JP2639501C - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTIONTechnical field   The present invention relates to an inspection system for use in the production of microcircuits, in particular a plurality of redundant circuits each.
Rear used in the manufacture of microcircuits of the type including an array of dies with patterns
The present invention relates to a real-time defect inspection system.Background of the Invention   Missing patterns in photomasks used in large-scale manufacturing of semiconductor devices and integrated circuits
Two very similar inspection systems for defect inspection are described in Watkins US Pat.
No. 0949 and Massen U.S. Pat. No. 3,614,232. this
These systems require simultaneous inspection of all dies, usually including a regular array of identical dies. Aperiodic defects, i.e. the same as in the other die in the array.
It detects defects in one die that is not repeated.   This involves a parallel laser emission of all dies of the photomask to be inspected from the laser.
Achieved by simultaneously illuminating
You. The spatial distribution of this composite diffraction pattern consists of two components, the first component of the die
The second component is the interference pattern of a single die of the array.
is there. The first and second patterns are sometimes die interference patterns and intra-die interference patterns.
It is called N. The light transmitted through the photomask is incident on the biconvex lens, and this lens
The spatial filter located at a distance equal to one focal length behind this lens
Distributed.   The spatial filter is a known error-free reference photomask corresponding to the photomask to be inspected.
It has a two-dimensional Fourier transform pattern of a disc. This filter is error-free Fourier transform
The part corresponding to the spatial frequency component of the conversion pattern is opaque, and the error-free Fourier transform
Parts not included in the turn are transparent. Both Watkins and Massen patents are lenses
Did not specify the design parameters of And the lens should be inspected in the Massen patent
Numerical aperture and magnification suitable to cover the area of the photomask
Is only stated.   The spatial frequency components corresponding to photomask defects to be inspected are mostly spatial
It passes through the filter and can be processed in two ways. Watkins system
Then, the light transmitted through the spatial filter is passed through a properly positioned biconvex lens for inspection.
Image of the photomask to be removed (does not contain any information blocked by the spatial filter
) Is formed. The imaging light that is not blocked by the spatial filter is
Appear at the position indicating the position where the defect exists in the workpiece. Massen's system is empty
The output signal that detects the transmitted light of the inter-filter by the photodetector and drives the “good / bad” alarm
Issue.   Watkins and Massen systems determine the presence of defects in an object pattern
Requires both inter-die interference pattern and intra-die interference pattern information
It is. Die-to-die interference pattern information is particularly problematic because of this information
Consists of light spots located in close proximity, resolved by a Fourier transform lens This is because it is extremely difficult to do so. Furthermore, the realization of such a lens is an inverse Fourier
Form an image of a test pattern from a Fourier-transformed light pattern using an E-transform lens
Difficult for inspection systems. The reason is that the design of each of these lenses is interchangeable.
A system that achieves both Fourier transform pattern and image forming functions in harmony
This is because it is necessary to achieve the overall design. This is why such a system design
Makes it increasingly difficult to obtain the resolution required for inter-die interference pattern information
. The above-mentioned lens design problem involves simultaneous inspection of the entire area of each die of the photomask array to be inspected.
Also occur in systems of the type to be inspected, thus making such systems unreliable
And make it useless.Summary of the Invention   Therefore, the object of the present invention is to provide a highly reliable defect detection used in the manufacture of microcircuits.
Inspection system.   Another object of the invention is to apply Fourier optics technology, but usually to the same die array.
Die to determine the presence of defects during the fabrication of microcircuits of the type
An object of the present invention is to provide an inspection system that does not use inter-interference pattern information.   Still another object of the present invention is to provide an aberration-free Fourier transform pattern from a microcircuit pattern.
A precise image corresponding to defects in the microcircuit pattern.
It is to provide a defect inspection system that can be formed.   Still another object of the present invention is to use the intra-die interference pattern information for the same die, usually.
Provide a defect inspection method for determining the presence of a defect in a microcircuit array pattern
It is in.   The present invention relates to a defect inspection method and system for manufacturing microcircuits.
Here, a tie including an array of circuit dies each having a number of redundant circuit patterns is described.
Real-time inspection system for inspecting surface defects of semiconductor wafers
I will explain. Such semiconductor wafers are, for example, random access memories, read
And a dedicated memory and a digital multiplier.   Two preferred embodiments of the inspection system include a Fourier transform laser located along the optical axis.
And the inverse Fourier transform lens are used to illuminate the patterned wafer
A spatial frequency component spectrum from the selected area, and selectively Can be filtered to generate image patterns of defects in the illuminated area of the wafer
To These lenses are aligned with the optical axis rather than the diffracted light across the wafer.
Due to the light diffracted by the Ephadie and other dies located in close proximity to this die
Collect the diffracted light. This collection limitation limits the applicability of inspection systems to many
Limited to dies with redundant circuit patterns, Fourier transform patterns and filters
Use of lenses that introduce off-axis aberrations that alter the characteristics of the filtered defect image
Can be used.   These lenses are relatively easy to manufacture. The reason is that the redundant circuit pattern
Spatially spaced at intervals of about 1.0 millimeter, typically repeating at 50 micron intervals
Generate several components, such spatial frequency components can be resolved by conventional optical elements
It is. Fourier transform and imaging area is light from wafer die only aligned with optical axis
It is preferably large enough to accommodate The spatial filter is such a die
Rejects the spatial frequency of the error-free Fourier transform of
Includes only turn information.   The wafer is positioned in the front focal plane of the Fourier transform lens and the wafer pattern
The surface is illuminated with a parallel laser beam. Fourier transform pattern on illumination wafer surface
Are formed at the back focal plane of the Fourier transform lens. Use the pre-fabricated spatial filter
The redundant circuit pattern of the illuminated die on the wafer is placed in the plane of the
Effectively block the light from the defect and allow the light emanating from the defect to pass.   The inverse Fourier transform lens is the light transmitted or reflected by the spatial filter
The light that has been diffracted and filtered by the illuminated wafer die
Performs a Lie transform. Whether the spatial filter is transmissive or reflective depends on this
It depends on the embodiment of the inspection system to be incorporated. The filtered image is a two-dimensional light
Projects onto the surface of the detector array, thereby addressing defects only in the die on the optical axis
Detect the presence of light. The photodetector array is centered on the optical axis and
Light sensing table large enough to cover the image plane area where the defect image corresponding to b
It has an area. Various wafer surfaces with multiple redundant circuit patterns
Inspection of all the defects in the part is performed by mounting the wafer on a two-dimensional translation stage.
As the stage is moved, the illumination area defined by the laser beam is cut across the wafer surface. Scan continuously from a to the die until the desired portion of the wafer is illuminated
To achieve. Movement and translation of the translation stage by using the time delay integration technique
And inspection of various parts of the wafer surface with multiple redundant circuit patterns.
It is possible to perform the scan in a raster sequence.   The present invention eliminates the need to manufacture a spatial filter using an error-free wafer.
It is advantageous. The reason is that defects present in such wafers are
This is because light of insufficient intensity is generated to expose the medium.   The present invention is applicable to a part of a test pattern having a large number of redundant circuit patterns.
A defect in the pattern to be inspected is detected using only the interference pattern information in (a). Of the present invention
Inspection methods can be used to inspect only a part of a large number of redundant circuit patterns.
No interference pattern information is needed and such partial inspections should be sufficiently statistically sampled.
Determining the defect distribution over the entire pattern to be inspected by performing ringing
It is based on the premise that it can be done.Detailed Description of the Preferred Embodiment   FIG. 1 shows about a quarter micron present in a periodic structure with a large number of redundant circuit patterns.
Inspection system 10 of the present invention designed to detect semiconductor wafer defects of
1 shows a first preferred embodiment of the present invention. Fig. 2 shows the inspection system 10
1 shows a preferred semiconductor wafer 12. Wafers 12 are generally of the same die 14
Includes a regular array, with at least about 20 dies along the X-axis 18 and Y-axis 20
Of redundant circuit patterns 16. Each die 14 is typically about 3 mm on a side.
It is a square of Torr. Reference photos A to C submitted separately are photographs of an example of one die 14.
At successively higher magnifications shows a number of repetitive circuit patterns in such a die.
These circuit patterns are rectangular as shown in Reference Photos A to C.
For simplicity, these circuit patterns 16 are square with a side of about 50 microns.
Shall be.   In FIG. 1, the inspection device 10 comprises a laser light source 22 which is 442.5
A substantially parallel beam 24 of a monochromatic light beam 24 of nanometer is projected on a lens 26,
Focuses this ray at a point 28 located in the back focal plane of the lens 26. Diverges from focal point 28
Light beam 30 strikes a small mirror 32 located at a small distance from focal point 28 and is relatively thin circular The light beam is reflected toward a Fourier transform lens section. This lens section 34 is the first
Although shown as a single lens element in the figure, as will be described later, five lens elements are used.
Will be revealed. The mirror 32 is located at the center of the Fourier transform plane defined by the lens section 34.
Block the area. The size of this region is small enough that other regions in the Fourier transform plane
The defect information at the location is hardly obstructed by the mirror 32.   The effective center of the Fourier transform lens section 34 is slightly smaller than one time the focal length from the mirror 32.
Project parallel light rays 36 onto the patterned surface of wafer 12
You. The wafer 12 is mounted in a chuck 38 that forms part of a two-dimensional translation stage 40.
. The wafer 12 is located in the object plane or front focal plane 42 of the lens section 34, and the parallel ray 36
Illuminate the patterned surface of the wafer 12. The parallel ray 36 is the surface of the wafer 12.
Illuminates an area of 20 mm in diameter. Diffracted in the illuminated area of wafer 20
The reflected ray 44 passes through the lens section 34 and is illuminated in the rear focal plane 46 of the lens section 34.
A Fourier transform pattern of the wafer surface is formed.   The Fourier transform pattern is a bright light source distributed in the expected manner in the back focal plane 46.
Including pot array. The illuminated area of the wafer 12 with a diameter of 20 mm
Gives a Fourier transform pattern with reasonable accuracy. The reason is that the wafer 12 has a lot of redundancy
This is because it is composed of a circuit pattern. However, lens section 34 is only 3 mm
An essentially astigmatic image of defects in a semiconductor wafer with an object field of Torr diameter
Design to form. The predetermined die is moved with respect to the illumination area by the translation stage 40.
Therefore, the entire predetermined die can be inspected for defects. Because of this,
A fairly large area of the wafer is illuminated and precise Fourier transformation of the redundant circuit pattern
In this case, a lens having a relatively small object field diameter is placed in the illumination area.
Introduces aberration into the Fourier transform pattern formed by condensing the light diffracted by
To a minimum.   A spatial filter 50 prepared in advance is located in the Fourier transform plane 46. Space fill
Data 50 was diffracted by a recording medium such as a photographic plate by all the dies 14 of the wafer 12.
It can be manufactured by exposure to light. This is achieved with the wafer 12 to be inspected
can do. The reason is that defect information carried by relatively low-intensity light is
The Fourier transform information which does not expose the photographic plate but is carried by relatively high intensity light is This is for exposing the photographic plate. The spatial filter 50 is a computer-based filter.
It can also be manufactured by using a method of generating a Fourier transform pattern.   The spatial filter 50 is the space of the error-free Fourier transform of the illuminated die 14 of the wafer 12
Rejects frequency, but emits light from defects in this die and other light near this die.
Passes the light diffracted by the die. Defect transport not blocked by spatial filter 50
The transmitted light beam 52 enters the inverse Fourier transform lens section 54. This lens classification is a single lens
Although shown as elements, it includes four lens elements as described below. Reverse foo
The Rier transform lens section 54 is a filter for the filtered light path of the illuminated wafer die 14.
Inverse Fourier transform the turn. Lens section 54 focuses from rear focal plane 46 of lens section 34.
It is located at a distance of one time the point distance. The elements of lens sections 34 and 54 are aligned along the same optical axis 48.
And the translation stage 49 moves the wafer die 14 across the optical axis 48.   The photodetector array 58 is centered on the optical axis 48 in the image plane 60 and resides in the die 14 on the optical axis.
An image of the existing defect is received. The image plane 60 is located in the back focal plane of the lens section 54. Les
The magnification of the lens section 54 is large enough that the resolution limit of the image substantially matches the pixels of the photodetector array 58.
To In particular, the photodetector array 58 is approximately 10 mm in a 30 mm diameter image section.
Assume that it has a light-sensing surface 62 that measures mx 10 millimeters. This is
To detect defects in the 3 mm diameter area of the wafer die 14 on the optical axis
A 10x magnification is appropriate for.   In order to detect the entire pattern surface of the wafer 12, the translation stage 40
Each part of the die 14 is sequentially moved to the optical axis 48 and illuminated with light generated from the light source 22. Stationary
The area of the light-sensing surface 62 of the photodetector array 58 determines the amount of detected light with respect to the optical axis 48.
Is limited to the area of a part of the image corresponding to only the wafer die 14 to be processed. Because of this,
Image information corresponding to any part of the illuminated off-optical axis of the wafer die 14 is obtained by a photodetector.
I can't reach Leh 58. The movement of the parallel moving stage 40 is performed continuously in a line-sequential raster system.
A time delay technique to collect defect image information for each die on the wafer 12 pattern surface
Perform surgery.   The Fourier transform lens section 34 and the inverse Fourier transform section 54 are part of one optical system.
It has a total of 10 elements as shown in FIG. Optical system 68 design
Is two important design parameters, the minimum spot diameter at the Fourier transform surface 46. “D₁”And the minimum spot diameter“ d_TwoComplicated by stringent requirements for "
become. The minimum spot diameter on the Fourier transform plane 46 is the largest expected circuit pattern.
Are needed to resolve the bright spots that occur in this plane. putter
If the square 16 is square, the required spot diameter d₁Is the following formula:         d₁<< λf₁/ c Where λ is the wavelength of light emitted from the laser light source 22, f₁Is the fruit of lens section 34
The effective focal length and c are the lengths of one side of the square pattern 16. 20 micron spot
Diameter d₁Can be achieved for c <300 microns.   The minimum spot diameter on the image plane 60 determines the smallest detectable defect size. minimum
Spot diameter d_TwoIs determined by the synergistic action of the lens sections 34 and 54 and the image magnification "m". d_Two/
Defects of diameter greater than m can be measured from the spatial extent of the image. d_Two/
Defects with a diameter smaller than m, i.e., defects with a resolution lower than_TwoHas an image spread equal to
Have an image intensity that decreases quadratically with increasing diameter. Sub-resolution defects
Inspection system 10 must achieve sufficiently low electronic or optical noise to detect
Need to be designed. The design parameters for the minimum spot diameter on the image plane are
In a preferred embodiment of stem 10 d_Two= 10 microns, m = 10 and d_Two/ m = 1 micron to diameter 3
It can be achieved over an image field of 0 millimeters.   3 and 4, the optical system 68 is a paraxial diffraction-limited optical system,
Accepts light diffracted into ± 15 ° -20 ° telesetonic ric cones and objects (ie,
Forming a periodic structure of the Fourier transform of the redundant circuit pattern of the wafer die 14)
Produces a detectable image of defects greater than 4 microns in diameter. The design of lens section 34
A paraxial diffraction-limited light pattern is formed on the Fourier transform surface 46 with asymmetric characteristics.
The reason for the asymmetry is that the diffraction angle of the optical system 68 is relatively large (± 15 ° to 20 °) and
This is because the surface to be imaged having a diameter of 3 mm is appropriately large. Lens section 54
Significantly longer focal length f to achieve 10x image magnification_TwoNeed. Lens classification 54
Design has an asymmetric characteristic and the entrance pupil is located close to the front lens element 72 of lens section 34.
To offset the residual aberration introduced by the lens section 34 and extend the length of the optical system 68.
And the diameter of the lens element incorporated therein is minimized.   The object surface of the lens section 54 is positioned so as to be substantially in contact with the Fourier transform surface 46, and the To obtain a compact spatial filter configuration. Between the lens section 54 and the Fourier transform surface 46,
For introducing the illumination beam through the Fourier transform plane and supporting the spatial filter 50
Sufficient space is required to accommodate the mechanical structure. Optical system 68
-f_Two/ f₁ Is designed to be an inverted image of an object having a magnification of. Where f_TwoIs Len
F at the effective focal length of_Two= 600mm, f₁Is the effective focal point of lens section 34
F at distance₁= 60 millimeters, which results in a magnification “m” of 10.   The lens section is designed to meet the following two performance requirements. First most important
The requirements are for any object on the object plane or 3 mm diameter object located in the front focal plane 42.
Light diffracted into a telecentric cone ± 15 ° -20 ° from point is small enough for aberration
To form a paraxial diffracted light image with extremely small geometric distortion.
To do it. The second requirement is ± 15 ° to 20 ° through a 20 mm diameter object.
° Plane waves propagating in the range have the smallest vignetting and 20 micron
An object of the present invention is to generate a light pattern having a spot diameter equal to or smaller than the spot diameter.   Since the residual aberration introduced into the lens section 54 by the lens section 34 is multiplied,
It is almost impossible to remove this residual aberration by compensation aberration in the lens section 54
It is. Therefore, as a design requirement, the lens section 34 is located on the Fourier transform surface 46.
For incident plane waves over a diffraction angle of ± 15 ° to 20 ° and an entrance pupil diameter of 20 mm.
(1) Being isopranatism (ie, the aberration is a small part of the Fourier transform image field)
The lens is kept constant over a linear
(2) Being essentially aplanat (ie, spherical aberration and
(3) be essentially anastigmat (ie, astigmatism)
(Having a flat image surface without aberration).   To generate a paraxial diffracted light image on image plane 60, lens section 34 must be
For a point-like object located within a 3 mm diameter area in the anterior focal plane 42
It is necessary to generate a plane wave. For this purpose, a main beam within a diffraction angle range of ± 15 ° to 20 ° is used.
Make the rays telecentric during the design of the lens section 34, i.e. parallel to the optical axis 48
It is necessary to make the residual aberration supplied to the lens section 54 extremely small so that it can be compensated.
You.   As a design approach to lens section 34, lens section 34 emanates from an object at infinity Which receives light within the range of ± 15 ° to 20 ° and the entrance pupil is located at the front focal plane 42
And As a result, the Fourier transform pattern is located at the back focal plane 46 of the lens.   In particular, lens section 34 comprises five elements coaxially positioned along optical axis 48.
Including. Element 72 is a biconvex lens, and element 74 is a positive meniscus lens.
The lens is located close to the entrance pupil of the lens system 68 to suppress spherical aberration. Biconcave
The lens 76 suppresses the curvature of field. Biconvex lens 78 and positive meniscus lens 80 are focused
It is located in the light beam to suppress astigmatism. Thus, a diffraction angle of ± 15 ° to 20 °
To achieve the aberration suppression imposed by the range, lens section 34 comprises five elements 72,
Requires 74, 76, 78 and 80, and these elements are made of high refractive index glass.   The design of the lens section 54 is such that the object is located at the front focal plane 42 of the lens section 34 at 3 mm.
Residual spherical aberration introduced by lens section 34 when located within an object of
Coma is to be canceled. The biconvex lens element 70 and the biconcave lens element 84
The residual spherical aberration introduced by the lens section 34 which is located close to the
cancel. The negative meniscus lens element 86 is the frame yield introduced by lens section 34.
Offset the difference. The weak positive refractive power meniscus lens element 88 is the refractive of the lens section 54
Dissipating forces allow lens elements 70, 84 and 86 to cancel residual aberrations from lens section 34
It is possible to introduce light spherical aberration and coma aberration. Lens element 88
Also serves to correct astigmatism in the image plane 60. Weak positive plano-convex lens
The zoom element 90 is located close to the image plane 60 to suppress geometric distortion of the image. But the element
A positive power of 90 weakens field curvature and astigmatism correction. Image quality at image plane 60
Is the higher order introduced by lens section 34 and such cancellation of field curvature and astigmatism.
Barely meet design goals due to limitations imposed by the presence of residual spherical aberration
You.   Tables I and II below show the design specifications for the optical system 68 and the spacing between adjacent elements.
It is.   Table I shows the design values of the elements of the lens section 34 and the spatial filter 50, and Table II shows the lens
The design value of the element of the section 54 is shown. Tables a to w correspond to the surfaces with letter symbols in FIG.
The surface “a” corresponds to the object plane 42, and the surface “w” corresponds to the image plane 60. Surface l₁And l_Two
Corresponds to the spatial filter. The radius of curvature and diameter are shown for each surface. These surfaces are flat surfaces a, l₁, L_Two, v and w are spherical.
The positive radius of curvature of the surface is such that the center of curvature is on the right side in FIG.
The radius of curvature indicates that the center of curvature is to its left. Unit of measurement is millimeter
The width in the width direction up to the next surface is measured from left to right in FIG.
You.   FIG. 5 shows a defect in a semiconductor wafer of the type inspected by the inspection system 10 described above.
FIG. 3 is a schematic diagram of a second preferred embodiment 100 of the inspection system of the present invention designed to detect.
is there. The inspection system 100 of this example substantially satisfies the same performance specifications as the inspection system 10.
It is designed as follows. Inspection system 100 was written to the liquid crystal layer spatial filter 102
Equipped with a folded Fourier transform optical system including a Fourier transform spectral pattern
You. The liquid crystal spatial filter 102 has a regular period of the pattern surface of the semiconductor wafer 12.
Scatters light at the spatial frequency associated with the spatial structure and defects in the pattern surface
Reflects light of the relevant spatial frequency. Is spatial filter 102 a defect in wafer 12?
It reflects the light emanating from it and provides the folded Fourier optics of the inspection system 100.   The inspection system 100 is a substantially parallel beam of 442.5 nanometer monochromatic light 106.
A light source 104 for generating a laser beam, the light beam being incident on a lens 108,
The lens focuses this beam at the center of the aperture of the pinhole spatial filter 112.
Focus on 0. The light beam generated by the laser 104 is linearly polarized in the plane of FIG.
Have been. The beam 114 diverging from the focal point 110 is a plate-type polarizing beam splitter.
At 116, this splitter converts the polarized light into a plane perpendicular to the plane of the paper in FIG.
It reflects, but transmits polarized light in the plane of the paper of FIG. Quarter wave plate 118
Receives the light beam 114 transmitted through the beam splitter 116 and turns it into circularly polarized light. Four
The circularly polarized light beam emitted from the half-wave plate 118 becomes a fairly thin circular beam and becomes a Fourier
Propagating toward the conversion lens section 120. This lens section 120 is a single lens in FIG.
Although shown as a lens element, it is realized by five lens elements as described later.   The effective center of the Fourier transform lens section 120 is focused by the pinhole spatial filter 112.
The parallel circularly polarized light beam 122 is placed at a distance of one time the point distance and the surface of the wafer 12 is patterned.
To project. The wafer 12 is mounted in a chuck 38 on a translation stage 40. Parallel rays
122 illuminates a 20 mm diameter area of the surface of the wafer 12. Wafer 12
It is located in the anterior focal plane or object plane 124 as described above for inspection system 10.   The circularly polarized light beam 126 diffracted in the illumination area of the wafer 12 is divided into a lens section 120 and a quarter.
The wave propagating through the quarter-wave plate 118, and transmitted from the quarter-wave plate 128
A linearly polarized light beam 128 is emitted. This light beam 128 is converted by the polarizing beam splitter 116
The light is reflected in the direction of the quarter wave plate 130, which turns the light beam 128 into circularly polarized light. The circularly polarized light beam 132 exiting the quarter-wave plate 130 is applied to the laser absorption layer of the spatial filter 120.
134. Ray 132 is applied to the back focal plane 136 of lens section 120 on the illumination wafer surface.
Form a Fourier transform pattern.   The spatial filter 102 is the error-free Fourier transform of the illuminated die 14 of the wafer 12
It absorbs and blocks inter-frequency, but reflects and transmits light emanating from defects in this die.
Spend. The spatial filter 102 differs from the spatial filter 50 in two main respects. No.
First, the error-free Fourier transform pattern is applied to the spatial filter 102 in the liquid crystal layer and the spatial filter.
In the filter 50, the data is written on the photosensitive emulsion on the photographic plate. Second, the spatial filter 102
Is a reflection type, and the spatial filter 50 is a transmission type.   The defect-carrying circularly polarized light beam 138 reflected by the spatial filter 102
, And the plate 130 converts the circularly polarized light beam 138 into a linearly polarized light in the plane of FIG.
Change to light ray 140. Ray 140 passes through beam splitter 116 and is inverse Fourier
The light enters the conversion lens section 142. This lens section is shown as a single lens in FIG.
, But includes five lens elements as described below. Inverse Fourier transform lens section
The component 142 performs an inverse Fourier transform of the filtered light pattern. Lens classification
142 is located at a distance of one time the focal length from the back focal plane 136 of the lens section 120. Les
The lens elements of lens sections 120 and 142 have two sections 14 in the plane of beam splitter 116.
6 and 148 are aligned along the same optical axis 144 which is bent. Translation stage
40 moves wafer die 14 across section 146 of optical axis 144.   The photodetector array 58 is centered on the optical axis 144 in the image plane 150, and the wafer on the optical axis is
An image of a defect present on the die is received. Image plane 150 is located at the back focal plane of lens section 142.
Place. Magnification of lens section 142 described for lens section 54 of inspection system 10
For the same reason, the same value as that of the lens section 54 is used. All pattern surface of wafer
Is accomplished in the same manner as described for the inspection system 10.   FIG. 6 is a cross-sectional view of the spatial filter 102, which
To form a laser smectic light valve on which a Fourier transform pattern is written.
You. The laser smectic light valve of the type used in the present invention was
Automation Technology Institute Conference in Montreal, Da
Kahn, Frederic J. et al. Paper “LARGE AREA, ENGIN-EERING D RAWING QUALITY DISPLAYS USING LASER ADDRESSED SMETIC LIQUID CRYSTAL LIGH
T VALVES ".   In FIG. 6, the spatial filter 102 comprises a pair of spaced glass substrates 160 and
And a liquid crystal material 164 between them. Laser absorption layer 166
It adheres to the inner surface of the glass substrate 160. A reflective electrode 168 is attached on the inner surface of the laser absorption layer 166.
Then, the transparent electrode 170 is attached to the inner surface of the glass substrate 162. Director alignment layer 172
It adheres to the inner surface of the reflective electrode 168 and the inner surface of the transparent electrode 170. Liquid contained in cell
The director 174 of the crystalline material 164 has a layered parallel order of the smectic phase. No error
The Fourier transform pattern is written in the spatial filter 102 as follows.   The thinly focused writing laser beam 176 passes through the glass
6 Focus on top. Laser absorption layer 160 absorbs incident laser light and converts it to heat
I do. This heat is rapidly conducted to a local area of the liquid crystal material 164, and the temperature of the local area is sufficiently increased.
This local part is heated to above the critical transition temperature. This temperature rise is about a few degrees Celsius
.   When the temperature of the liquid crystal material 164 exceeds the critical transition temperature, the director 174 is no longer
Without maintaining the layered parallel order of the smectic phase shown in Fig. 6, it is characteristic of ordinary isotropic liquids.
It has a random orientation. Turn off the focused laser beam or move to another writing position
When moved, the previously exposed local heat is dissipated to the glass substrates 160 and 162 and
The local cools rapidly to its ambient operating temperature. Glass substrates 160 and 162 are about 13 micron
The thickness is typically 100 to 500 times the thickness of the liquid crystal material 164. This place
The liquid crystal material 164 reorients the director 174 to a uniformly ordered smectic alignment.
Director 174 cools down at a rate that is insufficient to
Maintain disordered orientation.   The area of the spatial filter heated by the writing laser beam scatters the incident light,
The unheated areas of the spatial filter 102 do not scatter light. Therefore, the written area is empty
The incident light propagating in the direction of the arrow 178 in the inter filter is scattered. This light is scattered
Therefore, it is not focused by the lens section 142 (FIG. 5). Non-spatial filter 102
The writing area acts like a mirror.   The entire filter pattern is transparent to the transparent electrode 170 on the inner surface of the glass
It can be erased by applying an AC voltage between the reflective electrode 168 and the layer 166. it can. In this case, an electric field generated between the liquid crystal materials 164 causes the director 174 to generate an applied electric field.
Align light in parallel and thus scatter light perpendicular to the surfaces of glass substrates 160 and 162.
Not to be.   The error-free Fourier transform is converted into a spatial filter 102 by a laser beam scanning device.
The laser beam 176 is scanned along the surface of the glass substrate 160 and
By irradiating a fixed part, the writing area or light scattering point corresponding to the Fourier transform pattern
It can be written by forming.   The Fourier transform lens section 120 and the inverse Fourier transform lens section 142 are shown in FIG.
Thus, it is designed as a part of one optical system 200, and has a total of ten elements. Les
Lens section 120 is designed to satisfy the following two performance requirements. The first requirement is
From any point on the object with a diameter of 3 mm located at the object plane or the front focal plane 124
Light diffused in a ± 15 ° -20 ° telecentric cone with sufficiently small ray aberration
Parallel to enable the formation of very small geometric distortion paraxial diffraction-limited images
It is in. The second requirement is to use a long back focal length with the face of the beam splitter 116
To be able to fold the system. This optical system 200
The design is based on the minimum spot diameter “d” on the Fourier transform surface 136.₁At the image plane 150
Minimum spot diameter “d_TwoThe complexity is due to the stringent requirements for ". Spot diameter d₁
And d_TwoAnd design parameters for image magnification “m” are described for inspection system 10.
It is the same as solid.   To achieve a long back focal length, lens section 120 comprises five elements 202, 204, 206
, 208 and 210 are required.   Lens section 120 is in the form of a Berthele eyepiece. This is
Eyepieces have a long back focal length and a large aperture ratio
. Elements 204 and 206 are the strong negative refractions required on the input side of the Barcel eyepiece
A negative meniscus lens that gives power.   The lens section 120 forms a Fourier transform pattern from the incident plane wave, and
Cooperates with the lens section 142 to form an image ten times larger than the wafer 12 located at the front focal plane 124.
You. These two functions place severe constraints on the design of lens section 120. The reason is,
Dominant light of the fan beam associated with the diffraction energy producing the Fourier transform pattern This is because the line becomes a light ray constituting the on-axis image point of the enlarged image. Under these conditions
The magnified image shows that the chief ray is not exactly parallel to the optical axis 144, ie not telecentric
Sometimes exhibit extremely strong spherical aberration. The chief ray follows a sine relationship with the Fourier transform surface 136.
(Ie, the height of the intersection is the diffraction angle of the focal length of lens section 120)
If not equal to the sine factor), the magnified image shows very strong coma. This frame
If the magnitude of aberration and spherical aberration is not corrected in the lens section 120, the lens section 142
It is not possible to correct for.   The residual spherical aberration has a spherical aberration correction surface located downstream of the beam splitter 116.
Is corrected by the flat plate 212. The aspheric correction plate 212 removes substantially all of the spherical aberration.
However, due to its position and shape, coma aberration that is not very strong in the magnified image is introduced.
You. This coma is removed by the lens section 142.   Lens section 142 consists of five elements 214, 216, 218, 220 and 222,
The element is preferably mounted in a long tube. Elements 214, 216 and 218 are aspheric
It operates as a triplet located immediately after the correction plate 212. Elements 214, 216 and 21
8 bending mainly corrected for coma introduced by the aspherical correction plate 212
And select their power distribution and glass type in Petzval
Choose to suppress the song. Elements 220 and 222 correct system residual astigmatism
I do.   Tables III and IV show the design specifications of the optical system 200 and the spacing between adjacent elements.
is there. Table III shows the elements of lens section 120, quarter wave plates 118 and 130, beams
The design values of the splitter 116, the spatial filter 102, and the aspherical correction plate 212 are shown.
The design values of the elements of the lens section 142 are shown. Surfaces a to 11 are provided with the letters and symbols shown in FIG.
Surface "a" corresponds to the object plane, surface "ll" corresponds to the image plane 150
I do. Surfaces l-aa are quarter wave plates 118 and 130, beam splitters, spatial filters.
It corresponds to the filter 102 and the aspherical correction plate 212. Some surfaces have two symbols
, One symbol represents the surface on which light propagating toward the spatial filter 102 is incident.
, The other symbol represents the surface on which the light reflected by the spatial filter 102 is incident.
The radius of curvature and the diameter are shown for each surface, and these surfaces are plane surfaces a,
It is spherical except for the surface z which is aspherical with l-y, aa and ll. Surface positive song
rate The radius indicates that the center of the curvature is on the right side of the figure (FIG. 7), and the negative radius of curvature is
The center of the curvature is shown on the left side of the figure. Dimensions are in millimeters,
The axial distance to the next surface is measured in the positive direction from left to right in FIG.
You.   Beam splitter 116 undergoes an eccentricity of 45 ° rotation in a plane perpendicular to the plane of FIG.
. The aspheric correction plate 212 has an eccentricity of -2.1642 mm vertically downward in FIG.
receive. Eccentricity defines a new coordinate system (a coordinate system that has been moved and / or rotated).
A subsequent surface is defined in the coordinate system. Eccentric surface is local to new coordinate system
Aligned along a suitable mechanical axis. The new mechanical axis is changed by another eccentricity
Used until   Since the defect inspection techniques of the detection systems 10 and 100 are the same, the following description
Perform only for system 10. Referring to FIGS. 1 and 8, defects in wafer 12 are shown.
The presence of a predetermined threshold amount located in the inspection area 250 of the defect image field in the image plane 60.
It is determined by detecting the region of the light intensity exceeding. The inspection area 250 is shown in FIG.
Included in the outline 252 shown by the dashed line determined by the outer edge of the die adjacent to the periphery of the wafer 12.
It is an area to be filled. Since the lens section 54 gives 10 times magnification, the defect image in the image plane 60
The field of view has an area 100 times the area of the inspection area 250.   Determining the presence of defects in the wafer 12 requires the wafer to be stripped 1.0 mm wide.
And the parallel moving stage 40 is moved in a raster scanning manner to
The entire surface of the wafer 12 is scanned by the emitted 20 mm × 20 mm spot.
Achieved by tripe illuminating sequentially. Photodetector array 58 is centered on optical axis 48
And has a 10 mm x 10 mm light sensitive surface 62,
The photodetector array is on the 1.0mm x 1.0mm optical axis of the wafer 12.
Only the aberration-free inverse Fourier transform pattern corresponding to the illumination region is detected. (Len
The 10x magnification given by the size section 54 is 1.0 mm x 1.0 mm
The dimensions of the illumination wafer area correspond to the detection area of 10 mm x 10 mm.
Let )   The translation stage 40 moves the wafer 12 in a plane 42 by 20 mm × 20 mm.
An XY positioning table that can be positioned with respect to the light beam 36 of the camera. Translation
The upper or Y-stage 256 of the moving stage 40 supports the chuck 38 and places the wafer 12 in the plane 42.
And move it in the Y direction. The lower stage of the translation stage 40 or the X-stage 258 flattens the wafer 12.
It is moved in the X direction within the surface 42. A preferred example of the XY positioning table is Ken.
Jington Laboratories (Richmond, California, USA) Model 8500.   The control circuit (not shown) of the translation stage 40 divides each stripe region of the wafer with the optical axis 48.
And the wafer 12 is moved at a constant speed. Translation stage 40
Is the defect image in the image plane 60 for the position of the translation stage 40 and a known position on the wafer 12.
Position information indicating the position of the defect in the area. Detection of defects in the image is described below
This is done according to a time delay integration technique.   9A and 9B show the outline of the lower left corner of the wafer of FIG.
Raster scan of region 254 and translation stage 40 moving along these stripe regions
It is an enlarged view of each path. FIG. 10 shows the stripe region in FIG. 9B.
FIG. 4 is an enlarged view of a portion of the photodetector element forming a light sensitive surface 62 of a photodetector array 58;
The dimensions of the element 260 and the lens section 54 have a magnification of 10 times, so that
It shows a one-to-one correspondence between a pixel 262 having one dimension. Photo detector
Array 58 consists of 206336 photodetectors arranged in 430 rows × 512 columns as described below.
Have children.   9A, 9B and 10, the photodetector array 58 includes a light window 26 through which light to be detected passes.
With 4. The light window 264 has sides 266 and 268 that define the length and sides 270 and 270 that define the width.
And 272. In general, the light window 264 is fixedly arranged so as to be aligned with the optical axis 48.
You. Due to the movement of the wafer 12, the defect image field of view showing an enlarged image of the inspection area 250 closes the optical window.
Move across. However, for clarity, the following description
The 12 inspection zones 250 will be described as moving across the light window 264.   During normal scanning operation, the translation stage 40 moves the wafer 12 across the light window 264 in the X direction.
To move, side 272 of light window 264 aligns with line segment 278 of inspection area 250. Light window 26
The movement of the wafer 12 across 4 is from the effective starting position 276 in FIGS. 9B and 10.
A path segment 274a extending rightward is defined along the stripe region 254. Of light window 264
Sides 266 and 268 are parallel to the Y-direction, and these sides cause the detector array 58 to move in the X-direction.
Area 254 (section 9B) corresponding to the portion of the inspection area 250 traveling across
(Three shown in the figure) are determined.   After the line segment 282 of the inspection area 250 passes through the side 266 of the light window 264, the translation stage 40
Wafer 12 is moved along a retrace path segment 284a extending to the left in FIGS. 9B and 10. To a starting position 286 for scanning the second adjacent stripe region 254.
. During retrace, the Y-stage 256 divides the wafer 12 into the width of the stripe region 254 (ie, 1.0 mm).
), And the X-stage 258 moves the wafer through the path segment 27.
Move by a distance equal to the length of 4a. After returning, the translation stage 40 exits the wafer 12.
From the starting position 286, move along the path segment 274b in the X direction to
Scan region 254.   The above scanning and retrace operations are repeated until the entire inspection area 250 crosses the optical axis 48.
. However, the length of the scanning and retrace path segments may vary depending on the difference in the X-direction dimensions of the inspection area 250.
Make the difference.   Referring specifically to FIG. 10, the photodetector array 58 of the preferred embodiment includes a row 290 and a column 292.
Coupled RCA model 6220-044 including array 288 of photodetectors 260 arranged in a matrix
Device. Array 288 has 403 rows and 512 columns of photodetectors. Row 290
Are a group of elements 260 arranged in a straight line perpendicular to the inspection direction (ie, in the Y direction).
292 is a group of elements 260 arranged in a straight line parallel to the scanning direction (ie, in the X direction).
. Each row 290 and each column 292 has a size of 6.45 milliliters and 10.24 millimeters, respectively.
Have a length. Each photodetector 260 is 16 microns long and 20 microns wide.
Thus, the width of each stripe region 254 is equal to the total length of a row of 403 photodetectors.
. Each photodetector 260 emits from a portion of the inspection area 250 that is aligned therewith through a light window 264.
Incident on the potential well, and the intensity of the
A corresponding amount of charge or measured energy value is stored.   Each stripe area 254 in the inspection area 250 is detected by the enlargement processing of the lens section 54
It is divided into an array 294 of pixels 262 having the same dimensions as the element 260. Array 294
Of pixels 262 are arranged in rows 296 and columns 298, each row having 403 pixels, and each column having
It has a number of pixels determined by the length of the stripe region 264. Within the stripe area
Of light crosses the inspection area 250 across the light window 264 along each stripe region 254
And move the energy value corresponding to the light intensity of each pixel according to the method described below.
Detected by taking out.   The X-stage 258 moves the wafer 12 from the departure position 276 to the left in the X direction to the side of the light window 264.
268 is the inspection area The scanning process is started by accelerating the movement until it coincides with the 250 line segment 300.
X-stage 258 then moves wafer 12 along stripe region 254 at a predetermined constant speed.
Move.   The detector elements 260 in the first row 290a of the array 288 are aligned with the pixels in the first row 296a of the array 294.
When queuing, the following occurs: The potential of each of the photodetectors 260 in row 290a
A charge is generated in the luwer. The charge amount corresponds to the light intensity existing in the corresponding pixel
(The potential well of the photodetector 260 is not scanned before scanning the stripe region 254.
Does not accumulate). Row transfer clock supplied to each row 290 of array 288
The signal causes the charge from each photodetector element 260 in row 290a to be transferred to the next adjacent second row 290b.
Are transferred to the photodetectors in the same column 292. In this transfer, the photodetector and the pixel
(X-stage 258 moves wafer 12 along stripe region 254).
And continuous movement, so there is no noise resulting from aliasing between adjacent pixel rows.
There is a visible amount of image degradation). After the charge transfer from row 290a to row 290b, the light in row 290a
No accumulated charge exists in the potential well of the detection element.   Photodetectors 260 in second row 290b are aligned with pixels 262 in second row of array 294
And the following happens. Potential well of each photodetector 260 in rows 290a and 290b
A charge is generated inside. Charge generated in each photodetector 260 in row 290b is transferred before
Added to the charge. Therefore, the charge amount of the photodetector 260 in the row 290b is
2 of the energy value corresponding to the intensity of light present in the pixel in each column of row 296a of ray 294.
Represents double. A row transfer clock signal is output from each of the photodetectors 260 in rows 290b and 290a.
The charge is transferred to the next adjacent photodetector in the same column 292 in the third row 290c and 290b, respectively.
Transfer to   (1) Each light detection element 260 in the row 290 has the light intensity of the pixel 262 where each light detection element is aligned.
And (2) this energy value in the next adjacent row 2
The above-described process of transferring to the photodetectors in the same column 292 in 90 is performed by the row transfer clock signal.
Repeated for 255 cycles.   After the completion of 255 such line sequential transfers, the 256th or last line 290 of the array 288
The photodetectors in d align with the pixels in the first row 296a of array 294. In the first row 296a
The previous 255 accumulated energy values for each pixel in It is added to the 256th energy obtained by the child 260. 256th row transfer
Prior to the generation of the lock signal, 512 photodetectors 260 corresponding to pixels 262 in row 296a
Energy values are read out in series by the high-speed data transfer clock signal.
You. The accumulated energy value for each pixel 262 is converted to digital form and the threshold
The amount of light present at each pixel 262 as processed by the value detector corresponds to the corresponding amount in wafer 12.
Determine whether to indicate that the location is defective.   When the 256th cycle of the row transfer clock signal occurs, each pixel in the second row 296b
The accumulated energy value of 255 before 262 is used for each photodetector 26 in the last row 290d.
Added to the 256th energy value obtained by zero. Row transfer clock signal
Before the occurrence of the 257th cycle, 512 photodetectors corresponding to the pixels in row 296b
The contents of 260 are read and processed as described above.   Scanning of the stripe region 254 continues during successive cycles of the row transfer clock signal.
256 energy for each pixel 262 in row 296 and column 298 of array 294
The values are accumulated in the photodetectors 260 in the corresponding columns 292 and rows 290d of the array 288.   There are several features in the accumulation of energy values characteristic of the scanning process described above.
First, each photodetector in row 290a determines an energy value for one pixel 262.
Do not accumulate more than once. Second, currently aligned with pixel 262 in a particular row 296
The photodetector element 260 in row 290 is the next row aligned with pixel 262 in this particular row 296.
, The accumulated energy value is always larger by one time than that of the light detection elements 260 in the row 290 adjacent to.
Third, each photodetector in row 290d is exposed to the light present at pixel 262 where the element is aligned.
The corresponding energy value will be accumulated 256 times.   When the edge 282 of the inspection area 250 completely passes through the side 266 of the light window 264, a stripe is formed.
The scanning of the area 254 is completed, and the accumulated energy value of the pixel in the last row 296d of the array 294 is obtained.
Are read from the photodetectors in the last row 290d of the array 288. X-stage 258 is wafer
12 is decelerated and stopped at the stop position 302 (FIG. 11 shows the inspection area at this position).
Light window 264 for region 250 is shown in dashed lines). Then the X-stage 258 and the Y-stage 25
6 and the wafer 12 are returned along the path segment 234a to the starting position 286. This
During the period of time, the potential well of the photodetector 260 is cleared and the next adjacent stratum is cleared.
A preparation is made for scanning of the ip area 254. Running of the second and subsequent stripe regions 254 The search and retrace proceed as described above.   One skilled in the art can make many variations to the preferred embodiment of the invention described above.
It is clear that. For example, defect inspection of photomask instead of semiconductor wafer
can do. However, in this case, the inspection systems 10 and 100 are
Need to be changed to transmit through the photomask. As a second example, the inspection system
Instead of the plate-shaped beam splitter 116 used in the system 100, a cubic
An optical beam splitter can be used. Cube beam splitter reflects light
The background noise resulting from this is reduced, but the design values of the lens
There is a need to reduce the spherical aberration introduced by the muscle splitter.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing an optical system of a first preferred embodiment of a defect inspection system of the present invention, and FIG. 2 is suitable for defect inspection by the systems of FIGS. 1 and 6. And FIG. 3 is a simplified diagram showing the asymmetry of the Fourier transform and inverse Fourier transform lens systems incorporated in the defect inspection system of FIG. FIG. 4 is a diagram showing an optical element of the lens system shown in FIG. 3, FIG. 5 is a diagram showing an optical system of a second preferred embodiment of the defect inspection system of the present invention, and FIG. FIG. 7 is a sectional view of a spatial filter used in the defect inspection system of FIG. 7, FIG. 7 is a diagram showing optical elements of a Fourier transform and inverse Fourier transform lens system incorporated in the defect inspection system of FIG. 5, and FIG. The presence of defects in the semiconductor wafer and 9A is a partially enlarged view showing three stripe regions at the lower left corner of the semiconductor wafer shown in FIG. 8, and FIG. 9B is a partially enlarged view showing the stripe regions shown in FIGS. 8 and 9A. FIG. 10 is an enlarged view showing a raster scanning path by the scanning mechanism of FIG. 8 for detecting a defect image in a defect image plane relative to a photodetector. FIG. 10 shows a case where the magnification is 10 times. FIG. 2 is a diagram showing an array of pixels in a defect image plane and an array of photodetectors of a charge-coupled device used in the present invention. 10 Defect inspection system 12 Wafer 14 Die 16 Redundant circuit pattern 22 Laser light source 34 Fourier transform lens 40 Parallel movement stage 50 Spatial filter 54 Inverse Fourier transform transform lens 58 Photodetector array 100 Defect inspection system 102 Spatial filter 104 Laser light source 116 Beam splitter 118 λ / 4 plate 120 Fourier transform lens 130 λ / 4 plate 142 Inverse Fourier transform lens 250 Inspection area 254 Stripe area 260 Photodetector 264 Optical window 288 Photodetector array 262 pixels 294 pixels array

Claims (1)

[Claims] 1. A first lens and a second lens disposed along the optical axis, wherein the first lens generates a spatial frequency spectrum having frequency components that can be selectively filtered from the subject, and the second lens generates a spatial frequency spectrum in the subject. In an imaging system for generating an image of an existing defect, a die circuit pattern aligned with an optical axis and a proximity to the optical axis for detecting a defect in a subject including at least one die having multiple redundant circuit patterns. Circuit pattern putter
Illuminating a plurality of die circuit patterns including a down causes a light pattern representing substantially the Fourier transform pattern of the illuminated plurality of die circuit patterns to generate light patterns including the interference pattern information die, the optical path
An illuminated die that aligns the light collected by the first lens with the optical axis during the turn
With the circuit pattern and the illuminated die circuit pattern located close to the optical axis
Off-activities introduced into the light pattern when the light pattern is generated by limiting the diffracted light
Receiving an optical filter having a transparent portion and an opaque portion corresponding to the Fourier transform pattern of the error-free reference pattern corresponding to the plurality of die circuit patterns to minimize the cis aberration; A spatial frequency component not blocked by the optical filter, the number of circuit patterns being smaller than the number of circuit patterns in the die existing in a predetermined spatial region intersecting the optical axis. lens guiding the collection space frequency components by condensing the corresponding spatial frequency component
Defect inspection characterized by forming an image of a defect by limiting it to one containing almost no incident aberration, and processing the unrejected spatial frequency components in the die to determine the position and size of the defect in the die. Method. 2. By changing the position of the subject with respect to the position of the optical axis, and positioning a predetermined number of different die circuit patterns in a spatial region intersecting the optical axis, and processing the in-die spatial frequency components of these die circuit patterns. The method of claim 1, wherein: 3. Processing of unblocked intra-die spatial frequency components is performed by placing the light-sensitive surface of the photodetector about the optical axis, the light-sensitive surface corresponding to the focused spatial frequency component.
2. The method according to claim 1, wherein the area is smaller than the image area where the image of the defect appears . 4. The first and second lenses cooperatively receive light diffracted by the illuminated die circuit pattern and form an image from spatial frequency components corresponding to the die circuit patterns. Item 7. The method according to Item 1. 5. The first lens comprises a first lens section comprising a plurality of elements, the second lens comprises a second lens section comprising a plurality of elements, and the first and second lens sections comprise an asymmetric paraxial diffraction-limited lens. 5. The method according to claim 4, wherein said system comprises a system. 6. 2. The method of claim 1 wherein the illumination means generates substantially parallel light. The method of claim 1, further comprising: defining a plurality of adjacent stripes on the object, each including a series of adjacent dies, and applying the object and the illuminated parallel light to each stripe. Are illuminated a predetermined number of die circuit patterns at a center position with respect to the optical axis by moving relatively along the length of the optical circuit, and the corresponding die circuit patterns at the center position with respect to the optical axis are not blocked. A defect inspection method characterized by processing spatial frequency components. 7. 7. The method according to claim 6, wherein the object is movable, and the illumination parallel light is fixed with respect to the optical axis. 8. The method of claim 1 wherein the spatial frequency components to be collected correspond to less than all of the illuminated die circuit patterns. 9. An optical system for detecting a defect in an object of a type including at least one die having a plurality of redundant circuit patterns in a first portion, the optical system having an optical axis, and
One said a portion of the first portion of the subject, the die circuit patterns及aligned with the optical axis
A plurality of die circuit patterns including the die circuit patterns located proximate to a fine optical axis exist
A pattern generating means for generating a light pattern comprising illuminating means for illuminating a second portion, the die in the interference pattern information to a light pattern representing substantially the Fourier transform pattern of a plurality of circuit pattern is illuminated to standing Including a lens having a relatively small object field diameter,
The illuminated die circuit pattern aligned with the optical axis by
Generates a light pattern by condensing the light diffracted by the illuminated die circuit pattern
A pattern generating means for minimizing off-axis aberrations sometimes introduced into the light pattern , a transparent portion, and an opaque portion corresponding to a Fourier transform of an error-free reference pattern corresponding to the multiple die circuit patterns, an optical filter for blocking the spatial frequency component by receiving light pattern, the spatial frequency components not blocked by the optical filter with focusing means for focusing
A plurality of circuit paths in the die existing in a predetermined spatial region intersecting the optical axis.
Focuses spatial frequency components corresponding to fewer circuit patterns than turns
Condensing means for limiting the wave number component to one containing almost no lens-introduced aberration, and processing means for processing only the unblocked spatial frequency components in the die to determine the position and size of the defect in the die. An optical system for detecting defects. 10 ． The illuminating means emits substantially parallel light, and the optical system further changes the position of the subject with respect to the position of the illuminated parallel light to illuminate a plurality of different die circuit patterns with the illuminated parallel light. 10. The optical system according to claim 9 , further comprising positioning means positioned at the second portion of the sample to process the in-die spatial frequency components of these die circuit patterns. 11 ． The pattern generating means and the light condensing means are constituted by separate first and second lenses arranged along an optical axis intersecting a second part of the subject illuminated by the illuminating means. The optical system according to claim 9, wherein: 12 ． Said pattern generating means and the focusing means receives light diffracted by the plurality of die circuit patterns illuminated, the first and second from the spatial frequency component of each other for forming an image corresponding to these die circuit patterns 10. The optical system according to claim 9 , wherein the optical system comprises a lens. 13 ． The first lens comprises a first lens section comprising a plurality of elements, the second lens comprises a second lens section comprising a plurality of lens elements, and the first and second lens sections comprise paraxial diffraction-limited elements having asymmetric properties. 13. The optical system according to claim 12, comprising a lens system. 14 ． The first lens comprises a first lens section comprising a plurality of elements, the second lens comprises a second lens section comprising a plurality of lens elements, the first lens section forming a Fourier transform pattern and the second lens section. 13. The optical system according to claim 12 , wherein the optical system is adapted to provide an enlarged image of a defect in the plurality of die circuit patterns illuminated in cooperation. 15 ． The pattern generating means and the condensing means are configured by a folded Fourier transform optical system that receives light diffracted by a plurality of illuminated die circuit patterns and forms an image from spatial frequency components corresponding to the die circuit patterns. The optical system according to claim 9, wherein: 16 ． The optical system of claim 9, wherein the optical filter comprises an exposed photosensitive medium. 17 ． The optical system according to claim 9, wherein the Fourier transform pattern represents a Fourier transform image. 18 ． 10. The optical system according to claim 9 , wherein the subject is a semiconductor wafer. 19 . The optical system according to claim 9, wherein the second portion is significantly smaller than the first portion. 20 . The illuminating means emits substantially parallel light, and the processing means comprises a light detector having a light sensitive surface disposed about the optical axis, the light detectors being arranged in a first array of rows and columns. And a plurality of adjacent stripe areas each including a plurality of pixels arranged in a second array of rows and columns in the light pattern, and each of the plurality of light detection elements is provided. The element generates a measurement energy corresponding to the amount of light present in any one of these pixels, and the optical system further moves the subject with respect to the illuminating parallel light to cause the photodetector to emit the light. By scanning along the stripe region of the pattern, each photodetector in one row of the first array sequentially traverses each pixel in one row of the second array to obtain an energy value corresponding to the amount of light present in each pixel. Move to make Means for accumulating a total energy value proportional to the sum of energy values obtained for each pixel by all the photodetectors in one column of the first array; and calculating the light amount of the pixel from the total energy value. 10. An optical system according to claim 9 , further comprising: means for determining whether or not a defect in the object is represented. 21 . 21. The optical system according to claim 20, wherein the photodetector is a charge-coupled device. 22 . 2. The method of claim 1, wherein the parallel illumination light is fixed, and the moving means sequentially scans each stripe region across the light sensing surface of the photodetector.
20. The optical system according to 20 . 23 . 23. The optical system according to claim 22, wherein said moving means moves each stripe region continuously across the illumination parallel light. 24 . 21. The optical system according to claim 20, wherein said first array has a first row and has a total of N rows, further comprising position detecting means for detecting a position of the first array with respect to the stripe region. Cooperates with the accumulating means to accumulate the energy value of the pixel only once when each of the photodetectors in the first row of the first array is aligned with any one pixel of the second array, 21. The optical system according to claim 20, wherein when each of the photodetectors is aligned with an arbitrary pixel of the second array, the energy value of that pixel is accumulated N times.