JP2005003689A - Method and apparatus for inspecting defect in pattern on object to be inspected - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a defect inspection method and a defect-inspecting apparatus for detecting minute patterns with high resolution, and to provide a method for manufacturing a semiconductor substrate with high yields. <P>SOLUTION: A lamp house 24 for forming zone-like illumination, comprises an Xe lamp 3, an elliptic mirror 4, and a mask 5, allows a wafer 1 (object to be inspected), wherein a fine pattern is formed, placed on an XYZ stage 2 to be subjected to circular or elliptical polarization by a circular or ellipticably polarized light conversion element via a collimator lens 6, a filter 14 for adjusting the quantity of light, and a capacitance lens 7 by zone-like illumination light for irradiation via an objective lens 9. The image of the reflected light is detected by an image sensor 12a via half mirrors 8a, 8b, and a zoom lens 13. The wafer 1 is moved on the stage XYZ and is scanned by a sensor 12a for obtaining an image signal. The image output is converted by an A/D converter 15a and is compared with a reference image for detecting non-conformance as a defect. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a semiconductor substrate manufacturing method for manufacturing a semiconductor substrate such as a semiconductor wafer having a fine circuit pattern or a wiring pattern, a TFT liquid crystal substrate, a thin film multilayer substrate, and a printed circuit board at a high yield, a semiconductor wafer, and a TFT (Thin Film Transistor). ) Pattern on the inspection object that measures and detects the high-precision dimension of the inspection pattern consisting of a fine circuit pattern or wiring pattern formed on the inspection object such as a liquid crystal substrate, a thin film multilayer substrate, and a printed circuit board The present invention relates to a detection method and apparatus therefor, a pattern defect inspection method and apparatus therefor for performing a fine defect inspection on a pattern on the inspection object, and a microscope used therefor.

  Recently, for example, a pattern to be inspected consisting of a circuit pattern or a wiring pattern formed on a semiconductor wafer, a TFT liquid crystal substrate, a thin film multilayer substrate, a printed circuit board, etc. has been miniaturized in response to the need for high integration. Yes. By the way, the circuit pattern or the wiring pattern is further miniaturized as the degree of integration increases, so that the defects that must be detected are also increasingly minute. Detection of such minute defects has become a very important issue in determining the quality of these circuit patterns or wiring patterns when manufacturing circuit patterns or wiring patterns. However, since the miniaturization has further progressed and minute defect detection has reached the resolution limit of the imaging optical system in a pattern to be inspected such as a circuit pattern or a wiring pattern, an improvement in essential resolution is required. Yes. Japanese Laid-Open Patent Publication No. 5-160002 is known as a conventional technique for improving the resolution essentially. In this prior art, there is provided an illuminating means for irradiating a minute circuit pattern formed on a mask with a ring-shaped diffused illumination in which a large number of virtual point light sources are arranged by a light source spatial filter, and the illumination The diffracted light from the fine pattern that is transmitted or reflected through the mask that is diffusely illuminated almost uniformly by the means is sufficiently captured, and at least a part of the 0th-order diffracted light or the low-order diffracted light is shielded from the captured light. A light receiving means for receiving a circuit pattern formed through the optical pupil and obtaining an image signal, and an image signal obtained from the light receiving means and a mask. A pattern inspection apparatus including a comparison unit that inspects a pattern in comparison with pattern data, wafer pattern data, or data from a transfer simulator is described. This prior art also describes controlling the shapes of the light source spatial filter and the imaging spatial filter in accordance with the pattern shape data.

JP-A-5-160002

  Recently, for example, a pattern to be inspected consisting of a circuit pattern or a wiring pattern formed on a semiconductor substrate such as a semiconductor wafer, a TFT liquid crystal substrate, a thin film multilayer substrate, or a printed substrate has been miniaturized in response to the need for high integration. It is illustrated. By the way, since circuit patterns or wiring patterns are increasingly miniaturized (densified) with higher integration, defects that must be detected also become increasingly minute. However, in the above prior art, an annular diffuse illumination is applied to the fine pattern on the object to be inspected, and the diffracted light from the fine pattern is sufficiently taken into the aperture (pupil) of the objective lens. With this method, high-definition image signals are obtained and fine pattern defects are detected. However, it is highly reliable to detect fine defects corresponding to various fine patterns existing on the object to be inspected. It had the problem of not being fully considered. In addition, there is a problem that a semiconductor substrate such as a semiconductor wafer having a fine pattern, a TFT liquid crystal substrate, a thin film multilayer substrate, and a printed circuit board is not sufficiently considered in manufacturing a high yield by reducing the occurrence of defects. It was.

  An object of the present invention is to manufacture a semiconductor substrate that solves the above-described problems of the prior art and manufactures a semiconductor substrate having a fine pattern, such as a semiconductor wafer, a TFT liquid crystal substrate, a thin film multilayer substrate, and a printed substrate, at a high yield. It is to provide a method. Another object of the present invention is to detect a fine pattern with high reliability corresponding to various fine patterns existing on an object to be inspected such as a semiconductor wafer, a TFT liquid crystal substrate, a thin film multilayer substrate, and a printed circuit board. Another object of the present invention is to provide a pattern detection method and apparatus (microscope system) for an object to be inspected. Another object of the present invention is to provide a highly reliable fine defect in a fine pattern corresponding to various fine patterns existing on an object to be inspected such as a semiconductor wafer, a TFT liquid crystal substrate, a thin film multilayer substrate, and a printed circuit board. It is an object of the present invention to provide a method and an apparatus for inspecting a defect on a pattern on an inspection object.

  In order to achieve the above object, the present invention provides a semiconductor substrate manufacturing method for manufacturing a semiconductor substrate having a pattern by a manufacturing line constituted by various process apparatuses, and includes defect information on a pattern generated on the semiconductor substrate in advance. History data or database construction step of building history data or database showing causal relationship by accumulating history data or database showing defect cause or defect cause causing pattern defect in the production line; The semiconductor substrate that has reached a desired location on the production line is irradiated with condensed diffused illumination light formed from a number of virtual point light sources through the pupil of an objective lens. 1st order or diffracted light including 0th order diffracted light reflected from the pattern and entering the pupil of the objective lens The image of the pattern on the object to be inspected obtained by condensing the second-order diffracted light is received by the image sensor and converted into a pattern image signal, and the image signal of this pattern is compared with the image signal of the reference pattern A defect inspection process for detecting defect information of the pattern, and the defect information of the pattern detected by the defect inspection process and the history data or database indicating the causal relationship constructed in the history data or database construction process A defect occurrence analysis step for analyzing a defect occurrence cause or a defect occurrence factor causing a pattern defect in a production line upstream from a desired location of the production line, and a defect occurrence cause or a defect occurrence factor analyzed by this defect occurrence analysis step A process condition system for controlling process conditions in the upstream production line to eliminate Is a manufacturing method of a semiconductor substrate, characterized by a step.

  According to another aspect of the present invention, there is provided a semiconductor substrate manufacturing method for manufacturing a semiconductor substrate having a pattern by a manufacturing line constituted by various process apparatuses. In the semiconductor substrate manufacturing method, pattern defect information generated on the semiconductor substrate in advance and the pattern in the manufacturing line are detected. History data or database construction process for building history data or database showing causal relationship by accumulating history data or database showing defect cause or cause of defect causing defect, and desired location of said production line The annular diffused illumination light formed from a number of virtual point light sources controlled so as to be suitable for the pattern formed on the semiconductor substrate is condensed through the pupil of the objective lens. The objective lens reflected from the pattern on the semiconductor substrate A high-resolution image of the pattern on the object to be inspected obtained by condensing the first-order or second-order diffracted light including the 0th-order diffracted light incident on the pupil is received by the image sensor and converted into a pattern image signal. A defect inspection process for detecting defect information of the pattern by comparing the image signal of the pattern with the image signal of the reference pattern, the defect information of the pattern detected by the defect inspection process, and the history data or database construction process A defect occurrence analysis step of analyzing a defect occurrence cause or a defect occurrence factor that causes a pattern defect in a production line upstream of a desired location of the production line based on history data or a database showing a causal relationship constructed in The upstream manufacturing so as to remove the cause of the defect or the cause of the defect analyzed by this defect occurrence analysis process. Is a manufacturing method of a semiconductor substrate characterized by having a process condition control step of controlling the process conditions in the in.

  According to another aspect of the present invention, there is provided a semiconductor substrate manufacturing method for manufacturing a semiconductor substrate having a pattern by a manufacturing line constituted by various process apparatuses. In the semiconductor substrate manufacturing method, pattern defect information generated on the semiconductor substrate in advance and the pattern in the manufacturing line are detected. History data or database construction process for building history data or database showing the causal relationship by accumulating history data or database showing cause of defect or cause-and-effect relationship that causes defect, and desired location of the production line Condensed illumination light, which is formed from a large number of virtual point light sources controlled based on the image on the pupil of the objective lens, is condensed and irradiated through the pupil of the objective lens to the semiconductor substrate that has reached 0 reflected from the pattern on the semiconductor substrate and entering the pupil of the objective lens A high-resolution image of a pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including diffracted light is received by an image sensor and converted into a pattern image signal, and the image of this pattern A defect inspection process for detecting defect information of a pattern by comparing a signal with an image signal of a reference pattern, and the causal relationship established by the defect information of the pattern detected by this defect inspection process and the history data or database construction process A defect occurrence analysis step for analyzing a defect occurrence cause or a defect occurrence factor for generating a pattern defect in a production line upstream of a desired location of the production line based on history data or a database indicating In order to remove the analyzed cause of the defect or the cause of the defect, Is a manufacturing method of a semiconductor substrate characterized by having a process condition control step of controlling the scan condition.

  According to another aspect of the present invention, there is provided a semiconductor substrate manufacturing method for manufacturing a semiconductor substrate having a pattern by a manufacturing line constituted by various process apparatuses. In the semiconductor substrate manufacturing method, pattern defect information generated on the semiconductor substrate in advance and the pattern in the manufacturing line are detected. History data or database construction process for building history data or database showing the causal relationship by accumulating history data or database showing cause of defect or cause-and-effect relationship that causes defect, and desired location of the production line Polarized / annular diffuse illumination light, which is formed by further applying polarized light to the annular diffuse illumination light formed from a number of virtual point light sources, is collected through the pupil of the objective lens. Irradiate with light, reflect from the pattern on the semiconductor substrate and enter the pupil of the objective lens An image of the pattern on the object to be inspected obtained by condensing the first-order or second-order diffracted light including the 0th-order diffracted light that is emitted is received by an image sensor and converted into a pattern image signal. A defect inspection process for detecting defect information of a pattern by comparing a signal with an image signal of a reference pattern, and the causal relationship established by the defect information of the pattern detected by this defect inspection process and the history data or database construction process A defect occurrence analysis step for analyzing a defect occurrence cause or a defect occurrence factor for generating a pattern defect in a production line upstream of a desired location of the production line based on history data or a database indicating In order to remove the analyzed cause of the defect or the cause of the defect, Is a manufacturing method of a semiconductor substrate characterized by having a process condition control step of controlling the scan condition.

  In the method for manufacturing a semiconductor substrate, the present invention is characterized in that the polarized light in the defect inspection step is circular or elliptically polarized light. According to another aspect of the present invention, there is provided a semiconductor substrate manufacturing method for manufacturing a semiconductor substrate having a pattern by a manufacturing line constituted by various process apparatuses. Historical data or database construction that builds historical data or database that shows the causal relationship by accumulating historical data or database that shows the cause of the defect and foreign matter on the pattern and the cause of the defect or the cause of the defect or foreign matter. Irradiating the semiconductor substrate that has reached a desired location on the manufacturing line with a ring-shaped diffused illumination light formed from a number of virtual point light sources through a pupil of an objective lens. The 0th-order diffracted light reflected from the pattern on the substrate and entering the pupil of the objective lens The image of the pattern on the object to be inspected obtained by condensing the first-order or second-order diffracted light is received by an image sensor and converted into a pattern image signal. A defect inspection process for detecting defect information of a pattern in comparison with an image signal, and a dark field illumination is performed on the semiconductor substrate reaching a desired location on the manufacturing line to detect foreign matter on the pattern on the semiconductor substrate. Foreign matter on the pattern on the object to be inspected obtained by condensing scattered light that is reflected and incident on the pupil of the objective lens is received by the image sensor and converted into a signal indicating the foreign matter, and the foreign matter on the pattern Foreign matter inspection process for detecting information, defect information on the pattern detected by the defect inspection process, foreign matter information on the pattern detected by the foreign matter inspection process, and history data Is a defect that causes a pattern defect and a foreign matter on the pattern in the production line upstream of the desired location of the production line based on the history data or database showing the causal relationship constructed in the database construction process・ Defect / foreign matter generation analysis process for analyzing the cause of foreign matter generation, and process conditions in the upstream production line so as to remove the defect / foreign matter generation cause or defect / foreign matter generation cause analyzed by this defect / foreign matter generation analysis step A method for manufacturing a semiconductor substrate, comprising: a process condition control step for controlling.

  Further, the present invention condenses and irradiates a ring-shaped diffused illumination light formed from a large number of virtual point light sources onto a pattern on the object to be inspected through the pupil of the objective lens. A subject to be inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light that is reflected from a pattern on the subject to be inspected by an annular diffused illumination light and enters the pupil of the objective lens The pattern image on the object is received by the image sensor and converted into a pattern image signal. The pattern image signal is compared with the image signal of the reference pattern, and the pattern on the object to be inspected is erased by matching. A defect inspection method and apparatus for a pattern on an object to be inspected, wherein a defect is detected by mismatch. Further, the present invention condenses and irradiates a ring-shaped diffused illumination light formed from a large number of virtual point light sources onto a pattern on the object to be inspected through the pupil of the objective lens. A quantity of light that partially changes the intensity or quantity of the first-order or second-order diffracted light including the 0th-order diffracted light that is reflected from the pattern on the object to be inspected by the ring-shaped diffuse illumination light and enters the pupil of the objective lens A pattern image on the object to be inspected obtained by condensing first-order or second-order diffracted light including 0th-order diffracted light obtained through the control filter is received by an image sensor and converted into a pattern image signal; A pattern detecting method and apparatus for detecting an object to be inspected, wherein the pattern on the object to be inspected is detected based on the image signal of the converted pattern.

  Further, the present invention condenses and irradiates a ring-shaped diffused illumination light formed from a large number of virtual point light sources onto a pattern on the object to be inspected through the pupil of the objective lens. The incident light enters the pupil of the objective lens through a light amount control filter that controls the light amount of a part of the zero-order diffracted light reflected from the pattern on the object to be inspected by the annular diffused illumination light and incident on the pupil of the objective lens. The first or second order diffracted light including the zeroth order diffracted light is partially obtained by condensing the first or second order diffracted light including the zeroth order diffracted light obtained through a light amount control filter that changes the intensity or the light amount. The pattern image on the inspection object is received by the image sensor and converted into a pattern image signal. The image signal of the converted pattern is compared with the image signal of the reference pattern and inspected by matching. A defect inspection method and apparatus of the pattern on the inspected object, characterized in that to erase a pattern on elephant product to detect the defect by mismatch.

  Further, the present invention condenses and irradiates a ring-shaped diffuse illumination light formed from a large number of virtual point light sources onto a pattern on an object to be inspected through the pupil of the objective lens, The zonal diffused illumination light in the light source is controlled based on the image signal of the pupil obtained by receiving and converting the image of the image by the second image sensor. An image of the pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including 0th-order diffracted light that is reflected from the pattern on the object to be inspected and enters the pupil of the objective lens. A pattern on the object to be inspected, which is received by the first image sensor and converted into a pattern image signal, and a pattern on the object to be inspected is detected based on the converted image signal of the pattern Detection method It is a device of the benefactor. Further, the present invention condenses and irradiates a ring-shaped diffused illumination light formed from a large number of virtual point light sources through a pupil of the objective lens onto a pattern on the object to be inspected. The annular illumination light in the light source is controlled based on the image signal of the pupil obtained by receiving and converting the image of the locality distribution of the diffracted light including the zeroth order incident on the second image sensor. The first-order or second-order diffracted light including the 0th-order diffracted light that is reflected from the pattern on the object to be inspected and incident on the pupil of the objective lens is collected by the condensed and irradiated annular illumination light. The pattern image on the object to be inspected is received by the first image sensor and converted into a pattern image signal, and the pattern on the object to be inspected is converted based on the image signal of the converted pattern. Features that detect The detection method and apparatus of the pattern on the inspected object to.

  Further, the present invention condenses and irradiates a ring-shaped diffuse illumination light formed from a large number of virtual point light sources onto a pattern on an object to be inspected through the pupil of the objective lens, The zonal diffused illumination light in the light source is controlled based on the image signal of the pupil obtained by receiving and converting the image of the image by the second image sensor. An image of the pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including 0th-order diffracted light that is reflected from the pattern on the object to be inspected and enters the pupil of the objective lens. The light is received by the first image sensor and converted into a pattern image signal. The pattern image signal obtained from the first image sensor is compared with the image signal of the reference pattern, and the pattern on the object to be inspected is matched. The A defect inspection method and apparatus of the pattern on the inspected object, wherein detecting the defect by mismatch and removed by. Further, the present invention condenses and irradiates a ring-shaped diffused illumination light formed from a large number of virtual point light sources through a pupil of the objective lens onto a pattern on the object to be inspected. The annular illumination light in the light source is controlled based on the image signal of the pupil obtained by receiving and converting the image of the locality distribution of the diffracted light including the zeroth order incident on the second image sensor. The first-order or second-order diffracted light including the 0th-order diffracted light that is reflected from the pattern on the object to be inspected and incident on the pupil of the objective lens is collected by the condensed and irradiated annular illumination light. The pattern image on the object to be inspected is received by the first image sensor and converted into a pattern image signal, and the pattern image signal obtained from the first image sensor is used as a reference pattern image. Compare with signal A defect inspection method and apparatus of the pattern on the inspected object, wherein detecting the defect by mismatch erase the pattern on the inspected object by matching Te.

  Further, the present invention condenses and irradiates a ring-shaped diffuse illumination light formed from a large number of virtual point light sources onto a pattern on an object to be inspected through the pupil of the objective lens, The intensity or amount of light is controlled by the light amount control filter in the pattern detection optical system based on the image signal of the pupil obtained by receiving and converting the image of the image by the second image sensor. Light amount control for partially changing the intensity or light amount of first-order or second-order diffracted light including zero-order diffracted light that is reflected from the pattern on the object to be inspected by the belt-like diffuse illumination light and enters the pupil of the objective lens A pattern image on the object to be inspected obtained by condensing the first-order or second-order diffracted light including the 0th-order diffracted light obtained through the filter for use is received by the image sensor into a pattern image signal. And conversion is the detection method and apparatus of the pattern on the inspected object, characterized by detecting a pattern on the inspected object based on the image signal of the transformed pattern. Further, the present invention condenses and irradiates a ring-shaped diffuse illumination light formed from a large number of virtual point light sources onto a pattern on an object to be inspected through the pupil of the objective lens, The intensity or amount of light is controlled by the light amount control filter in the pattern detection optical system based on the image signal of the pupil obtained by receiving and converting the image of the image by the second image sensor. 0 incident on the pupil of the objective lens through a light amount control filter that controls the amount of light of a part of the 0th-order diffracted light reflected from the pattern on the object to be inspected by the band-like diffuse illumination light and incident on the pupil of the objective lens. Condensing first-order or second-order diffracted light including zero-order diffracted light obtained through a light amount control filter that partially changes the intensity or light amount of first-order or second-order diffracted light including second-order diffracted light The pattern image on the object to be inspected is received by the first image sensor and converted into a pattern image signal, and the pattern image signal obtained from the first image sensor is compared with the image signal of the reference pattern. A pattern defect inspection method and apparatus therefor, wherein a pattern on an inspection object is erased by coincidence and a defect is detected by mismatch.

  The present invention also provides polarized / annular diffuse illumination light formed by adding polarized light to an annular diffuse illumination light formed from a number of virtual point light sources on the object to be inspected through the pupil of the objective lens. 1 which includes 0th-order diffracted light that is focused on and irradiated to the pattern, reflected from the pattern on the object to be inspected by this condensed and irradiated annular illumination light, and incident on the pupil of the objective lens The image of the pattern on the object to be inspected obtained by condensing the second or second order diffracted light is received by the image sensor and converted into an image signal of the pattern, and the object to be inspected based on the image signal of this pattern A pattern detection method and apparatus for an object to be inspected, wherein the pattern is detected.

  The present invention also provides polarized / annular diffuse illumination light formed by adding polarized light to an annular diffuse illumination light formed from a number of virtual point light sources on the object to be inspected through the pupil of the objective lens. 1 which includes 0th-order diffracted light that is focused on and irradiated to the pattern, reflected from the pattern on the object to be inspected by this condensed and irradiated annular illumination light, and incident on the pupil of the objective lens The image of the pattern on the object to be inspected obtained by condensing the second or second order diffracted light is received by the image sensor and converted into an image signal of the pattern, and the image signal of this pattern is converted into the image signal of the reference pattern And a device for inspecting a defect on a pattern to be inspected, and a device for detecting the defect by non-matching.

  The present invention is also characterized in that in the pattern defect inspection method and apparatus therefor, the polarized light is circular or elliptically polarized light.

  Further, the present invention condenses and irradiates a ring-shaped diffused illumination light formed from a large number of virtual point light sources onto a pattern on the object to be inspected through the pupil of the objective lens. A subject to be inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light that is reflected from a pattern on the subject to be inspected by an annular diffused illumination light and enters the pupil of the objective lens The pattern image on the object is received by the image sensor and converted into a pattern image signal. The pattern image signal is compared with the image signal of the reference pattern, and the pattern on the object to be inspected is erased by matching. The defect information is detected by the mismatch, the dark field illumination is performed on the pattern on the inspection object, and the scattered light that is reflected from the foreign matter on the pattern on the semiconductor substrate and enters the pupil of the objective lens is collected. Obtained by light A defect inspection method for a pattern on an object to be inspected is characterized by detecting foreign matter information on the pattern by receiving the foreign matter on the pattern on the object to be inspected by an image sensor and converting it into a signal indicating the foreign matter And its device. The present invention also provides an illuminating means for applying substantially uniform ring-shaped illumination light in the detection field of the object to be inspected via the objective lens, and a reflected image from the object to be inspected as a detection image by a photoelectric conversion means. A pattern inspection apparatus and a method therefor, comprising: an image detection unit that performs image comparison, and an image comparison unit that compares the detected image with a reference image. According to the present invention, in the pattern inspection apparatus and method, the annular illumination light includes a large number of virtual point light sources.

  According to the present invention, in the pattern inspection apparatus and the method thereof, using the pupil image information detection means for detecting the image on the pupil of the objective lens and detecting the information of the pupil image, and the detected pupil image information It is characterized by comprising illumination control means for controlling the angle of incidence such as the annular illumination, so-called σ value. According to the present invention, in the pattern inspection apparatus and method, the pupil image information detection means detects either or both of the distribution and / or intensity of the 0th-order diffracted light and the ± 1st-order diffracted light entering the pupil of the objective lens. It is characterized by comprising. In the pattern inspection apparatus and method therefor, the annular illumination light may be partially attenuated by the intensity of the 0th order light and ± 1st order diffracted light detected by the pupil image information detection means, or The annular illumination light that does not generate ± first-order diffracted light is partially shielded. In the pattern inspection apparatus and method therefor, the present invention further includes a light intensity control unit (light quantity control filter) that partially controls the light intensity at either the pupil of the objective lens or a position conjugate with the objective lens. It is characterized by that. Further, the present invention is characterized in that, in the pattern inspection apparatus and method, the light intensity control unit is configured such that the intensities of the 0th-order diffracted light and the ± 1st-order diffracted light are the same or substantially equal. According to the present invention, in the pattern inspection apparatus and method, the annular illumination light is controlled so as to change the incident angle of the 0th-order diffracted light or to change the reflection angle of the 1st-order diffracted light. It is characterized by.

  Further, the present invention is a pattern inspection apparatus and method, comprising: a pattern density calculation unit that calculates a circuit pattern density from the detected image; and an illumination control unit that controls the illumination unit according to the calculated density. It is characterized by doing. According to the present invention, in the pattern inspection apparatus and the method thereof, there are provided means for analyzing the detected pupil image, and sensitivity control means for controlling the pattern defect inspection sensitivity of the image comparison means based on the analysis result. It is characterized by that. In the pattern inspection apparatus and method therefor, the present invention is characterized in that it is determined from the pupil image information whether the pattern has periodicity. In the pattern inspection apparatus and method, the present invention is characterized in that the presence or absence of foreign matter on the pattern is determined from the pupil image information. In the pattern inspection apparatus, the light intensity of the annular illumination light may be controlled to be a desired value. In the pattern inspection apparatus and method therefor, the ring-shaped illumination light irradiating means includes a plurality of ring-shaped light shielding plates arranged at a secondary light source position where the discharge lamp light is condensed, The light-shielding plate is switched and controlled. In the pattern inspection apparatus and method therefor, the ring-shaped illumination light irradiating means arranges a ring-shaped light shielding plate at a secondary light source position where a discharge lamp is condensed by an elliptical mirror, It is configured to be controlled by moving in the optical axis direction. According to the present invention, in the pattern inspection apparatus and method, the image detection means is provided with a variable optical magnification mechanism. Further, the present invention is characterized in that, in the pattern inspection apparatus and the method thereof, the pixel size of the detected image is selected so as to preserve the maximum and minimum values of the pattern by the optical magnification variable mechanism.

  According to the present invention, in the pattern inspection apparatus and method, the polarization state of illumination light is controlled. In the pattern inspection apparatus and method, the present invention is characterized in that the object to be inspected is irradiated with circularly polarized illumination or elliptically polarized illumination close to a circle. According to the present invention, in the pattern inspection apparatus and method, linearly polarized light is converted into circularly polarized light or elliptically polarized light to irradiate the inspection object, and reflected light from the inspection object is converted into linearly polarized light and detected. It is characterized by having constituted so. In the pattern inspection apparatus and method therefor, the present invention obtains linearly polarized light by a polarizing beam splitter, obtains circular or elliptical polarized light by a λ / 4 plate, irradiates the object to be inspected, and applies reflected light to the λ / 4 plate is linearly polarized light, and this linearly polarized light is detected through a polarizing beam splitter. In the pattern inspection apparatus and method, the present invention obtains linearly polarized light by a splitter coated with a dielectric multilayer film, obtains circular or elliptical polarized light by a λ / 4 plate, irradiates the object to be inspected, and reflects it. Light is linearly polarized through the λ / 4 plate and detected through a splitter. According to the present invention, in the pattern inspection apparatus and method, an optical element for controlling polarization is inserted between the objective lens of the infinity correction system and the imaging lens (second objective lens or tube lens). It is characterized by. In the pattern inspection apparatus and method therefor, the present invention is characterized in that a diffusing plate is inserted in the illumination light path to control the directivity of the illumination light according to the pattern. In the pattern inspection apparatus and method, the present invention is characterized in that at least the accumulation time of the one-dimensional image sensor is variable according to the detected light amount. Further, the present invention is characterized in that the pattern inspection apparatus and method have means for storing and analyzing an image of a predetermined position on the inspection object (sample) during the inspection.

  Further, the present invention detects a pupil image through an annular illumination means for illuminating substantially uniformly in the detection visual field of the object to be inspected through the objective lens and a lens that forms an image at the pupil position of the objective lens. Means for partially controlling the light intensity at a position conjugate to the pupil of the objective lens, and using the detected pupil image information, controlling the annular illumination or partially controlling the light intensity. The microscope system is configured to perform either one of the above.

  The functions of the above technical means are as follows. According to the configuration of the present invention, a semiconductor substrate, a TFT substrate, a thin film multilayer substrate, a printed substrate or the like has a fine pattern (for example, a pattern with a pitch of 1 μm or less (0.8 to 0.4 μm)). Defects can be detected with high resolution and high sensitivity, and a semiconductor substrate can be manufactured with high yield by reducing the occurrence of fine defects. According to the configuration of the present invention, the illumination light focused on the semiconductor substrate (object to be inspected) is incident obliquely by, for example, ring-shaped illumination, and the reflected light is used even when the pattern is fine. The 0th-order diffracted light and the 1st-order diffracted light (either + 1st-order diffracted light or -1st-order diffracted light) can be sufficiently taken into the aperture (pupil) of the objective lens. As a result, the image signal is received by the image sensor. Sufficient resolution is obtained, and the pupil plane image of the objective lens is monitored to recognize the 0th-order diffracted light and the 1st-order diffracted light (either the + 1st-order diffracted light or the −1st-order diffracted light) as reflected light, As a result, for example, by controlling the annular illumination, the image signal of the pattern can be detected with a sufficient resolution and a deep depth of focus, always under the optimum conditions corresponding to a fine pattern. It can be. Then, the locality distribution or intensity distribution of the 0th-order diffracted light and the 1st-order diffracted light (either the + 1st order diffracted light or the −1st order diffracted light) as reflected light is detected from the image of the Fourier transform plane (pupil plane), If the annular illumination is controlled in accordance with the locality distribution or intensity distribution (corresponding to the pattern density) of the detected diffracted light, the pattern density is not so high in a 4Mb DRAM memory device, for example. The ring-shaped illumination can be sufficiently inspected with a normal resolution. For example, in a 16 Mb DRAM memory device, the inspection can be performed with the annular-shaped illumination with higher resolution under a preset condition. Furthermore, for example, since the pattern density is high in the cell portion of the memory element, inspection may be performed with high resolution by ring-shaped illumination under preset conditions, and inspection sensitivity may be reduced in rough regions other than the cell portion. Can be inspected at high speed.

  Further, according to the configuration of the present invention, an annular illumination is applied to a fine pattern on the object to be inspected to detect a high resolution (high resolution) image signal from the fine pattern, and this high resolution Since the image signal is compared with the reference high-resolution image signal by chip comparison or cell comparison and the fine pattern is erased by matching, the fine pattern defect is detected, so the fine pattern defect is detected. It can be inspected with high reliability. Further, according to the configuration of the present invention, the ring-shaped illumination and the polarized illumination (especially circular or elliptically polarized illumination is excellent) are used in combination on a fine pattern on the object to be inspected. An image or image signal from a simple pattern can be detected with a high resolution (high resolution) and with a high contrast (brightness). Further, according to the configuration of the present invention, a ring-shaped illumination and a polarized illumination (especially a circular or elliptically polarized illumination is excellent) are used in combination with a pattern having fine various directions on the object to be inspected. As a result, it is possible to detect an image or image signal from a pattern having a fine variety of directions with high resolution (high resolution) and with high contrast (brightness). Further, according to the configuration of the present invention, a ring-shaped illumination and a polarized illumination (especially a circular or elliptically polarized illumination is excellent) are used in combination with a pattern having fine various directions on the object to be inspected. To detect an image signal having a high resolution (high resolution) and a high density difference (brightness) from this fine pattern having various directions, and a high density difference (brightness). ) By erasing a pattern having various directions by matching the image signal having a high resolution and a high contrast (brightness) with a reference high resolution. Since the defect on the pattern having directionality is detected, the defect on the pattern having various fine directions can be inspected with high reliability.

  Further, according to the configuration of the present invention, an image based on diffracted light generated from a fine pattern on the object to be inspected and incident in the pupil of the objective lens is detected, and annular illumination (for example, (Outer σ value, inner σ value) is controlled, and this controlled ring-shaped illumination is applied to the minute pattern on the object to be inspected, thereby achieving high resolution (high resolution) corresponding to the minute pattern. An image or image signal can be detected. Further, according to the configuration of the present invention, an image based on diffracted light generated from a fine pattern on the object to be inspected and incident in the pupil of the objective lens is detected, and annular illumination (for example, (Outer σ value, inner σ value) is controlled, and this controlled ring-shaped illumination is applied to the minute pattern on the object to be inspected, thereby achieving high resolution (high resolution) corresponding to the minute pattern. By detecting the image signal and comparing the high-resolution image signal with the reference high-resolution image signal and erasing the fine pattern by coincidence, defects on the fine pattern are detected. It is possible to inspect defects on simple patterns with high reliability.

  According to the configuration of the present invention, the 0th-order diffracted light and the first-order diffracted light (either the + 1st-order diffracted light or the −1st-order diffracted light) that are generated from a fine pattern on the object to be inspected and enter the pupil of the objective lens. Is recognized (observed or detected) by the pupil monitor of the objective lens, and ring-shaped illumination (for example, outer σ value, inner) is recognized based on the recognized locality distribution of the diffracted light (observed or detected). (sigma value) is controlled, and this controlled ring-shaped illumination is applied to a fine pattern on the object to be inspected, so that a high resolution (high resolution) image or image signal corresponding to the fine pattern is obtained. Can be detected. According to the configuration of the present invention, the 0th-order diffracted light and the first-order diffracted light (either the + 1st-order diffracted light or the −1st-order diffracted light) that are generated from a fine pattern on the object to be inspected and enter the pupil of the objective lens. Is recognized (observed or detected) by the pupil monitor of the objective lens, and ring-shaped illumination (for example, outer σ value, inner) is recognized based on the recognized locality distribution of the diffracted light (observed or detected). (sigma value) is controlled, and this controlled annular illumination is applied to the minute pattern on the object to be inspected, thereby detecting a high resolution (high resolution) image signal corresponding to the minute pattern. The defect on the fine pattern is detected by comparing the high resolution image signal with the reference high resolution image signal and erasing the fine pattern by coincidence. The high Can be inspected with confidence.

  Further, according to the present invention, an image showing the density of a fine pattern based on diffracted light that is generated from a fine pattern on the object to be inspected and enters the pupil of the objective lens is detected. High resolution corresponding to the fine pattern by controlling the strip illumination (for example, outer σ value, inner σ value) and applying this controlled annular illumination to the minute pattern on the object to be inspected By detecting a (high resolution) image signal, comparing this high resolution image signal with a reference high resolution image signal and erasing the fine pattern by matching, a defect on the fine pattern is detected. Therefore, it is possible to inspect defects on a fine pattern with high reliability.

  According to the present invention, a fine defect generated on a fine pattern in a semiconductor substrate having a fine pattern such as a semiconductor wafer, a TFT substrate, a thin film multilayer substrate, or a printed board can be inspected with high reliability. By feeding back the result to the semiconductor substrate manufacturing process, a semiconductor substrate having a fine pattern can be manufactured at a high yield. Further, according to the present invention, annular illumination is applied to a fine pattern on the object to be inspected to detect a high resolution (high resolution) image signal from the fine pattern, and the high resolution image signal is The defect on the fine pattern is detected by erasing the fine pattern by coincidence compared with the standard high resolution image signal, so the defect on the fine pattern can be inspected with high reliability There is an effect that can. According to the present invention, the annular illumination is applied to a fine pattern on the object to be inspected, and the 0th-order diffracted light and the 1st-order diffracted light (+1) that are generated from the fine pattern and enter the pupil of the objective lens. The first-order diffracted light and the -1st-order diffracted light) are attenuated by a filter that controls the amount of light provided at a position conjugate with the pupil of the objective lens. By receiving the reflected light, a high resolution (high resolution) image signal is detected from the fine pattern, and the high resolution image signal is compared with the reference high resolution image signal to form a fine pattern by matching. Since the defect on the fine pattern is detected by erasing, the defect on the fine pattern can be inspected with high reliability.

  Further, according to the present invention, a fine pattern is formed by applying a ring-shaped illumination and a polarized illumination (especially circular or elliptically polarized illumination is excellent) to a fine pattern on the object to be inspected. It is possible to detect an image or an image signal from the image with high resolution (high resolution) and with high contrast (brightness). Further, according to the present invention, a ring-shaped illumination and polarized illumination (especially circular or elliptically polarized illumination is excellent) are used in combination on a pattern having a fine variety of directions on the object to be inspected. As a result, it is possible to detect an image or an image signal from a pattern having a fine variety of directions with high resolution (high resolution) and high contrast (brightness). Further, according to the present invention, a ring-shaped illumination and a polarized illumination (especially circular or elliptically polarized illumination is excellent) are used in combination on a pattern having various fine directions on the object to be inspected. The image signal having high resolution (high resolution) and high density difference (brightness) is detected from the pattern having the fine various directions, and this high resolution and high density difference (brightness) is detected. Compared with the image signal having the high resolution of the reference and the image signal having a high contrast (brightness), the pattern having the finer directionality is erased by matching the image signal. Since the defect on the pattern having the pattern is detected, it is possible to inspect the defect on the pattern having various fine directions with high reliability.

  Further, according to the present invention, an image based on diffracted light that is generated from a fine pattern on the object to be inspected and is incident on the pupil of the objective lens is detected, and annular illumination (for example, outside σ) is detected based on the image. Value, inner σ value), and by applying this controlled annular illumination to the fine pattern on the object to be inspected, a high resolution (high resolution) image corresponding to the fine pattern or There is an effect that an image signal can be detected. Further, according to the present invention, an image based on diffracted light that is generated from a fine pattern on the object to be inspected and is incident on the pupil of the objective lens is detected, and annular illumination (for example, outside σ) is detected based on the image. High-resolution (high resolution) image signal corresponding to the fine pattern by applying the controlled annular illumination to the fine pattern on the object to be inspected. The high-resolution image signal is compared with the reference high-resolution image signal and the fine pattern is erased by coincidence to detect defects on the fine pattern. There is an effect that the upper defect can be inspected with high reliability.

  According to the present invention, the 0th-order diffracted light and the 1st-order diffracted light (either + 1st-order diffracted light or -1st-order diffracted light) that are generated from a fine pattern on the object to be inspected and enter the pupil of the objective lens The locality distribution is recognized (observed or detected) by the pupil monitor of the objective lens, and ring-shaped illumination (for example, outer σ value, inner σ value) based on the recognized locality distribution of the diffracted light (observed or detected) ) And applying the controlled annular illumination to the fine pattern on the object to be inspected, a high resolution (high resolution) image or image signal corresponding to the fine pattern is detected. There is an effect that can be. According to the present invention, the 0th-order diffracted light and the 1st-order diffracted light (either + 1st-order diffracted light or -1st-order diffracted light) that are generated from a fine pattern on the object to be inspected and enter the pupil of the objective lens The locality distribution is recognized (observed or detected) by the pupil monitor of the objective lens, and ring-shaped illumination (for example, outer σ value, inner σ value) based on the recognized locality distribution of the diffracted light (observed or detected) ) And applying this controlled ring-shaped illumination to a fine pattern on the object to be inspected, a high resolution (high resolution) image signal corresponding to the fine pattern is detected. By comparing the high-resolution image signal with the reference high-resolution image signal and erasing the fine pattern by matching, the defect on the fine pattern is detected. Degree of reliability There is an effect that can be inspected.

  According to the present invention, an image showing the density of a fine pattern based on diffracted light that is generated from a fine pattern on the object to be inspected and enters the pupil of the objective lens is detected. High resolution corresponding to a fine pattern by controlling the strip illumination (for example, outer σ value, inner σ value) and applying this controlled annular illumination to the minute pattern on the object to be inspected By detecting a (high resolution) image signal, comparing this high resolution image signal with a reference high resolution image signal and erasing the fine pattern by matching, a defect on the fine pattern is detected. Therefore, there is an effect that a defect on a fine pattern can be inspected with high reliability.

  An embodiment of detection of a pattern on an object to be inspected and inspection of a defect on the pattern according to the present invention will be described with reference to FIGS. An embodiment in which the defect inspection of the inspection object according to the present invention is applied to a semiconductor manufacturing process will be described with reference to FIG. First, in this embodiment, a description will be given of a case where annular illumination (annular diffuse illumination) is used as illumination means for performing illumination almost uniformly in the detection visual field of the object to be inspected via the objective lens.

[First embodiment]
FIG. 1 is a block diagram showing an embodiment of a pattern inspection apparatus using annular illumination according to the present invention. A pattern inspection apparatus according to the present invention includes an inspection target object (inspected pattern) 1 such as an LSI wafer for pattern inspection, an XYZθ stage 2 on which the inspection target object 1 such as an LSI wafer is placed, and a light source Disk-shaped mask 5 for forming an annular illumination (annular diffuse illumination) that forms an annular secondary light source formed from the Xe lamp 3, the condensing elliptical mirror 4, and a number of virtual point light sources (Secondary light source for ring-shaped illumination) secondary light source for ring-shaped illumination, illumination optical system composed of collimator lens 6, light quantity adjusting filter 14, and condenser lens 7, half mirrors 8a and 8b , Objective lens 9, convergent lens 11, zoom lens 13 provided with attenuation filter 38 on pupil plane 10b conjugate to pupil plane 10a of objective lens 9, and two-dimensional or one-dimensional image sensors 12a, 12b A digital image signal obtained from the pattern detection optical system composed of the A / D converters 15a and 15b and the A / D converter 15a for converting the image signals detected from the image sensors 12a and 12b into digital image signals. A delay memory 16 for storing and delaying, a comparison circuit 17 for comparing a delayed digital image signal stored in the delay memory 16 with a digital image signal obtained from the A / D converter 15a, and an A / D converter 15a The pupil of the objective lens 9 obtained via the A / D converter 15b from the edge detector 21 that detects the edge of the pattern from the digital image signal obtained from the image sensor 12b that detects the image of the pupil plane 10a of the objective lens 9 Based on the digital image signal of the surface 10a, a disk-shaped mask for forming an annular illumination as a secondary light source by the moving mechanism 19 From the CPU 21 that controls the disc 5 and the attenuation filter 38 by the moving mechanism 39, the comparison in the comparison circuit 17 based on the edge signal detected from the edge detector 21, and the control of the XYZθ stage 2 by the driver 45. It comprises an image processing and control system for configured defect detection. Reference numeral 18 denotes a defect determination output obtained from the comparison circuit 17. The defect determination output 18 is also input to the CPU 20 to add a defect occurrence position (coordinates) on the inspection object 1 and at least a plurality of units sampled from a predetermined manufacturing process for each unit of the inspection object 1. Each unit of the object to be inspected 1 is stored in a storage means (not shown). The CPU 20 outputs defect information 40 for each process unit of the inspection object 1 sampled from a predetermined manufacturing process, at least for each unit of the inspection object 1 stored in the storage means. The defect information 40 includes a defect occurrence position (coordinates) on the inspection object 1 obtained based on the defect determination output 18 and is classified based on the defect determination output 18 by the CPU 20 as shown in FIG. Types of defects (projection defect 231, chip defect 236, open defect 232, short defect 234, discoloration defect 233, contaminant 235, etc.) are included. It is not always necessary to clearly classify these defect types.

  In FIG. 1, a light house 24 is a ring-shaped secondary light source formed from an Xe lamp 3 as a primary light source, an elliptical mirror 4 for collecting light emitted from the Xe lamp 3, and a large number of virtual point light sources. And a disk-shaped mask 5 for forming the illumination. A disk-shaped mask (secondary light source for ring-shaped illumination) 5 is rotated stepwise in response to a command from the CPU 20, and different types of ring-shaped illumination (for example, shown in FIG. 3). A moving mechanism 19 for switching is provided. Since the circular mask 5 is an annular secondary light source formed from a large number of virtual point light sources, the annular secondary light source 5 provides annular diffuse illumination.

  Accordingly, the annular illumination emitted from the disc-shaped mask (secondary light source for annular illumination) 5 is applied to the pupil 10a of the objective lens 9 by the collimator lens 6 and the collimator lens 7 as shown in FIGS. As shown in FIG. 10, the incident illumination light 24 is condensed, and the collected incident illumination light is condensed by the objective lens 9 and placed on the XYZ stage 2 (the θ stage is not shown). Irradiation is performed on the inspection object 1 such as an LSI wafer. The light amount adjusting filter 14 adjusts the light amount irradiated onto the inspection object 1. Reference numeral 14 a denotes a drive mechanism that drives the light quantity adjustment filter 14 and is controlled by a command from the CPU 20. A zero-order reflected diffracted light (regularly reflected light) and primary and secondary reflected diffracted light are generated on the + and − sides from the pattern on the inspection object 1 such as the LSI wafer. Of the 0th-order reflected diffracted light (regularly reflected light) generated from the pattern on the inspection object 1 such as an LSI wafer and the first-order and second-order reflected diffracted light on the + and − sides, the objective lens 9 is reflected by the half mirror 8a and the half mirror 8b and is incident on the pupil 10b of the zoom lens 13. The reflected diffracted light is condensed on the image sensor 12a by the zoom lens 13. And imaged. The image sensor 12a receives a reflected diffracted light image that is generated from a pattern on the inspection object 1 such as an LSI wafer and is incident on the pupil 10a of the objective lens 9, and a reflected diffracted light image of the pattern on the inspection object 1 The image signal is output. The pupil 10a of the objective lens 9 and the pupil 10b of the zoom lens 13 are in a conjugate relationship. If necessary, by providing an attenuation filter 38 on the pupil 10 of the zoom lens 13, the 0th-order diffracted light incident on the pupil 10a of the objective lens 9 can be attenuated.

  On the other hand, the reflected diffracted light image incident on the pupil 10 a of the objective lens 9 is formed on the image sensor 12 b by the converging lens 11. Accordingly, the image sensor 12b receives the reflected diffracted light image incident on the pupil 10a of the objective lens 9 and outputs an image signal of the reflected diffracted light image incident on the pupil 10a of the objective lens 9, and from the image signal, the objective lens The state of the reflected diffracted light incident on the 9 pupil 10a can be detected. That is, as shown in FIGS. 13 and 14, if the periodicity of the pattern on the inspection object 1 such as an LSI wafer is different, the generation state of the first-order diffracted light with respect to the incident illumination light 24 changes, and the objective lens 9 The first-order diffracted light incident on the pupil 10a also varies. FIG. 13 shows a case where the density (periodicity) of the pattern on the inspection object 1 is high, and FIG. 14 shows a case where the density (periodicity) of the pattern on the inspection object 1 is low. Further, also from the relationship of equation (2) described later, when the pitch P (density (periodicity)) of the pattern on the inspection object 1 or the wavelength λ of the incident illumination light 24 changes, the incident angle Ψ of the incident illumination light 24 is changed. The diffraction angle θ of the first-order diffracted light also changes, and the first-order diffracted light incident on the pupil 10a of the objective lens 9 also changes.

  By the way, in the case of an LSI wafer, for example, in the inspection object 1, if the type of the LSI wafer changes, the pattern pitch P (density (periodicity)) also changes. In the LSI wafer, when the product type changes, for example, 256 MDRAM or 64 MDRAM, the pattern pitch P (density (periodicity)) also changes. Even if the type is the same, the pattern density (periodicity) may change when the process changes. For example, the pattern pitch P changes in the inspection object in the wiring process, the inspection object in the diffusion process, and the like. Further, the pattern pitch P changes in the memory section and the peripheral circuit section in one chip on the LSI wafer. Further, it is necessary to change the wavelength λ of the incident illumination light 24 in accordance with the cross-sectional structure of the inspection object 1. For example, since a film thickness variation occurs in the thin film forming the inspection object 1, the light reflected from the inspection object varies due to light interference in the thin film. In order to avoid the fluctuation of the reflected light, it is necessary to change the wavelength λ of the incident illumination light 24 in order to select the wavelength λ of the incident illumination light 24 that is unlikely to cause light interference in the thin film. For example, as shown in a later embodiment, the wavelength λ of the incident illumination light 24 can be changed by a wavelength selection filter using a light source that emits light of a plurality of types of wavelengths in the illumination optical system. Thus, when the pattern pitch P (density (periodicity)) on the inspection object 1 or the wavelength λ of the incident illumination light 24 changes, the diffraction angle θ of the first-order diffracted light with respect to the incident angle Ψ of the incident illumination light 24 also changes. The first-order diffracted light that changes and enters the pupil 10a of the objective lens 9 also fluctuates. In view of this, the diffracted light generated from the object to be inspected according to the type and cross-sectional structure of the object 1 to be inspected, and in particular, the first-order diffracted light is optimally incident on the pupil 10a of the objective lens 9. It is necessary to control σ of the strip-shaped secondary light source 5 for illumination, that is, the incident angle Ψ of illumination light applied to the object 1 to be inspected. Accordingly, the CPU 20 performs a Fourier transform image analysis on the digital Fourier transform image signal on the pupil (Fourier transform plane) 10a of the objective lens 9 obtained from the image sensor 12b via the A / D converter 15b, and this Fourier transform image analysis is performed. Edge density determination (periodic density determination of the pattern on the inspection object 1) is performed based on the converted image analysis result, and the image signal from the pattern on the inspection object 1 having a high density is faithfully received from the image sensor 12a. In order to obtain it, the digital image signal of the reflected diffracted light image incident on the pupil 10a of the objective lens 9 is optimized, that is, the 0th order diffracted light and the 1st order diffracted light from the pattern on the inspection object 1 are the objective. The moving mechanism 19 is driven and controlled so as to sufficiently enter the pupil 10a of the lens 9, and an optimal annular-shaped illumination secondary light source 5 (for example, shown in FIGS. 3 and 4). , The inner σ also includes the usual sources of zero.) Selects.

  The illumination is so-called Koehler illumination with no illumination unevenness. Although not shown, focusing is performed so that an image of reflected diffracted light from a pattern on the inspection object 1 such as an LSI wafer is clearly formed on the image sensor 12a. That is, the pattern (surface) on the inspection object 1 such as an LSI wafer is automatically focused on the above-described detection optical system. Then, while moving the X stage on which the inspection object 1 such as an LSI wafer is moved, the image sensor 12a scans and images the two-dimensional pattern on the inspection object 1 from the image sensor 12a. Image signal can be obtained. In this case, the movement of the X stage may be continuous feed, step feed, or repeat feed. The two-dimensional image signal of the pattern on the inspection object 1 obtained from the image sensor 12a in this way is A / D converted by the A / D converter 15a, and the two-dimensional digital image signal is stored in the delay memory 16. A chip or cell that is stored and repeated is delayed, and the delayed two-dimensional digital image signal is compared with the two-dimensional digital image signal output from the A / D converter 15a by the comparison circuit 17. Alternatively, the cells are compared and detected as a defect 18 due to a mismatch between the two digital image signals.

  The comparison circuit 17 is a known technique and will not be described in detail, but will be described briefly. The comparison circuit 17 is a two-dimensional grayscale image signal (digital image) obtained from each of the delay memory 16 and the A / D converter 15a for each of the patterns on the inspection object 1 that are originally formed to be the same. Signal) is compared, and the polarities of the gray image signals obtained by the differential processing are compared, and the two gray image signals to be compared are compared so that the number of pixels with mismatched polarities is less than or equal to the set value. The image is aligned, a difference image signal between the two gray image signals that have been aligned is detected, and the difference image signal is binarized with a desired threshold value to detect a defect. Note that the comparison processing in the comparison circuit 17 is described in detail in Japanese Patent Laid-Open No. Hei 3-209843 by the present inventors. The edge detector 21 detects an edge of a pattern on the inspection object 1 based on a two-dimensional digital image signal detected by the image sensor 12a and obtained via the A / D converter 15a. It is. The CPU 20 takes in the edge information of the pattern on the inspection object 1 detected by the edge detector 21 and feeds it back to the comparison circuit 17, whereby the position of the two grayscale image signals (digital image signals) in the comparison circuit 17. The chip comparison or the cell comparison can be performed by executing the alignment and further controlling the timing of reading from the delay memory 16.

  The comparison circuit 17 is not limited to the circuit described above, and it is needless to say that a comparison circuit having another configuration may be used. It is also possible to store a two-dimensional digital image signal at a specified position on the stage coordinates obtained from the A / D converter 15a in the delay memory 16, and read and analyze it by the CPU 20. In particular, when a defect in the inspection target object 1 is included as an inspection target, the feature of the defect can be analyzed, and therefore an optimal inspection condition can be found. Next, a disk-shaped mask (secondary light source for annular illumination) 5 for forming the annular illumination will be described with reference to FIGS. That is, FIG. 2 is a schematic explanatory view of a disk-like mask (in which a large number of ring-shaped illumination mask elements are arranged) in the embodiment shown in FIG. FIG. 3 is a diagram showing a specific example of the disk-shaped mask (a plurality of kinds of annular-shaped illumination mask elements arranged) shown in FIG. FIG. 4 is a diagram showing another specific example of the disc-shaped mask (a plurality of kinds of annular-shaped illumination mask elements arranged) shown in FIG. FIG. 5 is a diagram for explaining one annular illumination mask element shown in FIGS. 3 and 4. As shown in FIG. 2, the disk-shaped mask 5 is formed with, for example, many kinds of annular-shaped illumination mask elements 5-1, 5-2,. It is switched by the moving mechanism 19 to be driven. 3 shows in detail the various types of annular illumination mask elements 5-1, 5-2,... 5-n shown in FIG. 2 by 5a-1, 5a-2,. It is a figure. That is, 5a-1 in FIG. 3 indicates a ring-shaped mask element in which the space between the inner σ and the outer σ is transparent, and 5a-2 indicates a ring-shaped mask element in which the inner σ and the outer σ are larger than 5a-1. 5a-n denote ring-shaped mask elements in which a part of the ring-shaped transparent portion is shielded.

  FIG. 4 shows various types of annular illumination mask elements 5-1, 5-2,... 5-n shown in FIG. 2 as 5b-1, 5b-2, 5b-3, 5b-4, 5b. It is the figure shown in detail by -5, 5b-6, 5b-7, 5b-8, 5b-9, 5b-10. 5b-1 represents a ring-shaped mask element in which the inner σ is 0.6 and the outer σ is 1.0, and 5b-2 is the inner σ is 0.4 and the outer σ is 1.0. 5b-3 shows a ring-shaped mask element in which the inner σ is 0.2 and the outer σ is 1.0, and 5b-4 shows an inner σ. A ring-shaped mask element having a transparent area between 0.4 and an outer σ of 0.8 is shown. 5b-5 is a ring-shaped mask element having an inner σ of 0.2 and an outer σ of 0.8. 5b-6 is a ring-shaped mask element in which the inner σ is 0.4 and the outer σ is 0.6, and 5b-7 is a ring-shaped mask element. A ring-shaped mask element having a transparency between 0.6 is shown. From 5b-1 to 5b-7, an annular secondary light source for illumination is formed. 5b-8 represents a mask element formed of a circular transparent portion having a σ of 0.89, 5b-9 represents a mask element formed of a circular transparent portion having a σ of 0.77, and 5b-10 represents a mask element formed of a circular transparent portion. Fig. 5 shows a mask element formed with a 0.65 circular transparent part. From these 5b-8 to 5b-10, normal secondary light sources having different σ are formed. Note that σ of 1.0 is equal to the aperture (NA (Numerical Aperture)) (corresponding to the diameter of the pupil) of the objective lens 9. FIG. 5 is a view for showing specific dimensions of the ring-shaped mask element shown in FIG. That is, M is an opaque mask surface that shields light. The mask thickness t is set to 2.3 mm. The dimensions of the outer σ diameter and the inner σ diameter of each ring-shaped mask element from 5b-1 to 5b-7 are shown in the following (Table 1).

  In each of the ring-shaped mask elements, the inner side of the inner σ is made opaque. However, since the amount of light decreases if the inner σ is made opaque, the inner side of the inner σ is made translucent to reduce the amount of light. It is possible to correspond to the pattern on the inspection object 1 having a high density. In addition, it is sufficient that the inner σ and the outer σ are nearly transparent. Further, in the case of the above embodiment, the ring-shaped transparent portion is formed between the inner σ and the outer σ, but it is obvious that a large number of circular transparent portions may be arranged in a ring shape. As described above, a large number of types of secondary light sources can be formed by forming a large number of types of mask elements on the disc-shaped mask 5, so that appropriate incident illumination light can be applied to various objects 1 to be inspected. Can be obtained. As a result, 0th-order diffracted light and 1st-order diffracted light (either + 1st-order diffracted light or -1st-order diffracted light) obtained from various inspection objects 1 can be taken into the aperture (pupil) 10a of the objective lens 9, A two-dimensional image signal having a sufficient resolution for various inspection objects 1 can be obtained from the image sensor 12a. Next, the reason why a two-dimensional image signal can be obtained with sufficient resolution for a high-density pattern on the object to be inspected by using an annular illumination will be described. That is, in the case of normal illumination, the incident angle Ψ of the incident illumination light 30 shown in FIG. When the pitch P of the inspection object 1 is fine (in the case of a high-density pattern), the diffraction angle θ of the 0th-order diffracted light (m = 0) is the same as the incident angle ψ from the relationship of the following equation (Equation 2). However, since the pitch P is small for both the + 1st order diffracted light and the −1st order diffracted light, the diffraction angle θ increases and the light does not enter the pupil 10a of the objective lens 9. As a result, only the 0th-order diffracted light from the high-density pattern on the object to be inspected, that is, only the direct current component light, and an image from the pattern on the object to be inspected based on the diffracted light cannot be obtained.

  The above matters can be explained in more detail by the so-called Abbe's diffraction theory.

  That is, Abbe's diffraction theory is applied to the relationship between the incident angle Ψ relative to the optical axis of the incident illumination light 30 and the spatial frequency to be imaged. Whether the lattice pattern on the object to be inspected can be imaged by whether or not the first-order diffracted light from the lattice pattern on the object to be inspected can pass through the pupil 10a of the imaging system (objective lens 9). Determined. If the diffracted light 31 is focused on one point of the pupil plane 10a of the imaging system (objective lens 9) and this focused point is inside the pupil 10a, the diffracted light passes through the imaging system (objective lens 9). , Interferes with the 0th-order diffracted light on the imaging surface to form an image of a lattice pattern. In this configuration, there is an advantage that the depth of focus becomes deep. When the structure of the lattice pattern on the object to be inspected becomes fine (when the pitch P becomes fine), the angle θ formed by the first-order diffracted light with respect to the optical axis increases, and the angle θ becomes the value of the imaging system (objective lens 9). If it becomes larger than NA (Numerical Aperture: aperture), it cannot pass through the pupil 10a of the imaging system (objective lens 9), and the lattice pattern is not imaged. By the way, with simple oblique illumination, the resolution is improved in the plane connecting the optical axis and the light source, but the resolution is not improved in other planes. For this reason, in order to improve the resolution in an arbitrary direction, the ring-shaped illumination as described above is performed, and the first-order diffracted light is changed into an imaging system (objective lens 9) according to the direction of the pattern of the inspection object. It is necessary to prevent incident illumination light that does not enter the NA of the light from entering. An illumination method in which the 0th-order diffracted light does not enter the pupil 10a of the imaging system (objective lens 9) corresponds to so-called dark field illumination. Of course, if only the first-order diffracted light exists in the aperture of the imaging system (objective lens 9) by the dark field illumination in this way, the resolution is extremely low.

  In FIG. 1, the CPU 20 always uses first-order diffracted light and zero-order diffracted light even if the pattern of the inspection object 1 changes based on information detected by the image sensor 12 b that is a monitor of the pupil 10 a of the objective lens 9. Is controlled by driving the moving mechanism 19 so that the ring-shaped illumination light source composed of the disc-shaped mask 5 (secondary light source for ring-shaped illumination) is switched. Control strip lighting. Specifically, the CPU 20 drives and controls the moving mechanism 19 using the image of the Fourier transform surface (the surface of the pupil 10a of the objective lens 9) detected by the image sensor 12b, so that the disc-shaped mask 5 is 5a-. By switching to 1, 5a-2,... Or 5b-1, 5b-2,..., Etc., the first-order diffracted light 23 enters the pupil 10a of the objective lens 9 according to the pattern of the object 1 to be inspected. The ring-shaped illumination is controlled so as to block the incident illumination light that is not present or reduce the intensity. However, in the case where periodicity is not observed in the pattern of the object 1 to be inspected, that is, in the case of diffracted light (spread diffraction components are continuous) having various diffraction angles, an isolated pattern (periodic) 4), it is considered that the incident illumination light is not excessively obliquely incident, and an appropriate secondary light source for annular illumination (a mask element 5b-4 shown in FIG. 5b-5, 5b-6, 5b-7 (with a small outside σ)) or a normal circular illumination light source (mask elements 5b-8, 5b-9, 5b-10 shown in FIG. 4). select.

  The edge detector 21 differentiates the image signal of the pattern of the inspection object 1 detected by the image sensor 12a, and detects edge information of the pattern of the inspection object 1 by threshold processing. . Accordingly, the CPU 20 calculates the pattern width of the inspection object 1 by calculating, for example, the area surrounded by the pattern edge based on the edge information of the pattern of the inspection object 1 detected by the edge detector 21. Then, the density (pitch P) of the inspection target object 1 is calculated based on the pattern width of the inspection target object 1, and according to the calculated pattern density (pitch P) of the inspection target object 1, The moving mechanism 19 is driven and controlled, and the annular light source for illumination composed of the disc-shaped mask 5 is switched to control the annular illumination. For example, when the pattern density is high, the incident illumination light is incident more obliquely. Specifically, the CPU 20 compares the calculated pattern density of the inspection object 1 with a preset setting value, and controls the annular illumination according to the pattern density, that is, the density is high. Mask elements 5b-1, 5b-2, and 5b-3 (those having a large outer σ) shown in FIG. 4 are selected as ring-shaped illumination light sources so that the oblique incident components are increased. Of course, the control of the annular illumination performed by the CPU 20 is performed under predetermined conditions (information on the pattern type of the inspection object 1 mounted on the stage 2 input by the input means 32 or the manufacture of the inspection object 1). You may carry out by the information of the kind including the process of the to-be-inspected object 1 mounted in the stage 2 from the host computer which manages the process. That is, when the type of the object 1 to be inspected mounted on the stage 2 is, for example, a 4 Mb DRAM memory element, the pattern density is not so high and the pattern can be recognized with low resolution. Illumination (annular illumination close to circular illumination (mask elements 5b-4, 5b-5, 5b-6, 5b-7 shown in FIG. 4 by the disc-shaped mask 5 (those with a small outer σ))) Alternatively, in the case of a 16 Mb DRAM memory device using circular illumination (mask elements 5b-8, 5b-9, 5b-10 shown in FIG. 4), it is necessary to detect a high density pattern with high resolution. The zonal illumination is controlled so as to use the zonal illumination (mask elements 5b-1, 5b-2, and 5b-3 (those having a large outer σ) shown in FIG. 4) having high resolution under preset conditions. There is a need. In the circular illumination, if a mask element having a σ of about 0.5 smaller than the mask element 5b-10 (σ is 0.65) shown in FIG. 4 is used, an image of a deep groove or hole pattern is used. Can be received by the image sensor 12a to obtain a high-contrast image signal having a deep groove or hole pattern.

  Also, for example, since the pattern density is high in the cell portion of the memory element, high resolution is obtained by ring-shaped illumination under preset conditions, and inspection sensitivity is not lowered in rough regions other than the cell portion (incident illumination light It is also possible to use normal circular illumination (so as not to reduce the illumination intensity). In this way, by using various annular illuminations including circular illumination, the objective lens (imaging optical system) 9 can achieve high resolution and high sensitivity for various patterns (circuit patterns). In particular, it is possible to cope with a high-density pattern generated as the degree of integration increases. Further, it is not necessary to secure an NA (Numerical Aperture) of the objective lens (imaging optical system) 9 more than necessary so as not to sacrifice the depth of focus. Further, a means (attenuation filter 38) for partially controlling the light intensity is provided at a position 10b conjugate with the position of the pupil 10a of the objective lens 9, and a means (attenuation filter 38) for partially controlling the light intensity. When the 1st-order diffracted light does not enter the pupil 10a of the objective lens 9, the 0th-order diffracted light that enters the pupil 10a of the objective lens 9 and reaches the attenuation filter 38 is shielded or reduced in intensity. When the + 1st order diffracted light, the −1st order diffracted light, and the 0th order diffracted light enter the pupil 10a of the objective lens 9, the first order diffracted light and the 0th order are controlled by means for partially controlling the light intensity (attenuation filter 38). It is also possible to control the intensity so as to coincide with the next diffracted light. That is, based on the information detected by the image sensor 12b that is a monitor of the pupil 10a of the objective lens 9, the CPU 20 causes the image sensor 12a to generate 0th-order diffracted light and 1st-order diffracted light according to the pattern of the object 1 to be inspected. The transmission mechanism (attenuation rate) can be partially controlled by switching the means (attenuation filter 38) for partially controlling the light intensity by driving and controlling the moving mechanism 39 so as to receive light in a balanced manner.

  Next, as an object 1 to be inspected, an example of a lattice-like pattern in an LSI wafer pattern as shown in FIG. 6 will be described for detecting a lattice-like pattern with high resolution using annular illumination. . That is, FIG. 6 is a schematic diagram showing a lattice-like pattern composed of lines and spaces in the peripheral circuit portion of the LSI wafer pattern. In FIG. 6, reference numeral 101 denotes a pattern line (a wiring pattern including a gate) extending in the Y-axis direction. The lattice pattern lines 101 are repeated at a pitch P in the X-axis direction. In addition, a space (may be formed of an insulator or the like) is formed between the pattern lines 101. FIG. 7 shows the grid-like pattern shown in FIG. 6 by the ring-shaped incident illumination light 24 illuminating the grid-like pattern shown in FIG. 6 and the incident illumination light 24a incident on the XZ plane passing through the optical axis 33. FIG. 3 is a schematic explanatory view schematically showing a zero-order diffracted light 22a, a + 1st-order diffracted light 25a, and a −1st-order diffracted light 26a obtained by reflecting the light on the pupil 10a of the objective lens 9; FIG. 8 shows XZ light passing through the optical axis 33 through the optical axis 33 with the 0th-order diffracted light 22a, the + 1st-order diffracted light 25a, and the −1st-order diffracted light 26a obtained by reflecting the incident illumination light 24a shown in FIG. FIG. That is, as shown in FIGS. 7 and 8, on the image detected on the pupil 10a of the objective lens 9 by the image sensor 12b as a monitor, the 0th-order diffracted light 22a and the + 1st-order diffracted light 25a are point-like in the pupil 10a. It is observed as having an area. Reference numeral 34 denotes an illumination range on the lattice-shaped LSI wafer pattern by the annular illumination 24.

  As shown in FIG. 6, when the lattice-shaped LSI wafer pattern is oriented in the Y-axis direction, as shown in FIGS. 7 and 8, the 0th-order diffracted light 22a is incident illumination from the relationship of the formula (Equation 2) described later. The incident angle Ψ of the light 24a and the outgoing angle θ of the 0th-order diffracted light 22a are the same and occur at a position symmetrical to the incident illumination light 24a. The + 1st-order diffracted light 25a and the −1st-order diffracted light 26a Therefore, it occurs at the left and right positions. Since the lattice-like LSI wafer pattern is oriented in the Y-axis direction, the + 1st order diffracted light 25a and the −1st order diffracted light 26a are generated at the left and right positions in the pupil, but are not generated at the upper and lower positions or have occurred. Even weak. However, as can be seen from FIGS. 7 and 8, by applying the annular illumination 24, not only the 0th-order diffracted light 22 but also the + 1st-order diffracted light 25 or the −1st-order diffracted light 26 is always used as the pupil of the objective lens 9. The image sensor 12a can detect the image signal of the lattice-like LSI wafer pattern with high resolution.

  FIG. 9 shows an example of the ring-shaped incident illumination light 24 that illuminates the lattice-shaped pattern shown in FIG. 6 and the incident illumination light 24b incident on the YZ plane passing through the optical axis 33 from the lattice-shaped pattern shown in FIG. FIG. 6 is a schematic explanatory view schematically showing a zero-order diffracted light 22b, a + 1st-order diffracted light 25b, and a −1st-order diffracted light 26b obtained by reflection on the pupil 10a of the objective lens 9. That is, since the lattice-shaped LSI wafer pattern is oriented in the Y-axis direction, the 0th-order diffracted light 22b has an incident angle Ψ of the incident illumination light 24b and an exit angle of the 0th-order diffracted light 22b from the relation of the following formula (2). It becomes the same as θ and occurs at a position symmetrical to the incident illumination light 24b, and the + 1st order diffracted light 25b and the −1st order diffracted light 26b enter the pupil 10a of the objective lens 9 as shown in FIG. However, since the lattice-shaped LSI wafer pattern is oriented in the Y-axis direction, the + 1st-order diffracted light 25b and the −1st-order diffracted light 26b are weak even if they enter the pupil 10a of the objective lens 9, and the lattice-shaped LSI wafer pattern Therefore, the annular illumination in the Y-axis direction may be removed using the mask elements 5a-n shown in FIG. These first-order diffracted lights 23b are weaker than the 0th-order diffracted lights 22b. However, as shown in FIG. 9, both the + 1st-order diffracted lights 25b and the -1st-order diffracted lights 26b are used in the pupil 10a of the objective lens 9. In this case, even if the 0th-order diffracted light 22b enters the pupil 10a of the objective lens 9 and is received by the image sensor 12a, the resolution of the lattice-shaped LSI wafer pattern does not decrease so much.

  FIG. 10 shows a lattice shape shown in FIG. 6 when the outer σ and inner σ of the annular illumination with respect to the pupil 10a (NA) of the objective lens 9 are made larger than those shown in FIGS. Reflected from the lattice pattern shown in FIG. 6 by the ring-shaped incident illumination light 24 ′ for illuminating the pattern and the incident illumination lights 24′a and 24′b incident on the XZ plane and the YZ plane passing through the optical axis 33. The 0th-order diffracted lights 22′a and 22′b, the + 1st-order diffracted lights 25′a and 25′b, and the −1st-order diffracted lights 26′a and 26′b obtained on the pupil 10a of the objective lens 9 are schematically shown. FIG. As shown in FIG. 10, when the outer σ and inner σ of the annular illumination with respect to the pupil 10a (NA) of the objective lens 9 are made larger than those shown in FIGS. The folded light 25′b and the −1st order diffracted light 26′b are not incident on the pupil 10a, and when the 0th order diffracted light 22′b is received by the image sensor 12a, the resolution of the lattice pattern is lowered. Therefore, in order not to generate the 0th-order diffracted light 22′b oriented in the Y-axis direction, it can be realized by eliminating the annular illumination in the Y-axis direction using the mask elements 5a-n shown in FIG. . Further, a means (attenuation filter 38) for partially controlling the light intensity similar to the mask elements 5a-n shown in FIG. 3 is installed at a position 10b conjugate with the position of the pupil 10a of the objective lens 9 and directed in the Y-axis direction. By shielding the 0th-order diffracted light 22'b, the 0th-order diffracted light 22'b can be prevented from being received by the image sensor 12a. With this configuration, even if the outer σ and inner σ of the annular illumination are enlarged with respect to the pupil 10a (NA) of the objective lens 9, the lattice pattern is increased by the image sensor 12a. It can be detected with resolution.

  In the embodiment shown in FIG. 10, the outer σ and inner σ of the annular illumination with respect to the pupil 10a (NA) of the objective lens 9 are made larger than those shown in FIGS. However, even when the outer σ and inner σ of the annular illumination are respectively shown in FIGS. 7 to 9 and the pupil 10a (NA) of the objective lens 9 is reduced, the pupil 10a (NA) of the objective lens 9 is also reduced. The state of generation of the diffracted light incident on () is the same as that of the embodiment shown in FIG. Further, as apparent from the relationship of the formula (2) described later, when the pitch P of the lattice pattern becomes finer than that shown in FIGS. 7 to 9, the wavelength λ of the annular illumination light is changed from FIG. When the length is longer than that shown in FIG. 9, the state of generation of the diffracted light incident on the pupil 10a (NA) of the objective lens 9 is the same as in the embodiment shown in FIG. It is necessary to prevent the image sensor 12a from receiving light. Next, a diffraction phenomenon obtained from a lattice pattern by annular illumination will be described. First, the relationship between the σ value of arbitrary annular illumination and the incident angle Ψ with respect to the optical axis 33 will be described with reference to FIGS. 11 and 12. That is, FIG. 11 is an explanatory diagram showing the relationship between the σ value with respect to the optical axis 33 of the objective lens 9 and the incident angle Ψ in the incident illumination light 30 that illuminates the lattice pattern surface of the inspection object 1, and FIG. FIG. 6 is an explanatory diagram showing a relationship between an incident angle Ψ and an emission angle (diffraction angle) θ of the diffracted light 31. In FIG. 11, the following equation is established by the σ value of the annular illumination incident on the pupil 10 a of the objective lens 9 and the incident angle Ψ irradiated on the inspection object 1.

σ ₁ : σ ₂ = sin ψ ₁ : sin ψ ₂
sinψ ₂ = (σ ₂ / σ ₁ ) × sinψ ₁
In this study, the objective lens 9 having chromatic aberration corrected, having a magnification of 40 (x40) and NA = 0.8 is used. In the case of this objective lens 9, the maximum incident angle satisfies NA = sin Ψmax = 0.8 when σ = 1.0. That is, σ = 1.0 indicates the NA (aperture) (exit pupil) of the objective lens 9.

sinΨ = (σ / 1) × 0.8 = 0.8σ
Accordingly, the incident angle Ψ can be obtained from the relationship expressed by the following equation (1).

Ψ = sin ⁻¹ (0.8σ) (Equation 1)
Next, the relationship between the incident angle Ψ and the diffraction angle θ will be described.

  In FIG. 12, the diffraction angle θ of the m-th order diffracted light 31 has a relationship expressed by the following equation (Equation 2).

P = mλ / (sinΨ−sinθ)
sin θ = sin Ψ−mλ / P
θ = asin (sin Ψ− (mλ / P)) (Equation 2)
In the equation (2), λ is the wavelength (μm) of illumination light, θ is the diffraction angle (emission angle), P is the pattern pitch (μm), and m is the order of the diffracted light. In the formula (2), asin represents arcsin. Theoretical values (incident angle Ψ and diffraction angle) for the σ value of the annular illumination when the wavelength λ and the pattern pitch P of the object to be inspected are changed based on the above formulas (1) and (2). θ) is shown in the following (Table 2), (Table 3), (Table 4), and (Table 5).

  The above (Table 2) shows the case where the wavelength λ is 0.4 μm and the pattern pitch P is 0.61 μm, and the above (Table 3) shows the case where the wavelength λ is 0.6 μm and the pattern pitch P is 0.61 μm. (Table 4) shows the case where the wavelength λ is 0.4 μm and the pattern pitch P is 0.7 μm, and (Table 5) shows the case where the wavelength λ is 0.6 μm and the pattern pitch P is 0.7 μm. The incident angle Ψ and the diffraction angle θ are calculated based on the above formulas (1) and (2). In the LSI wafer pattern, the pattern pitch P = 0.61 μm corresponds to 256 Mb, and the pattern pitch P = 0.7 μm corresponds to 64 Mb. In the table,-indicates that calculation is not possible (theoretically, -1st order diffracted light is not generated). When the diffraction angle θ of the first-order diffracted light is 53.13 degrees or more, it does not enter the pupil 10a of the objective lens 9 with NA = 0.8.

  Ring-like illumination (σ value = 0.4, 0.6) 24 and a lattice pattern based on these theoretical values (FIG. 13 shows pattern pitch P = 0.61 μm, FIG. 14 shows pattern pitch P = 0.7 μm) FIG. 13 and FIG. 14 show the relationship with the + 1st order diffracted light 25, 25 ″ obtained by strengthening from FIG. 13A. FIG. 13A shows an annular illumination (σ value = 0.4, 0.6). The wavelength λ is in the range of 0.4 μm to 0.6 μm.) + 1st order diffracted light obtained by strengthening from a lattice pattern (corresponding to 256 Mb in an LSI wafer pattern) with a pattern pitch P = 0.61 μm by FIG. 13B shows an incident angle Ψ by an annular illumination with a σ value = 0.4, 0.6 (wavelength λ is in the range of 0.4 μm to 0.6 μm), Lattice pattern (LSI with pattern pitch P = 0.61 μm It is a diagram showing the range of the diffraction angle θ of the + 1st order diffracted light obtained from (corresponding to 256 Mb in the wafer pattern) .The diffraction range of the + 1st order diffracted light shown in FIG. This corresponds to the annular illumination with σ values of 0.4 and 0.6 in (Table 2) and (Table 3), and the region where the diagonal lines shown in FIG. A region of + 1st order diffracted light obtained by strengthening from a lattice-like pattern having a pattern pitch P = 0.61 μm by annular illumination with values = 0.4 and 0.6 (average wavelength of the annular illumination light (wavelength λ Corresponds to the region of the + 1st order diffracted light in the case where the objective lens 9 has a pattern pitch P = 0.61 μm and is strengthened from the lattice-like pattern in FIG. An annular zone of + 1st order diffracted light incident on the pupil 10a Region 25 is shown.

  FIG. 14A shows a grating with a pattern pitch P = 0.7 μm by annular illumination with a σ value = 0.4, 0.6 (wavelength λ is in the range of 0.4 μm to 0.6 μm) 24. + 1st order diffracted light 25 ″ obtained by strengthening from a pattern of a pattern (corresponding to 64 Mb in an LSI wafer pattern), FIG. 14B shows an annular illumination with σ value = 0.4, 0.6 (The wavelength λ is in the range of 0.4 μm to 0.6 μm.) Obtained from the incident angle Ψ by 24 and a lattice pattern (corresponding to 64 Mb in an LSI wafer pattern) with a pattern pitch P = 0.7 μm. 15 is a diagram showing the range of the diffraction angle θ of the + 1st order diffracted light, and the diffraction range of the + 1st order diffracted light (the range of the diffraction angle θ) shown in FIG. = Corresponding to the annular illumination of 0.4, 0.6. 14B is obtained by strengthening from the lattice pattern with the pattern pitch P = 0.7 μm by the annular illumination with the σ values = 0.4 and 0.6. 14 shows the region of the + 1st order diffracted light (corresponding to the region of the + 1st order diffracted light when the average wavelength of the annular illumination light (wavelength λ is 0.5 μm)), that is, FIG. An annular region 25 ″ of + 1st order diffracted light that is strengthened from a lattice-like pattern of P = 0.7 μm and is incident on the pupil 10a of the objective lens 9 is shown.

  Comparison of FIG. 13 and FIG. 14 shows that the diffraction angle θ of the first-order diffracted light increases as the pattern pitch P becomes fine, and it is clear that annular illumination is necessary. As described above, the annular illumination of σ values = 0.4 and 0.6 shown in FIGS. 13 and 14 can be realized by using the mask element 5b-6 shown in FIG. The above description is based on the above formulas (1) and (2), but when the inventor actually experimented, an image detected on the pupil 10a of the objective lens 9 by the image sensor 12b as a monitor. From FIG. 14, the result is almost the same (result shown in FIG. 13: pattern of pattern pitch P = 0.61 μm (corresponding to 256 Mb in the LSI wafer pattern)), result shown in FIG. 14: pattern pitch P = 0.7 μm A grid-like pattern (corresponding to 64 Mb in the LSI wafer pattern) was obtained.

  In the embodiment shown in FIGS. 7 to 14, the case of the lattice pattern repeated in the X-axis direction in the LSI wafer pattern as shown in FIG. 6 has been described. There also exists a lattice-like pattern composed of pattern lines 102 repeated in the Y-axis direction as shown. Therefore, the diffracted light 22a, 25a, 26a obtained from the lattice-like pattern composed of the pattern line 102 repeated in the Y-axis direction as shown in FIG. 15 by the annular illumination 24 is incident on the pupil 10a of the objective lens 9. Is as shown in FIG. Since the lattice-shaped pattern composed of the pattern lines 101 shown in FIG. 6 and the lattice-shaped pattern composed of the pattern lines 102 shown in FIG. 15 are rotated by 90 degrees, the state shown in FIG. The state is rotated 90 degrees. Therefore, the diffracted light generation state for the lattice pattern formed by the pattern lines 102 shown in FIG. 15 is obtained by rotating the diffracted light generation state shown in FIGS. 8 to 10 by 90 degrees. That is, the 0th-order diffracted light 22a and the + 1st-order diffracted light 25a obtained by reflecting the incident illumination light 24a incident on the YZ plane passing through the optical axis 33 from the lattice pattern shown in FIG. The first-order diffracted light 26a is incident on the pupil 10a of the objective lens 9 as shown in FIG. 16, and the incident illumination light 24a in the direction orthogonal to the pattern line 102 is effective for improving the resolution. Recognize. However, in the same manner as described with reference to FIG. 9, the incident illumination light incident on the XZ plane passing through the optical axis 33 in the annular illumination light 24 is reflected from the lattice pattern shown in FIG. Even if the + 1st order diffracted light 25b and the −1st order diffracted light 26b are incident on the pupil 10a of the objective lens 9, they are weaker than the 0th order diffracted light 22b and do not contribute to the improvement of the resolution. It is preferable to remove incident illumination light that is incident in the X-axis direction (XZ plane passing through the optical axis 33) from the annular incident illumination light 24.

  In any case, the CPU 20 determines the distribution state of the diffracted light generated from the lattice pattern by the annular illumination and entering the pupil 10a of the objective lens 9 on the pupil 10a (Fourier transform plane) of the objective lens 9. Can be detected based on an image signal obtained from the image sensor 12b that receives the image (the generation position and brightness of the 0th-order diffracted light 25a and the generation position and brightness of the + 1st-order diffracted light 25a). That is, the CPU 20 distributes the diffracted light incident on the pupil 10a of the objective lens 9 detected based on the image signal obtained from the image sensor 12b (the generation position of the 0th-order diffracted light 25a and its brightness, and the +1 next time. Based on the generation position of the folding light 25a and its brightness), the moving mechanism 19 is driven and controlled to select a mask element, and an annular illumination suitable for the lattice pattern (LSI wafer pattern) of the inspection object 1 is provided. Obtainable. As a result, a high-resolution image signal of the lattice pattern (LSI wafer pattern) of the inspection object 1 can be obtained from the image sensor 12a.

  Next, the operation of the means (attenuation filter 38) for partially controlling the light intensity provided at the position 10b conjugate with the position of the pupil 10a of the objective lens 9 will be described. That is, as shown in FIGS. 9 and 10, 0th-order diffracted lights 22b and 22′b that are incident on the pupil 10a of the objective lens 9 and need not be received by the image sensor 12a are shielded by the attenuation filter 38. Can do. In this case, the attenuation filter 38 serves as a spatial filter. Also, as shown in FIGS. 8 and 16, the intensity of the 0th-order diffracted light 22a that has entered the pupil 10a of the objective lens 9 is a position 10b conjugate with the position of the pupil 10a of the objective lens 9 as shown in FIGS. The image sensor 12a can receive the light by balancing the intensity of the 0th-order diffracted light 22a and the intensity of the + 1st-order diffracted light 25a incident on the pupil 10a of the objective lens 9, As a result, it is possible to detect an image of a lattice pattern of the inspection object 1 with high resolution and high contrast. The shape of the attenuation filter 38 has the same shape as the mask element 5a-1 shown in FIG. However, as shown in FIG. 3, it is not always necessary to have a ring shape, and any shape can be used as long as the light intensity at an arbitrary position can be controlled. However, when a ring shape is used as the shape of the attenuation filter 38, it is necessary to optimize the annular illumination 24 so that the 0th-order diffracted light 22a and the + 1st-order diffracted light 25a are not generated in the same ring-shaped region. There is.

  FIG. 17 shows that the 0th-order diffracted light 22a and the + 1st-order diffracted light 25a generated from the lattice pattern of the object 1 to be inspected by the annular illumination 24a are in a position conjugate with the pupil 10a of the objective lens 9 and the pupil 10a. It is a figure which shows the state which advances to the pupil 10b. FIG. 18 is a diagram showing the attenuation filter 38 disposed on the pupil 10b at a position conjugate with the pupil 10a of the objective lens 9. FIG. That is, it can be seen that the intensity of the 0th-order diffracted light 22a incident on the pupil 10a of the objective lens 9 is controlled by the attenuation filter 38 on the pupil 10b. FIG. 19 is a diagram schematically showing the attenuation filter 38a. (A) shows the transmission characteristic of the attenuation filter 38a, and (b) shows its shape. FIG. 20 is a diagram schematically showing another attenuation filter 38b, where (a) shows the transmission characteristics of the attenuation filter 38b, and (b) shows its shape. In this way, the transmission characteristics and the shape of the attenuation filter 38 may be optimized in accordance with the 0th-order diffracted light 22a and the + 1st-order diffracted light 25a generated from the lattice-like pattern of the inspection object 1 by the annular illumination 24a. .

  If the annular illumination can be optimized according to the lattice pattern (LSI wafer pattern) of the object 1 to be inspected, the attenuation filter 38 need not be installed. However, in order to optimize only with the ring-shaped illumination, it is necessary to prepare and select various types of ring-shaped illumination so as to correspond to various patterns on the inspection object 1. . Therefore, when the selection of the annular illumination is minimized, the intensity of the diffracted light is controlled by using the attenuation filter 38 or the like on the light receiving side, and the lattice pattern of the object 1 to be inspected by the image sensor 12a. It is desirable to detect the image with high resolution and high contrast. Regarding the attenuation filter 38, the CPU 20 also distributes the diffracted light incident on the pupil 10a of the objective lens 9 detected based on the image signal obtained from the image sensor 12b (the generation position of the 0th-order diffracted light 25a and its brightness). Further, based on the generation position and brightness of the + 1st order diffracted light 25a, the moving mechanism 39 is driven and controlled to select the attenuation filter 38, so that a lattice pattern (LSI wafer pattern) of the inspection object 1 is obtained. Suitable intensities of 0th-order diffracted light and 1st-order diffracted light can be obtained. As a result, a high-resolution image signal of the lattice pattern (LSI wafer pattern) of the inspection object 1 can be obtained from the image sensor 12a.

  In any case, since it is difficult to optimize only with the annular illumination so that various patterns on the inspection object 1 can be handled, the image sensor 12a receives light using the attenuation filter 38 as described above. It is necessary to control the detection sensitivity by controlling the intensity of the diffracted light and controlling the threshold value and the like in the image processing performed by the comparison circuit 17 or the CPU 20. As described above, the threshold value and the like in the image processing performed by the comparison circuit 17 or the CPU 20 are controlled by the image of the pupil 10b of the objective lens 9 detected by the image sensor 12b or the image of the pattern on the inspection target 1 detected by the image sensor 12a. Based on the above. For example, the CPU 20 stores, for example, a memory according to the locality distribution (the generation position including the spread and its intensity) of the diffracted light in the image of the Fourier transform plane (image of the pupil 10b of the objective lens 9) detected by the image sensor 12b. What is necessary is just to judge that it is an area | region which has a pattern with high repeatability like a cell, and to raise a detection sensitivity by controlling a threshold value etc. in the image processing which the comparison circuit 17 or CPU20 performs. Conversely, when it is determined that the region has a pattern with low repeatability, the detection sensitivity may be lowered by controlling the threshold value or the like in the image processing performed by the comparison circuit 17 or the CPU 20.

  In particular, when a defect in a pattern on the inspection object 1 is inspected by image processing performed by the comparison circuit 17 or the CPU 20, for example, an image of a Fourier transform plane detected by the image sensor 12b (an image of the pupil 10b of the objective lens 9). If the detection sensitivity is increased by controlling the threshold value or the like according to the locality distribution of the diffracted light in the image) (the generation position including the spread and its intensity), for example, a memory cell or the like can be obtained by annular illumination. In addition, it is possible to easily detect a defect in a region having a pattern with high repeatability. Next, another embodiment for changing the ring shape of the annular illumination emitted from the disk-shaped mask (secondary light source for annular illumination) 5 formed from a large number of virtual point light sources is shown in FIGS. 22 will be described. That is, FIG. 21 and FIG. 22 are diagrams schematically showing another embodiment for controlling the annular illumination. In FIG. 21, the lamp house 24 constituted by the Xe lamp 3, the elliptical mirror 4, and the disk-like mask 5 for forming the annular illumination is moved in the optical axis direction with respect to the collimator lens 6. The ring shape is changed to control the annular illumination. In FIG. 22, the collimator lens 6 is moved in the optical axis direction with respect to the lamp house 24 constituted by the Xe lamp 3, the elliptical mirror 4, and the disk-shaped mask 5 for forming the annular illumination. The ring shape is changed to control the annular illumination.

  In the embodiment of the lamp house 24 forming the annular light source for illumination 5 formed from a large number of virtual point light sources shown in FIGS. 1, 21, and 22, the Xe lamp 3 is arranged in the vertical direction. In the case where the Xe lamp 3 is arranged in the vertical direction as described above, the luminous flux in the optical axis direction is reduced. Therefore, the Xe lamp 3 is arranged in the horizontal direction to increase the luminous flux in the optical axis direction. It can also be configured as follows. Further, as the light source in the lamp house 24, not only the Xe lamp but also an Hg lamp, a halogen lamp or the like may be used. When a disk-shaped mask (annular illumination secondary light source) 5 formed from a large number of virtual point light sources is selected according to the pattern of the object 1 to be inspected, the annular illumination secondary light source Since the light amount of the annular illumination emitted from the image sensor 5 greatly fluctuates, the CPU 20 performs A / D conversion from the image sensor 12a so that the light amount received by the image sensor 12a shown in FIG. The light amount is controlled by controlling the light amount adjustment filter 14 such as an ND filter based on the image signal 41 obtained through the device 41.

[Second Embodiment]
Next, a microscope system (microscope observation system) used for pattern inspection or the like of the inspection object 1 using the ring-shaped illumination according to the present invention will be described with reference to FIG. In this embodiment, the microscope system (microscope observation system) according to the present invention is applied to pattern inspection of an inspection object 1 such as an LSI wafer pattern. FIG. 23 is an explanatory diagram schematically showing a microscope system (microscope observation system) according to a second embodiment of the present invention. In FIG. 23, the same reference numerals as those in FIG. In the microscope system using the annular illumination formed from a large number of virtual point light sources shown in FIG. 23, the description of the common part with the pattern inspection apparatus shown in FIG. 1 is omitted, and only the characteristic part is described. That is, in FIG. 23, TV cameras 12a ′ and 12b ′ are used as the image sensors 12a and 12b in FIG. 1, and the output images can be visually observed by the operator through the monitors 27a and 27b. It should be noted that 12a ′ and 12b ′ may be any one that can detect an image, and may be configured by an image sensor instead of a TV camera. That is, the TV camera 12a ′ detects a pattern image, and the TV camera 12b ′ detects an image on the pupil 10a of the objective lens 9, and displays each image on the monitors 27a and 27b. Further, a controller 46 is connected so that the sample stage 2 can be driven and controlled in the X, Y, Z, and θ (rotation) axis directions by a driver 45. The controller 46 detects the movement mechanism 19 based on the locality distribution image of the first-order diffracted light including the zeroth order that is detected by the TV camera 12b ′ and displayed on the monitor 27b and incident on the pupil 10a of the objective lens 9. Then, the lamp house 24 and the collimator lens 6 are driven and controlled, and an annular illumination or a normal circular illumination suitable for the pattern of the inspection object 1 is selected. When the disc-shaped mask 5 is driven and controlled by the moving mechanism 19, the mask element formed on the disc-shaped mask 5 may be selected. Further, when driving the lamp house 24 and the collimator lens 6 is controlled as shown in FIGS.

  Further, the controller 46 detects the movement mechanism 39 based on the locality distribution image of the first-order diffracted light including the 0th-order incident on the pupil 10a of the objective lens 9 detected by the TV camera 12b ′ and displayed on the monitor 27b. Is controlled to select an attenuation filter 38 suitable for the pattern of the inspection object 1. It should be noted that the attenuation filter 38 is not necessarily provided if the ring-shaped illumination secondary light source 5 can select appropriate or normal circular illumination for the pattern of the inspection object 1. Further, the controller 46 obtains from the pattern of the inspection object 1 by controlling the drive of the control mechanism 14b based on the pattern image of the inspection object 1 detected by the TV camera 12a ′ and displayed on the monitor 27a. The light amount control filter 14 is controlled so that the amount of light to be obtained becomes appropriate. Thus, according to the microscope observation system using the ring-shaped illumination, the pitch P (for example, 0.7 μm, 0.61 μm) of the lattice-like pattern as in the memory element of 64 Mb DRAM and 256 Mb DRAM as in the LSI wafer pattern. However, even if the wavelength becomes close to the wavelength λ (for example, 400 to 600 nm) of the illumination light, the high-density pattern is detected by the TV camera 12a ′ and displayed on the monitor 27a. The pattern image of the object 1 can be observed with high resolution and high contrast. In the circular illumination, if illumination is performed using a mask element having a σ of about 0.5, an image of a deep groove or hole is received by the TV camera 12a ′ and displayed on the monitor 27a with high contrast. be able to.

[Third embodiment]
Next, modifications of the first and second embodiments will be described.
That is, in the first and second embodiments, the explanation is based on the annular illumination and the circular illumination. However, the annular illumination includes so-called modified illumination (oblique illumination) (this modified illumination is , Which indicates an illumination condition in which at least zeroth-order diffracted light and first-order or second-order diffracted light are incident on the pupil 10a of the objective lens 9). The dark field illumination is not included in the modified illumination (oblique illumination) because normally the 0th-order diffracted light is not incident on the pupil 10a of the objective lens 9. In the first and second embodiments, the attenuation filter 38 attenuates the light transmittance as shown in FIGS. 19 and 20, but a method called a phase shift method, that is, a phase film May be used to attenuate the light amount of the 0th-order diffracted light 22a compared to the + 1st-order diffracted light 25a received by the image sensor 12a by shifting the phase of the 0th-order diffracted light 22a. That is, the means (attenuation filter) 38 for partially controlling the light intensity is provided at a position 10b conjugate with the position of the pupil 10a of the objective lens 9, but a phase plate may be placed at this position 10b. For example, the 0th-order diffracted light 22a received by the image sensor 12b can be weakened by advancing the phase of the 0th-order diffracted light 22a by Π / 2 with respect to the phase of the + 1st-order diffracted light 25a. Further, by providing the phase plate with an absorption characteristic, the 0th-order diffracted light 22a received by the image sensor 12b can be weakened.

  In the first and second embodiments described above, the framework of 0th order diffracted light and ± 1st order diffracted light is used. However, 0th order diffracted light (non-diffracted light) and diffracted light (± 1st order diffracted light or ± 2nd order diffracted light). Needless to say, it can also be applied to the framework. That is, even if the pitch P of the pattern becomes fine (the pattern becomes dense), the 0th-order diffracted light and the + 2nd-order diffracted light or -2nd-order diffracted light obtained from the pattern enter the pupil 10a of the objective lens 9. In this way, ring-shaped illumination may be applied. Usually, the first-order diffraction angle is smaller than the second-order diffraction angle, as shown by the relationship of the formula (2). However, the second-order diffraction angle may be smaller than the first-order diffraction angle depending on the pattern pitch P and the wavelength λ of the annular illumination. In the first and second embodiments, an example of the Xe lamp 3 (the dimensions are not specified) is shown as the light source in the light house 24. However, a large light source (that is, incoherent) is shown. A light source that emits light) or a point light source (that is, a light source that emits coherent light). Further, by selecting a light source, it is possible to obtain an appropriate σ value using only the primary light source (no mask element). Moreover, in the said 1st and 2nd Example, as wavelength lambda of ring-shaped illumination, although demonstrated by the normal wavelength of 400-600 nm, in each said Example, although the illumination wavelength is not described, The wavelength λ of the annular illumination may be a so-called i-line (about 365 nm) wavelength or, of course, a short wavelength such as excimer laser light (ultraviolet light). Needless to say, if the short-wavelength light such as excimer laser light (ultraviolet light) is used as the annular illumination light, the resolution is further improved.

  The control of the annular illumination in the first and second embodiments is performed for each type of pattern of the object to be inspected (for example, in the case of an LSI wafer pattern, for each process step or for each type of LSI wafer). You can go. Of course, the ring-shaped illumination may be dynamically controlled within one LSI wafer. In the case of the defect inspection of the pattern of the inspection object 1, the sensitivity may be controlled for each type of pattern of the inspection object in the same manner as the control of the annular illumination, or one LSI wafer. It may be done dynamically within.

  In addition, the adjustment of the secondary light source for annular illumination including the primary light source (Xe lamp 3) in the lamp house 24 in the first and second embodiments is performed by using a mirror wafer as the object 1 to be inspected. The ring-shaped illuminance distribution on the pupil 10a of the objective lens 9 (the distribution of the ring-shaped zero-order diffracted light 22 on the pupil 10a of the objective lens 9 detected by the image sensor 12b) may be made uniform. That is, for the annular illumination, the distribution of the ring-shaped zero-order diffracted light 22 on the pupil 10a of the objective lens 9 detected by the image sensor 12b using a mirror wafer as the object 1 to be inspected is uniform. For example, the position of the Xe lamp 3 and the position of the elliptical mirror 4 that form the secondary light source may be adjusted. In the first and second embodiments, the locality distribution (spreading) of the diffracted light (0th order diffracted light and + 1st order diffracted light) on the pupil 10a of the objective lens 9 detected by the image sensors 12b and 12b ′. It has been described that image information (monitor information) based on the included position and brightness is used for condition control of each unit by the CPU 20 or the controller 46. That is, as the condition control of each part, illumination such as ring-shaped illumination control (for example, control of inner σ and outer σ values and incident range shown in FIGS. 3 and 4) and light amount control by the light amount adjustment filter 14. There are control of conditions, control of the amount of light detected by the attenuation filter 38, and control of detection sensitivity in the comparator 17 and the like. Further, the CPU 20 or the controller 46 determines whether the area is a repetitive portion in the memory element or the like from the image information based on the locality distribution of the diffracted light on the pupil 10a of the objective lens 9 detected by the image sensors 12b and 12b ′. As a result, it is possible to control the ring-shaped illumination including the appropriate circular illumination according to whether the region is the identified repetitive portion region or the other region. .

[Fourth embodiment]
Next, an example of inspecting a defect in the memory cell portion as an LSI wafer pattern by optically improving the resolution using an annular illumination will be described with reference to FIGS. That is, FIG. 24 is a diagram showing defects in the memory cell portion of the LSI wafer pattern. FIG. 25 is a diagram showing the relationship between this LSI wafer pattern and detection pixels obtained from the A / D converter 15a. FIG. 26 is a diagram illustrating an image signal waveform obtained by receiving light from the high-density pattern with high resolution by the image sensor 12a by the annular illumination. FIG. 27 is a diagram for explaining sampling performed by the A / D converter 15a on the image signal shown in FIG. In the memory cell portion of the LSI wafer pattern, as shown in FIG. 24, various defects (that is, protrusions 231, open 232, discoloration 233, short circuit 234, chipping 235, contaminants 236, etc.) exist on the pattern. In order to detect the defects 231 to 236 with high reliability, the image sensor 12a must first be able to detect the pattern as an image signal with high resolution. By the way, by using annular illumination, the image sensor 12a can obtain a high-resolution image signal shown in FIG. 26B for the pattern shown in FIG. FIG. 26A is an enlarged view of a part of the pattern shown in FIG. 24, and FIG. 26B shows the position along the horizontal axis and the image sensor 12a along the vertical axis in the pattern AA ′. The waveform of the brightness in the obtained image signal (pattern detection signal) is shown. FIG. 26 shows a case where a high-resolution image signal in which turn edge information appears is obtained from the image sensor 12a using annular illumination. By the way, when ring-shaped illumination is performed, + first-order diffracted light having a wide range of diffraction angles is obtained from the protrusion defect 231, the chip defect 236, the contaminant 235, and the like. An image signal different from the pattern can be obtained from the image sensor 12a. Further, when ring-shaped illumination is performed, the open defect 232 and the short defect 234 do not generate a + 1st order diffracted light component in the X-axis direction, so that an image signal different from the pattern can be obtained from the image sensor 12a. In addition, when ring-shaped illumination is performed, for example, the generation state of the 0th-order diffracted light is changed from the color change defect 233 as compared with a region having no color change defect, and an image signal indicating the color change defect 233 can be obtained from the image sensor 12a. it can.

  FIG. 25 shows a case where detection pixels sampled by the A / D converter 15a are larger than the LSI wafer pattern shown in FIG. As shown in FIG. 25, when the detection pixels 241 to be sampled in the A / D converter 15a are large, there are two pattern edges in one detection pixel 241, and the pattern edge information is lost. It will be. In order to prevent the edge information of the pattern from being lost in this way, the size of the detection pixel 241 sampled in the A / D converter 15a may be reduced. When the size of the detection pixel 241 is reduced, the sampled digital image signal information obtained from the A / D converter 15a is increased, the amount of defect detection image processing performed by the comparison circuit 17 and the like is increased, and the time for defect detection is increased. It will take. Therefore, as shown in FIG. 27, in the A / D converter 15a, the pattern AA ′ is sampled with a detection pixel size (size) that preserves the minimum and maximum values of the pattern brightness. What is necessary is just to convert into the digital image signal which shows light and shade (brightness). That is, the CPU 20 reduces the pattern brightness from the digital image signal 41 (shown in FIG. 27A) obtained from the A / D converter 15a by reducing the pixel size to be initially sampled for the A / D converter 15a. The minimum value (indicating the edge information of the pattern) and the interval between the maximum values are calculated, and the detection pixel size is set such that these values are divisible. The A / D converter 15a samples the density (brightness) sampled with a relatively large detection pixel size without losing pattern edge information by sampling based on the detection pixel size 42 set in the CPU 20. The digital image signal shown (shown in FIG. 27B) can be obtained. Thereby, it is possible to reduce the information of defect detection image processing performed in the comparison circuit 17 and the like, and to perform defect inspection at high speed and with high reliability.

Referring to FIG. 27A, the CPU 20 first reduces the pixel size to be sampled with respect to the A / D converter 15a, and from the digital image signal 41 obtained from the A / D converter 15a, x ₁ = _{x 2} = x 3 = a detection pixel size between _{x 4 x} 1 _{is, x 2, x 3, x} 4 selects the _x 1/2, the minimum value and _x 5 is large spacing between maxima, _{x 6} parts sets a detection pixel size such that the _{3x 1} / _2,2x 1. A signal 42 corresponding to the detected pixel size set in this way is provided to the A / D converter 15a.

  FIG. 27B shows a digital image sampled by the A / D converter 15a based on the signal 42 provided from the CPU 20 with respect to the waveform of the image signal in the AA ′ section shown in FIG. The waveform of the signal is shown. As apparent from FIG. 27B, the A / D converter 15a obtains a digital image signal in which the minimum and maximum values indicating the edge of the pattern are stored from the image signal output from the image sensor 12a. Thereby, in the comparison circuit 17 or the like, the digital image signal delayed by the cell interval or the chip interval in the delay memory 16 and the digital image signal directly obtained from the A / D converter 15a are subjected to cell comparison or chip comparison. It is possible to inspect the defect at high speed and with high reliability by deleting the image signal indicating the edge of the pattern obtained at the resolution. Since the image signal indicating the edge of the pattern is repeated at the cell interval or the chip interval, the image signal indicating the pattern edge is erased when detected by the comparison circuit 17 or the like when performing cell comparison or chip comparison. Can do. As a result, the signal 18 indicating a defect can be detected as a mismatch in the cell comparison or chip comparison in the comparison circuit 17 or the like.

Incidentally, CPU 20 is in order to be sampled between the minimum value and the maximum value as the detection pixel size can be set to _{x =} x 1/4 or _x 1/8. Accordingly, A / D converter 15a can obtain the digital image signal shown by sampling shading (brightness) at the set detection pixel size (x = x _1/4 or x _1/8) . In this case, since the sampling interval is narrow, the high-resolution image signal obtained from the image sensor 12a can be faithfully converted into a digital image signal.

  Further, the CPU 20 reduces the pixel size to be initially sampled with respect to the A / D converter 15a, and based on the digital image signal 41 obtained from the A / D converter 15a, as shown in FIG. By controlling the zoom lens 13 by 43, the magnification of the image received by the image sensor 12a can be changed. As a result, in the A / D converter 15a, the detected pixel size to be sampled can be changed according to the magnification by the zoom lens 13 even if the signal 42 for determining the detected pixel size is kept constant. Accordingly, when it is desired to change the magnification by the zoom lens 13, it can be controlled by a command from the CPU 20. Further, a high-resolution image signal obtained by receiving, by the image sensor 12a, the 0th-order diffracted light 22a and the + 1st-order diffracted light 25a, which are generated from the lattice pattern by the ring-shaped illumination and are incident on the pupil 10a of the objective lens 9. Sampling in the A / D converter 15a will be described with reference to FIGS. 28, 29, and 30. FIG. The lattice pattern (line and space wafer pattern) in FIGS. 28 to 30A is a repetitive pattern having a pitch P = 0.84 μm composed of 0.42 μm wide lines and 0.42 μm wide spaces. is there.

  The digital image signal waveform indicating the sampled shading (brightness) shown in FIG. 28B shows the case where the detection pixel size is 0.0175 μm, and the edge information of the grid pattern is clearly detected. I understand that. That is, it indicates that the high-resolution image signal obtained by receiving light by the image sensor 12a has been faithfully converted into a digital image signal that shows light and shade (brightness) by the A / D converter 15a. The digital image signal waveform indicating the sampled shading (brightness) shown in FIG. 29B shows the case where the detected pixel size is 0.14 μm, and edge information (minimum value, maximum value, etc.) of the lattice pattern. It can be seen that the extreme value information) is detected in a stored state. The digital image signal waveform indicating the sampled shading (brightness) shown in FIG. 30B shows the case where the detection pixel size is 0.28 μm, and edge information (minimum value, maximum value, etc.) of the lattice pattern. It can be seen that a part of the extreme value information is detected and detected. Therefore, for a repetitive pattern (lattice pattern) having a pitch P = 0.84 μm composed of 0.42 μm wide lines and 0.42 μm wide spaces, the A / D converter 15a receives a signal 42 from the CPU 20. Therefore, it is necessary to convert the detection pixel size set by (1) to a digital image signal indicating lightness (brightness) by sampling with a size of about 0.3 μm or less. Thereby, the edge information (extreme value information such as the minimum value and the maximum value) of the lattice pattern is stored in the digital image signal indicating the light and shade (brightness) obtained from the A / D converter 15a. In the comparison circuit 17 or the like, it is possible to detect the defect (i.e., protrusion 231, open 232, discoloration 233, short circuit 234, chip 235, contaminant). 236). That is, when sampling is performed in the A / D converter 15a so that the minimum value and maximum value of the pattern are stored, the pattern information is not impaired even with a large detection pixel size, and the comparison circuit 17 or the like is high-speed and high. Accurate defect inspection and the like can be performed. In the embodiments shown in FIGS. 28 to 30 described above, the digital image signal waveform indicating the density (brightness) sampled by taking a one-dimensional grid pattern as an example has been described. Needless to say, this is true.

[Fifth embodiment]
In the said Example, although the Example which inspects the defect on the pattern formed in the to-be-inspected target object 1 and the Example of the microscope observation system which observes the pattern formed in the to-be-inspected target object 1 were demonstrated. The present invention can be applied to foreign matter inspection existing on the pattern formed on the inspection target object 1 and dimension measurement of the pattern formed on the inspection target object 1. That is, as described in the fourth embodiment, by obtaining from the A / D converter 15a a digital image signal showing a light and shade (brightness) faithful to the pattern formed on the inspection object 1, for example, The CPU 20 can measure the dimension of the pattern with high accuracy. Further, as described in the fourth embodiment, the foreign matter existing on the pattern (LSI wafer pattern) formed on the inspection object 1 can be detected in the same manner as the defect inspection. That is, from the foreign matter, first-order or higher-order diffracted light having a spread with various diffraction angles enters the pupil 10a of the objective lens 9 in the same manner as the projection defect 231 and the chip defect 236. On the other hand, the 0th-order diffracted light and the + 1st-order diffracted light enter the pupil 10a of the objective lens 9 from the pattern, and different image signals are detected from the image sensor 12a. By doing so, foreign matter can be detected. Of course, foreign matter on the mirror wafer can be detected in the same manner.

  Further, as will be described later in the embodiments, the pattern information can be erased by using the spatial filter 309 instead of being erased by cell comparison or chip comparison. Further, a foreign object can be detected based on an image signal on the pupil 10a of the objective lens 9 detected by the image sensor 12b. That is, the foreign matter on the mirror surface wafer can be directly detected from the image signal on the pupil 10a of the objective lens 9 detected by the image sensor 12b. Even for a foreign object existing on the pattern (LSI wafer pattern) formed on the inspection object 1, the locality distribution of the diffracted light incident on the pupil 10a of the objective lens 9 is different between the foreign object and the pattern. It can be detected by erasing the pattern information from the image signal on the pupil 10a of the objective lens 9 detected by the sensor 12b. That is, similar to the delay memory 16, by storing a reference image signal on the pupil 10a obtained from a normal pattern in which there is no foreign matter detected by the image sensor 12b, the stored pupil is stored in a comparison circuit or the like. By comparing the reference image signal on 10a with the image signal on the pupil 10a obtained from the pattern to be inspected, which is actually detected by the image sensor 12b, the pattern information can be erased to detect foreign matter. In addition, the locality distribution information of the diffracted light on the pupil 10a obtained from a normal pattern or the locality distribution information obtained by inverting this (spatial filter (shown by 309 in FIGS. 31 and 33)), from the pattern to be inspected. By masking (shading) the locality distribution information of the diffracted light on the obtained pupil 10a, foreign matter can be detected by erasing the pattern information.

[Sixth embodiment]
Next, a specific configuration of the optical system in the pattern inspection apparatus according to the present invention will be described with reference to FIGS. 31 and 32 are specific configuration diagrams of the optical system in the pattern inspection apparatus of the embodiment shown in FIG. 1, FIG. 31 is a plan view thereof, and FIG. 32 is a front view thereof.

  The configuration of the optical system in the pattern inspection apparatus shown in FIGS. 31 and 32 is basically the same as the configuration of the optical system in the pattern inspection apparatus shown in FIG. 1, and the same reference numerals in FIG. Since there is, detailed description is omitted again. A characteristic part of the present embodiment will be described.

In the present embodiment, as the TV camera for image observation, the TV ₁ for annular illumination (bright field illumination), the TV ₂ for dark field illumination, and the TV camera for pupil observation for observing the pupil 10a of the objective lens 9 are dark. A TV ₄ for field illumination is added. Accordingly, the TV camera TV ₁ for annular illumination (bright field illumination) and the TV camera TV ₂ for dark field illumination are used for image observation. An annular illumination (bright field illumination) TV ₃ as a pupil observation television camera for observing the pupil 10a of the objective lens 9 is the same as the image sensor 12b. That is, the image on the pupil 10a (local distribution of diffracted light obtained from the pattern on the inspection object 1 by the annular illumination) captured by the annular illumination (bright field illumination) TV ₃ (12b). Based on this, the CPU 20 detects the defect detection sensitivity or A / D converter in the ring-shaped illumination filter (disk-shaped mask: aperture stop) 5 or the zero-order diffracted light amount control pupil filter (attenuation filter) 38 or the comparison circuit 17. The detection pixel size by sampling in 15a or the magnification by the zoom lens 13 is selected and controlled. Further, based on the image on the pupil 10a imaged by the TV ₄ for dark field illumination (distribution of scattered light obtained from the pattern on the inspection object 1 by dark field illumination described later), the CPU 20 performs the spatial filter 309 or the comparison circuit. The detection sensitivity of the foreign matter in the image sensor or the like, or the detection pixel size by sampling in the A / D converter for A / D converting the image signal obtained from the linear image sensor 308 is selected and controlled. As a result, the pattern on the inspection object 1 can be optimized. The dichroic mirror 325 transmits the light of the image of the pupil 10a by the diffracted light obtained from the object 1 to be inspected by annular illumination with a wavelength of 600 nm or less, and dark field illumination with a wavelength of 780 to 800 nm described later. The light of the image of the pupil 10a by the scattered light obtained from the inspection object 1 is reflected. Reference numeral 326 denotes a mirror.

Further, the lamp house 24 'within a primary light source 3 and 4 shown in FIG. 1, constituted by two types of Hg-Xe lamp _{L 1} and Xe lamp _{L 2,} switches the two kinds of primary light sources The mirror 317 can be switched. Hg-Xe lamp L ₁ has a line spectrum, are capable of high brightness lighting in a short wavelength range, Xe lamp L ₂ is capable white light illumination. That is, in the case where annular illumination including circular illumination is performed by the annular illumination filter (secondary illumination light source: disk-shaped mask: aperture stop) 5, the Hg-Xe lamp L ₁ is used. it can be illuminated by switching between the white light illumination with high brightness illumination with short-wavelength range using a Xe lamp L _2. Reference numeral 318 denotes an integrator that equalizes the intensity of light emitted from the Hg—Xe lamp L ₁ or the Xe lamp L ₂ . Reference numeral 341 denotes an illuminance monitor that monitors illuminance fluctuations in the primary light sources L ₁ and L ₂ . The light quantity control filter 14 is controlled according to the fluctuation of the illuminance in the primary light sources L ₁ and L ₂ monitored by the illuminance monitor 341, and the conversion level in the A / D converter 15 is corrected. Reference numeral 316 denotes a wavelength selection filter that selects the wavelength of the annular illumination light to, for example, 600 nm or less. A field stop 319 shields light other than the annular illumination, including circular illumination. Reference numeral 301 denotes a mirror.

  The dichroic mirror 320 provided at the tip of the second objective lens 303 transmits light of diffracted light obtained from the object 1 to be inspected by ring-shaped illumination having a wavelength of 600 nm or less, and has a wavelength of 780 to 800 nm described later. The scattered light obtained from the object 1 to be inspected is reflected by dark field illumination with a wavelength. Reference numeral 321 denotes a half mirror. Reference numerals 325 and 326 denote lenses. A spatial image by scattered light generated from the pattern on the inspection object 1 is formed at the position of the spatial filter 309 by dark field illumination with a wavelength of 780 to 800 nm. Reference numerals 322, 324 and 327 are mirrors. Reference numeral 323 denotes a half mirror.

Further, so as to be able to detect a foreign substance with high sensitivity, a semiconductor laser light source L ₃ the inspected object by focusing the laser beam emitted by (LSI wafer) dark field illumination optical system for illuminating obliquely to 1 (304, 305, 306, 307). 304 is a beam expander for expanding the beam diameter of the laser light emitted by the semiconductor laser light source L ₃ (beam expanding optics). Reference numerals 305 and 306 denote mirrors that reflect laser light. Reference numeral 307 denotes a condensing lens that condenses the laser beam having an enlarged beam diameter and irradiates the inspection object 1 from an oblique direction. The 0th-order diffracted light (regular reflection light) generated from the inspection object 1 by darkfield illumination by the darkfield illumination optical system (304, 305, 306, 307) does not enter the pupil 10a of the objective lens 9 and is inspected. Only scattered light (first-order or higher-order diffracted light) generated from a foreign object existing on the object 1 is incident on the pupil 10a of the objective lens 9, and is received by the image sensor 308 to output a signal to detect the foreign object. can do. Reference numeral 309 denotes a spatial filter that blocks and erases scattered light (first-order or higher-order diffracted light) that is generated from the edge of the pattern on the inspection object 1 and enters the pupil 10a of the objective lens 9 by the dark field illumination. Is. The wavelength of the laser beam emitted from the semiconductor laser light source L ₃ has a wavelength different from any wavelength of the illumination of the annular by the lamp house 24 '(bright field illumination), for example, 780～800Nm.

Further, an automatic focus control optical system is installed in order to detect the pattern on the inspection object 1 with high accuracy as an image signal by the image sensor 12a. The automatic focus control optical system includes a light source 310, a filter 311 having a wavelength of 600 to 700 nm, an automatic focus (A / F) pattern 315, and the A / F pattern 315 on the inspection object 1. constituted by the projection lens 314 for projecting, a half mirror 312 and 313, the sensor _{S 1} which are positioned upstream and downstream of the focus plane, the sensor _{S 2.} Derived from the contrast signal and the sensor S ₂ obtained from the sensor S ₁ detects the contrast signal of the A / F pattern 315 projected on the inspected object 1 by the projection lens 314 by the sensor S ₁ and the sensor S ₂ The surface (pattern surface) of the object 1 to be inspected is moved relative to the objective lens (imaging optical system) 9 by finely controlling the object 1 to be inspected up and down as indicated by arrows so that the contrast signals to be matched. Will be focused. A dichroic mirror 302 provided in the optical path of the detection optical system reflects light for autofocusing with a wavelength of 600 to 700 nm, and light for annular illumination (bright field illumination) with a wavelength of 600 nm or less. And light for dark field illumination having a wavelength of 750 nm or more.

  FIGS. 33 and 34 more specifically show the configuration of the optical system in the pattern inspection apparatus shown in FIGS. 31 and 32. FIG. 33 is a plan view thereof, and FIG. 34 is a front view thereof. That is, since an objective lens 9 having an infinity correction system is used, a second objective lens (also referred to as a tube lens) 303 having a long focal length (for example, f = 200 mm) is necessary. The linear image sensor 12a for annular illumination (bright field illumination) and the linear image sensor 308 for dark field illumination are configured by a TDI (Time Delay & Integration) time delay integration type image sensor. Further, the characteristic part of the present embodiment will be described. That is, in this embodiment, between the objective lens 9 and the second objective lens 303, a polarization beam splitter (PBS (Polarization Beam Splitter)) 8a ′, a λ / 4 plate (¼ wavelength plate) 51, and Is to install. Since the light between the objective lens 9 and the second objective lens 303 is parallel light, even if the optical elements 8a ′ and 51 are inserted, there is little deterioration in aberrations and the like. Functions of the PBS 8a ′ and the λ / 4 plate 51 are as shown in FIGS. That is, the circular illumination light or the annular illumination light 330 passes through the P-polarized light 331 as it is and is reflected by the S-polarized light 332 and reaches the λ / 4 plate 51 by the PBS 8a ′. The S-polarized light 332 reaching the λ / 4 plate 51 has a component whose phase is delayed by 90 degrees (the refractive index of the extraordinary ray and the ordinary ray is not equal, and the extraordinary ray has a longer optical path length. A phase difference Π / 2 is generated between the extraordinary ray and the ordinary ray, and the amplitudes thereof are equal), which is converted into circularly polarized light or elliptically polarized light 334 and irradiated onto the wafer which is the inspection object 1 through the objective lens 9. Diffracted light (reflected light) incident on the pupil 10 a of the objective lens 9 reaches the λ / 4 plate 51 again, and circularly polarized light or elliptically polarized light becomes P polarized light 333. Since the P-polarized light 333 passes through the PBS 8a ′, the P-polarized light 333 passes through the second objective lens 303 and reaches the image sensor 12a as a detector. That is, in FIG. 36, the incident light beam (S-polarized light) 332 converted into linearly polarized light by the PBS 8a ′ is at an angle ω with respect to the quarter wave plate 51 (the linear polarization plane of the incident light beam and Is exactly 45 ° (+ or −), it indicates that linearly polarized incident light 332 can be converted to circularly polarized light 334 (or vice versa). When the angle ω is other than 45 °, the linearly polarized light is converted to elliptically polarized light 334 (or vice versa).

  Next, FIG. 37 and FIG. 38 show the effect of this embodiment. The object 1 to be inspected is a 256 MDRAM line and space pattern (a high-density grid pattern with a pitch P of 0.61 μm). FIG. 37 shows the brightness (detection intensity) from the pattern received by the image sensor 12a with respect to the rotation direction of the pattern. The circularly polarized light / annular illumination (including elliptically polarized light / annular illumination) in the figure is based on the embodiment shown in FIGS. The linearly polarized illumination in the figure corresponds to illumination without the λ / 4 plate 51. Moreover, the half mirror in the figure indicates illumination using a normal half mirror instead of PBS 8a ′ (illumination using the λ / 4 plate 51). From the relationship shown in FIG. 37, even if the high-density pattern on the object 1 to be inspected has various rotation directions such as a memory cell pattern as shown in FIG. If illumination (including elliptically polarized light and annular illumination) is applied, an image signal having high brightness (detection intensity) can be obtained from the image sensor 12a without being greatly affected by the pattern directionality. In addition, by applying ring-shaped illumination, both linearly polarized illumination (S-polarized illumination) and half-mirror illumination (illumination using the λ / 4 plate 51) can be inspected compared to normal circular illumination alone. An image signal having high brightness (detection intensity) can be obtained with respect to a pattern formed on the object 1 having a specific direction such as a peripheral circuit portion. It can also be seen that half mirror illumination (illumination using the λ / 4 plate 51) is superior to linear polarization illumination (S polarization illumination). As described above, the fact that an image signal having high brightness (detection intensity) can be obtained from the image sensor 12a indicates that high-efficiency illumination can be realized for a high-density pattern.

  FIG. 28 shows the contrast (the ratio between the minimum value and the maximum value indicating the resolution) from the pattern received by the image sensor 12a with respect to the rotation direction of the pattern. The circularly polarized light / annular illumination (including elliptically polarized light / annular illumination) in the figure is based on the embodiment shown in FIGS. The circularly polarized light in the figure is only circularly polarized illumination, and the linearly polarized illumination corresponds to S-polarized illumination without the λ / 4 plate 51. In the case of circularly polarized light or annular illumination, the contrast does not have a constant value with respect to the pattern angle because it is not completely circularly polarized during the experiment but is elliptically polarized. is there. In order to reduce the ellipticity and make it a perfect circle, linearly polarized incident light is introduced into the optical element (λ / 4 plate 51) (the electric vector oscillation direction is aligned parallel or perpendicular to the incident surface), and the object to be inspected It may be made circularly polarized using a phase plate before irradiating 1.

  From the relationship shown in FIG. 38, even if the high-density pattern on the object 1 to be inspected has various rotation directions such as a memory cell pattern as shown in FIG. If illumination (including elliptically polarized light and ring-shaped illumination) is applied, an image signal having high contrast (high resolution) can be obtained from the image sensor 12a without being greatly affected by the directionality of the high-density pattern. . That is, by using the circularly polarized illumination and the annular illumination in combination, an image signal having a high contrast can be obtained from the image sensor 12a at all times without depending on the direction of the high-density pattern. Fine defects and the like on a high density pattern can be detected. Further, even when only circularly polarized illumination is combined with ordinary circular illumination (circularly polarized illumination only), the direction of a high-density pattern from the image sensor 12a is not significantly influenced as compared with ordinary circular illumination. An image signal having high contrast (high resolution) can be obtained. However, by using the circularly polarized illumination and the annular illumination in combination, an image signal having a higher contrast (high resolution) than that of the circularly polarized illumination alone can be obtained.

  Further, by applying ring-shaped illumination, even with linearly polarized illumination (S-polarized illumination), a high-density pattern formed on the object 1 to be inspected has a specific direction such as a peripheral circuit portion. In contrast, an image signal having high contrast (high resolution) can be obtained. However, in a normal VLSI (VLSI, ULSI) or the like, there are vertical and horizontal high-density patterns. Therefore, it is difficult to always irradiate linearly polarized light according to the direction of the high-density pattern. If the specified wiring pattern has the specified directionality, the linearly polarized illumination is controlled by adjusting the polarization state of the linearly polarized illumination to match the direction of the wiring pattern. By illuminating only the wiring pattern, it is possible to detect an image signal having high contrast from the image sensor 12a.

  In the above embodiment, the PBS 8a ′ has been described, but the same effect can be obtained even if it is constituted by a half mirror coated with a dielectric multilayer film. In addition, linearly polarized light may be obtained by using a polarizing plate instead of PBS 8a ′. In this case, light passing through the polarizing plate is attenuated and brightness (detection intensity) is reduced, but the contrast is PBS 8a ′. It will improve as well.

[Seventh embodiment]
In this embodiment, in the embodiment shown in FIGS. 31 to 34, a diffusion plate is inserted at the position of the field stop 319 or the annular illumination filter (aperture stop) 5 (position conjugate with the pupil 10a of the objective lens 9). It is a thing. The diffuser diffuses light. This diffusion plate is designated by sand number 800 or the like. This increases the diffusibility of the illumination light by using only the annular illumination light to the object to be inspected 1 or the combined use of the polarized illumination and the annular illumination, and minute irregularities on the surface of the metal wiring pattern or the like. Even if there is a change in brightness, bright uniform reflected light is obtained, and the surface of the metal wiring pattern or the like is made uniform brightness by the image sensor 12a or the TV camera TV ₁ for bright field observation through the objective lens 9 or the like. It can be detected or observed as an image. This diffuse illumination can be performed simultaneously with the same optical system without contradiction with annular illumination and polarized illumination. The degree of diffusion is selected according to the pattern on the inspection object 1.

  In the embodiment shown in FIGS. 31 to 34, the image sensor 12a is composed of a TDI (Time Delay & Integration) time delay integration type image sensor. If the brightness is low and the brightness (detection intensity) is not sufficient, control may be performed to increase the accumulation time of the image sensor. In this way, the accumulation time of the image sensor may be appropriately determined according to the pattern of the inspection object 1. Further, the accumulation time of the image sensor may be determined according to the illumination condition for the pattern of the inspection object 1.

[Eighth embodiment]
Next, the defect determination output 18 output from the comparison circuit 18 of the apparatus shown in FIG. 1 and the defect information 40 output from the CPU 20 are input to analyze, for example, the cause of the defect in the semiconductor manufacturing process. The production of non-defective semiconductor chips at a high yield by eliminating the cause of the occurrence of defects will be described with reference to FIG. That is, reference numeral 380 denotes a semiconductor production line. Reference numeral 381 denotes a transfer path of the semiconductor wafer 1a. Reference numeral 382 denotes a CVD apparatus for executing a CVD film forming process for forming an insulating film in the semiconductor manufacturing process. Reference numeral 383 denotes a sputtering apparatus that executes a sputtering process for forming a wiring film in the semiconductor manufacturing process. Reference numeral 384 denotes an exposure apparatus that performs an exposure process for performing resist coating, exposure, development, and the like in the semiconductor manufacturing process. Reference numeral 385 denotes an etching apparatus for performing an etching process for patterning in the semiconductor manufacturing process. Thus, the semiconductor wafer is manufactured through various manufacturing processes.

  By the way, 390 receives the defect determination output 18 output from the comparison circuit 18 of the apparatus shown in FIG. 1 and the defect information 40 output from the CPU 20 to input each of the process apparatuses 382, 383, 384, This is a computer for analysis that analyzes the cause of defect occurrence or the cause of defect occurrence in the production line 380 made of 385. The analysis computer 390 includes an interface 391 for inputting the defect determination output 18 output from the comparison circuit 18 of the apparatus shown in FIG. 1 and the defect information 40 output from the CPU 20, a CPU 392 for executing processing such as analysis, analysis, etc. 1, a control circuit 394, 395, 396, 397, an output device 398 such as a printing device that outputs an analysis result of the cause of a defect, a display device 399 that displays various data, and the device shown in FIG. For example, an input device (consisting of a keyboard, a disk, etc.) 401 for inputting data relating to each process apparatus 382, 383, 384, 385, data relating to the semiconductor wafer 1a to be sent to the production line 380, etc., semiconductor wafer 1a Defects generated above and the respective process devices 382, 383, 38 385 related to a defect occurrence cause or a defect occurrence factor analyzed by the CPU 392, an external storage device 402 storing a history data or a database of the cause of the defect or a cause-and-effect relationship with the defect occurrence factor. It comprises an interface 403 that provides information 410 to each process device 382, 383, 384, 385, and a bus line 400 that connects them. Therefore, the CPU 392 in the analysis computer 390 receives the input defect determination output 18 and the defect information 40, the defects generated on the semiconductor wafer 1a stored in the external storage device 402, and the respective process devices 382, 383, 384, and 385. The defect in the production line 380 including the process devices 382, 383, 384, and 385 is determined based on the history data or database of the cause of the occurrence of the defect or the cause-and-effect relationship with the defect occurrence factor. The generated defect occurrence cause or defect occurrence factor is analyzed, and information 410 relating to the analyzed defect occurrence cause or defect occurrence factor is provided to each of the process devices 382, 383, 384 and 385. Each process apparatus 382, 383, 384, 385 provided with information 410 regarding the cause of the defect or the cause of the defect is a non-defective product by removing the cause of the defect or the cause of the defect by controlling various process conditions including cleaning. The semiconductor wafer 1a can be sent to the next process, and as a result, the semiconductor can be manufactured with a high yield. 1 is sampled in units of the semiconductor wafer 1a or lots in the manufacturing line 380 from before and after processes where defects are likely to occur.

  In addition, the CPU 392 in the analysis computer 390 receives foreign matter information obtained and input from the CPU 20 based on the foreign matter signal detected from the image sensor 308, and foreign matter generated on the semiconductor wafer 1 a stored in the external storage device 402. Each process device 382, 383, 384 is based on the foreign matter generation cause or the cause-and-effect relationship history data or database that generated the foreign matter in the production line 380 comprising the process devices 382, 383, 384, 385. 385, the cause of the occurrence of foreign matter or the cause of the occurrence of foreign matter in the production line 380 is analyzed, and the analyzed information 410 regarding the cause of foreign matter occurrence or the cause of foreign matter generation is sent to each process device 382, 383, 384, 385. provide. Each process apparatus 382, 383, 384, 385 provided with the information 410 regarding the cause of foreign matter generation or the cause of foreign matter generation is defective by controlling various process conditions including cleaning to remove the cause of foreign matter generation or the cause of foreign matter generation. A good semiconductor wafer 1a having no defect can be sent to the next process, and as a result, a semiconductor can be manufactured at a high yield.

It is a block diagram which shows one Example of the defect inspection apparatus of the pattern on the to-be-inspected target concerning this invention. FIG. 2 is a schematic explanatory diagram of a disk-like mask (annular secondary light source) in the embodiment shown in FIG. 1. FIG. 3 is a schematic explanatory diagram specifically showing mask elements in the disc-like mask shown in FIG. 2. FIG. 3 is a schematic explanatory view showing a mask element in the disk-like mask shown in FIG. 2 more specifically. It is explanatory drawing for showing the specific dimension of the mask element shown in FIG. It is explanatory drawing which shows the grid | lattice-like pattern repeated in the X-axis direction which is one Example of an LSI wafer pattern. The objective lens showing zero-order diffracted light and first-order diffracted light that are generated in the X-axis direction from the lattice-shaped pattern and incident on the pupil of the objective lens by irradiating the lattice-shaped pattern shown in FIG. It is XY top view on a pupil. It is XZ sectional drawing which shows the 0th-order diffracted light and 1st-order diffracted light which generate | occur | produce from the annular illumination shown in FIG. 7, and a grid | lattice-like pattern, and inject on the pupil of an objective lens. The objective lens showing zero-order diffracted light and first-order diffracted light generated in the Y-axis direction from the lattice-like pattern and incident on the pupil of the objective lens by irradiating the lattice-shaped pattern shown in FIG. It is XY top view on a pupil. A zero-order diffracted light and a first-order diffracted light that are generated in the X-axis method and the Y-axis direction from the lattice-shaped pattern and incident on the pupil of the objective lens by applying different annular illuminations to the lattice-shaped pattern shown in FIG. It is XY top view on the pupil of the objective lens which shows these. It is explanatory drawing which shows the relationship between (sigma) value and incident angle (PSI) in an objective lens. It is explanatory drawing which shows the relationship between incident angle (PSI) and diffraction angle (theta). Generation of + 1st order diffracted light when ring-shaped illumination is applied to a lattice pattern of P = 0.61 μm with a wavelength of λ = 0.4 to 0.6 μm, σ of 0.60 to 0.40 It is a diagram which shows a condition. Generation of + 1st order diffracted light when annular illumination is applied to a lattice pattern of P = 0.7 μm with a wavelength of λ = 0.4 to 0.6 μm, σ of 0.60 to 0.40 It is a diagram which shows a condition. It is explanatory drawing which shows the grid | lattice-like pattern repeated in the Y-axis direction which is one Example of an LSI wafer pattern. An objective lens showing zero-order diffracted light and first-order diffracted light generated in the Y-axis direction from the lattice-shaped pattern and incident on the pupil of the objective lens by irradiating the lattice-shaped pattern shown in FIG. It is XY top view on a pupil. It is a figure which shows the YZ cross section which shows the state which attenuates 0th-order diffracted light with an attenuation | damping filter (filter for light quantity control). FIG. 17 is an XY plan view on the pupil conjugate with the pupil of the objective lens showing the state shown in FIG. is there. It is a figure which shows the cross-sectional shape of the attenuation | damping filter (filter for light quantity control) when the transmittance | permeability is set to about 0, and the transmittance | permeability characteristic. It is a figure which shows the cross-sectional shape of the attenuation | damping filter (filter for light quantity control) when the transmittance | permeability is set to about 0.2, and the transmittance | permeability characteristic. In an annular illumination according to the present invention, it is a diagram showing an embodiment in which the light house is controlled in the optical axis direction with respect to the collimator lens. In an annular illumination according to the present invention, it is a diagram showing an embodiment of controlling the collimator lens in the optical axis direction with respect to the light house. It is a block diagram which shows one Example of the microscope system which concerns on this invention. It is a figure which shows the various defects in the wafer pattern which concerns on this invention. It is explanatory drawing which shows the relationship with the detection pixel dimension in a certain location on the wafer pattern shown in FIG. It is a figure which shows the image signal waveform corresponding to the pattern in the certain location shown in FIG. 25, and the brightness which represented this pattern faithfully. It is a figure which shows the image signal corresponding to the sampled brightness obtained by sampling with respect to the image signal corresponding to the brightness shown in FIG. It is a figure which shows the image signal waveform corresponding to the brightness obtained faithfully when a detection pixel dimension is set to 0.0175 micrometer with respect to the grid-like repetitive pattern which consists of a 0.42 micrometer-width line and space. It is a figure which shows the image signal waveform corresponding to the brightness | luminance with which the detection pixel dimension was set to 0.14 micrometer and the extreme value was preserve | saved with respect to the grid | lattice-like repeating pattern which consists of a 0.42 micrometer width line and space. It is a figure which shows the image signal waveform corresponding to the brightness which does not preserve | save an extreme value by making a detection pixel dimension into 0.28 micrometer with respect to the repetitive pattern of the grid | lattice which consists of a 0.42 micrometer width line and space. It is the top view which showed the optical system in one Example of the defect inspection apparatus of the pattern on the to-be-inspected object based on this invention shown in FIG. FIG. 32 is a front view of FIG. 31. It is the top view which showed the Example shown in FIG. 31 more concretely. It is a front view of FIG. It is a figure which shows one Example of the optical system which performs circularly polarized illumination. It is a figure for demonstrating conversion from the linearly polarized light to a circular or elliptically polarized light by a quarter wavelength plate. It is an experiment example which shows the detection intensity corresponding to the brightness with respect to the pattern angle in the experiment example which controlled the polarization state in the annular illumination. It is a figure which shows the contrast with respect to the pattern angle in the experiment example which controlled the polarization state in annular illumination. Using the analyzing computer according to the present invention, the cause of the defect or the cause of the defect is analyzed, and the analyzed cause of the defect or the cause of the defect is fed back to various process devices in the production line to manufacture a semiconductor substrate at a high yield. It is a figure for demonstrating what to do.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... LSI wafer (to-be-inspected object), 2 ... XYZ stage 3 ... Xe lamp, 4 ... Condensing elliptical mirror 5, 5a, 5b ... Disc mask (annular illumination filter)
6 ... Collimator lens, 7 ... Condenser lens 8a, 8b ... Half mirror, 8'a ... Polarizing beam splitter (PBS)
8'a, 8'b, 8'c ... Dichroic mirror 9 ... Objective lens, 10a ... Pupil of objective lens, 10b ... Pupil conjugate to 10a 11 ... Converging lens, 12, 12a, 12b, 308 ... Image sensor 13 ... Zoom lens, 14 ... Filters 15a, 15b for adjusting the light quantity, A / D converter, 16 ... Delay memory, 17 ... Comparison circuit 18 ... Defect output, 19 ... Moving mechanism, 20 ... CPU
DESCRIPTION OF SYMBOLS 21 ... Edge detector 24, 24 '... Lamp house 25 ... Driver 26 ... Controller 27a, 27b ... Monitor 38 ... Attenuation filter (filter for light quantity control) 39 ... Moving mechanism 51 ... λ / 4 plate, 302, 320, 325 ... Dichroic mirror 303 ... Second objective lens, 309 ... Spatial filter 315 ... Automatic focusing pattern, 316 ... Wavelength selection filter 317 ... Switching mirror, 319 ... Field stop, TV ₁ ... Television for image observation bright field Camera TV ₂ ... TV camera for image observation dark field TV ₃ (12b) ... TV camera for pupil observation bright field (image sensor)
TV ₄ ... TV camera for pupil observation dark field L ₁ ... Hg-Xe lamp light source, L ₂ ... Xe lamp, L ₃ ... Semiconductor laser L ₃ , 304 to 307 ... Dark field illumination optical system S ₁ for detecting foreign matter S ₁ ... sensor, S ₂ ... Sensor.

Claims (38)

  A ring-shaped diffused illumination light formed from a large number of virtual point light sources is condensed and applied to the pattern on the object to be inspected through the pupil of the objective lens. The pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including 0th-order diffracted light that is reflected from the pattern on the object to be inspected by light and enters the pupil of the objective lens. A method for detecting a pattern on an object to be inspected, wherein the image is received by an image sensor and converted into an image signal of a pattern, and a pattern on the object to be inspected is detected based on the image signal of the pattern.   A ring-shaped diffused illumination light formed from a large number of virtual point light sources is condensed and applied to the pattern on the object to be inspected through the pupil of the objective lens. The pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including 0th-order diffracted light that is reflected from the pattern on the object to be inspected by light and enters the pupil of the objective lens. The image is received by the image sensor and converted into a pattern image signal. The image signal of this pattern is compared with the image signal of the reference pattern, the pattern on the object to be inspected is erased by coincidence, and the defect is detected by mismatch. A defect inspection method for a pattern on an object to be inspected.   A ring-shaped diffused illumination light formed from a large number of virtual point light sources is condensed and applied to the pattern on the object to be inspected through the pupil of the objective lens. First-order or second-order diffracted light including zero-order diffracted light reflected from a pattern on the object to be inspected by light and entering the pupil of the objective lens is obtained through a light amount control filter that partially changes the intensity or light amount. A pattern image on the object to be inspected obtained by condensing the first-order or second-order diffracted light including the 0th-order diffracted light is received by an image sensor and converted into a pattern image signal, and the converted pattern A method for detecting a pattern on an object to be inspected, comprising detecting a pattern on the object to be inspected based on the image signal.   A ring-shaped diffused illumination light formed from a large number of virtual point light sources is condensed and applied to the pattern on the object to be inspected through the pupil of the objective lens. Including 0th-order diffracted light that has entered the pupil of the objective lens through a light amount control filter that controls the amount of light of a part of the 0th-order diffracted light that is reflected from the pattern on the object to be inspected and incident on the pupil of the objective lens On the object to be inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light obtained through a light amount control filter that partially changes the intensity or light amount of the first-order or second-order diffracted light The pattern image is received by the image sensor and converted into a pattern image signal, and the converted pattern image signal is compared with the reference pattern image signal to match the pattern on the object to be inspected. Pattern defect inspection method on the object to be inspected, characterized in that to erase the down detecting defects by mismatch.   A ring-shaped diffused illumination light formed from a large number of virtual point light sources is condensed and applied to the pattern on the object to be inspected through the pupil of the objective lens, and the image on the pupil of the objective lens is secondly reflected. The zonal diffused illumination light in the light source is controlled based on the pupil image signal obtained by receiving and converting by the image sensor, and the object is inspected by the condensed radiant diffused illumination light. An image of a pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light that is reflected from the pattern and incident on the pupil of the objective lens is a first image sensor. And detecting the pattern on the object to be inspected based on the image signal of the converted pattern, and detecting the pattern on the object to be inspected.   A zero-order incident on the objective lens pupil by irradiating the annular diffuse illumination light formed from a large number of virtual point light sources to the pattern on the object to be inspected through the pupil of the objective lens. The diffusing light including the diffracted light is controlled by receiving an image signal of a pupil obtained by receiving and converting the image of the locality distribution of the diffracted light by the second image sensor, and the condensed irradiation The object obtained by condensing first-order or second-order diffracted light including zero-order diffracted light that is reflected from the pattern on the object to be inspected by the ring-shaped diffuse illumination light and enters the pupil of the objective lens The pattern image on the inspection object is received by the first image sensor and converted into a pattern image signal, and the pattern on the inspection object is detected based on the converted pattern image signal. Inspected vs. Detection method of Butsujo of pattern.   A ring-shaped diffused illumination light formed from a large number of virtual point light sources is condensed and applied to the pattern on the object to be inspected through the pupil of the objective lens, and the image on the pupil of the objective lens is secondly reflected. The zonal diffused illumination light in the light source is controlled based on the pupil image signal obtained by receiving and converting by the image sensor, and the object is inspected by the condensed radiant diffused illumination light. An image of a pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light that is reflected from the pattern and incident on the pupil of the objective lens is a first image sensor. Is received and converted into a pattern image signal, the pattern image signal obtained from the first image sensor is compared with the image signal of the reference pattern, and the pattern on the object to be inspected is erased by matching. Pattern defect inspection method on the object to be inspected, characterized in that detected more defects.   A zero-order incident on the objective lens pupil by irradiating the annular diffuse illumination light formed from a large number of virtual point light sources to the pattern on the object to be inspected through the pupil of the objective lens. The diffusing light including the diffracted light is controlled by receiving an image signal of a pupil obtained by receiving and converting the image of the locality distribution of the diffracted light by the second image sensor, and the condensed irradiation The object obtained by condensing first-order or second-order diffracted light including zero-order diffracted light that is reflected from the pattern on the object to be inspected by the ring-shaped diffuse illumination light and enters the pupil of the objective lens The pattern image on the inspection object is received by the first image sensor and converted into a pattern image signal, and the pattern image signal obtained from the first image sensor is compared with the image signal of the reference pattern. By match Pattern defect inspection method on the object to be inspected, characterized in that to detect a defect by mismatch erase the pattern on the inspected object.   A ring-shaped diffused illumination light formed from a large number of virtual point light sources is condensed and applied to the pattern on the object to be inspected through the pupil of the objective lens, and the image on the pupil of the objective lens is secondly reflected. The intensity or amount of light is controlled by the light amount control filter in the pattern detection optical system based on the image signal of the pupil received and converted by the image sensor. The first-order or second-order diffracted light including the 0th-order diffracted light that is reflected from the pattern on the object to be inspected and enters the pupil of the objective lens is obtained through a light amount control filter that partially changes the intensity or the light amount. A pattern image on the object to be inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light is received by an image sensor and converted into a pattern image signal. Pattern detecting method of the pattern on the inspected object, characterized by detecting a pattern on the inspected object based on the image signal.   A ring-shaped diffused illumination light formed from a large number of virtual point light sources is condensed and applied to the pattern on the object to be inspected through the pupil of the objective lens, and the image on the pupil of the objective lens is secondly reflected. The intensity or amount of light is controlled by the light amount control filter in the pattern detection optical system based on the image signal of the pupil received and converted by the image sensor. 1 includes 0th-order diffracted light that has entered the pupil of the objective lens through a light amount control filter that controls the amount of light of a part of the 0th-order diffracted light that is reflected from the pattern on the object to be inspected and is incident on the pupil of the objective lens. Inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light obtained through a light amount control filter that partially changes the intensity or light amount of second-order or second-order diffracted light The pattern image on the figurine is received by the first image sensor and converted into a pattern image signal, and the pattern image signal obtained from the first image sensor is compared with the image signal of the reference pattern to match. A method for inspecting a defect on a pattern to be inspected by erasing a pattern on the object to be inspected and detecting a defect by mismatch.   Polarized / annular diffuse illumination light, which is formed by further applying polarized light to the annular diffuse illumination light formed from a number of virtual point light sources, is collected on the pattern on the object to be inspected through the pupil of the objective lens. The first-order or second-order light including the 0th-order diffracted light that is reflected from the pattern on the object to be inspected and incident on the pupil of the objective lens by the condensed and irradiated annular-shaped diffuse illumination light. The image of the pattern on the inspection object obtained by condensing the diffracted light is received by the image sensor and converted into an image signal of the pattern, and the pattern on the inspection object is detected based on the image signal of this pattern A method for detecting a pattern on an object to be inspected.   12. The method for detecting a pattern on an object to be inspected according to claim 11, wherein the polarized light is circular or elliptically polarized light.   Polarized / annular diffuse illumination light, which is formed by further applying polarized light to the annular diffuse illumination light formed from a number of virtual point light sources, is collected on the pattern on the object to be inspected through the pupil of the objective lens. The first-order or second-order light including the 0th-order diffracted light that is reflected from the pattern on the object to be inspected and incident on the pupil of the objective lens by the condensed and irradiated annular-shaped diffuse illumination light. The pattern image on the object to be inspected obtained by condensing the diffracted light is received by the image sensor and converted into a pattern image signal, and the pattern image signal is compared with the image signal of the reference pattern to match A method for inspecting a defect on a pattern to be inspected by erasing a pattern on the object to be inspected and detecting a defect by mismatch.   14. The defect inspection method for a pattern on an inspection object according to claim 13, wherein the polarized light is circular or elliptically polarized light. A ring-shaped diffused illumination light formed from a large number of virtual point light sources is condensed and applied to the pattern on the object to be inspected through the pupil of the objective lens. The pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including 0th-order diffracted light that is reflected from the pattern on the object to be inspected by light and enters the pupil of the objective lens. The image is received by an image sensor and converted into a pattern image signal. The image signal of this pattern is compared with the image signal of the reference pattern, the pattern on the object to be inspected is erased by coincidence, and defect information is obtained by mismatch. Detect
Inspected object obtained by applying dark field illumination to the pattern on the object to be inspected, and collecting scattered light that is reflected from the foreign matter on the pattern on the semiconductor substrate and enters the pupil of the objective lens A defect inspection method for a pattern on an object to be inspected, wherein foreign matter on a pattern on an object is received by an image sensor and converted into a signal indicating the foreign matter to detect foreign matter information on the pattern.   A light source that emits an annular diffuse illumination light formed from a large number of virtual point light sources, and an annular diffuse illumination light emitted from the light source through a pupil of an objective lens to a pattern on the object to be inspected An illumination optical system that condenses and irradiates, and a zero-next-order incident on the pupil of the objective lens after being reflected from the pattern on the object to be inspected by the annular diffuse illumination light that is condensed and irradiated by the illumination optical system A detection optical system that receives an image of a pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including folding light, and converts the image into a pattern image signal. An apparatus for detecting a pattern on an object to be inspected.   A light source that emits an annular diffuse illumination light formed from a large number of virtual point light sources, and an annular diffuse illumination light emitted from the light source through a pupil of an objective lens to a pattern on the object to be inspected An illumination optical system that condenses and irradiates, and a zero-next-order incident on the pupil of the objective lens after being reflected from the pattern on the object to be inspected by the annular diffuse illumination light that is condensed and irradiated by the illumination optical system A detection optical system that receives an image of a pattern on an object to be inspected obtained by condensing first-order or second-order diffracted light including folding light, and converts the image into a pattern image signal, and the detection optical system Comparing means for comparing the image signal of the pattern obtained from the image sensor with the image signal of the reference pattern, erasing the pattern on the object to be inspected by matching, and detecting a defect by mismatching Defect inspection apparatus of a pattern on the inspected object to.   A light source that emits an annular diffuse illumination light formed from a large number of virtual point light sources, and an annular diffuse illumination light emitted from the light source through a pupil of an objective lens to a pattern on the object to be inspected The illumination optical system that condenses and irradiates, and the next time it is reflected from the pattern on the object to be inspected by the annular diffuse illumination light that is condensed and irradiated by the illumination optical system and enters the pupil of the objective lens An object to be inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light obtained through a light amount control filter that partially changes the intensity or light amount of first-order or second-order diffracted light including folded light An apparatus for detecting a pattern on an object to be inspected, comprising: a detection optical system that receives an image of a pattern on an object by an image sensor and converts the image into a pattern image signal.   A light source that emits an annular diffuse illumination light formed from a large number of virtual point light sources, and an annular diffuse illumination light emitted from the light source through a pupil of an objective lens to a pattern on the object to be inspected The illumination optical system that condenses and irradiates, and the next time it is reflected from the pattern on the object to be inspected by the annular diffuse illumination light that is condensed and irradiated by the illumination optical system and enters the pupil of the objective lens The first-order or second-order diffracted light including the 0th-order diffracted light that has entered the pupil of the objective lens is obtained through a light amount control filter that partially changes the intensity or light amount through a light amount control filter that controls the light amount of a part of the folded light. A detection optical system that receives an image of a pattern on an object to be inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light by an image sensor and converts the image into a pattern image signal; Comparing means for comparing the image signal of the pattern obtained from the image sensor of the detection optical system with the image signal of the reference pattern, erasing the pattern on the object to be inspected by matching, and detecting defects by mismatching A defect inspection apparatus for a pattern on an object to be inspected.   19. The apparatus for detecting a pattern on an object to be inspected according to claim 16, wherein the detection optical system includes an optical magnification varying means.   20. The defect inspection apparatus for a pattern on an object to be inspected according to claim 17 or 19, wherein the detection optical system includes an optical magnification varying means.   A light source that emits an annular diffuse illumination light formed from a large number of virtual point light sources, and an annular diffuse illumination light emitted from the light source through a pupil of an objective lens to a pattern on the object to be inspected An illumination optical system that condenses and irradiates, and a zero-next-order incident on the pupil of the objective lens after being reflected from the pattern on the object to be inspected by the annular diffuse illumination light that is condensed and irradiated by the illumination optical system A pattern detection optical system for receiving an image of a pattern on an object to be inspected obtained by condensing first-order or second-order diffracted light including folded light by a first image sensor and converting it into an image signal of the pattern; A pupil detection optical system that receives an image on the pupil of the objective lens by a second image sensor and converts it into a pupil image signal, and a pupil obtained from the second image sensor of the pupil detection optical system. Before based on image signal Detection device of the pattern on the inspected object, characterized in that a control means for controlling the diffuse illumination light zonal in the light source.   A light source that emits an annular diffuse illumination light formed from a large number of virtual point light sources, and an annular diffuse illumination light emitted from the light source through a pupil of an objective lens to a pattern on the object to be inspected An illumination optical system that condenses and irradiates, and a zero-next-order incident on the pupil of the objective lens after being reflected from the pattern on the object to be inspected by the annular diffuse illumination light that is condensed and irradiated by the illumination optical system A pattern detection optical system for receiving an image of a pattern on an object to be inspected obtained by condensing first-order or second-order diffracted light including folded light by a first image sensor and converting it into an image signal of the pattern; A pupil detection optical system that receives an image of the locality distribution of the diffracted light including the zeroth order incident on the pupil of the objective lens and converts the image into a pupil image signal by the second image sensor; The second image center of the detection optical system Detection device of the pattern on the inspected object, characterized in that a control means for controlling the diffuse illumination light of the annular in the light source based on the pupil image signals obtained from the service.   The apparatus for detecting a pattern on an object to be inspected according to claim 22 or 23, wherein said pattern detection optical system comprises an optical magnification varying means.   A light source that emits an annular diffuse illumination light formed from a large number of virtual point light sources, and an annular diffuse illumination light emitted from the light source through a pupil of an objective lens to a pattern on the object to be inspected An illumination optical system that condenses and irradiates, and a zero-next-order incident on the pupil of the objective lens after being reflected from the pattern on the object to be inspected by the annular diffuse illumination light that is condensed and irradiated by the illumination optical system A pattern detection optical system for receiving an image of a pattern on an object to be inspected obtained by condensing first-order or second-order diffracted light including folded light by a first image sensor and converting it into an image signal of the pattern; A pupil detection optical system that receives an image on the pupil of the objective lens by a second image sensor and converts it into a pupil image signal, and a pupil obtained from the second image sensor of the pupil detection optical system. Before based on image signal The control means for controlling the annular illumination light in the light source and the image signal of the pattern obtained from the first image sensor of the pattern detection optical system are compared with the image signal of the reference pattern to be inspected. Comparing means for erasing a pattern on an object and detecting a defect by mismatching, a defect inspection apparatus for a pattern on an object to be inspected.   A light source that emits an annular diffuse illumination light formed from a large number of virtual point light sources, and an annular diffuse illumination light emitted from the light source through a pupil of an objective lens to a pattern on the object to be inspected An illumination optical system that condenses and irradiates, and a zero-next-order incident on the pupil of the objective lens after being reflected from the pattern on the object to be inspected by the annular diffuse illumination light that is condensed and irradiated by the illumination optical system A pattern detection optical system for receiving an image of a pattern on an object to be inspected obtained by condensing first-order or second-order diffracted light including folded light by a first image sensor and converting it into an image signal of the pattern; A pupil detection optical system that receives an image of the locality distribution of the diffracted light including the zeroth order incident on the pupil of the objective lens and converts the image into a pupil image signal by the second image sensor; The second image center of the detection optical system A control means for controlling the annular diffused illumination light in the light source based on an image signal of the pupil obtained from the sensor, and a pattern image signal obtained from the first image sensor of the pattern detection optical system as a reference pattern A defect inspection apparatus for a pattern on an object to be inspected, comprising: comparing means for erasing a pattern on the object to be inspected by coincidence and detecting a defect by inconsistency in comparison with the image signal.   27. The defect inspection apparatus for a pattern on an object to be inspected according to claim 25 or 26, wherein said pattern detection optical system comprises an optical magnification varying means.   A light source that emits an annular diffuse illumination light formed from a large number of virtual point light sources, and an annular diffuse illumination light emitted from the light source through a pupil of an objective lens to a pattern on the object to be inspected The illumination optical system that condenses and irradiates, and the next time it is reflected from the pattern on the object to be inspected by the annular diffuse illumination light that is condensed and irradiated by the illumination optical system and enters the pupil of the objective lens An object to be inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light obtained through a light amount control filter that partially changes the intensity or light amount of first-order or second-order diffracted light including folded light A pattern detection optical system that receives an image of a pattern on an object by an image sensor and converts it into an image signal of the pattern, and an image on the pupil of the objective lens is received by a second image sensor into an image signal of the pupil Strange And a control means for controlling intensity or light quantity by a light quantity control filter in the pattern detection optical system based on a pupil image signal obtained from a second image sensor of the pupil detection optical system. An apparatus for detecting a pattern on an object to be inspected.   A light source that emits an annular diffuse illumination light formed from a large number of virtual point light sources, and an annular diffuse illumination light emitted from the light source through a pupil of an objective lens to a pattern on the object to be inspected The illumination optical system that condenses and irradiates, and the next time it is reflected from the pattern on the object to be inspected by the annular diffuse illumination light that is condensed and irradiated by the illumination optical system and enters the pupil of the objective lens The first-order or second-order diffracted light including the 0th-order diffracted light that has entered the pupil of the objective lens is obtained through a light amount control filter that partially changes the intensity or light amount through a light amount control filter that controls the light amount of a part of the folded light. A pattern in which an image of a pattern on an object to be inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light is received by a first image sensor and converted into a pattern image signal. From a detection optical system, a pupil detection optical system that receives an image on the pupil of the objective lens by a second image sensor and converts it into a pupil image signal, and a second image sensor of the pupil detection optical system An image signal of a pattern obtained from a control means for controlling intensity or light quantity by a light quantity control filter in the pattern detection optical system based on an obtained pupil image signal, and a first image sensor of the pattern detection optical system A pattern defect on the object to be inspected, comprising comparing means for comparing the image signal of the reference pattern with the reference pattern and deleting the pattern on the object to be inspected by coincidence and detecting a defect by inconsistency Inspection device.   A light source that emits a ring-shaped diffuse illumination light formed from a large number of virtual point light sources, and a polarization that is formed by further adding polarization to the ring-shaped diffuse illumination light emitted from the light source by a polarization conversion optical element An illumination optical system that collects and irradiates a ring-shaped diffuse illumination light on a pattern on an object to be inspected through the pupil of the objective lens, and a ring-shaped diffuse illumination light that is condensed and irradiated by the illumination optical system. An image of the pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including 0th-order diffracted light that is reflected from the pattern on the object to be inspected and enters the pupil of the objective lens. An apparatus for detecting a pattern on an object to be inspected, comprising: a detection optical system that receives light by an image sensor and converts the light into a pattern image signal.   27. The apparatus for detecting a pattern on an object to be inspected according to claim 26, wherein the polarization conversion optical element in the illumination optical system is a circle or an elliptical polarization conversion optical element that adds a circular or elliptical polarization as the polarized light. .   A light source that emits a ring-shaped diffuse illumination light formed from a large number of virtual point light sources, and a polarization that is formed by further adding polarization to the ring-shaped diffuse illumination light emitted from the light source by a polarization conversion optical element An illumination optical system that collects and irradiates a ring-shaped diffuse illumination light on a pattern on an object to be inspected through the pupil of the objective lens, and a ring-shaped diffuse illumination light that is condensed and irradiated by the illumination optical system. An image of the pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including 0th-order diffracted light that is reflected from the pattern on the object to be inspected and enters the pupil of the objective lens. A detection optical system that receives light by the image sensor and converts it into a pattern image signal, and compares the pattern image signal obtained from the image sensor of the detection optical system with the image signal of the reference pattern and matches the image signal on the object to be inspected. Defect inspection apparatus of a pattern on the inspected object, characterized by comprising comparison means for detecting a defect by mismatch erase the pattern.   29. A defect inspection of a pattern on an object to be inspected according to claim 28, wherein the polarization conversion optical element in the illumination optical system is a circle or an elliptical polarization conversion optical element that adds a circular or elliptical polarization as the polarization. apparatus.   A first illumination optical system for condensing and irradiating a ring-shaped diffused illumination light formed from a large number of virtual point light sources onto a pattern on an object to be inspected through a pupil of an objective lens; and the object to be inspected A second illumination optical system that performs dark field illumination focused on a pattern on the object, and a ring-shaped diffused illumination light focused and irradiated by the first illumination optical system on the object to be inspected An image of a pattern on the object to be inspected obtained by condensing first-order or second-order diffracted light including zero-order diffracted light that is reflected from the pattern and enters the pupil of the objective lens is received by an image sensor. A first detection optical system that converts the image signal into a pattern, and scattered light that is reflected from the pattern on the inspection object illuminated by the second illumination optical system and enters the pupil of the objective lens Putter on the object to be inspected obtained by focusing A second detection optical system that receives light from the foreign matter on the image sensor and converts it into a signal indicating the foreign matter, and an image signal of a pattern obtained from the image sensor of the first detection optical system is used as a reference pattern Comparison means for erasing the pattern on the object to be inspected by coincidence with the image signal and detecting defect information by inconsistency, and foreign matter for detecting foreign substance information based on the signal obtained from the second detection optical system A defect inspection apparatus for a pattern on an object to be inspected, comprising: a detecting unit.   A light source that emits a ring-shaped diffuse illumination light formed from a large number of virtual point light sources, and a polarization that is formed by further adding polarization to the ring-shaped diffuse illumination light emitted from the light source by a polarization conversion optical element A first illumination optical system that focuses and irradiates a ring-shaped diffused illumination light on a pattern on the object to be inspected through the pupil of the objective lens, and is focused on the pattern on the object to be inspected. A second illumination optical system that performs dark field illumination, and a pupil of the objective lens that is reflected from the pattern on the object to be inspected by the annular diffused illumination light that is condensed and irradiated by the first illumination optical system. A pattern image on the object to be inspected obtained by condensing first-order or second-order diffracted light including 0th-order diffracted light incident on the image is received by an image sensor and converted into a pattern image signal. Detection optical system and second illumination light Image of light from foreign matter on the pattern on the object to be inspected obtained by condensing the scattered light that is reflected from the pattern on the object to be inspected illuminated by the system and enters the pupil of the objective lens A second detection optical system that receives light by the sensor and converts it into a signal indicating a foreign substance, and an image signal of a pattern obtained from the image sensor of the first detection optical system is compared with an image signal of a reference pattern by matching Comparing means for erasing the pattern on the object to be inspected and detecting defect information due to mismatch, and foreign matter detecting means for detecting foreign matter information based on a signal obtained from the second detection optical system A defect inspection apparatus for a pattern on an object to be inspected.   Illumination means for applying illumination light substantially uniformly in the detection visual field of the object to be inspected via the objective lens, image detection means for converting reflected light from the object to be inspected into a detection image by photoelectric conversion, and the detection image A defect inspection apparatus for a pattern on an object to be inspected, comprising image comparison means for comparing the reference image with a reference image.   37. The defect inspection apparatus for a pattern on an object to be inspected according to claim 36, further comprising pupil image detection means for detecting an image on the pupil of the objective lens by forming an image with a lens. 38. The defect inspection apparatus for a pattern on an inspection object according to claim 36 or 37, further comprising a polarization state control means for controlling a polarization state of the illumination light.

Images