JP2003035678A - Defect-inspection optical system and surface-defect inspection apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a defect-inspection optical system and a surface-defect inspection apparatus wherein a recess defect and a protrusion defect can be discriminated and inspected with high accuracy regarding an irregular defect on the surface of a face plate. SOLUTION: A light receiving system which comprises n pieces of arranged light receiving elements and in which images formed along a direction at right angles to a main scanning direction become the arrangement direction of the light receiving elements is installed, and an upper-limit peak and a lower-limit peak are obtained regarding the recess defect and the protrusion defect in detection signals of the light receiving elements in the arrangement direction of the n pieces of light receiving elements. According to the relationship between the upper-limit peak and the lower-limit peak in the arrangement direction of the light receiving elements, the recess defect and the protrusion defect are discriminated, and the defect is detected.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a defect detection optical system and a surface defect inspection apparatus, and more particularly to a surface defect inspection apparatus for a surface plate such as a magnetic disk or its glass substrate (glass substrate) or wafer. The present invention relates to a defect detection optical system and a surface defect inspection apparatus capable of accurately detecting a concave defect and a convex defect by distinguishing between them.

[0002]

2. Description of the Related Art A hard magnetic disk used as a recording medium of a computer system is a substrate disk (substrate) as a material, or a magnetic disk coated with a magnetic film (for convenience, these are collectively referred to as a magnetic disk or simply a disk. In this step, the defects existing on the surface and their sizes are inspected. In recent years, the size of the disk has become 3.3 inches or less, and the recording density thereof has been dramatically increased by the adoption of the GMR head. In this type of disc, a glass disc having a smaller coefficient of thermal expansion is used from an aluminum substrate, and the thickness thereof is as thin as about 0.6 mm to 0.8 mm.

A conventional magnetic disk surface defect inspection apparatus uses a sensitivity calibration disk to detect the projection angle of a laser beam of a light projecting system and the light reception in order to detect well the size of a concave defect and a convex defect. Detects the light receiving angle of the system, the voltage applied to the light receiver, the gain of the amplifier built in the signal processing circuit that receives the detection signal from the light receiver and detects defects, the threshold voltage for noise removal, the laser output of the laser light source, etc. Optimal setting is made for each factor related to sensitivity. Sensitivity calibration discs can be prepared by using a disc-shaped defect of known size, an actual disc with sample defects such as pit defects and scratch defects, or an actual disc with protrusions of a specific height. Has been done. As an invention of this kind, the applicant of the present invention, JP-A-10-32
There is an application for No. 5713 "Surface defect inspection method and inspection apparatus".

However, in the above-mentioned conventional defect detection system, by comparing the level of the reflected light or scattered light detected by the photodetector with the reference level, a concave defect or a convex defect (including foreign matter), etc. Therefore, the detection of the size of the uneven defect to be detected is substituted by the received light level, which is not accurate.

Recently, it has been required to measure the shape of irregularities and improve the classification accuracy. However, with the technique as described above,
There arises a problem that accurate classification cannot be performed. Therefore, as a technique for accurately detecting a concave defect and a convex defect, the photodetector is an array sensor composed of multiple APD elements, and a staggered stripe pattern is provided in front of the photosensor so as to correspond to the elements, and the amount of light received by adjacent photodetectors is increased. The applicant has applied for a technique for detecting a concave defect and a convex defect from the difference of Japanese Patent Application No. 11-358769 "defect detection optical system and surface defect inspection device", and also has a staggered stripe pattern. As a technique not used, Japanese Patent Application No. 2000-397091 "defect detection optical system and surface defect inspection device" has also been filed by the applicant.

[0006]

However, in the above-mentioned prior art, in order to obtain the detection signals having different waveforms for the concave defect and the convex defect, two rows of array sensors composed of APD multi-elements are arranged in parallel. It is necessary. Then, the concave defect and the convex defect are distinguished by which of the left and right light receiving elements in the two rows first receives the light and generates the detection signal.
In order to detect various kinds of defects with high accuracy, the number of elements of the light receiver is arranged in a range of ten to several tens. Therefore, the above-mentioned prior art has a drawback that a large number of circuits for receiving signals from adjacent elements in the width direction are required. The purpose of this invention is
In order to solve the problems of the conventional technology,
It is an object of the present invention to provide a defect detection optical system and a surface defect inspection apparatus capable of accurately detecting a concave defect and a convex defect with respect to an uneven defect on the surface of a face plate.

[0007]

The features of the defect detection optical system and the surface defect inspection apparatus of the present invention for achieving such an object are that a light beam having a width in a direction perpendicular to the main scanning direction is used. A light projecting system that irradiates and relatively scans the face plate, and n (where n is an integer of 2 or more) light receiving elements in which light receivers are arranged, and n images of the scanning position of the face plate are received. And a lower limit peak value of the n detection signals of the n light receiving elements at a certain scanning position. When the difference between and is a predetermined value or more, the defect is detected by determining whether it is a concave defect or a convex defect according to the relationship between the upper limit peak and the lower limit peak in the arrangement direction of the light receiving elements.

[0008]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention has n light receiving elements arranged as described above, and the image formation along a direction perpendicular to the main scanning direction is in the arrangement direction of the light receiving elements. A concave defect or a convex defect is detected by providing a light receiving system. Such a concave defect or a convex defect causes unevenness in the reflected light of the image at the scanning position in the arrangement direction of the n light receiving elements due to the respective lens effects. This sparse and dense state causes an upper limit peak and a lower limit peak in the detection signals of the respective light receiving elements in the arrangement direction of the n light receiving elements.

Therefore, when the difference between the upper limit peak value and the lower limit peak value of the n detection signals of the n light receiving elements at a certain scanning position is a predetermined value or more, it is a concave defect or a convex defect. As a result, the defect can be detected by determining whether the defect is a concave defect or a convex defect according to the relationship between the upper limit peak and the lower limit peak in the arrangement direction of the light receiving elements. As a result, the concave and convex defects on the surface of the face plate are accurately distinguished and detected.

[0010]

1 is a block diagram of an embodiment of a surface defect inspection apparatus to which the present invention is applied, and FIG. 2 is a direction parallel to the paper surface (θ direction) in the drawing of a detection optical system and a paper surface in the drawing. And FIG. 3 is an explanatory view of the development of the light emitting and receiving system with one direction (R direction) perpendicular to the direction, FIG. 3 is an explanatory view of the relationship between the image formed on the light receiving surface of the inspection area and the APD array sensor, and FIG. FIG. 5 is an explanatory diagram of the detection circuit, FIG. 5 is an explanatory diagram of a dense / dense state due to a lens effect of specularly reflected light of a concave defect, and FIG.
FIG. 6 is an explanatory diagram of a detection waveform in the lens effect in FIG. In FIG. 1, 100 is a surface defect inspection apparatus, and 50 is a detection optical system thereof. Reference numeral 51 is a projection system of the detection optical system 50, and the projection system 51 is a laser light source 5
The laser beam L _T from the beam expander 51
2 direction, one direction perpendicular to the paper surface in the drawing (R direction)
By expanding the laser beam, the focus position is moved toward you. The beam expanded here passes through the cylindrical lens 513 and the focusing lens 514, and is offset by Δd (FIG. 2) in the beam waist in the radius R direction.
(See (a)), the focal position is shifted toward the front and focused on the surface of the disk 1. As a result, the beam spot _Sp expands to an elliptical shape (see FIG. 2B).

In FIG. 1, a disk 1 to be inspected is mounted on a spindle (not shown) of a rotating mechanism and rotated by driving a motor. The disk 1 moves in the X-axis direction that coincides with the radial direction (R direction). This X
Spot S _P by the movement in the axial direction, moves in the direction of the disk 1 R, thereby the surface of the disk 1 is scanned spirally. In this case, in order to minimize the scanning time, the spot S _P is an elliptic shape having a minor axis and a major axis, as shown as FIG. 2 (b) to the lower, set at right angles to the major axis to the scanning direction Then, a method of widening the scanning width is adopted.

In FIG. 1, in a direction parallel to the plane of the drawing (the disc rotation direction, that is, the θ direction), the light beam that has passed through the cylindrical lens 513 passes through the focusing lens 514 as shown in FIG. It is focused on S in a dot shape (refer to ♦ in FIG. 2B). The projection angle is about 30 ° with respect to the normal line on the disk 1 in the θ direction, as shown in FIG. Then, the inspection point S becomes a light having a constant width W (for example, about 150 μm to about 200 μm) in a direction perpendicular to the paper surface (R direction), and this width W indicates the array range of the light receiving elements on the image plane. It is long enough to cover.

The light receiving system 52 is a high resolution condenser lens 521.
With this lens, the inspection point S of the disc 1
The regular reflection light from is received, and it is converted into parallel light and guided to the imaging lens 522. The imaging lens 522 forms an image of the inspection point S on the light receiving surface of the APD array sensor 523. The APD array sensor 523, as shown in FIG.
The light receiving elements are arranged in a number n (for example, 23) along the image formation in the radius R direction of the disk 1 (direction perpendicular to the paper surface of FIG. 1). An image of the inspection point S is formed on the APD array sensor 523 so that the width W (about 150 μm to about 200 μm) described above covers the range of the array length of the light receiving elements of about 120 μm. At this time, the width of each light receiving element in the arrangement direction of each element is, for example, 0.5 mm on the light receiving element and about 5 μm on the surface of the disk 1.

FIG. 2A is a development explanatory view of the light emitting and receiving system for explaining a direction parallel to the paper surface (θ direction) and one direction perpendicular to the paper surface (R direction), and the upper side is parallel to the paper surface. It is the θ direction, and the lower side is the R direction perpendicular to the paper surface. As shown in FIG. 2A, in the θ direction, the laser beam L _T from the laser light source 511 passes through the beam expander 512, the cylindrical lens 513, and the focusing lens 5.
The surface of the disk 1 is focused by 14 as a spot at the inspection point S as a spot (see ♦ in FIG. 2B). The specular reflection light from the inspection point S is a high-resolution condenser lens 521,
The light is received by each light receiving element of the APD array sensor 523 via the imaging lens 522.

On the other hand, as shown in the lower side of FIG. 2A, in the radius R direction, the laser beam L _T from the laser light source 511 passes through the beam expander 512, the cylindrical lens 513, and the focusing lens 514 to the disc. An elliptical shape having a width W is focused on the inspection point S with a position in front of the surface of No. 1 by an offset Δd as a focal point (see FIG. 2B). The specularly reflected light from the inspection point S passes through the high-resolution condenser lens 521 and the imaging lens 522, and AP
Each light receiving element 523a, 523 of the D array sensor 523
The image of the inspection point S is received along the arrangement direction of b, 523c, ..., 523n.

As shown in FIG. 1, each of the light receiving elements 523a to 523 of the APD array sensor 523 arranged in the R direction by n pieces.
Detection signal (output signal) obtained from each of 23n
Are input to the defect detection unit 400. Here, data relating to the defect F and its depth or height is generated. The defective data is input to the data processing device 410 including the MPU 411, the memory 412 and the like. Then, in the data processing device 410, the area of the defect F is calculated, and the size is classified according to the calculated area together with the depth or the height. Then, the result is printed out by the printer (PR) 45 together with the position of the defect on the disc 1. Further, they are displayed together with the position on the disk on the display (CRT) 413 and the like, and the count value is also displayed separately.

Here, each of the n light receiving elements 523a to 523 is provided.
The detection signal output by 3n and the offset Δd in the radius R direction
The relationship with (see FIG. 2A) will be described with reference to FIG. Due to the relationship between the size and the inspection point S of the detected concave defect or convex defect, a lens effect due to the defect appears in the reflected light. Therefore, the offset Δd is determined so that this lens effect appears. For example, 100 μm to 2
Assuming that the detection of a recess defect of a size of 00 μm is the main, the inspection point S has an elliptical major axis of 200 μm as described above.
The offset Δd is set so as to be approximately the same.

At this time, each of the light receiving elements 523a ...
At 523n, as shown in FIG. 3, the imaged image of the inspection region S is imaged in the R direction so that the center of the image is the center of the inside of the light receiving region of each element and does not protrude from the light receiving region. That is, the width of the image formed in the inspection area S is equal to or smaller than the width D of each light receiving element. Therefore, when no defect is detected, such a light receiving state occurs, and the detection signal is at the maximum level. Since the peripheral portion of the concave or convex defect is an inclined surface, the above-mentioned Japanese Patent Application No.
As described in No. 0-397091 “Defect detection optical system and surface defect inspection apparatus”, when a concave defect or a convex defect is detected, the scanning light is reflected light from the inclined surface in the peripheral portion of each defect. Received when entering a defect and when exiting a defect. Therefore, on the inclined surface in the peripheral portion of the defect, the imaged image of the inspection area S is swung to the right or left in FIG. When shaken, the level of the detection signal of each light receiving element is lowered according to the shake state. As a result, when a concave defect or a convex defect is detected, a detection signal having two peaks is generated. The size and depth or height of the defect can be detected from the distance and level of these two peaks.

On the other hand, at the bottom of the concave defect or the convex defect, the detection signal is in the state shown in FIG. At this time, the relationship between the concave defect and the light receiving elements 523a to 523n is as shown in FIG. 5 due to the lens effect of the above defect. That is, the concave defect F has a function of focusing the received light and reflecting it like a concave mirror or a convex lens. Therefore, considering the relationship between the entire inspection point S and the concave defect F in the arrangement direction of the light receiving elements, the reflected light passing through the concave defect F and the reflected light from the surroundings of the concave defect F are In the arrangement direction of the light-receiving elements, the density becomes inward and the outside becomes coarse, so that the density appears. In the figure, since the center of the concave portion defect F substantially coincides with the optical axis L, the light receiving elements 523a to 523n.
Among these, the light receiving level of the light receiving element near the central portion becomes large, and the light receiving level of the light receiving element at the peripheral portion decreases. As a result, when the levels of the respective detection signals of the detection signals of the light receiving elements 523a to 523n are sampled with the element array direction as an axis, the levels are higher than the reference level on both sides of the high upper limit peak as shown in FIG. 6A. Also, a waveform with two lower lower limit peaks is generated. Even if the center of the concave defect F is deviated from the optical axis L, the detected waveform has a characteristic that the waveform has upper and lower peaks only slightly shifting in the arrangement direction, but the concave defect or the convex defect does not change much. Is larger than the width of the inspection area S and a part thereof enters the inspection area S, or a part thereof enters the inspection area S due to a concave defect or a convex defect smaller than the width of the inspection area S. In this case, one of the two lower limit peaks is missing.

In contrast to the concave defect, the convex defect has a dense outer side and a rough inner side due to its lens effect. As a result, as shown in FIG. 6B, a reverse polarity waveform having two upper limit peaks slightly lower than the reference level on both sides of the high lower limit peak is generated. Of course, where there is no concave defect or convex defect, the waveform of the level of the detection signal taken in the array direction of the light receiving elements has some irregularities, but is substantially flat. The level of the detection signal in this flat area is the level indicated as + DC on the horizontal axis of the waveform graphs of FIGS. 6A and 6B, which is the detection reference level. Therefore, it is possible to detect a concave defect or a convex defect and to distinguish it from a signal obtained by taking the level of the detection signal of each light receiving element in the arrangement direction of the light receiving elements. The classification detection will be described.

Now, returning to FIG. 1, n light receiving elements 523 are provided.
The detection signals of a, 523b, ..., 523n are respectively provided to the inverting amplifiers 4 provided correspondingly.
.. 401n are inverted and amplified to become positive peak signals, which are input to upper / lower limit peak detection / sample hold circuits 402a, 402b ,. Here, the detection signals having two peaks are detected corresponding to each light receiving element (each detection channel) by receiving the reflected light from the inclined portions on both side surfaces of the concave defect or the convex defect. The detected peak values are A / D converted by the A / D conversion circuits (A / D) 403a, 403b, ... 403n, and the respective converted values are stored in the defect memory 4
It is stored in 04. Further, upon receiving the reflected light from the bottom surface of the concave defect and the reflected light from the upper surface of the convex defect, detection signals having an upper limit peak and a lower limit peak are detected for each light receiving element (each detection channel). The upper limit / lower limit peak detection / sample hold circuits 402a to 402n are
With reference to the detection reference level shown as + DC on the horizontal axis of the waveform graphs of FIGS. 6 (a) and 6 (b), an upper limit peak exceeding this is detected, and a lower limit peak is detected when it is less than this. . Upper / lower limit peak detection
The sample hold circuits 402a to 402n receive the clock CLK from the clock generation circuit 406 and hold the peak value as a sample value in response to the clock CLK.

The outputs of the respective inverting amplifiers 401a to 401n are the defect detection circuit 405a and the peak difference detection circuit 40.
5b and are input. Each of the defect detection circuits 405a has a hysteresis comparator, compares the outputs of the inverting amplifiers 401a to 401n with a predetermined value Vth under an OR condition, and the output level of any one of the inverting amplifiers rises to a predetermined value or more. Occasionally, the defect detection signal DET is generated,
When the output level of the amplifier falls below a certain value (corresponding to the vicinity of the peak value), the defect detection signal DET stops. The falling edge (trailing edge) when DET is stopped is sent to the defective memory 404 as the write control signal WR. Further, the write control signal WR is slightly delayed and input as a reset signal RS to each of the upper limit / lower limit peak detection / sample hold circuits 402a to 402n, and the values up to that point are reset. As a result, the defect detection signal DE
The peak value in the period from the time when T occurs to the time when it ends is from the A / Ds 403a to 403n to the defective memory 4.
It is sent to 04 and stored as defect data. The level at this time represents the depth of the concave defect,
The height of the convex defect is represented. Further, the distance between the peaks corresponds to the size of the defect.

The peak difference detection circuit 405b is shown in FIG.
The detection signals of the waveforms shown in (a) and (b) are detected. The peak difference detection circuit 405b is configured to detect the detection signal DET from the defect detection circuit 405a and each inverting amplifier 401a.
˜401n, the delay detection circuit 405a operates for a certain period of time after the detection signal DET of the defect detection circuit 405a is generated (delay circuit and timer circuit are not shown). A detection signal is generated when the difference between the outputs of the above is greater than a certain value, and the write control signal C is sent to the defective memory 404. This write control signal C is delayed a little like the previous write control signal WR and input to each upper / lower limit peak detection / sample hold circuit 402a to 402n as a reset signal RS, and the values up to that point are reset. To be done. By operating the peak difference detection circuit 405b for a certain period with a slight delay, it is possible to match the generation timing of the detection signal on the bottom surface of the concave defect or the upper surface of the convex defect. As a result, the data of the peak values of the upper and lower limits of the detection signal of each light receiving element at the bottom of the concave defect or the upper surface of the convex defect held by the upper / lower peak detection / sample hold circuits 402a to 402n is stored in the defect memory 404. Memorized in.

As shown in FIG. 4, the peak difference detecting circuit 405b is composed of two comparators 405c and 405d and an AND circuit 405e which receives the detection outputs of the two comparators 405c and 405d. Comparator 4
05c compares the output of the OR condition of each of the inverting amplifiers 401a to 401n with the threshold THh of FIG. 6 and generates a detection signal A when the output exceeds the threshold THh. The comparator 405d is
A signal obtained by inverting the output of the OR condition of each of the inverting amplifiers 401a to 401n is received via an inverter 405f, and this inverted signal is compared with the threshold value THL of FIG. 6, and a detection signal B is generated when it is lower than the threshold value THL. The AND circuit 405e is
When the detection signals A and B are generated, the detection signal C is generated. The detection signal C is used to detect the waveforms shown in FIGS. 6A and 6B.

The upper / lower limit peak detection / sample and hold circuits 402a to 402n hold the A / D signal.
The converted defect data is stored at the updated address of the defect memory 404 together with the position data of R and θ. This address is updated each time it is written. The position coordinate data of R and θ is generated by the scanning position coordinate generation circuit 407. The write control signal WR is the defect detection signal DET.
Upper / lower limit peak detection / sample hold circuit 4 at the timing of upper / lower limit peak value detection, not at the falling edge of
02 (upper / lower limit peak detection / sample hold circuit 4
02a to 402n).

The scanning position coordinate generation circuit 407 receives the index signal indicating the rotation reference position of the disk 1 and the angle signal θ corresponding to the rotation from the rotary encoder 23, and further receives the current coordinates in the radius R direction from the data processing device 410. Upon receiving the position R, the position coordinate data of R and θ is generated and sent to the defect memory 404 as position data. As a result, the defect memory 404 outputs the level value of the detection signal corresponding to the depth or height of the defect and the respective positions where two detection signals are generated for one concave defect or convex defect each time the defect is detected. Remember. The clock generation circuit 406 includes A / Ds 403a to 403n,
Upper / lower limit peak detection / sample hold circuit 402a
~ 402n, clock CLK to the data processing device 410, etc.
Is sent.

Next, the area calculation and size classification of defects of the data processing device 410 will be described. In FIG. 1, the data processing device 410 includes an MPU 411 and a memory 412,
It is composed of a CRT display 413, an interface 414, etc., which are interconnected by a bus 415. When the spiral scan of the disk 1 is completed, the defect data obtained from the defect memory 404 by the MPU 411 is read and stored in the memory 412 via the interface 414 and the bus 415. Then, the memory 412 has a concave / convex defect detection / determination program 412.
a, a detection signal distance calculation program 412b, a continuity determination processing program 412c, a defect area calculation program 412d, a defect size classification program 412e, and a height / depth classification program 412f are stored. In addition, H connected through the interface 414
An external storage device 416 such as a DD (hard disk device) stores various data files for classification.

Concavo-convex defect detection / determination program 412a
Is executed by the MPU 411, and the execution of the MPU 411 causes the MPU 411 to read out the detected peak values of the n light receiving elements 523a, 523b, ..., 523n at the same position coordinates of R and θ to determine the upper limit peak value and the lower limit peak value. Those having a difference in value equal to or larger than a predetermined reference value Vp are extracted as the concave defect in FIG. 6A or the convex defect in FIG. 6B. Since the detection is performed through the inverting amplifiers 401a to 401n, the detection waveform is inverted here, the detection waveform of FIG. 6A becomes a convex defect, and the detection waveform of FIG. 6B becomes a concave defect. Become. However, in the relationship of FIG. 5 due to the relationship between the position of the defect and the inspection area S, the defect F is displaced in the vertical direction of the drawing. Therefore, the detected waveforms in FIGS. 6 (a) and 6 (b) are vertically shifted, and there may be one upper limit peak and one lower limit peak. Therefore, first, in the array of n light receiving elements, one having two upper limit peaks and one lower limit peak between them is defined as a concave defect in FIG. 6A, and two lower limit peaks and one upper limit peak between them. The defect is determined to be the convex defect in FIG. Further, in the case of one upper limit peak and one lower limit peak, the central peak value is larger than both sides,
The absolute value of the peak value of the two peaks is compared, the higher peak is set as the true value, and the unevenness is determined depending on whether it is the upper limit or the lower limit. According to the characteristics of FIGS. 6A and 6B, according to the inverted signal, when the true value is the upper limit, it becomes a convex defect,
When the true value is the lower limit, it becomes a concave defect. That is, except for the case where there is one upper limit peak and one lower limit peak, it is determined whether the defect is a concave defect or a convex defect depending on the relationship between the upper limit peak and the lower limit peak in the arrangement direction of the light receiving elements. Then, the data is stored in the memory 412 together with the detected position coordinates of R and θ and the flag of the determination result as to whether each is a concave defect or a convex defect.

Here, the peak difference detection circuit 405 is used.
In FIG. 6B, the detection signals having the waveforms shown in FIGS. 6A and 6B are detected, and the two-stage detection is performed in which the detection signal corresponding to the waveform is extracted in the program processing. this is,
This is because the detection level of the peak difference detection circuit 405b is made a little weaker and the concave portion defect or the convex portion defect is selected with higher accuracy in the program processing from among them to improve the detection accuracy and reduce the load of the program processing. is there. If the detection accuracy of the peak difference detection circuit 405b is high, it is not necessary to extract the difference between the upper limit peak value and the lower limit peak value that is equal to or larger than the predetermined reference value Vp. In this case, the detection signal C of the peak difference detection circuit 405b may be stored in the defect memory 404 together with the defect data, and the defect data obtained according to the detection signal C may be determined to be a concave defect or a convex defect. . On the contrary, the inverting amplifier 40 is provided on the entire surface of the disc.
The detection signals of 1a, 401b, ... 401n are directly A / D4
03A to 403n, A / D converted data is stored in the defect memory 404 together with the detection position coordinates, and the concave / convex defect detection / judgment program 412a is executed to judge concave and convex defects. Good. In this case, the defect detection circuit 405a and the peak difference detection circuit 4
05b, and the upper / lower limit peak detection / sample hold circuits 402a to 402n are not necessary. However, the processing load of the data processing device 410 increases.

Detected signal distance calculation program 412b
Is executed by the MPU 411, and by this execution, the MPU 411 reads the defect data of the defect memory 404 into the memory 412, calculates the distance L from the detection coordinates of two adjacent peak values, and the distance L is the reference value. When it is within S, the data is determined to be one defect. Further, the center coordinates of the two peak values are further calculated, and the calculated distance L (corresponding to the length of the defect) and the center coordinate value thereof are sequentially stored. Then, similar processing is performed for adjacent peak values for the next defect data, and when this processing ends, the continuity determination processing program 412c is called. When the reference value S is exceeded, one of them is stored as defect data with a length of zero and its coordinate value is stored as the central coordinate.

The continuity determination processing program 412c is M
It is executed by the PU 411, and by executing this, the MPU
When the distance L is calculated with respect to all the defect data, 411 is the center coordinate, the length (distance L), and the detection width of one light receiving element (5 μm in this embodiment) as continuity determination processing.
It is determined whether or not the distance L overlaps with a width of 5 μm in the upper, lower, left and right directions, and the continuity of defects with adjacent defect data is determined. Then, the continuous data is grouped and stored as one piece of defect data by assigning numbers to the defects sequentially detected from the innermost circumference or the outermost circumference of the disk, and calling the defect area calculation program 412d.
In the defect detection here, one or more of the detection of the light receiving elements arranged in the R direction according to the size of the defect occur at the same coordinate in the θ direction, so that the continuity of the defect is
It is possible to determine that the detected coordinates in the θ direction are within the range of the defect distance L and that the defects in which the coordinates in the R direction are adjacent are continuous. Therefore, the continuity may be determined by tracing defects having the same detection coordinates in the θ direction with the width of L along the R direction. In this way, the determination of continuity becomes easier.

By the way, here, since defects are detected corresponding to n light receiving elements, the detection coordinates of the defects in the R direction are the coordinates of the defect detection positions stored in the defect memory 404 in the R direction. Furthermore, it can be subdivided and captured according to the positions of the n light receiving elements. Therefore, as the detection coordinates in the R direction, the detection positions of n light receiving elements are added to make the position accuracy n times higher. As a result, the resolution of defect detection in the R direction can be improved to correspond to each light receiving element. In addition, when a plurality of adjacent light receiving elements out of n in the R direction detect one defect, the position of the light receiving element in the center of the plurality of light receiving elements is set as the detection coordinate in the R direction of the defect.

The defect area calculation program 412d uses MP
It is executed by U411, and by executing it, MPU4
The area calculation process 11 calculates the area of one grouped defect. In addition, the defect of one peak which does not become two detection signals has an area of 5 μm.
If m is an isolated defect and has no continuity in the R direction,
The area is calculated by the length L × 5 μm. Then, the defect size classification program 412e is called. The defect size classification program 412e is executed by the MPU 411, and the MPU 411 executes the defect size classification program 412e according to the classification criteria as the size classification determination process for the area calculated as described above. , And the classification result is stored in association with the defect number by classifying into five minimum levels. After that, the height / depth classification program 412f is called.

The height / depth classification program 412f is M
When executed by the PU 411, the MPU 411 calculates the average value of the values of the defect data of the defect memory 404 stored in the memory 412 in the order of the defect numbers (the value consisting of the absolute value of the peak value of the detection signal in one grouped defect). Calculate and use this average value as height or depth for large, medium,
It is classified into small and stored in the memory 412 corresponding to the defect number. If one defect does not have a plurality of defect data, the average value is not taken. Most of the defect data has two or more peak values for one concave defect and one convex defect, so that the depth and height of each defect can be calculated from the average value. The detection accuracy of these is improved. Note that the above processing is a simple one in which each program is sequentially executed, and therefore the description in the flowchart is omitted.

As described above, in the embodiment, the example of the laser beam is used as the irradiation light for irradiating the inspection surface of the disk. However, when the laser beam is used, in particular,
It is preferable to use an S-polarized laser beam. However, the present invention is not limited to this laser beam, and needless to say, the irradiation light may be white light. Further, in the embodiments, the description has been centered on the disk surface defect inspection apparatus, but the present invention is not limited to the disk, and it goes without saying that it can be applied to the inspection of wafers, LCD substrates, and the like. Further, in the example,
Although the scanning is performed in the Rθ direction, the scanning may of course be XY two-dimensional scanning.

[0036]

According to the present invention, a light receiving system is provided which has n light receiving elements arranged and whose image formation along the direction perpendicular to the main scanning direction is the light receiving element arranging direction. The upper limit peak and the lower limit peak are obtained for the concave defect or the convex defect in the detection signals of the respective light receiving elements in the arrangement direction of the n light receiving elements. Then, according to the relationship between the upper limit peak and the lower limit peak in the arrangement direction of the light receiving elements, it is determined whether the defect is a concave defect or a convex defect, and the defect is detected. as a result,
Concavo-convex defects on the surface of the face plate can be accurately distinguished and detected from concave defects and convex defects.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of a surface defect inspection apparatus to which the present invention is applied.

FIG. 2 is a development explanatory diagram of a light emitting / receiving system in a direction (θ direction) parallel to the paper surface in the drawing and one direction (R direction) perpendicular to the paper surface in the drawing in the detection optical system.

FIG. 3 is a diagram showing an image formed on the light receiving surface of the inspection area and AP.
It is explanatory drawing with the relationship of a D array sensor.

FIG. 4 is an explanatory diagram of a peak difference detection circuit.

FIG. 5 is an explanatory diagram of a coarse and dense state due to a lens effect of specularly reflected light of a defect in a recess.

6 is an explanatory diagram of a detection waveform in the lens effect in FIG.

[Explanation of symbols]

1 ... Disk, 50 ... Detection optical system, 51 ... Projection system, 51
1 ... Laser light source, 512 ... Beam expander, 513
... Cylindrical lens, 514 ... Focusing lens, 521 ... High-resolution condensing lens, 522 ... Imaging lens, 523 ... APD array sensor, 52 ... Light receiving system, 40
0 ... Defect detection section, 401a, 401b, 401n ... Inversion amplifier, 402a, 402b, 402n ... Upper / lower limit peak detection / sample hold circuit, 403a, 403
b, 403n ... A / D conversion circuit (A / D), 404 ... Defect memory, 405a ... Defect detection circuit, 406 ... Clock generation circuit, 407 ... Scan position coordinate generation circuit, 410 ... Data processing device, 411 ... MPU, 45 ... Printer (P
R), 412 ... Memory, 412a ... Concavo-convex defect detection / determination program 412a, 412b ... Detection signal distance calculation program, 412c ... Continuity determination processing program, 41
2d ... Defect area calculation program, 412f ... Height / depth classification program, 413 ... CRT display, 414
... interface, 415 ... Bus, 416 ... external storage device, 100 ... surface defect inspection apparatus, L T _... laser beam,
_SP ... Beam spot.

   ─────────────────────────────────────────────────── ─── Continued front page    F term (reference) 2F065 AA49 AA61 BB03 CC03 FF01                       FF17 FF44 FF67 GG04 HH04                       JJ18 LL09 MM03 QQ02 QQ03                       QQ23 QQ25 SS06 SS13 UU07                 2G051 AA71 AB07 AB20 BA10 BB01                       BB09 BC07 CA03 CD05 DA08                       EA03 EA08 EA11                 5D112 AA02 AA24 BA03 JJ03

Claims (7)

[Claims] 1. A defect detection optical system in which a surface of a face plate is scanned by a light beam, reflected light from the surface by the light beam is received by a light receiver, and the light receiver generates a detection signal for defect detection. , A light projecting system for irradiating the light beam having a width in a direction perpendicular to the main scanning direction to relatively scan the face plate, and n light receiving devices arranged (where n is 2 or more) A light receiving system for forming an image of the scanning position of the face plate on the n light receiving elements and forming an image along the orthogonal direction in the array direction. When the difference between the upper limit peak value and the lower limit peak value of the n detection signals of the n light receiving elements at the scanning position is equal to or more than a predetermined value,
A defect detection optical system characterized by determining whether a defect is a concave defect or a convex defect according to the relationship between an upper limit peak and a lower limit peak in the arrangement direction of the light receiving elements and detecting these defects. 2. The face plate is a disk, the relationship between the upper limit peak and the lower limit peak is a positional relationship in the arrangement direction of the light receiving elements, and the image formation is an elliptical or elliptical shape. The defect detection optical system according to claim 1, wherein a major axis thereof corresponds to an array direction of the light receiving elements, and a width of the image formation in the main scanning direction is equal to or smaller than a width of the light receiving elements. . 3. The defect detection optical system according to claim 2, wherein the disk is a glass substrate, and the scanning is spiral scanning. 4. A surface defect detection device in which a surface of a face plate is scanned with a light beam, and light reflected by the light beam from the surface is received by a light receiver, and the light receiver generates a signal for defect detection. , A light projecting system that irradiates the light beam having a width in a direction perpendicular to the main scanning direction to relatively scan the face plate, and n (where n is 2 or more) in which the light receivers are arranged. (Integer) light-receiving element and forms an image of the scanning position of the face plate on the n light-receiving elements, and a light-receiving system in which the image formation along the perpendicular direction is the array direction, When the difference between the upper limit peak value and the lower limit peak value of the n detection signals of the n light receiving elements at the position is a predetermined value or more, the upper limit peak and the lower limit peak in the arrangement direction of the light receiving elements Concave defect or convex defect depending on the relationship with A surface defect detecting device characterized by determining whether or not these defects are detected. 5. The face plate is a disk, the image formation is an elliptical or elliptical shape, and the major axis thereof corresponds to the arrangement direction of the light receiving elements, and the width of the image formation in the main scanning direction is The surface defect inspection apparatus according to claim 4, wherein the width of the light receiving element is smaller than or equal to the width. 6. A position coordinate generating circuit for generating information on a scanning position of the disk, n amplifiers for amplifying n light receiving signals from each of the n light receiving elements, and n of these amplifiers. N sampling and holding circuits for respectively receiving the detection signals obtained from the amplifiers and holding the upper limit peak value and the lower limit peak value, and respectively receiving the respective held peak values and receiving them. N A / D conversion circuits for A / D conversion, and a data processing device that receives the digital value of the peak value from these n A / D conversion circuits and receives the information of the scanning position from the position coordinate generation circuit 6. The surface defect inspection apparatus according to claim 5, further comprising: and which determines whether the defect is a concave defect or a convex defect in the data processing device. 7. The surface defect inspection apparatus according to claim 6, wherein the disk is a glass substrate, and the scanning is spiral scanning.