where the threshold value is changed will be described with reference to FIG. FIG. 15A shows a case where the threshold setting is changed for each of the areas 2102, 2103, 2104, and 2105 that are equidistant from the center 2101 of the wafer 1. This is used when the process state in a region equidistant from the center 2101 of the wafer 1 is close. For example, when a viscous film such as an oxide film is applied to the wafer 1, the film is expanded by rotating the wafer 1 around the center 2101. At this time, since the film thicknesses applied at equal distances from the rotation center of the wafer 1 are substantially the same, in order to reduce erroneous detection of foreign matters due to changes in film thickness, the film thickness is interlocked with the distance from the center 2101. The threshold value becomes effective.
[0085]
FIG. 15B shows a case where the threshold setting is changed for each chip 2106 manufactured on the wafer 1. This is effective when a semiconductor chip is manufactured and a test chip is built or when it is desired to change the detection sensitivity in units of chips. For example, when a test chip is built on the wafer 1, since the chip does not become a product, the need for inspection is low, but the test chip has a different circuit pattern from the product chip, so that a normal circuit pattern is used when comparing chips. Is mistakenly detected, and as a result, it cannot be distinguished from a case where a large amount of foreign matter is generated. Therefore, when the detection sensitivity of the test chip is lowered or the test chip is excluded from the inspection, the threshold setting is changed for each chip.
[0086]
Further, FIG. 15C shows a case where the threshold setting is changed in a specific area 2107 in the chip. This can be done when there is an area that does not need to be inspected in the chip, for example, when removing foreign substance information detected in an area that does not significantly affect the yield, such as a dicing area, or by changing the detection sensitivity for each area in the chip. It may be used when extracting the important foreign matter as described above.
An embodiment of the region setting method is shown in FIG. FIG. 16 is a diagram showing a screen for setting a threshold value of a region at a certain distance from the wafer center in the inspection region processing units 1308a to 1308n. FIG. 16A includes an inspection target wafer layout 2201, an area setting button 2202, and a sensitivity setting button 2203.
[0087]
First, when an area setting button 2202 is pressed to select an area for setting a threshold, another screen as shown in FIG. 16B is displayed. On this screen, an input place 2204 for inputting a set distance from the wafer center is provided, and a radius and the like from the wafer center are input. For example, FIG. 16C shows an example in which a value half the wafer radius is input. FIG. 16C is an example in which the setting area 2205 set in the input place 2204 is clearly shown. Here, as an explicit method, it is only necessary to distinguish between the wafer layout 2201 and the setting area 2205, and the pattern and color of the setting area 2205 may be changed and displayed.
[0088]
Next, when the sensitivity setting button 2203 is pressed to set the sensitivity of the setting area 2205, another screen as shown in FIG. Sensitivity selection buttons 2206, 2207, and 2208 are prepared on this screen. Here, the sensitivity selection button 2206 is a button for selecting a mode for inspection with high sensitivity, the sensitivity selection button 2207 is a button for performing inspection with medium sensitivity, and the sensitivity selection button 2208 is a button for performing inspection with low sensitivity. . Note that FIG. 16D illustrates an example in which a mode for inspection with high sensitivity is selected. In order to change the sensitivity of the setting area 2205, one of the sensitivity selection buttons 2206, 2207, and 2208 may be selected.
[0089]
The content set by the above procedure is stored as an inspection condition file in the inspection region processing unit 1308 or the overall control unit 50. In the case of the overall control unit 50, the inspection condition file may be stored in the storage unit 53. The sensitivity for each setting region saved as an inspection condition file is fed back to the threshold value calculation unit 1411 of the threshold value calculation processing unit 1303 and stored as a threshold value map in the data storage unit 1412. As a result, in the foreign matter detection processing circuits 1307a to 1307n, at the time of inspection, the setting area 2205 is inspected with the detection sensitivity set in FIG. 16D, and the other areas are inspected with normal sensitivity. Here, as a sensitivity changing method, the threshold value set in the signal processing system 40 may be changed. For example, in the high sensitivity inspection mode, a threshold value that is normally set is set, in the medium sensitivity inspection mode, a value that is twice the threshold value set in the high sensitivity inspection mode is set as the threshold value in the intermediate inspection mode, and in the low sensitivity inspection mode. A value four times the threshold set in the high sensitivity inspection mode may be set as the threshold in the low sensitivity inspection mode.
In the above-described method, the example in which each setting is performed on a separate screen has been described. However, the screen need not necessarily be a separate screen, and may be set on the same screen.
[0090]
Next, the integrated processing unit 1309 and the inspection result display unit 1311 will be described. The integration processing unit 1309 integrates the foreign object detection results processed in parallel by the pixel merge circuits 1305 and 1306, integrates the feature amount calculated by the feature amount calculation circuit 1310 and the foreign object detection result, and displays the result on the result display unit 1311. Has the function to send. This inspection result integration process is desirably performed by a PC or the like in order to easily change the processing contents.
[0091]
First, the feature amount calculation circuit 1310 will be described. The feature amount is a value representing the feature of the detected foreign matter or defect, and the feature amount calculation circuit 1310 is a processing circuit that calculates the feature amount. As the feature amount, for example, the amount of reflected diffracted light (scattered light amount) (Dh, Dl) from the foreign matter or defect obtained by high angle illumination and low angle illumination, the number of detected pixels, the shape of the foreign matter detection region, and the direction of the principal axis of inertia , The detection location of foreign matter on the wafer, the circuit pattern type of the ground, the threshold value when foreign matter is detected, and the like.
[0092]
After detecting the foreign matter through the signal processing described above, an example of the inspection result display unit 1311 will be described. FIG. 17 is composed of an inspection map 2301 indicating position information of the foreign matter on the wafer and inspection information 2302. The inspection information 2302 displays an inspection date, a wafer type name, a process name, a wafer ID, and a detected number. It is an example. Thereby, the generation | occurrence | production state of the foreign material on a wafer can be understood at a glance.
[0093]
The means for outputting the detection result may be displayed on a display such as a CRT or printed as a hard copy. As a method for storing the inspection result, it may be recorded on a hard disk, magnetic recording medium, magneto-optical recording medium, optical recording medium, LSI memory, LSI memory card, or the like.
[0094]
Next, an example of DFC in the integrated processing unit 1309 will be described.
That is, since the foreign substance detection signal merged with various pixels is input to the integration processing unit 1309, the foreign substance is classified as “large foreign substance”, “fine foreign substance”, “low foreign substance” as shown in FIG. It becomes possible to classify as. FIG. 18 is a table showing the relationship between classification criteria and classification results. FIG. 18 shows an example using the detection result merged with 1 × 1 pixels and the detection result merged with 5 × 5 pixels. That is, from the foreign matter detection processing circuits 1307a and 1307c, the inspection result with 1 × 1 pixel and the inspection result with 5 × 5 pixels are obtained by the signal processing circuit. Using these results, classification is performed according to FIG. That is, if a foreign object is detected at 1 × 1 pixel or 5 × 5 pixel, it is classified as “large foreign object”. In addition, if it is detected at 1 × 1 pixel but not detected at 5 × 5 pixel, it is “fine foreign matter”. Further, it is not detected at 1 × 1 pixel, but it is detected at 5 × 5 pixel is “high”. Classify as "low foreign matter".
[0095]
FIG. 19 shows an embodiment of displaying the inspection result including the classification result. The display of the inspection result includes position information 2501 of foreign matters obtained from the detection location determination processing units 1602a and 1602c, category information 2502 of classification results obtained from the integration processing unit 1309, and the number of foreign matters 2503 for each category. In this embodiment, the position of a foreign object is indicated by the position information 2501 of the foreign object, and the classification category is also displayed by a display symbol. The contents of the classification category of each symbol are shown in the classification result category information 2502. The number of foreign objects 2503 for each category represents the number classified into each category. Thus, by changing the display for each category, there is an advantage that the distribution of each foreign substance can be known at a glance.
[0096]
Next, an embodiment of the foreign matter size measuring method according to the present invention will be described. This method is based on the fact that there is a proportional relationship between the size of the foreign matter and the amount of light detected by the photodetector 26. That is, when the foreign matter is small, the detected light quantity D is proportional to the sixth power of the foreign matter size G according to Mie's scattering theory. Therefore, the feature amount calculation circuit 1310 can measure the foreign substance size by the following equation (9) based on the detected light amount D, the foreign substance size G, and the proportionality coefficient ε, and can provide the measurement result to the integrated processing unit 1309.
[0097]
G = ε × D ^(1/6) (9)
Note that the proportionality coefficient ε may be obtained in advance from the amount of light detected from a foreign substance with a known size and input.
[0098]
Next, an example of a method for calculating the detected light amount D will be described with reference to FIG. 20A shows a digital image signal (image signal obtained by A / D converting the signal of the photodetector 26) obtained from the data storage unit 1302 for the fine foreign matter detected by the foreign matter detection processing circuit 1307. FIG. It is the image of the minute foreign material part created based on. A minute foreign matter portion 2601 indicates a signal of a minute foreign matter. FIG. 20B shows A / D conversion values (tone values for each pixel) of the minute foreign matter portion 2601 and the neighboring pixels in FIG. In this example, A / D conversion is performed with 8 bits, and the foreign matter signal unit 2602 indicates a detection signal from a minute foreign matter. Here, “255” in the central portion of the foreign matter signal portion 2602 indicates that the analog signal is saturated, and “0” portions other than the foreign matter signal portion 2602 indicate signals from other than fine foreign matter. . As a method of calculating the detected light amount D of the minute foreign matter, the sum of the pixel values of the foreign matter signal portion 2602 shown in FIG. For example, in the example of FIG. 20B, the detected light amount D of the minute foreign matter 2601 is “805” which is the sum of the pixel values.
[0099]
Next, another embodiment of a method for calculating the detected light amount D will be described. The idea of this embodiment is to improve the calculation accuracy of the detected light quantity by correcting the saturated portion of the foreign substance signal portion 2602 in FIG. 20B by Gaussian distribution approximation. The correction method will be described with reference to FIG. FIG. 21 is a diagram representing the Gaussian distribution three-dimensionally. FIG. 21 shows that y = y ₀ , The signal is saturated, and the method described below uses y = y in FIG. ₀ The lower part, ie V ₃ This is a method of calculating the detected light amount of the entire Gaussian distribution when the detected light amount of the part is obtained. First, the volume of the entire Gaussian distribution in FIG. ₁ , Y = y ₀ The volume of the upper part is V ₂ , Y = y ₀ The volume of the lower part is V ₃ And Further, it is assumed that the cross-sectional shape is obtained by the following equation (10) on the x axis of the Gaussian distribution of FIG.
y = exp (−x ² / 2 // σ ² (10)
At this time, V ₁ Is expressed by the following equation (11) by integrating around the y-axis.
V ₁ = 2 × π × σ ² (11)
In addition, V ₂ Is expressed by the following equation (12).
V ₂ = 2 × π × σ ² (Y ₀ Xlog (y ₀ ) + 1-y ₀ (12)
Note that “log” in the above equation indicates that the natural logarithm is calculated. Where the volume ratio V ₁ / V ₃ If CC is rewritten as CC, CC can be calculated by the following equation (13). Therefore, the following equation (14) is calculated from the above equations (11) and (12).
CC = V ₁ / (V ₁ -V ₂ (13)
CC = 1 / (y ₀ × (1-log (y ₀ ))) (14)
Here, assuming that the signal width of the saturated portion is SW, the following equation (15) is obtained, and therefore, CC can be represented by the following equation (16).
y ₀ = Exp (-SW ² / 2 // σ ² (15)
CC = exp (SW ² / 2 // σ ² ) / (1 + SW ² / 2 // σ ² (16)
Therefore, the detected light quantity obtained as shown in FIG. ₃ The volume V of the entire Gaussian distribution ₁ Can be calculated by the following equation (17). ₁ May be the corrected detected light amount D. It is necessary to calculate the signal width SW of the saturated portion.
V ₁ = V ₃ Xexp (SW ² / 2 // σ ² ) / (1 + SW ² / 2 // σ ² )
(17)
The method for calculating the detected light amount D has been described above. However, when the field of view of the variable magnification detection optical system 20 is wide, an error may occur due to lens distortion in the field of view. In this case, correction according to lens distortion in the field of view may be added.
[0100]
FIG. 22 shows an embodiment of display of the inspection result including the measured minute foreign matter size. The display of the inspection result includes foreign object position information 2701 obtained from the detection location determination processing unit 1602 and foreign object size histogram 2702 obtained from the integrated processing unit 1309. In this embodiment, the foreign object position information 2701 shows an approximate size of the foreign object by changing the size of the display symbol together with the position of the foreign object. The histogram 2702 is an example in which the size of the detected foreign matter is represented by a bar graph. By displaying the histogram in this way, there is an advantage that the size distribution of the foreign matter generated on the wafer can be immediately known.
[0101]
In this embodiment, the value of the signal sum of the foreign object signal unit 2602 is used as the detected light amount. However, the signal sum does not necessarily have to be the maximum value of the foreign object signal unit 2602. As an advantage, when the maximum value is used, the electric circuit scale can be reduced, and when the signal sum is used, the sampling error of the signal can be reduced, and a stable result can be obtained.
Note that the screens shown in FIGS. 15 to 17, 19, and 22 may be displayed on the display unit 52 provided in the overall control unit 50.
[0102]
Next, another embodiment of foreign matter or defect classification performed by the integrated processing unit 1309 will be described with reference to FIGS. FIG. 23 shows a sequence for classifying foreign matters based on the result of the inspection performed twice by the integrated processing unit 1309.
[0103]
First, the wafer 1 is inspected under the first inspection condition (S221). In the first inspection, the coordinate data of the foreign matter obtained from the foreign matter detection processing circuit 1307 and the feature amount of each foreign matter obtained from the feature amount calculation circuit 1310 are stored in a storage device (not shown) (S222). Next, the wafer 1 is inspected under a second inspection condition different from the first inspection condition (S223), and the foreign matter coordinate data and the feature amount obtained from the foreign matter detection processing circuit 1307 in the second inspection. The feature amount of each foreign object obtained from the calculation circuit 1310 is stored in a storage device (not shown) (S224). At this time, as the second inspection condition, for example, when the first inspection condition is irradiated with illumination light from an angle close to the wafer surface (in the case of low-angle illumination), the second inspection condition is the wafer surface. The condition for irradiating illumination light from an angle close to the normal line (high angle illumination condition) may be selected.
[0104]
Next, the obtained coordinate data of the first inspection result and the coordinate data of the obtained second inspection result are compared (S225). Are classified (S226). Here, as one embodiment of the method for determining that the coordinate data is close, the coordinate data obtained from the first inspection result is expressed as x. ₁ And y ₁ , X is the coordinate data obtained from the second inspection result ₂ And y ₂ If the comparison radius is r, the data corresponding to the following equation (18) may be determined to be the same.
(X ₁ -X ₂ ) ² + (Y ₁ -Y ₂ ) ² <R ² (18)
Here, r may be 0 or a value considering an error associated with the apparatus. As a measuring method, for example, the value of the left side of the equation (18) is calculated with the coordinate data of several foreign objects, and the value calculated by the equation (19) is set to r from the average value and the standard deviation value. good.
r ² = Average value + 3 x standard deviation (19)
Furthermore, a method for classifying foreign substances from foreign substance information regarded as the same object will be described with reference to FIG. In FIG. 24, the horizontal axis indicates the amount of scattered light (Dl), which is the characteristic amount obtained by the first inspection (low angle illumination), and the vertical axis indicates the second inspection (high angle illumination). 6 is a graph in which a scattered light amount (Dh) that is a characteristic amount is set. In FIG. 24, a point 3501 is a point plotted according to each feature amount of the foreign matter regarded as the same object. In this embodiment, one point indicates one foreign object. A classification line 3502 is a classification curve for classifying the foreign matter detected in the inspection. FIG. 24 shows an example in which the area is divided into two areas, that is, an area 3503 and an area 3504 by a classification line 3502. As a classification method, in FIG. 24, when the detected foreign matter is plotted in the region 3503, it is classified as “large foreign matter”, and when it is plotted in the region 3504, it is classified as “small foreign matter”.
[0105]
Here, the classification line 3502 needs to be determined in advance. As a method for determining in advance, it is only necessary to plot several detection objects that are known to be large or small in advance on the graph of FIG. 24 and set the classification line 3502 so that the detection objects can be correctly divided. Alternatively, the feature amount obtained from the foreign object may be calculated by simulation, and the classification line 3502 may be set based on the result.
In this embodiment, the example in which the inspection is performed twice has been described. However, when the type of feature amount (for example, the number of detected pixels: corresponding to the defect area Q) is increased, the classification performance can be improved. The feature amount (detection pixel number) of the foreign matter may be acquired by performing the inspection three times or more.
[0106]
Next, still another embodiment of the classification of foreign matters or defects performed by the inspection result integration processing unit 1309 will be described with reference to FIG. FIG. 25 shows a sequence of an embodiment in which inspection is performed once and classification is performed using feature amounts calculated under three kinds of optical conditions.
[0107]
First, the wafer 1 is inspected under the first optical condition (S241), and the foreign matter coordinate data obtained from the foreign matter detection processing circuit 1307 and the feature amount of each foreign matter obtained from the feature amount calculation circuit 1310 are also obtained. Save it (S242). Next, the optical conditions of the foreign matter inspection apparatus of the present invention are changed. This is, for example, the illumination angle or illumination direction of the illumination optical system. Further, the magnification of the detection optical system may be changed, and the optical filter may be changed. The condition with the above changes is defined as the second optical condition.
[0108]
After changing the optical condition to the second optical condition, the wafer 1 is moved to the position of the stored coordinates of the foreign matter by the transport system 30 and detected by the photodetector 26 under the second optical condition. Based on the detected image signal obtained by the / D conversion, the feature amount calculation circuit 1310 calculates the feature amount of the foreign matter (S243). Further, the same processing is performed when the feature amount is calculated under the third optical condition (S244). At this time, it is desirable that the first optical condition, the second optical condition, and the third optical condition are different from each other.
[0109]
The concept of the classification method will be described with reference to FIG. FIG. 26 shows a feature amount space in which three types of feature amounts are set on three axes. As the contents of the three axes, for example, the feature amount 1 is a feature amount (for example, scattered light amount (Dh)) from a defect acquired under the first optical condition (for example, high angle illumination), and the feature amount 2 is the second amount. The feature amount (for example, the scattered light amount (Dl)) from the defect acquired under the optical condition (for example, low angle illumination), and the feature amount 3 is the third optical condition (for example, the high angle illumination and the first optical condition). 2 is a feature amount (for example, the number of detected pixels: a planar area Q of the defect) obtained from the defect acquired under the optical condition 2 (low-angle illumination). In this feature amount space, (classification category number-1) classification boundaries are set. Since FIG. 26 is an example in which three types of classification are performed from three types of feature amounts, it is sufficient that there are two or more classification boundaries.
[0110]
In particular, as three types of feature amounts, the amount of scattered light (detected light amount) (Dh) from a defect by high-angle illumination, the amount of scattered light (detected light amount) (Dl) from a defect by low-angle illumination, and the defect at the time of the above high-angle illumination By using the number of detected pixels and the number of detected pixels at the time of low-angle illumination, it is possible to classify at least three categories (foreign matter defect, scratch defect, and circuit pattern defect). In addition, since the number of detected pixels (defect planar area Q) is taken as the feature amount, the category of the foreign object defect can be classified into a large foreign object and a small foreign object as shown in FIG. It becomes.
[0111]
In addition, the three feature quantities are the amount of scattered light from the defect at high imaging magnification, the amount of scattered light from the defect at low imaging magnification, and the number of detected pixels of the defect, so that at least the category of large foreign matter and small foreign matter Can be easily classified into categories.
[0112]
FIG. 26 shows an example in which classification boundaries 4501 and 4502 are set. As a classification method, first, the above three feature quantities are plotted in the feature quantity space of FIG. 26 (S245 shown in FIG. 25). Then, the foreign substances belonging to the areas divided by the classification boundaries 4501 and 4502 are classified as a category (for example, a foreign object defect) a, a category (for example, a flaw defect) b, and a category (for example, a circuit pattern defect) c, respectively (shown in FIG. 25). S246). FIG. 26 shows an example in which about 30 defects are classified into category a, category b, and category c, and the display symbols of the defects classified into the respective categories are changed. In other words, those classified into category a (for example, foreign object defects) are “◯”, those classified into category b (for example, flaw defects) are “▲”, and those classified into category c (for example, circuit pattern defects) are Displayed with “×”.
[0113]
Next, a classification boundary setting method will be described with reference to FIG. FIG. 27 shows a two-dimensional feature amount space in which three types of feature amounts are set on one axis. The feature amount space 4601 is a graph for classification based on the relationship between the feature amount 1 and the feature amount 2, and the feature amount spaces 4602 and 4603 are obtained from the relationship between the feature amount 1 and the feature amount 3, and the feature amount 2 and the feature amount 3, respectively. It is a graph for classifying.
[0114]
As a classification boundary setting method, first, feature quantities of foreign substances whose classification categories are known are plotted in the feature quantity spaces 4601, 4602, and 4603. Here, when plotting in the feature amount space, a display symbol or the like is changed for each category to express a difference in category. For example, in FIG. 27, category a is displayed as “◯”, category b as “▲”, and category c as “x”.
[0115]
Next, in each feature amount space 4601, 4602, 4603, classification boundaries 4604, 4605, 4606 are set at portions where categories can be divided. Here, when a plurality of categories overlap, it is not necessary to set a classification boundary. For example, in the feature amount space 4601, the category a is distributed at a position distant from the other categories b and c. Therefore, the classification boundary 4604 is set to classify the category a and the other categories b and c. Since the distributions of category b and category c overlap, it is not always necessary to set a classification boundary. When classifying a foreign object, the feature amount space 4601 is used to classify the category a or another category. Similarly, classification boundaries 4605 and 4606 are set in the feature amount spaces 4602 and 4603, and the classification boundaries are used when classifying foreign objects.
[0116]
The classification boundary setting method has been described above. In this example, the case where the classification boundary is divided into two areas has been described. However, when the distribution of three or more categories is clearly divided, a plurality of classification boundaries are set in order to divide into multiple areas. May be. Further, the classification boundary may be set by a straight line or a curve. The classification area may be set manually by the user or automatically calculated and set. In the case of manual setting, there is an advantage that the user can arbitrarily determine, and in the case of automatic setting, setting error by a person can be reduced. Here, as an automatic setting method, for example, in one feature amount space, the centroid of each category distribution is calculated, and a straight bisector of straight lines connecting the centroids may be used as a classification boundary. Further, the separation rate of each category may be displayed together in each feature amount space.
[0117]
An example in which the separation rate is displayed is shown in FIG. In FIG. 28, a display 4701 is a display of the separation rate. Here, the separation rate may indicate, for example, how much foreign matter of the same category is included in the region separated by the separation boundary. The advantage of displaying the separation rate is that the user can easily grasp the separation performance.
In this embodiment, the case of using feature amounts calculated under three types of optical conditions has been described. However, it is not always necessary to limit the number of features to three types. It can be used when a plurality of feature quantities can be acquired under various optical conditions.
[0118]
Next, another embodiment relating to the display of the test result obtained from the signal processing system 40 displayed on the display means 52 by the overall control unit 50 will be described.
FIG. 29 includes detected foreign matter or defect position information 3801, the number of detected foreign matters or defects 3802, and a detected foreign matter or defect size histogram 3803. In this embodiment, a case where a scratch is detected as a defect is shown.
[0119]
Specifically, the position information 3801 indicates the position of a foreign substance or a flaw on the wafer. In the present embodiment, an example is shown in which foreign matters are indicated by ◯ and scratches are indicated by ▲. The detected number 3802 is the number of detected foreign matters or scratches. Further, a graph 3803 is a histogram of the number of detected foreign objects or scratches and their size. By displaying the detected object in the defect inspection apparatus of the present invention in this way, the distribution of foreign matters or defects can be known at a glance.
[0120]
FIG. 30 includes an inspection map 3901 indicating the detection position of a detection object (foreign matter or defect), a histogram 3902 of the size of the detection object, and a review image 3903 of the foreign matter. In this embodiment, the inspection map 3901 and the histogram 3902 are examples in which all or part of the detected detected objects are displayed. Further, the review image 3903 is sampled for each size of the detected object, and the review image of the detected object is displayed. In this embodiment, six review images of foreign matters having a size of 0.1 μm or more and less than 1 μm are displayed. A case where six review images of foreign matters are displayed is shown.
[0121]
Here, the review image 3903 may be an image obtained by reflected diffracted light from a foreign substance detected by the detector 26, or may be an image obtained by an optical microscope using a white light source described later or a review device using a white light source. In the case of displaying an image by laser light, if the image is left in the storage devices 53, 1302, etc. during the inspection, there is an advantage that it can be displayed immediately after the inspection and the detected object can be confirmed quickly. When displaying an image by an optical microscope, an observation image may be taken after the inspection based on the coordinates of the sampled detection object, and a clearer image can be obtained as compared with an image by a laser beam. In particular, when observing foreign matters or defects of less than 1 μm, a high-resolution microscope using ultraviolet light as a light source is desirable.
[0122]
Further, the position of the detected object displayed on the review image 3903 may be displayed together on the inspection map 3901, and the detection number of the detected object may be displayed on the review image 3903 together. Further, in this embodiment, the case where the number of review images to be displayed is six has been described, but it is not necessary to limit the number to six, and all the detected foreign matters or defects may be displayed. You may display only the number of ratios.
[0123]
FIG. 31 shows an example in which detected objects are classified and displayed as foreign objects and scratches, and the correct answer rate of the classification is also displayed. FIG. 31 includes a detected number 4001 of each classified category, an inspection map 4002 indicating the detected position of the detected object, and a detected object confirmation screen 4003. The detected object confirmation screen 4003 further includes the present invention. The defect inspection apparatus includes a confirmation screen portion 4004 for a detection object classified as a foreign object, a confirmation screen portion 4005 for a detection object classified as a flaw, and a classification accuracy rate display portion 4006. The confirmation screens 4004 and 4005 are further composed of a detection object observation screen 4007 and a classification correct answer determination unit 4008.
[0124]
In this embodiment, the detected objects are classified into two categories. In the inspection map 4002, the symbol “1” is displayed as a foreign object, and the symbol “2” is displayed as a scratch.
[0125]
Next, a method for calculating the classification accuracy rate will be described. First, after inspection with the defect inspection apparatus of the present invention, the detection object 4007 is displayed on the observation screen portions 4004 and 4005, respectively. At this time, which of the observation screens 4004 and 4005 is displayed is displayed based on the result of classification by the defect inspection apparatus of the present invention. Next, the user of the defect inspection apparatus of the present invention inputs the category determined by the user into the classification correct answer determination unit 4008 associated with each observation screen 4007. In this example, as an input method, a case where a check box of a category determined by the user is checked is shown. In the foreign matter confirmation screen section 4004, 5/6 is checked as a foreign matter (category “1”). In this example, 6 is checked as a scratch (category “2”). Further, in the scratch confirmation screen section 4005, all are determined as scratches (category “2”).
[0126]
After the above check, the correct answer rate is displayed on the classification correct answer display unit 4006. This value displays, for example, the rate at which the classification result in the defect inspection apparatus of the present invention matches the classification result of the user. After this, for the detected object in which the classification result in the defect inspection apparatus of the present invention and the user's classification result do not match, the classification condition is set to improve the classification accuracy using the feature amount of the detected object. It may be updated.
[0127]
[Overall control unit 50]
Next, inspection condition (inspection recipe) setting performed in the overall control unit 50 and the like will be described with reference to FIGS. FIG. 32 is a diagram illustrating a flow for setting inspection conditions (inspection recipes). First, inspection conditions (inspection recipes) set in the overall control unit 50 before performing inspection include chip layout setting (S211) that matches the inspection target, rotation alignment of the inspection target (S212), and inspection region setting. (S213), optical condition setting (S214), optical filter setting (S215), detected light amount setting (S216), and signal processing condition setting (S217). Note that S218 is execution of actual inspection.
[0128]
Next, each setting executed by the overall control unit 50 will be described. First, the chip layout setting (S211) is to set the chip size and the presence / absence of a chip on the wafer in the signal processing system 40 and the like in the overall control unit 50 by CAD information or the like. This chip size needs to be set because it is a distance for performing comparison processing. Next, in the rotation alignment setting (S212), the alignment direction of the chips on the wafer 1 placed on the stage and the pixel direction of the photodetector 26 controlled by the overall control unit 50 with respect to the transport system 30 are parallel. In other words, this is a setting for rotating the wafer 1 in order to make the rotational deviation substantially “0”. By performing this rotation alignment, the repetitive patterns on the wafer 1 are arranged in the uniaxial direction, so that the chip comparison signal processing can be easily performed. Next, the inspection area setting (S213) is to perform setting of an inspection place on the wafer and setting of detection sensitivity in the inspection area, which are controlled by the overall control unit 50 for the signal processing system 20. By performing this inspection area setting (S213), each area on the wafer can be inspected with optimum sensitivity. The setting method is as described in the description of FIG.
[0129]
Next, in the optical condition setting (S214), the overall control unit 50 controls the illumination optical system 10 and the variable magnification detection optical system 20 to select the direction and angle of the illumination light applied to the wafer, or to change the magnification. The magnification of the detection optical system 20 is selected. As a selection method, for example, an optical condition setting window as shown in FIG. 33 may be used. The optical condition setting screen includes an illumination direction condition 3001 of the illumination optical system, an illumination angle condition 3002 of the illumination optical system, and a detection optical system condition 3003. In FIG. 33, two types of illumination direction conditions 3001, three types of illumination angle conditions 3002, and two types of detection optical system conditions 3003 can be selected. The user of this foreign substance inspection apparatus may select an appropriate condition by looking at the contents of the conditions 3001, 3002, and 3003. For example, if the object to be inspected 1 is a wafer in a metal film deposition process and it is desired to inspect the foreign matter on the surface with high sensitivity, the “depot process” of the illumination direction condition 3001 is selected, and further the illumination angle condition 3002 The condition “surface foreign matter” is selected, and the detection optical system condition 3003 is selected as “high sensitivity inspection”. FIG. 33 shows an example in which these selections are made.
[0130]
Next, the optical filter setting (S215) is to set the optical filter 25 such as the spatial filter 22 and the polarization element shown in FIG. 1 which the overall control unit 50 controls the detection optical system 20 and the like. Since this spatial filter 22 is a filter for shielding reflected diffracted light from a repetitive pattern manufactured on a wafer, it is better to set it for a wafer having a repetitive pattern, but a wafer without a repetitive pattern. It is not necessary to set for. The polarizing element 25 is effective when it is used when the edge of the wiring pattern is etched at a right angle.
[0131]
Next, the detection light amount setting (S216) is a step of adjusting the light amount incident on the photodetector 26, which is controlled by the overall control unit 50 with respect to the illumination optical system 10 or the variable magnification detection optical system 20. The reflected and scattered light from the circuit pattern manufactured on the wafer varies in components scattered depending on the pattern shape. Specifically, when the wafer surface is flat, little scattered light is generated and most of the light is specularly reflected. On the other hand, when the unevenness of the wafer surface is large, a lot of scattered light is generated. Therefore, the reflected and scattered light from the circuit pattern varies depending on the state of the wafer surface, that is, the device manufacturing process. However, since the dynamic range of the photodetector 26 exists, it is desirable to adjust so that the amount of light matched to this dynamic range is incident. For example, it is desirable to adjust so that the amount of reflected and scattered light from the circuit pattern of the wafer is about 1/10 of the dynamic range of the photodetector 26. Here, as a method of adjusting the amount of light incident on the photodetector 26, the amount of light output from the laser light source 11 may be adjusted, or may be adjusted by the ND filter 24.
[0132]
Next, the signal processing condition setting (S217) is to set a foreign object detection condition that is controlled by the overall control unit 50 for the signal processing system 40.
After the above setting is completed, if the inspection is performed in the inspection step (S218), the user can perform the inspection under desired conditions.
However, as a method of setting the contents described in the present embodiment, for example, it may be input manually from the design information of the inspection target, or input using the input assist function attached to the foreign matter inspection apparatus of the present invention. Alternatively, information may be acquired from a host system via a network.
[0133]
Further, among the above-described settings, the inspection area setting (S213), optical condition setting (S214), optical filter setting (S215), detected light amount setting (S216), and signal processing condition setting (S217) are not necessarily set depending on the object to be inspected. There is no need to change, and a constant value may be used regardless of the object to be inspected. When a constant value is set, the time for setting the inspection condition can be shortened, but it is desirable to tune each condition in order to achieve high sensitivity. The inspection area setting (S213) is not necessarily performed before the optical condition setting (S214), and may be set before the inspection process (S218).
[0134]
An example of a screen for setting the contents described above is shown in FIG. FIG. 34 includes a condition setting sequence 4301, detailed conditions 4302 for each setting content, a setting content display change button 4303, and a help button 4304.
Next, details will be described. First, a condition setting sequence 4301 shows a flow of setting inspection conditions in the foreign matter inspection apparatus of the present invention. The user may set conditions in order from “chip layout setting” in the condition setting sequence 4301.
[0135]
A feature of the condition setting sequence 4301 is that the flow of condition setting is indicated by an arrow 4305 so that the user can set in the shortest order without making a mistake in the setting order. Another feature is that it is divided into items that must be set and items that do not necessarily have to be set, that is, items that have default values. By dividing the display, if the minimum setting items are known and the user needs the inspection results immediately, it is only necessary to set and inspect only the mandatory setting items. Since it is only necessary to set conditions for items that do not exist, the degree of condition setting can be changed according to the user's request. For example, the button 4306 indicates that the item should be set by indicating the frame in triplicate, and the button 4307 indicates that the setting is low by indicating the frame in single. It is the Example shown. Another feature is to clearly indicate which item the user is currently setting. For example, the button 4308 is distinguished from the buttons 4306 and 4307 by adding a shadow to the button. Thus, by clearly indicating the current location, there is an advantage that the number of remaining setting items can be understood at a glance.
In this embodiment, an option condition setting 4309 is added to the sequence described with reference to FIG. The contents of this optional condition setting 4309 are, for example, setting of conditions for the size measurement function of foreign matter and setting of classification conditions for foreign matter and defects.
[0136]
Next, a detailed condition 4302 is a screen for setting details of each condition item. As an item input or selection method, a place for inputting with a keyboard like the input box 4310 may be provided, or a method for selecting an input item with an icon like the input icon 4311 may be used. Note that the input icon 4311 is an example for each of the three types of input items, and another window appears when the corresponding icon is pressed, and detailed condition setting is performed. Furthermore, a method of selecting a necessary item like an input check box 4312 may be used.
[0137]
A setting content display change button 4303 is a button for changing or customizing display items. For example, when there is an item that the user always wants to set or an item for which the number of setting contents is to be increased, the setting contents display change button 4303 can be changed to make the user-friendly screen. Inspection conditions can be set quickly. Further, a help button 4304 is a button for outputting information for helping the user when the user cannot understand the setting method and setting contents. As a technique, the contents of each setting item may be voice-guided or the operation method may be shown as a moving picture such as MPEG. Further, it may be possible to communicate with a designer of a manufacturer who manufactured the foreign matter inspection apparatus of the present invention online through a network or a telephone line.
[0138]
[Embodiment with microscope]
Another embodiment relating to the defect inspection apparatus according to the present invention will be described with reference to FIG. In the present embodiment, the difference from the configuration shown in FIG. 1 is that an observation microscope 60 is provided in parallel. The observation microscope 60 moves the detected foreign matter (including false information) on the wafer 1 to the position of the detection optical system of the observation microscope 60 by moving the stages 31 and 32, and observes this image. It is.
[0139]
An advantage of arranging the observation microscope 60 in parallel is that the detected foreign matter can be observed immediately without moving the wafer to the review apparatus. By immediately observing the detected object in the inspection apparatus, the cause of the foreign matter can be quickly identified. The observation microscope 50 may be a microscope that uses visible light (for example, white light) as a light source, or a microscope that uses ultraviolet light as a light source. In particular, in order to observe a minute foreign substance having a level of 0.1 μm, a high-resolution microscope, for example, a microscope using ultraviolet light is desirable. In addition, when a visible light microscope is used, the color information of the foreign matter can be obtained and the foreign matter can be easily recognized.
[Embodiment when the defect inspection apparatus of the present invention is connected to a system]
A system in which the defect inspection apparatus according to the present invention is connected to a semiconductor device manufacturing line will be described with reference to FIG. This system includes a semiconductor device manufacturing process 3101 and an inspection process 3102, a foreign substance inspection apparatus group 3103, a database 3104, and a measurement apparatus group 3105 for managing the process state. Here, the inspection apparatus group 3103 is, for example, a defect inspection apparatus according to the present invention. The measuring device group 3105 includes, for example, a device for determining whether a chip is good or defective by an electrical test, a device for performing component analysis, a device for measuring the thickness of a film applied to the device, and a wiring formed on the device. A device for measuring the width of the film, a device for measuring the conductivity between circuit patterns, and a device for observing foreign matter and defects. Further, the database 3104 stores inspection results in the inspection apparatus group 3103, measurement results in the measurement apparatus group 3104, information on device manufacturing processes, past defect cases, and the like.
[0140]
Next, the operation will be described. First, a semiconductor device is manufactured through each manufacturing process along a device manufacturing process 3101. The inspection process 3102 set in the middle of the device manufacturing process 3101 is inspected for foreign matters or defects, and if the inspection result is abnormal, the inspection result is matched with the information in the database 3104, and the countermeasure method for abnormality is the device manufacturing process. 3101 is fed back. The inspection results used here are the contents displayed by the inspection results of the defect inspection apparatus of the present invention and the data obtained by the defect inspection apparatus.
[0141]
【The invention's effect】
According to the present invention, diffracted light from a circuit pattern on a substrate such as an LSI pattern can be reduced, and a minute foreign matter of 0.1 μm level or a target object in which a repeated pattern and a non-repeated pattern are mixed There is an effect that a defect, a foreign substance that short-circuits between wirings, or a defect or a thin film-like foreign substance can be inspected at high speed and with high accuracy.
[0142]
Further, according to the present invention, there is an effect that the classification and size of the detected foreign matter or defect can be measured.
[Brief description of the drawings]
FIG. 1 is a schematic configuration diagram showing an embodiment of a defect inspection apparatus according to the present invention.
FIG. 2 is a plan view showing the illumination optical system shown in FIG.
FIG. 3 is a diagram for explaining the function of a conical curved lens used in an illumination optical system.
4 is an explanatory diagram of variable operations of the variable magnification optical system shown in FIG. 1;
FIG. 5 is a schematic diagram showing a mechanism for moving a lens in a variable magnification optical system.
FIG. 6 is a configuration diagram showing another embodiment of the variable magnification optical system.
FIG. 7 is a diagram for explaining another embodiment of the photodetector.
8 is a block diagram showing a signal processing system shown in FIG.
FIG. 9 is a configuration diagram illustrating a threshold value calculation processing unit.
FIG. 10 is a configuration diagram illustrating a pixel merge circuit.
FIG. 11 is a configuration diagram illustrating a foreign object detection processing unit.
FIG. 12 is an explanatory diagram of a detection location specifying process.
FIG. 13 is an explanatory diagram of a detection location specifying process.
FIG. 14 is an explanatory diagram of a detection location specifying process.
FIG. 15 is a diagram for explaining types of inspection area settings;
FIG. 16 is a diagram for explaining an inspection area setting method;
FIG. 17 is a diagram showing a display example of inspection results.
FIG. 18 is a diagram for explaining a foreign matter or defect classification method;
FIG. 19 is a diagram showing a display example of inspection results when foreign matters or defects are classified.
FIG. 20 is a diagram for explaining a foreign matter or defect size measurement method;
FIG. 21 is a diagram for explaining another embodiment relating to a method of calculating the amount of scattered light from a foreign substance or a defect.
FIG. 22 is a diagram illustrating a display example of an inspection result when the size of a foreign object or a defect is measured.
FIG. 23 is a diagram showing a sequence of another embodiment relating to classification of foreign matter or defects.
FIG. 24 is a diagram illustrating a classification graph used for classification of foreign matters or defects.
FIG. 25 is a diagram showing a sequence of still another example relating to the classification of foreign matters or defects.
FIG. 26 is a diagram for explaining a method of classifying from a plurality of types of feature amounts of foreign matters or defects.
FIG. 27 is a diagram for explaining a method of setting a classification boundary;
FIG. 28 is a diagram illustrating a display example when displaying a classification rate.
FIG. 29 is a diagram illustrating an example of a display in which a classification result of a foreign matter or a defect and a size measurement result are written together.
FIG. 30 is a diagram illustrating an example of a display in which a size measurement result of a foreign matter or defect and an observation image of the foreign matter or defect are written together.
FIG. 31 is a diagram showing an example of a display in which the classification accuracy rate of a foreign object or a defect is written together with the inspection result.
FIG. 32 shows an inspection condition setting sequence in the defect inspection apparatus according to the present invention.
FIG. 33 is a diagram for explaining an optical condition setting screen.
FIG. 34 is a diagram for explaining a classification graph of another embodiment relating to the classification of foreign matters or defects.
FIG. 35 is a schematic configuration diagram of an embodiment provided with an observation microscope.
FIG. 36 is a diagram showing an embodiment of a system in which a defect inspection apparatus according to the present invention is connected to a production line.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Wafer (substrate to be inspected), 10 ... Illumination optical system, 11 ... Laser light source, 12 ... Concave lens, 13 ... Convex lens, 14 ... Conical curved lens, 15 ... Mirror, 20 ... Variable magnification detection optical system, 21 ... Objective Lens, 22 ... Spatial filter, 23 ... Imaging lens, 24 ... ND filter, 25 ... Optical filter, 26 ... Optical detector, 30 ... Transport system, 31 ... x stage, 32 ... y stage, 33 ... z stage, 34 DESCRIPTION OF SYMBOLS ... Sample mounting stage, 35 ... Stage controller, 40 ... Signal processing system, 50 ... Overall control unit, 60 ... Observation microscope, 201 ... Slit beam, 202 ... Chip, 203 ... Pixel direction, 701 ... Illumination position, 702 ... Mirror, 801 ... Branch optical element, 802, 804 ... Mirror, 805 ... Branch optical element, 806, 807, 808 ... Mirror, 809, 810 ... Mirror element, 1 01, 1202 ... movable lens, 1203 ... fixed lens, 1204 ... moving mechanism, 1301 ... A / D converter, 1302 ... data storage unit, 1303 ... threshold value calculation processing unit, 1305, 1306 ... pixel merge circuit, 1307 ... foreign matter detection Processing circuit, 1308... Inspection region processing unit, 1309... Integration processing unit, 1310... Feature amount calculation circuit (feature amount calculation unit), 1311... Result display unit, 1401. Deviation calculation value section, 1409 ... average value calculation section, 1410 ... coefficient calculation section, 1411 ... threshold calculation section, 1412 ... data storage section, 1601 ... comparison circuit, 1602 ... detection location determination processing section, 1708, 1709 ... difference signal, 1908, 1909 ... differential signal, 2008, 2009 ... differential signal, 2202 ... area setting button, 2203 ... sensitivity Constant button, 2205 ... setting area, 2502 ... classification result of the category information, 2503 ... the number of foreign matter of each category.

Claims (10)

An illumination optical system configured to irradiate an illumination light beam emitted from an illumination light source by switching between a high inclination angle and a low inclination angle with respect to the surface of the substrate to be inspected;
An objective lens that collects the reflected and scattered light from the substrate to be inspected, an imaging optical system that forms an image of the reflected and scattered light collected by the objective lens, and a reflected and scattered light imaged by the imaging optical system A detection optical system having a photodetector that receives the light and converts it into an image signal;
An A / D converter that converts an image signal obtained from the photodetector of the detection optical system into a digital image signal when illuminated at a high tilt angle with the illumination optical system and when illuminated at a low tilt angle with the illumination optical system And a difference signal between similar circuit patterns between different chips using the digital image signal converted by the A / D conversion unit, and a detection threshold value and the detection threshold value for the obtained difference signal When detecting a defect using a smaller verification threshold, a defect is detected by changing the detection threshold and the verification threshold according to the distance from the center on the surface of the substrate to be inspected. A defect detection processing unit, a feature amount calculation unit for calculating a feature amount of the defect detected from the defect detection processing unit, a defect detected from the defect detection processing unit when illuminated at the high inclination angle, and the low inclination When illuminated at an angle Integrated processing for acquiring a feature amount of a defect identified with a defect detected from the defect detection processing unit from the feature amount calculation unit, and classifying the defect by category based on the acquired feature amount of the defect And a signal processing system having a section.   The defect inspection apparatus according to claim 1, wherein in the integrated processing unit, the feature amount of the defect acquired from the feature amount calculation unit for classifying the defect by category is a detected light amount obtained from the defect. In the defect detection processing unit, further, that the merging in the converted neighboring pixel digital image signal by the A / D conversion unit, configured to detect the defect by using the merged digital image signal The defect inspection apparatus according to claim 1, wherein:   4. The defect inspection apparatus according to claim 1, wherein the imaging optical system is configured by a variable magnification imaging optical system capable of imaging at different imaging magnifications. 5. .   In the illumination optical system, the illumination light beam is an illumination state on the substrate to be inspected, and is formed into a slit beam composed of substantially parallel light in the longitudinal direction, and the longitudinal direction is substantially perpendicular to the traveling direction of the scanning stage. The defect inspection apparatus according to any one of claims 1 to 3, wherein   The defect inspection apparatus according to claim 1, wherein the illumination light source is a laser light source in the illumination optical system.   4. The defect inspection apparatus according to claim 1, wherein the detection optical system includes a spatial filter that blocks reflected diffracted light from a repetitive pattern on the inspection target substrate. 5. The illumination light beam emitted from the illumination light source is irradiated to the surface of the substrate to be inspected by the illumination optical system at the first inclination angle, and the reflected scattered light from the irradiated substrate to be inspected is collected by the objective lens Then, an image is formed by the imaging optical system, the imaged reflected and scattered light is received by a photodetector and converted into a first image signal, and the converted first image signal is A / D converted. A first digital image signal is converted by the detector, and a difference signal between similar circuit patterns between different chips is obtained using the converted first digital image signal , and the obtained difference signal is detected. the verified and the detection threshold value according to the distance from the center of the object substrate on the surface when detecting the defect by using the small verification threshold than threshold and the defect detection threshold to detect the defect by changing the threshold value, attached to the detected defect A first step of calculating on the basis of the feature quantity in the first digital image signal,
The illumination light beam emitted from the illumination light source is irradiated to the surface of the substrate to be inspected by the illumination optical system at a second inclination angle different from the first inclination angle, and from the irradiated substrate to be inspected. The reflected and scattered light is collected by the objective lens and imaged by the imaging optical system. The imaged reflected and scattered light is received by the photodetector and converted into a second image signal. 2 is converted into a second digital image signal by an A / D converter, and a difference signal between similar circuit patterns between different chips is obtained by using the converted second digital image signal. When detecting a defect using a detection threshold and a verification threshold smaller than the detection threshold with respect to the obtained difference signal, the distance is determined according to the distance from the center on the surface of the substrate to be inspected. detecting a defect by changing the and the verification thresholds and detection threshold A second step of calculating on the basis of characteristic amounts of the detected defect in the second digital image signal,
The feature amount calculated in the first step for the defect in which the defect detected in the first step and the defect detected in the second step are identified and calculated in the second step And a classification step for classifying the defects into categories based on the feature amounts.   In the classification step, the feature amount calculated in the first step and the feature amount calculated in the second step for classifying defects by category are detected light amounts obtained from the defects. The defect inspection method according to claim 8. In the first step, further, the merge with neighboring pixels the converted first digital image signal by the A / D converter to detect the defect by using the merged digital image signal,
In the second step, further, that the merged with neighboring pixels the converted first digital image signal by an A / D converter, for detecting the defect by using the merged digital image signal The defect inspection method according to claim 8, wherein:

Images