JP6895768B2 - Defect inspection equipment and defect inspection method - Google Patents

Description

The present invention relates to a defect inspection device. Further, the present invention is a defect inspection device suitable for defect inspection of a glass substrate, a mask blank (a substrate with a thin film), a mask blank with a resist (a substrate with a resist), and a defect inspection of a large substrate for a display device such as an FPD. Regarding.

Usually, in a defect inspection device, scattered light from a defect on the substrate surface or a defect inside the glass is detected by illuminating from a specific direction with a laser beam or the like.
For example, there is a defect inspection device that irradiates a substrate surface with a laser beam and detects scattered light from a defect on the substrate surface with a detection optical system.
Further, for example, laser light is irradiated (introduced) inside the glass substrate, and the entire inside of the glass is irradiated by repeating total internal reflection inside the glass substrate to emit scattered light from defects on the surface of the glass substrate and defects inside the glass. There is a defect inspection device that detects with a detection optical system.

Japanese Patent No. 3422935

Lighting that is not from a specific direction is essential for high-precision defect inspection. Illumination from a specific direction may cause a significant decrease in defect detection power for scratches that are strongly direction-dependent. In high-precision defect inspection, high defect detection ability for defects such as scratches with strong direction dependence is desired.

In addition, for thin film defects of the type in which the generation of scattered light from defects is small or the generation of scattered light is extremely small, the inspection method for detecting scattered light causes a significant decrease in defect detection power or undetectability. there is a possibility. For example, there is a possibility that a significant decrease in defect detection power may occur for a half pinhole that does not have a clear edge and generates little scattered light. Further, for example, a dent (gradation) on a gently curved surface of a thin film cannot be detected because the generation of scattered light is extremely small. In high-precision defect inspection, high defect detection ability is desired for such thin film defects in which the generation of scattered light from the defects is small or the generation of scattered light is extremely small.

Further, there is a problem of ensuring defect detection ability for defects having low contrast such as thin scratches and white haze (white stin). In high-precision defect inspection, high defect detection ability for these low-contrast defects is desired.

In high-precision defect inspection, there are many defects having different scattered light characteristics as described above.

There is a problem that inspection time is required for high-precision defect inspection. For example, in the detection of weak scattered light, the inspection accuracy can be improved by increasing the exposure time (inspection time) and increasing the sensitivity.
In high-precision defect inspection, a method that does not sacrifice inspection speed (inspection time) is desired.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a defect inspection apparatus capable of achieving stable and high defect detection power regardless of the direction of defects.
Another object of the present invention is to provide a defect inspection apparatus capable of achieving a significant improvement in defect detection ability for thin film defects of a type in which the generation of scattered light from defects is small or the generation of scattered light is extremely small.
Furthermore, an object of the present invention is to provide a defect inspection apparatus capable of achieving high defect detection ability for defects having low contrast.

In order to solve the above-mentioned problems, the present inventor has carried out diligent research and development. As a result, the application of ring lighting was examined. However, although the ring illumination currently provided and the ring illumination simply manufactured by improving it can be illuminated from various directions, high defect detection power can be obtained even when applied to a defect inspection device. However, it was found that the defect detection ability was insufficient and the expected effect could not be obtained.
The present inventor has carried out more diligent research and development. As a result, we have succeeded in developing a ring lighting that is far superior to the ring lighting currently provided. Further, by using this exceptionally excellent ring illumination, it is possible to simultaneously illuminate from all directions of 360 °, and it is possible to detect defects regardless of the direction of defects, and it is currently provided. It was found that the high defect detection power, which is much more stable than the case of using ring lighting, can be exhibited and the expected effect can be obtained.

In addition to the above, in the present invention, both forward scattering and backscattering are performed by double-sided simultaneous darkfield illumination (reflection / transmission composite darkfield ring illumination) including both reflective darkfield ring illumination and transmitted darkfield ring illumination. Can be detected at the same time. Thereby, for example, stable detection can be performed even for a type of defect in which scattered light having a bias in forward scattering or backscattering occurs. Therefore, the defect detection ability is improved as compared with the case where either ring illumination is used.
In addition to this, in the present invention, it is possible to achieve high defect detection ability for low-contrast defects such as thin scratches and white haze (white stin) by the configuration of remarkably excellent ring illumination and double-sided simultaneous dark field.
When comparing the reflected dark-field ring illumination and the transmitted dark-field ring illumination, the latter has higher defect detection power. This is because, according to the scattering theory, the forward scattering and the backscatter are generally stronger than the forward scattering. Light that is incident from the back surface side of the substrate by the transmitted dark field ring illumination and is transmitted through refraction becomes forward scattering.

The present inventor has determined that a dent (recess) in a film that generates extremely little scattered light and a dent (gradation) in a gently curved surface of a thin film cannot be detected in a double-sided simultaneous dark field because the generation of scattered light is extremely small. I found it. The present inventor has found that by using spot illumination from the back surface side, a portion of a film dent (recess) where scattered light is extremely little generated can be detected as a difference in transmittance. Furthermore, the present inventor uses vertical transmittance illumination with good parallelism and high brightness (high illuminance) as spot illumination from the back surface side, so that the dents on the gently curved surface of the thin film, which are particularly difficult to detect, are thin. It was found that the locations appear as differences in transmittance and that stable and high defect detection power can be exhibited for such defects.
In the present invention, in the determination of whether or not there is a defect, the two-dimensional image is differentially processed, and only the portion having a gradient of shading is left and that portion is determined to be a defect. Even if the thin film is dented, the difference in brightness becomes a gradient of light and shade, so it can be detected.
The above is the same for thin film defects of the type in which the generation of scattered light from the defects is small or the generation of scattered light is extremely small. For example, it is possible to achieve a significant improvement in defect detection power for half pinholes that do not have clear edges and generate less scattered light, and dents (recesses) in the film that generate less scattered light.

In the present invention, both the reflective dark-field ring illumination and the transmitted dark-field ring illumination are provided (or a spot illumination from the back surface side is added), and all of these illuminations are used at the same time. It was found that defects caused by (corresponding to) lighting can be detected at one time, and it is possible to efficiently inspect whether or not there are defects in one inspection. In the present invention, different types of defects as described above, which are problematic in high-precision defect inspection, can be detected at one time in one inspection. Therefore, the efficiency is high, and the total inspection time can be shortened as compared with the case where different types of defects are individually detected. In the present invention, it can be said that the inspection is highly accurate in that the types and numbers (opportunities) of detectable defects are relatively large.
The greatest merit of the present invention is that it is possible to inspect for defects at high speed and with high accuracy. It will be possible to put in all the methods that can detect all defects and inspect them.

The present invention has the effect of detecting defects that cannot be detected even over time. For example, scratches with strong direction dependence and dents (gradation) on a gently curved surface of a thin film are particularly good examples.

It should be noted that the auto-focusing technology for maintaining the in-focus state and the lighting technology for satisfactorily detecting defects are different. No matter how good the lighting or its light source is, the performance expected to be out of focus in the image formation cannot be achieved. Therefore, it is most preferable to apply the ring illumination and the spot illumination from the back surface side according to the present invention in addition to the technique of constantly focusing the focused position.
Further, in order to realize high-speed and high-precision inspection, high-precision autofocusing technology is required, and there is coaxial autofocus (AF) as a technology for doing so. Even if the coaxial AF is always focused, the expected effect cannot be obtained unless the lighting is excellent. Therefore, from the ring lighting according to the present invention designed to obtain the expected effect and the back side. Spot lighting is required.

The present invention has the following configurations.
(Structure 1)
It has an optical system for defect inspection and a lighting system for defect inspection.
The illumination system is a ring illumination in which a plurality of LEDs are arranged in an annular shape and spot light from the plurality of LEDs is collected in a region centered on the focal position of the optical system.
In the ring illumination installed on the side of the optical system of the substrate to be inspected, each of the spot lights by the plurality of LEDs is irradiated to the surface of the substrate at a sharp angle, and the reflected light is the objective lens. Reflective darkfield ring illumination configured to prevent direct entry into
In the ring illumination installed on the side of the substrate opposite to the optical system, each of the spot lights by the plurality of LEDs is irradiated at a sharp angle to the back surface of the substrate and transmitted through the inside of the substrate through refraction. A transmitted dark-field ring illumination that prevents transmitted light from directly entering the objective lens, and
A defect inspection device comprising.

(Structure 2)
It has an optical system for defect inspection and a lighting system for defect inspection.
The illumination system is a ring illumination in which a plurality of LEDs are arranged in an annular shape and spot light from the plurality of LEDs is collected in a region centered on the focal position of the optical system.
In the ring illumination installed on the side of the optical system of the substrate to be inspected, each of the spot lights by the plurality of LEDs is irradiated to the surface of the substrate at a sharp angle, and the reflected light is the objective lens. Reflective darkfield ring illumination configured to prevent direct entry into
In the ring illumination installed on the side of the substrate opposite to the optical system, each of the spot lights by the plurality of LEDs is irradiated at a sharp angle to the back surface of the substrate and transmitted through the inside of the substrate through refraction. A transmitted dark-field ring illumination that prevents transmitted light from directly entering the objective lens, and
Illumination installed on the side of the substrate opposite to the optical system, which is a transmission spot illumination coaxial with the optical axis of the optical system.
A defect inspection device comprising.

(Structure 3)
The defect inspection apparatus according to the configuration 1 or 2, wherein the optical axis of the optical system and the central axis of the annulus in the ring illumination are made coaxial with each other.

(Structure 4)
It has an imaging optical system including an objective lens, an imaging lens, and an image sensor.
The image sensor is a TDI sensor.
The imaging optical system includes a TDI camera.
It has a means for relatively moving the substrate and the TDI camera in a certain direction at a constant speed.
By matching the moving direction and speed of the imaging region on the substrate with the direction and speed of charge transfer of the CCD in the TDI sensor, the imaging region is repeatedly exposed and photographed by the number of vertical stages of the CCD. The defect inspection apparatus according to any one of configurations 1 to 3.

(Structure 5)
The defect inspection apparatus according to any one of configurations 1 to 4, wherein the defect inspection apparatus includes a high-precision imaging optical system having a focal likelihood of less than ± 0.1 mm.

(Structure 6)
Using the objective lens and using a laser beam, an autofocus means is constructed and at the same time.
The defect inspection apparatus according to any one of configurations 4 or 5, further comprising means for avoiding reflection of the laser beam used in the autofocus means on the image pickup device.

(Structure 7)
It has an optical system for defect inspection and a lighting system for defect inspection.
The illumination system is a ring illumination in which a plurality of LEDs are arranged in an annular shape and spot light from the plurality of LEDs is collected in a region centered on the focal position of the optical system.
The ring illumination is installed on the side of the optical system of the substrate to be inspected.
The distance from the tip of the ring illumination on the substrate side to the surface of the substrate is defined as the working distance d of the ring illumination.
Let r be the radius of the ring illumination.
From the viewpoint of mechanical size limitation, the range of d and r is defined.
When the light beam of the LED is incident on the substrate, the angle formed by the light beam and the surface of the substrate is set as the irradiation angle α of the ring illumination, and the dark field is requested so that the illumination light does not directly enter the objective lens. From the viewpoint, the range of α is defined,
The range of the p is defined from the viewpoint of increasing the irradiation density p and limiting the upper limit of the irradiation angle α from the dark field request.
A defect characterized in that the values of d, r and α are determined from the mathematical formula d = rTan (α) determined from the positional relationship while considering the irradiation density p, and the apparatus is manufactured based on these values. Manufacturing method of inspection equipment.

(Structure 8)
It has an optical system for defect inspection and a lighting system for defect inspection.
The illumination system is a ring illumination in which a plurality of LEDs are arranged in an annular shape and spot light from the plurality of LEDs is collected in a region centered on the focal position of the optical system.
The ring illumination is installed on the side of the substrate to be inspected opposite to the optical system.
The distance from the tip of the ring illumination to the back surface of the substrate is defined as the working distance d of the ring illumination.
Let r be the radius of the ring illumination.
From the viewpoint of mechanical size limitation, the range of d and r is defined.
When the light ray of the LED is incident on the substrate, the angle formed by the light ray and the back surface of the substrate is set as the irradiation angle α of the ring illumination, and a dark field request for preventing the illumination light from directly entering the objective lens. From the viewpoint and from the viewpoint of requesting the incident near the Brewster angle, the range of α is defined.
From the viewpoint of increasing the irradiation density p by setting the irradiation density by the ring illumination as p, and from the viewpoint that the upper limit of the irradiation angle α is limited by the request for dark field and the request for incident near the Brewster angle, the above p. Define the range,
Let the thickness of the glass be g
A defect inspection device characterized in that the values of d, r and α are determined from the following mathematical formula (3) determined from the positional relationship while considering the irradiation density p, and the device is manufactured based on these values. Manufacturing method.

According to the present invention, it is possible to provide a defect inspection apparatus capable of achieving stable and high defect detection power independent of the direction of defects.
Further, according to the present invention, it is possible to provide a defect inspection apparatus capable of achieving a significant improvement in defect detection ability for thin film defects of a type in which the generation of scattered light from defects is small or the generation of scattered light is extremely small.
Further, according to the present invention, it is possible to provide a defect inspection apparatus capable of achieving high defect detection power for defects having low contrast.

It is a schematic diagram for demonstrating the main part of the defect inspection apparatus of this invention. It is a schematic diagram for demonstrating ring illumination. It is a figure for demonstrating the housing of a ring lighting. It is a figure for demonstrating the relational expression of the working distance d of ring illumination, the radius r of ring illumination, and the irradiation angle α of ring illumination. It is a figure for demonstrating the optimum design of the surface side ring illumination. It is a figure for demonstrating the optimum design of the back side ring illumination. It is a figure for demonstrating the XYZ drive system of the defect inspection apparatus of this invention. It is a schematic diagram for demonstrating the astigmatism method. It is a figure which shows the structure of the coaxial autofocus module. It is a figure for demonstrating means for avoiding the reflection of a laser beam used for an autofocus function on an image sensor.

FIG. 1 is a schematic view for explaining a main part of the defect inspection apparatus of the present invention.
In FIG. 1, the direction perpendicular to the paper surface is the X-axis, the vertical direction of the paper surface is the Y-axis, and the left-right direction of the paper surface is the Z-axis.
An imaging optical system 100 is arranged on the surface side (right side of the drawing) of the object to be inspected 1. The imaging optical system 100 is configured to be driveable in the X-axis, Y-axis, and Z-axis directions by the XYZ driving means.
The imaging optical system 100 includes an objective lens 11, an imaging camera (for example, a TDI camera) 10 including an imaging lens 12, and an autofocus module (coaxial AF module) arranged between the objective lens and the imaging lens. 20 and lighting means 31. The illumination means 31 is attached to the objective lens 11.
An illumination system 101 having an illumination means 32 and a spot illumination means 33 is arranged on the back surface side (left side of the drawing) of the object to be inspected 1. The illumination system 101 is configured to be integrally driveable in the X-axis, Y-axis, and Z-axis directions by the XYZ drive means as an integral part (or completely synchronously and integrally) with the imaging optical system 100.

The lighting means 31 is ring lighting. In the ring illumination, a plurality of LEDs are arranged in an annular shape, and each of the spot light generated by the plurality of LEDs is a region centered on the focal position of the imaging camera (for example, a TDI camera) 10 (image acquisition region, inspection region, imaging region). ), After adjusting the directivity and optical axis of the plurality of LEDs, they are fixed to the annular member. In the ring illumination which is the illumination means 31, the central axis of the ring in the ring illumination and the optical axis O of the imaging optical system 100 are aligned and coaxial.
The ring illumination, which is the illumination means 31, has a configuration in which LEDs are arranged in an annular shape in three layers (three rows) along each of the three concentric circles (see FIG. 2). Blue LEDs are used for the outer and inner LED annulus, and yellow or orange LEDs are used for the middle LED annulus. When inspecting glass substrates and mask blanks (thin film substrates), all three rows (blue LED and yellow LED) are usually used. When inspecting a resisted mask blank, the middle LED annulus (yellow LED only) is used. This is to avoid exposure of the resist.

An illuminating means 32 and a spot illuminating means 33 are provided on the back surface side (left side of the drawing) of the object to be inspected 1.
The lighting means 32 is ring lighting.
The ring illumination, which is the illumination means 32, has a configuration in which LEDs are arranged in an annular shape in three layers (three rows) along each of the three concentric circles (see FIG. 2). Blue LEDs are used for the outer, middle, and inner LED annulus rows. When inspecting glass substrates and mask blanks (thin film substrates), all three rows (blue LEDs) are usually used. In the ring illumination which is the illumination means 32, the central axis of the ring in the ring illumination and the optical axis O of the imaging optical system 100 are aligned and coaxial.
As the spot illumination means 33, transmitted spot illumination (for example, a spotlight of parallel light) is used, and a blue LED is used. The spot illumination means 33 includes pinholes in a mask blank (a substrate with a thin film), half pinholes that do not have clear edges and generate less scattered light, dents (recesses) in a film that generate less scattered light, and scattering. It is effective in detecting dents (graduations) on gently curved surfaces of thin films that generate extremely little light. The spot illumination means 33 aligns the central axis of spot illumination (the optical axis of the LED) with the optical axis O of the imaging optical system 100 and is coaxial.
The spot illuminating means 33 includes a mode in which vertical transmission illumination is performed by an LED light source or a lamp light source having good parallelism and high brightness (high illuminance), a surface emitting light source, and the like.

In the present invention, the LED includes an embodiment in which the LED has a plurality of layers (1 row), 2 layers (2 rows), 4 layers (4 rows) or more in an annular shape.
Further, the present invention includes an embodiment in which the center of the illumination region formed on the substrate by the ring illumination and the central axis of the annulus in the ring illumination do not match (an eccentric illumination by the ring illumination). ..

The working distance d (working distance: WD) of the ring illumination, which is the illumination means 32, can be adjusted according to the plate thickness of the object 1 to be inspected. The working distance is the installation distance of the ring illumination with respect to the substrate surface, and more specifically, the distance from the tip of the ring illumination on the substrate side to the substrate surface.

When inspecting the glass substrate, it is preferable that the lighting means 31 and the lighting means 32 are used, and it is preferable that both of these lights are used at the same time. This is because defects caused by (corresponding to) both lights can be detected at one time, and it is possible to efficiently inspect whether or not there is a defect in one inspection. Inspection of the glass substrate detects optical defects such as scratches, foreign matter, foreign matter inside the glass, and veins.
When inspecting a mask blank (a substrate with a thin film), it is preferable that the lighting means 31, the lighting means 32, and the spot lighting means 33 are used, and it is preferable that all of these lights are used at the same time. This is because defects caused (corresponding) to each of all the lights can be detected at once, and it is possible to efficiently inspect whether or not there are defects in one inspection. In the inspection of the mask blank (board with thin film), pinholes, half pinholes, foreign substances, etc. are detected.
When inspecting the resisted mask blank, the middle LED annulus (yellow LED only) in the illuminating means 31 is used. This is to avoid exposure of the resist. In the inspection of the mask blank with resist, in addition to substrate defects, resist pinholes, resist half pinholes, foreign substances and the like are detected.

Examples of the object 1 to be inspected include a glass substrate, a mask blank (a substrate with a thin film), a mask blank with a resist, and the like. Examples of the mask blank include mask blanks used for manufacturing binary masks, gray tone masks (gradation masks), phase shift masks, and the like.

The inspection of the mask blank includes a mode of inspecting in the state of a single-layer film and a mode of inspecting in the state of a two-layer film or a laminated film having three or more layers. Further, in the case of a two-layer film or a laminated film having three or more layers, the inspection of the mask blank includes an aspect of inspecting each time a film is formed.

The object 1 to be inspected includes a large substrate for a display device such as an FPD and a medium-sized / small-sized substrate.
In the present invention, typical display devices (display devices) such as FPDs (flat panel displays) are liquid crystal display devices, plasma display devices, EL display devices, organic EL display devices, LED display devices, and DMD display devices. Is.

The light-shielding mask 40 is inserted into the optical system as a measure against stray light. The light-shielding mask 40 cuts light (stray light) that does not contribute to the image formation of the defect image. The light-shielding mask 40 may be installed in a size suitable for a position suitable for measures against stray light while changing the size. The light-shielding mask 40 may use a light-shielding plate or a diaphragm.
The light-shielding mask 40 is installed between the objective lens 11 and the imaging lens 12, for example. The light-shielding mask 40 is installed, for example, between the objective lens 11 and the object to be inspected 1, between the object to be inspected 1 and the illuminating means 31, and between the object to be inspected 1 and the illuminating means 32. The light-shielding mask 40 can be installed at all of these locations, at any of these locations, and at any location in the optical path other than the above.

In the present invention, the image sensor 13 is typically a solid-state image sensor such as a CCD, TDI, CMOS, or VMIS.
The gantry 200 is a vibration isolation table. The vibration isolation table has a function of locking the vibration isolation function, and when the robot attaches / detaches the substrate to / from the device, the space position of the device can be fixed by locking the vibration isolation function.

Next, the LED lighting according to the present invention will be described.

In ring lighting, the wavelength of the LED, the spread angle of the LED, the diameter (dimension) of the LED, the irradiation angle of the LED, and the like are designed. These values are determined from the viewpoint of improving the defect detection ability.
The wavelength of the LED is designed, for example, in the visible range. For example, the wavelength of the LED is designed to be blue (465 nm), yellow (592 nm), orange (610 nm), or the like.
The LED spread angle can be selected from, for example, a narrow irradiation angle type (half-value angle ± 5 degrees), a normal type (half-value angle ± 20 degrees), a power LED type (half-value angle ± 60 degrees), and the like.
The diameter (dimensions) of the LED is designed, for example, in the range of several mm. For example, the diameter (size) of the LED is designed to have a narrow angle of 5.0 mm, an ultra-narrow angle of 3.1 mm, or the like.
The irradiation angle of the LED is designed to be in the range of, for example, 10 to 40 degrees.

The ring lighting housing is designed to have an inner diameter R1, an outer diameter R2, a thickness t, and the like (see FIG. 3). These values are determined from the viewpoint of improving the defect detection ability. The housing of the ring illumination is an annular shape, and an inclined portion 35 for mounting the LED is formed on the inner peripheral surface of the annulus. The housing dimension R3 (thickness of the body) is designed in consideration of strength (see FIG. 3).
In the ring lighting housing, chamfering (not shown), shading plate mounting holes (not shown), and the like can be designed. These can be formed, for example, on the surface of the inclined portion 35 on the inspection substrate side.

In ring lighting, the number of optical axes of each LED specified in the design is positioned (direction control is performed) and fixed to the housing. For example, each LED 37 is embedded in a flexible printed circuit board 36 attached to an inclined portion 35 on the inner peripheral surface of the housing 34 (see FIG. 3).
Since the directivity (spread range) of the LED is larger than the positioning accuracy, it is advisable to adjust the positioning accuracy in consideration of this.

In the present invention, the directivity (spread range) of the LED differs depending on the diameter (size) of the LED. The diameter (size) of the LED itself is not a very important parameter, and it is preferable to design with more emphasis on the directivity.
When the irradiation angle of the LED is, for example, an ultra-narrow angle: ± 5.5 ° or ± 4.0 °, it is possible to control so that the light is applied only to the target location.
The diameter (size) of narrow-angle or ultra-narrow-angle type LEDs is, for example, 5.5 mm or 3.8 mm, but if the design is made with more emphasis on directivity, for example, a narrow angle of 5.0 mm or , Ultra-narrow angle 3.1 mm, etc.

For example, when the diameter (size) of the LED is an ultra-narrow angle of 3.1 mm, the LED elements are in the closest arrangement and the pitch b is 4 mm (see FIG. 3). The number of LED elements arranged is approximately 120 elements. This makes it possible to simultaneously illuminate from all directions of 360 °.

In the present invention, by combining the various designs and various adjustments described above, we have succeeded in developing a ring lighting that is far superior to the ring lighting currently provided. This ring lighting is a lighting system capable of supporting not only the next generation but also two generations ahead, three generations ahead, and the next generation after that. It was confirmed that this ring lighting had no problems with heat generation, life, and durability.

In the present invention, the ideal arrangement of dark-field illumination obtained from the light-collecting characteristics of the imaging optical system and the light-emitting characteristics of a plurality of LED elements serving as light sources has been formulated. In the present invention, the optimum design has been succeeded due to the mechanical size limitation and the lighting efficiency.

In the present invention, the limiting range and the inequality are defined from the three requirements (limiting factors) of the mechanical requirement, the dark field requirement (optical requirement), and the efficiency requirement (irradiance density), and the LED design is determined. At this time, the design of the LED is determined within the limit range and the range of the inequality in consideration of ease of manufacture, production cost, and the like. Although each value described below differs depending on the individual machine, an example of each value will be described.

(Surface side: Reflective darkfield ring illumination)
The first point of the mechanical requirement (dimensional limitation) is the working distance d (working distance: WD) of the ring illumination, which is the installation distance of the ring illumination with respect to the substrate surface, and more specifically, the substrate side of the ring illumination. Is the distance from the tip of the substrate to the surface of the substrate. In order to secure the clearance with the substrate holding portion, the working distance d> 15 mm of the ring illumination.
The second point of the mechanical requirement (dimensional limitation) is the outer shape (radius) of the ring illumination. Since the arrangement pitch when three TDI cameras are arranged in contact with each other in succession is set to 130 mm, the outer shape of the ring illumination is limited to 128 mm. This is because when the imaging optical system equipped with the ring illumination is scanned from one end of the substrate to the other for inspection, the outer peripheral portion of the ring illumination is arranged on the substrate holding mechanism (particularly usually on the back surface side) at the end of the substrate. This is a request from the need to keep the external dimensions of the ring illumination within a predetermined range so that collisions and contact with the rotary substrate holding unit) and the frame of the device do not occur. The radius of the ring illumination is r <62 mm.

The dark field request (optical request) is the irradiation angle α of the ring illumination. The irradiation angle α of the ring illumination is the angle formed by the light beam and the surface of the substrate when the light ray of the LED is incident (irradiated) on the substrate. In darkfield illumination, prevent the illumination light from entering the objective lens directly.
In order to realize dark field illumination, it is necessary to design the illumination including the imaging lens system. Since a lens with a high numerical aperture (NA) is used as the imaging lens, it is necessary to irradiate the substrate surface with an acute angle in order to obtain dark field illumination. Therefore, in order to secure the working distance, it is necessary to increase the outer diameter of the ring. The upper limit of the irradiation angle α is determined by the magnification and NA of the objective lens in the imaging optical system. For example, when the magnification of the objective lens is 1 and the NA is 0.5, the upper limit of the irradiation angle α is 50 °. Further, in order to reduce stray light, the upper limit of the irradiation angle α is preferably 30 ° or less from the viewpoint of setting the irradiation angle to a safer angle.

The efficiency requirement is the irradiation density p by ring illumination. The higher the irradiation density (brightening the illumination), the shorter the time for detecting defects and the higher the inspection efficiency. Therefore, the higher the irradiation density, the better. From the viewpoint of increasing the irradiation density, the larger the irradiation angle α is, the better, but the upper limit of the irradiation angle α is limited by the above-mentioned dark field request. This also limits the irradiation density p. From the viewpoint of the irradiation density p, the irradiation angle α is preferably 20 ° or more.

From the positional relationship shown in FIG. 4, there is a relationship of d = rTan (α) (Equation 1) in the surface side ring illumination. FIG. 5 (1) shows Equation 1. Since the irradiation density p is inversely proportional to the square of the distance from the LED light source, there is a relationship of (Equation 2) in FIG. This 3D graph is shown in FIG. 5 (2). These 3D graphs are superimposed and shown in FIG. 5 (3). In FIG. 5 (3), the dark-colored graph (curved surface) shows Equation 1, and the light-colored graph (curved surface) shows Equation 2.
From the 3D drawings of r, d, and α shown in FIG. 5 (3), considering the irradiation density p and considering the ease of manufacture, the manufacturing cost, and the like, as shown in FIG. 5 (4), r: 54.0 mm. , Α: 25.5 °, d: 25.0 mm.
In addition, in the reflection ring illumination on the surface side, in order to maintain the dark field request, the reflection ring illumination needs to be interlocked with the objective lens.

(Back side: transmitted dark field ring illumination)
The transmissive ring illumination on the back surface side is the same as the reflection ring illumination on the front surface side with respect to each relational expression (inequality) of the mechanical requirement, the dark field requirement (optical requirement), and the efficiency requirement (irradiance density). ..
In the transmitted ring illumination, the equation d is different from the above-mentioned reflective ring illumination in that the thickness g of the glass and the point of being refracted light from the back surface are different from the above-mentioned reflective ring illumination, and the equation d also changes accordingly. From the positional relationship shown in FIG. 4, the transmission ring illumination has the relationship of (Equation 3) in FIG. This 3D graph is shown in FIG. 6 (1). In FIG. 6 (2), the dark-colored graph (curved surface) shows Equation 3, and the light-colored graph (curved surface) shows Equation 2. Note that FIG. 6 (3) is an enlarged view of FIG. 6 (2).
In transmission illumination, a request for incident near the Brewster's angle, that is, a request for α: 33.0 ° is added. Brewster's angle is the angle of incidence at which the reflectance of P-polarized light becomes zero. By utilizing the Brewster's angle phenomenon, the return loss for effective illumination reaching the glass surface can be made extremely small.
From the 3D drawings of r, d, and α shown in FIG. 6 (2), considering the irradiation density p and considering the ease of manufacture, the manufacturing cost, etc., as shown in FIG. 6 (3), r: 50.0 mm. , Α: 33.0 °, d: 20.0 mm.
It is necessary to link the transmission dark field ring illumination on the back surface side with the magnification of the imaging lens. Due to the need for such interlocking, a dark field transmitted by backside illumination is not common in microscopes and the like. For the same reason, there is no example of simultaneous dark field for both reflection and transmission.

In the present invention, based on the values determined in the above design (with some numerical changes), for example, the surface side (reflection side) ring illumination has an irradiation angle α of 25 °, a housing outer diameter R2 of 128 mm, and a housing. The final design has a thickness t of 22 mm, 20 mm, and the like.
The back surface side (transmission side) ring illumination is finally designed so that the irradiation angle α is 31 °, the outer diameter R2 of the housing is 125 mm, and the thickness t of the housing is 22 mm or 19 mm.

In the present invention, with respect to the above-mentioned specific surface-side reflection dark-field ring illumination, the working distance d (working distance: WD) of the ring illumination is changed to, for example, 18 mm, 20 mm, 22 mm, 25 mm, 27 mm and the like. By examining the brightness distribution (brighter is better) and the light intensity profile (flatter and higher light intensity is better) in the illumination region having a radius on the substrate, the optimum ring illumination working distance d can be obtained. .. As a result, uniform illumination can be created within the field of view of the imaging camera. The same applies to the transmitted dark field ring illumination on the back surface side.

In the present invention, as can be seen from FIG. 3, an illumination region having a radius on the substrate is formed by a collection of spot lights by a plurality of LEDs. This is the same as a spotlight, and only the spotted area is illuminated brightly, and the surrounding area is not illuminated, so that it becomes dark. This also applies to a collection of spot lights by a plurality of LEDs (a collection area of spot lights formed by overlapping a large number of spot lights).
In the present invention, it is preferable that the radius of the illumination region formed on the substrate by the collection of spot light by a plurality of LEDs is in the range of, for example, 1 mm to 20 mm (for example, 10 mm). This range preferably corresponds to a region (image acquisition region, inspection region, imaging region) centered on the focal position of an imaging camera (for example, a TDI camera).

In the present invention, the region (image acquisition region, inspection region, imaging region) centered on the focal position of the imaging camera (for example, TDI camera) is relatively narrowed in order to improve the inspection accuracy (imaging). In the case of using only the region near the optical axis having good aberration characteristics of the lens system), the radius of the illumination region formed on the substrate by the collection of spot light by a plurality of LEDs is also relatively narrowed accordingly. You can go.
For example, by making the inclined portion 35 of the inner peripheral surface of the housing 34 a curved surface (or a part of a polygon inscribed in a circle that approximates the curved surface), the radius of the illumination region is also relative. Can be narrowed down.
Further, the inclined portion 35 of the inner peripheral surface of the housing 34 is inscribed in a curved surface having a light-collecting characteristic such that spot lights from a plurality of LEDs overlap at substantially one point (a region of one spot) or a circle similar to the curved surface. It can also be part of a polygon.

In the present invention, the working distance d (working distance: WD) of the ring illumination is configured to be finely adjustable. For example, when assembling the device, it has a fine adjustment mechanism that first fixes the device at a determined value and then fine-tunes (for example, fine-tunes in ± 1 mm units) to fix the device. The fine-tuning mechanism is also used during the use of the device. In these cases, adjust so that the lighting effect is maximized. As for the illumination effect, it is preferable that the illuminance is uniform in the field of view and the irradiation density is high (bright). In particular, it is preferable that the illuminance is uniform and the irradiation density is high (bright) within the range of the field of view of TDI (the field of view corresponding to the longitudinal direction of TDI).
The working distance d of the ring illumination is naturally determined at the optimum position on the front surface side (reflection side) in combination with the imaging lens, but on the back surface side (transmission side), it is set to the thickness of the glass (for example, 8 mm to 13.5 mm). Since the optimum position changes accordingly, it is a mechanism that can be dynamically changed on the stage.

In the present invention, as a measure against stray light, light-shielding masks that cut illumination light that does not contribute to image formation may be installed on the front surface side (reflection side) ring illumination and the back surface side (transmission side) ring illumination.
As for the stray light due to the lighting system, the stray light of the reflected light from the substrate is taken into consideration in the surface side (reflection side) ring illumination. In the back surface side (transmission side) ring illumination, the reflected light on the back surface of the substrate is taken into consideration, the refraction inside the glass substrate is taken into consideration, and the stray light of the transmitted light from the inside of the glass substrate is taken into consideration.

The defect inspection device of the present invention preferably uses a TDI camera.

The defect inspection apparatus of the present invention includes an imaging optical system including an objective lens, an imaging lens, and an imaging element, the imaging element being a TDI sensor, and the imaging optical system including a TDI camera.
It has a means for relatively moving the substrate and the TDI camera in a certain direction at a constant speed.
By matching the moving direction and speed of the imaging region (imaging object) on the substrate with the direction and speed of charge transfer of the CCD in the TDI sensor, the imaging region (imaging object) is repeated by the number of vertical stages of the CCD. It is preferable to have a means for exposing and photographing.
Here, by matching the moving direction and speed of the imaging region on the substrate with the direction and speed of charge transfer of the CCD in the TDI sensor, the means for repeatedly exposing and photographing the imaging region by the number of vertical stages of the CCD is, for example, , It can be performed by a control device (control circuit, CPU, software, etc.) built in the TDI camera.

A normal line sensor is a array of CCDs. In the TDI (Time Delay Integration) sensor (element), a plurality of rows of CCDs (one row) arranged on the line are further arranged in a direction perpendicular to the direction along the line. By integrating and exposing the images obtained by the plurality of rows of CCDs, it becomes possible to obtain an image with high sensitivity.
For an object to be photographed that moves at a constant speed and in a certain direction, by matching the moving direction / speed of the object to be photographed with the direction / speed of charge transfer of the CCD, the number of objects (one row) is equal to the number of vertical stages of the CCD. The same area on the object corresponding to the CCD) is repeatedly exposed and photographed.
The image taken by the CCD in the first row is directly transferred to the CCD in the second row. In the CCD in the second row, the imaging obtained in the CCD in the second row is added to the imaging sent from the front row, accumulated, and further transferred to the CCD in the third row. In the n-th row CCD, the imaging obtained by the n-th row CCD is added to the accumulated imaging up to the (n-1) row and transferred to the (n + 1) row. That is, in the TDI sensor in which x-row CCDs are arranged, the obtained imaging is multiplied by x and accumulated. When performing integral exposure of x rows, x times the amount of light and noise reduction of √x (root x) can be expected.
As a result, extremely high sensitivity can be obtained, and both high speed and high sensitivity can be achieved.
With the conventional CCD, only a low-luminance image with insufficient sensitivity due to the high resolution can be obtained, but by integral exposure, a bright and clear image can be obtained while having a high resolution. Compared with the conventional method of slowing down or stopping the moving speed of the object to be photographed in order to make up for the lack of brightness, the processing can be performed at high speed and in a short time.
In TDI, not only camera technology but also camera technology because of the nature of shooting an object while moving. Total solution capabilities including optical technology and transfer technology are required.

TDI can take panning shots like a copy machine. It is not a step-and-repeat method in which a certain area is imaged in a stopped state, moved to an adjacent area, and the step of imaging in a stopped state is repeated as in a normal area sensor. TDI can create a state where the total throughput is 3 to 4 times. In the step-and-repeat method, the heavy imaging optical system is repeatedly driven and stopped (acceleration and deceleration) at high speed, so that vibration occurs, but such vibration is unlikely to occur in TDI.

The TDI sensor is, for example, vertically long (24 mm × 1.5 mm (128 steps)) and 1.5 mm wide.
The present invention is also applicable to line sensors. With the line sensor, panning can be taken in the same way as TDI. The TDI sensor has 128 stages and can take 128 times as much light as the line sensor. The inspection speed can be increased 128 times as much as the amount of light can be taken.
In the case of a CCD area sensor, it is, for example, 15 mm square. You cannot take a panning shot. The sensitivity is also not good.
In the present invention, the TDI sensor is preferable. It is difficult to inspect defects two generations ahead with line sensors and area sensors.

FIG. 7 is a diagram for explaining an XYZ drive system and the like of the defect inspection apparatus of the present invention. (Holding means) for mounting and fixing the inspected body (board) 1 to be inspected for defects in a vertically upright state inside the main body frame 201 installed on the gantry 200 (not shown, will be described later). To move the head portion 300 including the imaging optical system 100 (optical system for performing defect inspection) and the observation optical system (optical system for performing magnified observation of defects), and the head portion 300 in three directions of XYZ. The X-axis, Y-axis, and Z-axis stages 211, 212, 213, and the like are arranged.
The X-axis stage 211 drives the head portion 300 in the X-axis direction by moving the Y-axis stage 212 in the X-axis direction (left-right direction on the paper surface).
The Y-axis stage 212 drives the head portion 300 in the Y-axis direction by moving the Z-axis stage 213 in the Y-axis direction (vertical direction on the paper surface).
The Z-axis stage 213 drives the head portion 300 in the Z-axis direction by moving the head portion 300 in the Z-axis direction (direction perpendicular to the paper surface).

In the defect inspection apparatus of the present invention, it is preferable that, for example, three TDI cameras are continuously arranged in contact with each other in the Y-axis direction of FIG. 7, for example, in the head portion 300. As a result, it is possible to scan (scan) with three (three) TDI cameras in the X-axis direction and inspect the area for three at once. Compared to the case of one TDI camera, the inspection speed is three times faster (inspection time is one-third). The number of TDI cameras can be increased or decreased as appropriate to any number.
In the apparatus of the present invention, in the head portion 300, for example, a TDI camera and an observation optical system (microscope) can be arranged adjacent to each other in the X-axis direction of FIG.

The X-axis, Y-axis, and Z-axis stages 211, 212, and 213 preferably have a linear lock mechanism that acts on the direction of movement during scanning by an imaging optical system (for example, a TDI camera). This is a mechanism for high-precision observation, and when a very fine vibration remains when viewing an image with a microscope, the image is blurred, so that very fine vibration is eliminated by locking the stage.

In the apparatus of the present invention, it is preferable to use a material having high rigidity for the base portion (frame) of the main body. The gantry preferably has a vibration isolating function that reduces the influence of vibration.
Further, in the apparatus of the present invention, it is preferable that the object to be inspected (board) 1 is fixed and the imaging optical system 100 is moved to a desired position on the body to be inspected (board) 1 to perform defect inspection. In the case of a structure in which the object to be inspected (board) 1 is moved, a moving space corresponding to that amount is required. Further, since the body to be inspected (board) 1 is very heavy, vibration is likely to occur when the body to be inspected (board) 1 is moved, inverted, or stopped.

Further, the observation optical system and the imaging optical system 100 are provided with an illumination device for illuminating the defect position and its surroundings during the defect inspection. Since it may be difficult to observe actual defects such as foreign matter defects and scratches depending on the type of lighting due to their shape and material, the lighting is reflected illumination such as coaxial epi-illumination, reflection dark-field ring illumination, and differential interference. Equipped with illumination, transmitted illumination (for example, coaxial vertical transmitted illumination), transmitted dark field ring illumination, color filter for changing the quality of light, polarizing filter, etc., various defects can be observed more clearly. It is desirable to equip.

Further, in the observation optical system and the imaging optical system 100, in order to carry out defect inspection in a good field of view with correct focus, auto is performed so that the distance between the object to be inspected (substrate) 1 and the tip of the lens is kept constant. It is desirable to have a means of focusing. For example, the types of autofocus include those that utilize laser reflection and those that utilize the image contrast of the focus surface, but in the case of the object to be inspected (substrate) 1, there is no part with contrast. Therefore, the one utilizing the reflection of the laser is preferable in the present invention.
In addition to the astigmatism method, the knife edge method and the like can be used as the autofocus means.

Further, the observation optical system for observing the defective portion (not shown, equipped in the head portion 300) is, for example, a low-magnification lens for grasping a rough position of the defective portion and observing a minute defect. It is composed of a high-magnification lens and some medium-magnification lenses that are used in between. As for these optical lenses, a method of switching a plurality of single lenses by using a revolver or the like is preferable.
Further, in the present invention, it is preferable that the optical lens used for the optical system has a long working distance as much as possible in the space between the optical system and the object (substrate) 1 to be inspected. A working distance of at least several mm, preferably 4 mm or more (working distance of the lens) is preferable.

The wavelength of the light source used in the lighting device in the observation optical system is preferably in the range of 380 to 800 nm. If the light in the ultraviolet region having a wavelength smaller than 380 nm is included, an optical component corresponding to the ultraviolet region is required and the cost becomes high. In addition, if light in the infrared region larger than 800 nm is included, it has heat, so there is a risk of adversely affecting the object to be inspected and the observation device. The wavelength of the light source is more preferably 400 to 750 nm from the same viewpoint. The wavelength band is preferably selected by providing a wavelength filter in the light source device.

Further, when it is desired to further select the wavelength or wavelength band of light because the defect may be easily manifested depending on the type of defect depending on the wavelength or wavelength band used, the space between the light source and the object to be inspected, or A wavelength filter may be provided between the object to be inspected and the observation device.

By using a light source having good parallelism as the light source, the manifestation of defects can be made more stable. Moreover, by using a light source having high brightness (high illuminance), the amount of light received by the light receiving optical system increases, so that the size of the observable pinhole increases. In addition, observable defects extend to defects such as half pinholes and recesses in thin films.

The present invention is suitable for a defect inspection apparatus including a high-precision imaging optical system. This is because high-precision defect inspection requires inspection with a high-precision imaging optical system. The defect inspection apparatus of the present invention is suitable for a defect inspection apparatus including a high-precision imaging optical system having a focal likelihood of less than ± 0.1 mm, and a high-precision imaging optical system having a focal likelihood of less than ± 0.05 mm. Suitable for defect inspection equipment equipped with, and suitable for defect inspection equipment equipped with a high-precision imaging optical system with a focal likelihood of less than ± 0.03 mm, and a high-precision imaging optical system with a focal likelihood of less than ± 0.02 mm. Suitable for defect inspection equipment.

High-precision focus adjustment is essential for high-precision defect inspection. It is difficult to hold a large substrate vertically with high accuracy, and an inclination of about 0.05 ° occurs. This inclination is a deviation of about 0.9 mm at a length of 1000 mm. If the optical axes of the laser displacement meter and the imaging optical system are separated by 100 mm, a measurement error of about 0.09 mm occurs. Since the focal likelihood of a high-precision imaging optical system is about ± 0.03 mm, the image can be acquired in a defocused state (a state exceeding the focal likelihood), and the defect detection power may be significantly reduced. There is sex. This problem can be dealt with by the autofocus function described below.

The defect inspection apparatus of the present invention preferably has an autofocus function using the astigmatism method.

FIG. 8 is a schematic diagram for explaining the astigmatism method.
In the astigmatism method, a cylindrical lens (cylindrical lens, semi-cylindrical lens) is generally used. When a cylindrical lens is inserted in the optical path that is reflected by the surface of the object 1 to be inspected and returns to the 4-segment photodetector (PD), the cylindrical lens has a lens effect only in the X-axis direction in the figure, so the focal position in the X-axis direction. Astigmatism occurs due to the shift of the focal position in the Y-axis direction, and the shape of the beam changes as 1 (vertical ellipse) → 2 (circular) → 3 (horizontal elliptical) depending on the distance on the optical axis. (Upper view of FIG. 8).
Here, when the beam is received by the photodetector divided into four A, B, C, and D clockwise from the top, the balance of the incident light amounts of A to D changes in each of the cases 1 to 3.
In the case of 1, the amount of incident light of A and C is large (the figure below 1 in FIG. 8).
In case of 2, A. B. C. The four incident light amounts of D are equal.
In the case of 3, the amount of incident light of B and D is large.

If the optical system is adjusted so that the four incident light amounts are equal and the beam shape is circular when the laser beam is in focus on the surface of the object 1 to be inspected, (A + C)-(B + D) By controlling the position of the imaging optical system 100 so that (A + C)-(B + D) = 0 from the calculation result of), the state in which the laser beam is always in focus on the surface of the object 1 to be inspected is maintained. Can be kept. (A + C)-(B + D) is called a focus error signal (FE). The position of the imaging optical system 100 can be moved by a drive mechanism (for example, a linear motor stage), and can be controlled at high speed and accurately.
When FE> 0, the focus is off to the front (the object to be inspected 1 is close).
When FE = 0, it is in focus (focusing).
When FE <0, the focus is shifted to the back side (the object 1 to be inspected is far away).

FIG. 9 is a diagram showing a configuration of a coaxial autofocus module.
The coaxial autofocus module includes a laser light source (laser diode) 21, a diaphragm 22, a polarizing beam splitter (PBS) 23, a 1/4 wave plate 24, a reflecting element (prism mirror) 25, a condenser lens 26, a cylindrical lens 27, and 4 It is composed of a split photodetector (light detector) 28.

The laser beam emitted from the laser light source (laser diode) 21 is incident on the polarizing beam splitter (PBS) 23 via the aperture 22, and the transmitted light is transmitted through the 1/4 wavelength plate 24 and is transmitted through the reflecting element (reflecting element (PBS). It is incident on the prism mirror) 25, reflected by the reflecting element (prism mirror) 25, transmitted through the objective lens 11 along the optical axis O of the imaging optical system 100, incident on the surface of the object 1 to be inspected, and reflected. Will be done. This reflected light passes through the objective lens 11 along the optical axis O of the imaging optical system 100, is incident on the reflecting element (prism mirror) 25, and the reflected light is transmitted through the 1/4 wavelength plate 24. It is incident on the polarizing beam splitter (PBS) 23, and the reflected light is sequentially transmitted through the condenser lens 26 and the cylindrical lens 27 and is incident on the quadrant photodetector 28.

The present invention is a defect inspection device having an autofocus function using an astigmatism method coaxial with the imaging optical system and avoiding reflection of a laser beam (reference light) used for the autofocus function on an image sensor. It is preferable to apply.
FIG. 10 is a diagram for explaining means for avoiding reflection of the laser beam used for the autofocus function on the image sensor.
Specifically, FIG. 10 (1) shows the positional relationship between the image acquisition region on the substrate to be inspected by the imaging optical system (objective lens) of the imaging camera (for example, a TDI camera) and the spot of the laser beam on the substrate to be inspected. It is a figure for demonstrating. FIG. 10 (1) is a view of the image acquisition area viewed from the front (a view seen through an objective lens).
In FIG. 10 (2), the image sensor (for example, TDI sensor) region and the laser beam are reflected by the substrate to be inspected in the light receiving plane including the image sensor (for example, TDI sensor), and the reflected light is transmitted through the imaging optical system. It is a figure for demonstrating the positional relationship with the position which forms an image on a light receiving plane including an image sensor (for example, a TDI sensor). FIG. 10 (2) is a view of the light receiving plane viewed from the back.
The TDI sensor is a sensor in which CCDs are arranged in a row in the horizontal direction, and a plurality of CCDs are arranged in a row in the vertical direction. By integrating and exposing the images obtained by the plurality of rows of CCDs, it becomes possible to obtain an image with high sensitivity.

As shown in FIG. 10, there is an image acquisition area A on the substrate to be inspected captured by the objective lens (FIG. 10 (1)) and a corresponding area B projected by the imaging lens (FIG. 10 (2)). )). These regions are the regions of the visual field of the imaging optical system 100.
In the present invention, as shown in FIG. 10 (2), an image sensor T that uses a part of the region B projected by the imaging lens is used.
Then, in the present invention, as shown in FIG. 10 (1), the laser beam is applied to the region excluding the image acquisition region T'corresponding to the image sensor in the image acquisition region A on the substrate to be inspected captured by the objective lens. The spot S of is located. As a result, as shown in FIG. 10 (2), the image S'of the spot of the laser beam is positioned in the region other than the image sensor T.
In this way, for example, as shown in FIG. 10 (1), the position of the spot S of the laser beam (the focusing position of the laser beam) is slightly shifted from the optical axis O (on the surface of the image acquisition region A). By translating the image S'to the outside of the image acquisition area T'), the image S'of the spot of the laser beam is prevented from being reflected on the image sensor T.

For this purpose, for example, in the coaxial autofocus module shown in FIG. 9, the laser light source (laser diode) 21 or the reflecting element (prism mirror) 25 is slightly tilted from the state where the laser beam passes through the optical axis O. ..
The term "slightly" means that the image S'of the spot of the laser beam can be prevented from being reflected on the image sensor T. Since the image S'of the spot of the laser beam expands according to the out-of-focus, it is possible to avoid reflection on the image sensor and influence on the image sensor (for example, noise) in consideration of this. .. In addition, the size of the laser beam spot image S'changes according to the position where the laser light source (laser diode) is placed and the thickness (beam diameter) of the laser light source. The arrangement of the laser light source (laser diode) and the aperture 22 are adjusted so as to avoid reflection and influence on the image pickup element (for example, noise).

In the present invention, high defect detection power can be obtained, which is due to synergistic effects such as the application of the lighting system, the application of the high NA optical system, the application of the TDI camera, and the application of real-time autofocus according to the present invention.
In the present invention, as described above, by using a remarkably excellent ring illumination, it is possible to simultaneously illuminate from all directions of 360 °, and in addition, it is possible to detect defects regardless of the direction of defects. Compared to the case of using the ring lighting currently provided, it is possible to exert a remarkably stable and high defect detection power, and the expected effect can be obtained.
In the present invention, by using a high-precision imaging optical system (brighter with high NA, high resolution, and minimum focal likelihood), the accuracy of defect detection is higher than that without such an optical system. It is highly accurate in that the detectable defect size is relatively small.
Further, in the present invention, high-sensitivity inspection can be supported by applying a TDI camera. In addition to this, the present invention can support high-sensitivity inspection by applying extremely accurate focus adjustment at the image acquisition position.
In the present invention, the accuracy of the two-dimensional image created by the TDI sensor is improved by acquiring data with the TDI sensor while illuminating with the lighting system according to the present invention and constantly focusing. This makes it possible to further improve the defect detection accuracy.
Further, in the present invention, high-speed inspection can be supported by applying a TDI camera. In addition to this, in the present invention, high-speed inspection can be supported by real-time autofocus control.
The present invention solves the above-mentioned problem of focal likelihood in an inspection using a high-precision imaging optical system (high NA, bright, high resolution, and minimum focal likelihood), and thus high-precision imaging optics. The accuracy of defect detection is higher than that when the system is used but the real-time autofocus of the present invention is not used (in this case, the focus may not be achieved due to the problem of focus likelihood).
In the present invention, the application of real-time autofocus enables extremely accurate focus adjustment at the image acquisition position, and can exhibit stable and high defect detection power independent of the substrate orientation.

Hereinafter, the present invention will be described in more detail with reference to Examples.
(Example 1)
[Manufacturing of defect inspection equipment]

A TDI camera having a focal likelihood of ± 0.02 mm was used in the defect inspection apparatus shown in FIG. 1, and the coaxial autofocus module shown in FIG. 9 was incorporated. The magnification of the objective lens of the TDI camera was 1x, and the NA was 0.3.
The inclination of the substrate is 0.05 degrees in the X-axis direction of FIG. 1, and if the substrate is separated by 100 mm in the X-axis direction of FIG. 1, a deviation of about 0.09 mm in the Z-axis direction occurs. The scanning direction of the TDI camera is the X-axis direction.

The illuminating means are the illuminating means 31 and the ring illuminating means 32 described with reference to FIG. 2, the dark field ring illuminating of the reflection on the front surface side described with reference to FIG. All of the field ring illumination and the spot illumination means 33 described with reference to FIG. 1 were equipped.

The reflective dark-field ring illumination (front side), which is the illumination means 31, has a configuration in which LEDs are arranged in a ring shape in three layers (see FIG. 2) along each of the three concentric circles (see FIG. 2). .. Blue LEDs (wavelength 465 nm) were used for the outer and inner LED annulus, and orange LEDs (wavelength 610 nm) were used for the middle LED annulus.
The diameter (size) of the LED was an ultra-narrow angle of 3.1 mm. The number of LED elements arranged was 120.
The housing shown in FIG. 3 was used. The irradiation angle α was 25 ° (dark field illumination), the outer diameter R2 of the housing was 128 mm, and the thickness t of the housing was 20 mm.
The working distance d of the reflection dark field ring illumination (surface side), which is the illumination means 31, was set to 25.0 mm. The operating distance d is the distance from the tip of the LED lighting on the substrate side to the surface of the substrate.

The transmissive dark-field ring illumination (back side), which is the illumination means 32, has a configuration in which LEDs are arranged in a ring shape in three layers (see FIG. 2) along each of the three concentric circles (see FIG. 2). .. Blue LEDs (wavelength 465 nm) were used for the outer, middle, and inner LED annulus rows.
The diameter (size) of the LED was an ultra-narrow angle of 3.1 mm. The number of LED elements arranged was 120.
The housing shown in FIG. 3 was used. The irradiation angle α was 31 ° (dark field illumination), the outer diameter R2 of the housing was 125 mm, and the thickness t of the housing was 19 mm.
The working distance d of the transmissive dark field ring illumination (back surface side), which is the illumination means 32, was set to 20.0 mm. The operating distance d is the distance from the tip of the LED lighting on the substrate side to the surface of the substrate.

The spot illumination means 33 used a parallel light spotlight (a high-intensity (high-luminance) LED light source) and used a blue LED (see FIG. 1).

(Comparative Example 1)
In Comparative Example 1, a commercially available ring illumination currently provided was used. Moreover, a general spot lighting means was used. Other than that, it was the same as in Example 1.
The diameter (size) of the LED was 120 mm. The number of LED elements arranged was 120 elements. The irradiation angle α was approximately 30 °. The combined illumination includes bright-field light and dark-field light, and the contrast of the defective part is reduced.
The spot illumination means 33 used a spotlight (blue LED) that was not parallel light and did not have high brightness (high illuminance).

(Comparative Example 2)
In Comparative Example 1, a ring illumination simply produced by improving the currently provided ring illumination was used. Moreover, a general spot lighting means was used. Other than that, it was the same as in Example 1.
The diameter (size) of the LED was 120 mm. The number of LED elements arranged was 120 elements. The irradiation angle α was approximately 25 °. The combined illumination includes bright-field light and dark-field light, and the contrast of the defective part is reduced.
In Comparative Example 2, more defects were overlooked than in Example 1.
The spot illumination means 33 used a spotlight (blue LED) having high brightness (high illuminance) but not parallel light.

[Defect inspection]
Defect inspection was performed using the defect inspection devices of Example 1, Comparative Example 1 and Comparative Example 2.
When inspecting the glass substrate, both the illuminating means 31 and the illuminating means 32 were used at the same time.
When inspecting the mask blank (the substrate with the thin film), all of the lighting means 31, the lighting means 32, and the spot lighting means 33 were used at the same time.
When inspecting the resisted mask blank, the middle LED annulus (orange LED only) in the illuminating means 31 was used.

In Example 1, in the inspection of the glass substrate, optical defects such as scratches, foreign substances, foreign substances inside the glass, and veins could be detected very well, and high defect detection power was obtained.
In Comparative Example 1, defects could hardly be detected.
In Comparative Example 2, high defect detection power was confirmed in the region near the optical axis of the imaging lens system, but high defect detection power was not obtained in the other visual field regions, and the defect detection power was insufficient and the expected effect was obtained. I couldn't get it.

In the first embodiment, in the inspection of the mask blank (thin film-attached substrate), pinholes, half pinholes having no clear edges and less scattered light generation, dents (recesses) in the film that generate less scattered light, and the like. , It was possible to detect dents (gradation) on the gentle curved surface of a thin film that generates extremely little scattered light.
In Comparative Example 1, defects could hardly be detected.
In Comparative Example 2, high defect detection power was confirmed in the region near the optical axis of the imaging lens system, but high defect detection power was not obtained in the other visual field regions, and the defect detection power was insufficient and the expected effect was obtained. I couldn't get it.

With respect to the foreign matter on the thin film, high defect detection power was obtained in Example 1.
In Comparative Example 1, defects could hardly be detected.
In Comparative Example 2, a high defect detecting ability was not obtained, and the defect detecting ability was insufficient to obtain the expected effect.

In Example 1, defects such as foreign substances could be detected very well in the inspection of the mask blank with resist, and high defect detection power was obtained. In Comparative Example 2, a high defect detecting ability was not obtained, and the defect detecting ability was insufficient to obtain the expected effect.

(Example 2)
In the second embodiment, the coaxial autofocus module was not operated in the first embodiment. Other than that, it was the same as in Example 1.
As a result, there was a large difference between the part where the defect could be detected and the part where the defect could not be detected.

(Example 3)
In the second embodiment, the coaxial autofocus module was not operated in the first embodiment. Further, in Example 1, a line sensor was used instead of the TDI camera. The inspection time was 20 times that of Example 1. Other than that, it was the same as in Example 1.
As a result, defect detection was almost impossible.

(Example 4)
In the fourth embodiment, in the first embodiment, only the dark field ring illumination transmitted on the back surface side was turned on.
In Example 4, the detection of foreign matter was inferior to that in Example 1.

(Example 5)
In the fifth embodiment, in the first embodiment, only the dark field ring illumination of the reflection on the surface side was turned on.
In Example 5, the detection of scratches and pinholes was inferior to that in Example 1.

In Example 4, there was a difference in the types of defects that were good at as compared with Example 5.
This is because, according to the scattering theory, the forward scattering and the backscatter are generally stronger than the forward scattering, and the light is incident from the back surface side of the substrate by the transmission dark field ring illumination and is transmitted. This is because the light produced is backscattered.

(Example 6)
In the sixth embodiment, only the spot lighting means was turned on in the first embodiment.
In Example 6, the detection of foreign matter and half pinhole was inferior to that in Example 1.

1 Substrate to be inspected 10 Imaging camera (for example, TDI camera)
11 Objective lens 12 Imaging lens 13 Image sensor 20 Autofocus module 31 Lighting means 32 Lighting means 33 Lighting means 100 Imaging optical system 200 Stand 300 Head unit

Claims (10)

A defect inspection device capable of inspecting defects of an object to be inspected, which is composed of a glass substrate, a mask blank, or a mask blank with a resist.
It is a defect inspection device that receives scattered light generated by defects existing in the object to be inspected by an optical system for defect inspection and detects defects based on the received scattered light.
It has an optical system for defect inspection and a lighting system for defect inspection.
In the lighting system, a plurality of LEDs are arranged in an annular shape, and the directions of the plurality of LEDs are arranged toward a region centered on the focal position of the optical system, so that the spot light generated by the plurality of LEDs can be detected. Each is a ring illumination that gathers in the area centered on the focal position of the optical system.
In the ring illumination installed on the side of the optical system of the inspected object , each of the spot lights by the plurality of LEDs is irradiated sharply on the surface of the inspected object, and the inspected object emits the spot light. Reflected dark-field ring illumination that prevents reflected light from entering the objective lens directly,
The ring illumination installed on the opposite side of the object to be inspected from the optical system, in which spot light from the plurality of LEDs is irradiated sharply to the back surface of the substrate to irradiate the inside of the object to be inspected. A transmitted dark-field ring illumination that prevents the transmitted light transmitted through refraction from directly entering the objective lens.
A defect inspection device comprising. A defect inspection device capable of inspecting defects of an object to be inspected, which is composed of a glass substrate, a mask blank, or a mask blank with a resist.
It is a defect inspection device that receives scattered light generated by defects existing in the object to be inspected by an optical system for defect inspection and detects defects based on the received scattered light.
It has an optical system for defect inspection and a lighting system for defect inspection.
In the lighting system, a plurality of LEDs are arranged in an annular shape, and the directions of the plurality of LEDs are arranged toward a region centered on the focal position of the optical system, so that the spot light generated by the plurality of LEDs can be detected. Each is a ring illumination that gathers in the area centered on the focal position of the optical system.
The ring illumination installed on the side of the optical system of the inspected object , in which spot light from the plurality of LEDs is irradiated at a sharp angle to the surface of the inspected object from the inspected object. The reflected dark-field ring illumination is configured so that the reflected light does not enter the objective lens directly.
The ring illumination installed on the opposite side of the object to be inspected from the optical system, and spot light from the plurality of LEDs is irradiated to the back surface of the object to be inspected at a sharp angle, and the object to be inspected is inspected. Transmission dark-field ring illumination that prevents transmitted light transmitted through refraction inside the body from directly entering the objective lens.
Illumination installed on the opposite side of the object to be inspected from the optical system, which is a transmission spot illumination coaxial with the optical axis of the optical system.
A defect inspection device comprising. The defect inspection apparatus according to claim 1 or 2, wherein the optical axis of the optical system and the central axis of the annulus in the ring illumination are made coaxial with each other. It has an imaging optical system including an objective lens, an imaging lens, and an image sensor.
The image sensor is a TDI sensor.
The imaging optical system includes a TDI camera.
It has a means for relatively moving the object to be inspected and the TDI camera at a constant speed in a constant direction.
By matching the moving direction and speed of the imaging region on the object to be inspected with the direction and speed of charge transfer of the CCD in the TDI sensor, the imaging region is repeatedly exposed and photographed by the number of vertical stages of the CCD. The defect inspection apparatus according to any one of claims 1 to 3. The defect inspection apparatus according to any one of claims 1 to 4, wherein the defect inspection apparatus includes a high-precision imaging optical system having a focal likelihood of less than ± 0.1 mm. Using the objective lens and using a laser beam, an autofocus means is constructed and at the same time.
The defect inspection apparatus according to claim 4, further comprising means for avoiding reflection of the laser beam used in the autofocus means on the image pickup device. The LED used in the lighting system is any one of claims 1 to 6, wherein the LED is selected from a blue LED (wavelength: 465 nm), a yellow LED (wavelength: 592 nm), and an orange LED (wavelength: 610 nm). Defect inspection device described in. The LED used for the reflection dark-field ring illumination is a blue LED (wavelength: 465 nm) and a yellow LED (wavelength: 592 nm) or an orange LED (wavelength: 610 nm), and the transmission dark-field ring illumination. The defect inspection apparatus according to any one of claims 1 to 7, wherein the LED used in the above is a blue LED (wavelength: 465 nm). Defect inspection method characterized by using a defect inspection equipment according to any of claims 1 to 8, inspecting said inspection object. Depending on the type of the object to be inspected, the LED used for the reflected dark-field ring illumination and the LED used for the transmitted dark-field ring illumination are appropriately selected to inspect the object to be inspected. The defect inspection method according to claim 9.

Images