JP2002022415A - Fine protrusion inspecting apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To inspect the height of a fine protrusion at a high accuracy and a high speed. SOLUTION: Projection optical system means radiates a light beam so as to locate the surface of a sample under test at a focused position and deflects a beam spot formed on the sample surface to linearly reciprocally scan over an inspecting region at a specified period. Detecting optical system means generates two semi-circular light beams with mutually different regions shielded from light beams reflected from the sample surface, and they are received separately by a first and second optical sensor means to obtain height information from detected signals thereof. Thus the height information are obtained, based on the detected signals from the two systems of optical sensor means, hence the detection accuracy is improved and, if a thermal variation of each optical element constituting the projection optical system means or a deflection error in the deflection causes a spot scan deviation (Δt), a high accuracy measurement can be made, without being influenced thereby.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention measures the height of minute projections such as ball grid arrays (BGA), wafer bumps, IC bumps, and spacers used for liquid crystal display (LCD) substrates and filters. The present invention relates to a micro-projection inspection apparatus for detecting a defect or detecting the defect.

[0002]

2. Description of the Related Art FIG. 1 shows an optical inspection apparatus for optically inspecting the appearance of a defect such as a ball or a bump provided for connecting wiring such as an IC package.
There is the one shown in FIG. The inspection apparatus shown in FIG. 1A receives a laser beam 12 emitted from a semiconductor laser 11 via a half mirror 13 and an objective lens 14 so that the surface of a sample 15 to be inspected is at a focal point (0 position). The test sample 15 is irradiated. When a convex portion (projection) such as a ball or a bump exists on the surface of the sample 15 to be inspected, the laser beam 12 is reflected at a plus position (+) according to the size. Conversely, when a concave portion (dent) exists on the surface of the sample 15 to be inspected, the laser beam 12 is reflected at a minus position (-) corresponding to the size.

Light reflected from surfaces having different heights enters a photosensor 17 via an objective lens 14, a half mirror 13, and a pinhole 16. At this time, the pinhole 16 and the surface position (0 position) of the sample 15 to be inspected have a conjugate relationship. The light reflected at the in-focus position (0 position) passes through the entire pinhole 16 and is incident on the photosensor 17, and the other light reflected at the plus position (+) and the minus position (-) is part of the light. Passes through the pinhole 16 and enters the photosensor 17. Since only the light passing through the pinhole 16 is incident on the photosensor 17,
The height of the surface of the test sample 15 can be measured based on the magnitude of the electric signal detected by the photosensor 17. That is, when the laser beam 12 is reflected at the in-focus position (0 position), the output of the photo sensor 17 shows the maximum as shown in FIG. 1B, and the other plus position (+)
When the light is reflected at the minus position (-), the output is smaller. Therefore, the height of the surface of the test sample 15 can be measured based on the magnitude of the output value of the photosensor 17. Then, based on the height information thus measured, the inspection device performs an inspection as to whether or not a defect exists in the ball, the bump, or the like.

However, the inspection apparatus shown in FIG. 1 can measure the position of the laser beam reflected from the in-focus position (0 position) and its height. It was not possible to recognize whether it was at +) or at the minus position (-). FIG. 2 shows that the laser light is in the plus position (+) of the sample to be inspected.
FIG. 11 is a diagram showing a conventional technique of an inspection device capable of detecting which side of a negative position (−) is reflected.
The inspection apparatus shown in FIG. 2A converts the laser light emitted from the semiconductor laser 21 into a collimator lens 23 and a half mirror 24.
Then, the test sample 26 is irradiated via the objective lens 25 such that the surface of the test sample 26 is at the focal point (0 position). When a convex portion (projection) such as a ball or a bump exists on the surface of the sample 26 to be inspected, the laser beam is reflected at a plus position (+) corresponding to the size. Conversely, when a concave portion (dent) exists on the surface of the sample 26 to be inspected,
The laser light is reflected at a minus position (-) according to the size.

The light reflected from the surfaces having different heights is reflected by an objective lens 25, a half mirror 24, and an image forming lens 2.
7. The light enters the photosensors 2B and 2C via the half mirror 28 and the pinholes 29 and 2A. At this time, position 2
The pinholes 29 and 2A are set so that D and 2E are conjugate with the surface position (0 position) of the sample 15 to be inspected.
And the photo sensors 2B and 2C are arranged. The light reflected at the plus position (+) passes through all the pinholes 2A and enters the photosensor 2C, and the light reflected at the other in-focus position (0 position) and at the minus position (-) is part of the light. Passes through the pinhole 2A and enters the photosensor 2C.
On the other hand, the light reflected at the minus position (−) passes through the entire pinhole 29 and enters the photosensor 2B, and the light reflected at the other in-focus position (0 position) and the plus position (+) is Part of the light passes through the pinhole 29 and enters the photosensor 2B.

The photo sensor 2B receives the light passing through the pinhole 29 and outputs a sensor output Pa. The photo sensor 2C receives the light passing through the pinhole 2A and outputs a sensor output Pb. Sensor output P of photo sensor 2C
By subtracting the sensor output Pa of the photosensor 2B from b, the height information ε at the plus position (+) or the minus position (−) on the surface of the sample 26 can be measured. That is, the laser light 22 is focused on the in-focus position (0
2B, the sensor output Pa of the photosensor 2B and the sensor output Pb of the photosensor 2C are equal to each other, and the height information ε is “0” as shown in FIG. 2B.
become. When the laser beam 22 is reflected at the minus position (-), the sensor output Pa of the photosensor 2B becomes sufficiently larger than the sensor output Pb of the photosensor 2C, so that the height information ε takes a negative value, and the minus position (−) ) Can be recognized as reflected. On the other hand, when the laser beam 22 is reflected at the plus position (+), the sensor output Pb of the photosensor 2C becomes sufficiently larger than the sensor output Pa of the photosensor 2B, so that the height information ε becomes a positive value,
It can be recognized that the light was reflected at the plus position (+). Based on the height information ε, the displacement of the height of the surface of the sample 26 to be inspected can be measured. The inspection device inspects, based on the measured height information, whether there is a defect in the ball or the bump, or the state of a concave portion (dent) in the sample to be inspected 26.

[0007] Such an optical inspection apparatus can detect only a portion corresponding to the width of the beam diameter of the laser light,
The beam diameter is increased according to the size of the minute projections such as balls and bumps. However, if the beam diameter is increased, defects such as chipping around the balls and bumps cannot be sufficiently detected. There is. Therefore, conventionally, the beam diameter of the laser beam is reduced to about 3 to 4 [μm], and the chip or the wafer on which the chip is formed is two-dimensionally scanned in the XY directions.

[0008]

However, a ball or bump having a size of about 300 to 400 [μm] is inspected by using the above-described inspection apparatus having a beam diameter of about 3 to 4 [μm]. In this case, the laser beam is reciprocated in the main scanning direction (X direction) by a distance of about 300 to 400 [μm], and gradually moved in the sub-scanning direction (Y direction) by a distance of about 3 to 4 [μm] corresponding to the beam diameter. In order to move the laser beam, the chip or the wafer on which the chip is formed and the laser beam must be relatively moved. How to move the laser light in the main scanning direction with high precision and high speed leads to a drastic reduction in inspection time. However, the XY stage on which a chip or a wafer on which the chip is formed is mounted as described above is used. There is a technical limit to moving at high speed under such conditions, and there has been a problem that much time is required for inspection.

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and has as its object to provide a microprojection inspection apparatus capable of performing a high-precision, high-speed inspection of a microprojection. I do.

[0010]

According to a first aspect of the present invention, there is provided a micro-projection inspection apparatus which irradiates a light beam so that a surface of a sample to be inspected is at a focal point and a predetermined beam on the sample to be inspected. A light projecting optical system means for deflecting the inspection area so as to linearly reciprocately scan in a predetermined cycle, and a semicircular light beam whose diameter is a straight line drawn by the optical axis of the reciprocally scanned light beam, A detection optical system means for generating two light beams whose different regions are shielded from the light beam reflected from the surface of the sample to be inspected to form an image, and the two light beams generated by the detection optical system means; A light receiving region provided corresponding to one of the light beams and divided into at least two portions so that the diameter of the semicircular light beam is a boundary, and outputting a detection signal corresponding to each light receiving region. First optical sensor means for performing A light-receiving area is provided corresponding to the other of the two light beams generated by the detection optical system means, and has at least two divided light receiving regions such that the diameter of the semicircular light beam is a boundary, A second optical sensor unit that outputs a detection signal corresponding to each light receiving area, a stage unit on which the sample to be inspected is mounted, the stage unit, the light projecting optical system unit, and the detection optical system unit. Driving means for relatively moving, and control means for controlling the driving means and calculating height information of minute projections formed on the test sample based on the detection signal output from the optical sensor means It is provided with.

The light projecting optical system irradiates a light beam such that the surface of the sample to be inspected is at the focal point, similarly to the conventional inspection apparatus shown in FIGS. Further, the light projecting optical system deflects the light beam so that the beam spot formed on the surface of the sample to be inspected by the light beam reciprocally scans the inspection area linearly at a predetermined cycle. When the light beam is deflected, it is difficult to receive the light beam with the optical sensor through the pinhole as in the related art. Therefore, in the present invention, the detection optical system means is configured by applying a conventionally known knife edge method as a detection principle of the focus error detection optical system. In the knife-edge method, a half-circular light beam is generated by blocking half of the light beam reflected from the surface of the sample to be inspected with the knife edge, and the semi-circular light beam is applied along the edge of the knife edge. The light is received by an optical sensor means having a light receiving area divided into two. To simply calculate the height information of the minute projections formed on the sample to be inspected, use one of the two optical sensor means to receive the light beam reflected from the surface of the sample to be inspected through the knife edge do it. However, due to heat fluctuation of each optical component constituting the light projecting optical system means or a deflection error at the time of deflection, the scanning deviation of the beam spot (Δ
When t) occurs, the influence of the scanning shift affects the optical sensor means divided into two, and accurate height information cannot be obtained. Therefore, in the present invention, two semicircular light beams whose different regions are shielded are generated from the light beam reflected from the surface of the sample to be inspected, and the two light beams are separately separated by the first and second optical sensor means. Light is received, and height information is obtained based on the detection signal. As described above, since the height information is obtained based on the detection signals from the two systems of optical sensor means, there is an effect that the detection accuracy is improved. Further, even if spot scanning deviation (Δt) occurs due to thermal fluctuation of each optical element constituting the light projecting optical system means or deflection error at the time of deflector, high-precision height measurement can be performed without being affected by the deviation. There is an effect that can be performed.

According to a second aspect of the present invention, in the micro-projection inspection apparatus according to the first aspect, the detection optical system means reflects a part of the light beam reflected from the surface of the sample to be inspected, and removes the rest. A half mirror means for generating the two light beams by transmitting the light beam; and a first means for generating the semicircular light beam by blocking one side of one of the light beams generated by the half mirror means. A shielding plate means and the shielded semicircular light beam in an area different from the area shielded by the first shielding plate means by shielding one side of the other light beam generated by the half mirror means A first imaging lens means for imaging the semicircular light beam generated by the first shielding plate means, and a second shielding plate means. Is obtained by constituting thus the generated said semicircular light beam and a second imaging lens means for imaging. This is a specific limitation of the configuration of the detection optical system means described in claim 1. The light beam is split into two beams by a half mirror means, and each light beam is divided into first and second light beams. The first and the second are shielded by a shielding plate means.
The image forming lens means forms an image on the first and second optical sensor means.

According to a third aspect of the present invention, in the micro-projection inspection apparatus according to the first aspect, the detection optical system means includes a straight line drawn by an optical axis of the light beam reflected from the surface of the sample to be inspected and a tip. The light beam reflected from the surface of the sample to be inspected is reflected by each slope so that different areas are shielded from each other.
Knife edge prism means for generating a semicircular light beam; first imaging lens means for imaging one of the semicircular light beams reflected by the knife edge prism means; and the knife edge And second imaging lens means for imaging the other of the semicircular light beams reflected by the prism means. This specifically restricts the configuration of the detection optical system means according to the first aspect, and divides the light beam into two by the knife-edge prism means and simultaneously separates the different regions of the two light beams. The image is shielded, and the image is formed on the first and second optical sensor means by the first and second image forming lens means.

According to a fourth aspect of the present invention, there is provided a micro-projection inspection apparatus, wherein an acousto-optical deflector is used to deflect the light beam so as to linearly reciprocally scan at the predetermined period. It is. There are mechanical beam deflecting devices and optical beam deflecting devices for deflecting the light beam, but this is limited to the optical beam deflecting device using an acousto-optic deflector.

According to a fifth aspect of the present invention, in the inspection apparatus for minute projections according to the first aspect, a polygon mirror is used to deflect the light beam so as to linearly reciprocally scan at the predetermined period. . This is limited to one that mechanically deflects a light beam using a polygon mirror.

According to a sixth aspect of the present invention, in the micro-projection inspection apparatus according to the first aspect, the zero-order light component of the light beam reflected from the surface of the sample to be inspected is cut, and a high-order component other than the zero-order light component is cut. A light receiving element means for receiving the next reflected scattered light or the like is provided. The microprojection inspection device according to claim 1 detects the height of the microprojection,
Since the zero-order light component of the reflected light beam can be cut, and based on that, foreign matter, dirt, defects such as scratches and chips, etc. on the surface of the sample to be inspected can be detected. Provided separately.

According to a seventh aspect of the present invention, in the micro-projection inspection apparatus according to the first aspect, the control means includes a step of reflecting the sum of the detection signals output from the optical sensor means from the surface of the sample to be inspected. The position information of the minute projection formed on the sample to be inspected is obtained by comparing the value of the reflected light luminance signal with a predetermined value, and controlling the stage means based on the position information. It was made. This is because when detecting a defect such as the height or shape of a microscopic projection having a three-dimensional structure, it is necessary to clearly distinguish a three-dimensional object, such as a pattern, present around the microscopic projection from the microprojection. For example, when a plurality of minute projections are arranged on a test sample in a complicated manner, without performing complicated coordinate setting management for each of the minute projections, the detection signal output from the optical sensor means is not required. The coordinate position of the minute projection is measured in advance with high accuracy based on the sum.

According to an eighth aspect of the present invention, in the micro-projection inspection apparatus according to the first aspect, the control means is configured to control the control means based on the detection signal output from the two divided light receiving areas of the first optical sensor means. A signal output from a light receiving area corresponding to an unshielded area of the light beam is used as a first detection signal, and a signal output from a light receiving area corresponding to the light beam blocked area is used as a first detection signal. Is a second detection signal, the detection signal output from the light receiving area divided into two of the second optical sensor means, from the light receiving area corresponding to the unshielded area side of the light beam When the output signal is a third detection signal, and the output signal from the light receiving area corresponding to the light-shielded area side of the light beam is a fourth detection signal, the first detection signal Second inspection A first value obtained by dividing the sum of the value obtained by subtracting the signal and the value obtained by subtracting the fourth detection signal from the third detection signal by the total sum of the first to fourth detection signals is used as the test target. The height is calculated as height information of the minute projections formed on the sample. This specifically shows a method of calculating height information performed by the control means,
The calculation formula of the height information ε shown in FIG. 12 ｛(AB) + (C
−D)｝ / (A + B + C + D).
Here, the first detection signal corresponds to A, the second detection signal corresponds to B, the third detection signal corresponds to C, and the fourth detection signal corresponds to D.

According to a ninth aspect of the present invention, in the micro-projection inspection apparatus according to the eighth aspect, the control means determines the value obtained by subtracting the third detection signal from the first detection signal and the second detection signal.
A value obtained by subtracting from the first value a value obtained by dividing the sum of the detection signal and the value obtained by subtracting the fourth detection signal from the sum of the first to fourth detection signals. The height information is calculated as height information of minute projections formed on the sample to be inspected. This control means is intended specifically showing the method of calculating height information performed by, those corresponding to the calculation formula epsilon-epsilon ₁ shown in Figure 12 (a). Here, epsilon is the first value calculated in claims 8, wherein the epsilon ₁ of calculation formula {(A-C) + ( B-D)}
/ (A + B + C + D) is specifically shown.

According to a tenth aspect of the present invention, in the micro-projection inspection apparatus according to the eighth aspect, the control means is configured to determine that the value obtained by subtracting the third detection signal from the first detection signal and the second detection signal are used. Masking the first value when a value obtained by dividing a sum of a value obtained by subtracting the fourth detection signal from the signal by a total sum of the first to fourth detection signals is larger than a predetermined value; Is calculated as height information of the minute projections formed on the sample to be inspected. This specifically shows a method of calculating the height information performed by the control means, and corresponds to FIG. 12B, and the calculation formula ｛(AC) + (B− D)｝
/ The epsilon ₁ obtained from (A + B + C + D ) is compared with a predetermined value, epsilon ₁ is greater than a predetermined value, and to mask the output of the height information calculated by the claims 8 by operating the gate circuit epsilon Things.

According to an eleventh aspect of the present invention, in the micro projection inspection apparatus according to any one of the first to tenth aspects, the control means is formed on the sample to be inspected based on the calculated height information. The defect inspection of the minute projection is performed. In this method, the minute projections on the adjacent chips are compared adjacently based on the calculated height information of the minute projections, thereby performing a defect inspection of the minute projections.

According to a twelfth aspect of the present invention, in the micro-projection inspection apparatus according to the sixth aspect, the control means compares the value of the detection signal output from the light receiving element means with a predetermined value to thereby detect the object. The position information of the minute projection formed on the inspection sample is obtained. An inspection apparatus for microprojections according to claim 6, which cuts off a zero-order light component of a reflected light beam and detects a defect such as a foreign substance, dirt, a scratch or a chip on the surface of the sample to be inspected based on the cut. However, in this method, the coordinate position of the minute projection is measured with high accuracy using such an inspection optical system.

According to a thirteenth aspect of the present invention, in the inspection apparatus of the seventh aspect, the controller outputs the reflection light output from the optical sensor means when the fine projection has a spherical shape. By comparing the value of the light luminance signal with a predetermined value, at least one of the center position information and the size information of the spherical minute projection is obtained. When the minute projection has a spherical three-dimensional structure, it is necessary to measure at least one of the center position and the size of the spherical projection with high accuracy. For this purpose, at least one of the center position and the size of the spherical projection is measured with high accuracy based on the sum of the detection signals output from the optical sensor means.

[0024]

Embodiments of the present invention will be described below with reference to the accompanying drawings. FIG. 3 is a diagram showing a schematic configuration of the micro-projection inspection apparatus of the present invention. FIG. 4 is a perspective view showing the inspection optical system in a three-dimensional manner to make FIG. 3 easier to understand. Accordingly, in FIGS. 3 and 4, the same components are denoted by the same reference numerals. The micro-projection inspection apparatus shown in FIG. 3 inspects bump defects based on height information of bumps formed on a plurality of chips in a wafer. This microprojection inspection apparatus includes an inspection stage unit, an inspection optical system, and a signal processing unit.

The inspection stage section of the micro-projection inspection apparatus according to this embodiment includes an XY stage moving mechanism 300, a wafer mounting table 302, and a stage driving circuit 304. The XY stage moving mechanism 300 moves the wafer mounting table 302 on which the wafer 306 to be inspected is mounted in the XY directions. The stage driving circuit 304 is controlled by the image processing / control device 500, and by controlling the driving of the XY stage moving mechanism 300, the beam spot is adjusted to the wafer 3 on the wafer mounting table 302 as shown in FIG.
06 is controlled to relatively scan in the X and Y directions. The stage drive circuit 304 first moves the wafer 306 such that the beam spot main scans in the X direction along the arrow 51 corresponding to the bump formation lines X1 and X2 of the chips 60 on the wafer 306 as shown in FIG. . After the end of the main scanning, the stage driving circuit 304 moves the wafer 306 so that the beam spot performs sub-scanning in the Y direction along the arrow 52 to a position corresponding to the next bump formation line X3, X4. After the main scanning is completed, the stage driving circuit 304 moves the bump forming lines X3 and X4 of the chip 60 along the arrow 53 opposite to the above-described main scanning.
The wafer 306 is moved so that the beam spot performs main scanning in the X direction corresponding to. After completion of the main scanning, the stage driving circuit 304 sets the bump forming line X of the next chip.
The wafer 306 is moved so that the beam spot sub-scans in the Y direction along the arrow 54 to a position corresponding to 1, X2. The above-described scanning method uses the bump forming line X
In this case, the stage driving circuit 304 can be inspected at the same time or when only one row of bump forming lines is provided around the chip 60. When the bump forming lines X1 and X2 cannot be inspected simultaneously, Moves the wafer 306 so that main scanning is performed along the respective bump formation lines X1 and X2. In this way, when the main scanning corresponding to the bump forming lines X1 to X4 is completed for all the chips on the wafer 306, the stage driving circuit 304 rotates the wafer 306 rightward or leftward by about 90 degrees. , Bump forming lines Y1 to Y
Similarly, the wafer 306 is moved so that the beam spot performs main scanning and sub-scanning.

The inspection optical system of the micro-projection inspection apparatus according to this embodiment includes a projection optical system for irradiating the laser beam to the wafer 306 and a detection optical system for guiding the laser beam reflected by the wafer 306 to the optical sensor. It is basically composed of The light projecting optical system includes a semiconductor laser light source 400, an S-polarizing plate 401,
Collimating lens 402, acousto-optic deflector (AOD: A
costo-Optical Deflector) 40
4. Deflector drive circuit 405, fθ lens 406, lens 408, polarizing beam splitter 410, quarter wave plate 412, and objective lens 414.

The semiconductor laser light source 400 emits laser light of a predetermined frequency. The S-polarized plate 401 transmits only the S-polarized component of the laser light emitted from the semiconductor laser light source 400. The collimating lens 402 is used for the S polarizing plate 40.
The laser beam of the S-polarized component transmitted through 1 is converted into a parallel light beam. The acousto-optic deflector 404 is a collimating lens 40
The parallel beam of the S-polarized light component transmitted through
Is converted into polarized light reciprocating in the Y direction. The deflector driving circuit 405 is connected to the image processing / control device 50.
0, and supplies a drive signal of a predetermined frequency (about 4 MHz) to the acousto-optic deflector 404. fθ lens 4
06 and the lens 408 are connected to the detection area 30 of the wafer 306.
8 so as to optimally focus the laser beam at each scanning position. Polarizing beam splitter 410
Reflects the S-polarized component laser light and transmits the P-polarized component laser light. Accordingly, the polarization beam splitter 410 reflects the S-polarized component laser light transmitted through the fθ lens 406 and the lens 408 toward the wafer 306. The laser beam reflected by the polarization beam splitter 410 is converted into a circularly polarized component laser beam by a quarter-wave plate 412, and focused on the wafer 306 via an objective lens 414.

The detection optical system includes an objective lens 414, a quarter-wave plate 412, a polarizing beam splitter 410, half mirrors 416 and 418, shielding plates 420 and 422, imaging lenses 424 and 426, a two-divided optical sensor 428, 43
0, a spatial filter 432, an imaging lens 434, a slit 436, and a photomultiplier 438.
The reflected laser light of the circularly polarized component reflected by the wafer 306 is
The light passes through the objective lens 414 and the quarter-wave plate 412. At this time, the reflected laser light of the circularly polarized component is one quarter.
By passing through the wave plate 412, the light is converted into a reflected laser light having a P-polarized component, and is introduced into the polarizing beam splitter 410. The polarization beam splitter 410
The reflected light of the P-polarized component from the quarter-wave plate 412 is transmitted to the half mirror 416 on the subsequent stage. The half mirror 416 is disposed at an angle of 45 degrees with respect to the optical axis of the P-polarized component reflected laser light from the polarization beam splitter 410, and about half of the reflected laser light is disposed on the side surface of the optical axis. The reflected laser light is reflected to the bump defect detection optical system side, and about half of the reflected laser light is transmitted to the wafer surface defect detection optical system side.

The bump defect detecting optical system includes a half mirror 4
18, shielding plates 420, 422, imaging lenses 424, 42
It comprises 6 and 2 split optical sensors 428, 430. The half mirror 418 is disposed at an angle of 45 degrees with respect to the optical axis of the reflected laser light from the half mirror 416, and approximately half of the reflected laser light is disposed on a side surface of the optical axis. About half of the reflected laser light reflected from the half mirror 416 is transmitted to the two-divided optical sensor 430. The shielding plates 420 and 422
Half mirror 418 and split optical sensors 428 and 430
And shields different one-side half regions of the cross section of the reflected laser light reflected by the half mirror 416. As shown in FIG. 3, the shielding plate 422 shields a lower half area of the laser light transmitted through the half mirror 418, that is, a lower half area of the laser light reflected by the half mirror 416. The right half area of the laser light reflected by the half mirror 418, that is, the upper half area of the laser light reflected by the half mirror 416 is blocked. The shielding plates 420 and 422 may be arranged between the imaging lenses 424 and 426 and the two-divided optical sensors 428 and 430.

The imaging lenses 424 and 426 are
The images corresponding to the laser beams not blocked by 0,422 are formed on the two-divided optical sensors 428,430, respectively. The two-split optical sensors 428 and 430 receive the reflected laser light from the half mirror 416 and output detection signals A to D corresponding thereto. 2-split optical sensor 42
8, 430 are the same Y as the detection area 308 of the wafer 306.
The light receiving area is constituted by a photosensor in which the light receiving area is divided into two along the direction (deflection direction). Therefore, the two-part optical sensor 4
28 and 430 output independent detection signals A to D, respectively. In the case of FIG. 3, the Y direction (deflection direction) of the detection area 308 of the wafer 306 and the two-divided optical sensors 428 and 430
The deflection direction in which the image formed above reciprocates is the depth direction of the drawing.

The wafer surface defect detection optical system includes a spatial filter 432, an imaging lens 434, a slit 436, and a photomultiplier 438. Spatial filter 4
Reference numeral 32 denotes a photomultiplier 438 which is disposed on the optical axis of the transmitted laser light from the half mirror 416, cuts the 0th-order light component of the transmitted laser light, and removes higher-order reflected and scattered light other than the 0th-order light component. Pass to the side. Imaging lens 434
The image corresponding to the laser beam passing through the spatial filter 432 is transmitted through the slit 436 to the photomultiplier 438.
Image on top. The slit 436 is formed by a gap formed along the same Y direction (deflection direction) as the detection area 308 of the wafer 306. Photo Multiplier 438
Is for receiving the laser light that has passed through the slit 436 among the light condensed by the imaging lens 434, and outputs a current corresponding to the intensity of the laser light.

The signal processing section includes an image processing / control device 50
0, amplifiers 502A to 502D, analog-digital (A / D) converters 504 and 505, an amplifier 506, a defect detection circuit 508, and an A / D converter 510. The amplifiers 502A to 502D include a two-divided optical sensor 4
The detection signals A to D output from the terminals 30 and 428 are amplified and supplied to the A / D converters 504 and 505. A /
The D converters 504 and 505 are connected to the amplifiers 502A to 502, respectively.
The analog detection signals A to D output from D are converted into digital signals and supplied to the image processing / control device 500.
The amplifier 506 amplifies the detection current output from the photomultiplier 438 and supplies it to the defect detection circuit 508. If defects such as foreign matter, dirt, scratches or chips are present on the wafer 306, the photomultiplier 438
Is instantaneously incident on the laser beam, and accordingly, a pulse-like current is output from the photomultiplier 438. The defect detection circuit 508 supplies a defect detection signal corresponding to a defect on the surface of the wafer 306 detected based on the pulse-like current to the A / D converter 510. The A / D converter 510 converts the analog defect detection signal output from the defect detection circuit 508 into a digital defect detection signal and supplies the digital defect detection signal to the image processing / control device 500.

The image processing / control device 500 comprises a microcomputer system including a microprocessor unit (CPU), a program memory (ROM), a working memory (RAM), a hard disk device (HDD), an external interface, etc., not shown. And execute various processes. Although not shown, the image processing / control device 500 includes:
A keyboard, a monitor screen, and the like are connected to construct a graphical user interface (GUI) with the operator. The memory (R in the image processing / control device 500
OM, RAM, HDD) include a bump height shape detection program, a wafer scanning program, a defect detection program,
Various programs such as a focusing program are stored.

The bump height shape detection program is executed in response to the operation of a predetermined inspection start key on the keyboard. By executing the bump height shape detection program, the image processing / control device 500 executes each program in the order of the focusing program, the wafer scanning program, and the defect detection program. The focusing program adjusts the fθ lens 406, the lens 408, and the objective lens 414 based on the detection signals A to D output from the two-divided optical sensors 428 and 430, and
Is performed so that the focal point of the laser beam is located on the surface of the laser beam. The wafer scanning program includes the stage driving circuit 30
5 and a control signal is supplied to the deflector drive circuit 405, and FIG.
The main scanning and the sub-scanning are repeatedly executed as shown in FIG.
/ D converters 504 and 505 output detection signals A to
The defect detection signal output from the D and A / D converter 510 is temporarily stored in a memory area. The defect detection program detects bump height information based on detection signals A to D output from the A / D converters 504 and 505, detects bump defects based on the information, and performs A / D conversion. Vessel 5
10 based on the defect detection signal output from
Defect positions such as foreign matter, dirt, scratches and chips on wafer 6
06. Details of how the height information is detected based on the detection signals A to D will be described later.

FIG. 5 is a view of the wafer 306 mounted on the wafer mounting table 302 as viewed from above. As is clear from the figure, a plurality of rectangular chips are regularly provided on the wafer 306. In the chip 60, a plurality of bumps are regularly arranged on bump forming lines X1 to X4 and Y1 to Y4 provided along the outer periphery of the chip as shown in FIG. In FIG. 6, the shape of the bump is shown as a square. In FIG. 6, two bump formation lines (X1 and X
2), (X3, X4), (Y1, Y2), (Y3, Y
4) shows the case where there is, but there is a case where the number is usually one or more.

FIG. 7 is a diagram showing a scanning state of a laser beam on a bump on a bump forming line. In FIG. 7, bumps 72 and 73 are formed on pads 70 and 71 along the bump formation line X1. The detection area 308 in the Y direction in FIG. 4 corresponds to the vertical direction in FIG.
The distance is 0.36 [mm]. The diameter of the beam spot 74 of the laser beam applied to the surface of the wafer 306 is about 3.6 [μm], and the beam spot 74 scans the detection area 308 by reciprocating at a frequency of 4 [MHz]. Further, when the detection area 308 is scanned downward in the beam spot 74, the bump formation line X1
, The scanning operation is repeated by scanning the detection region 308 upward again by the predetermined pitch and moving by the predetermined pitch again. The bump forming line X1
When the movement in the direction is continuous, zigzag scanning is performed,
The bump formation line X1 is synchronized with one scan in the Y direction.
If the movement in the direction is performed, a rectangular scan as shown in FIG. 7 is obtained.

This microprojection inspection apparatus employs a knife edge method as a focus error detection optical system for scanning a sample to be inspected and measuring a height displacement based on reflected light corresponding to surface irregularities. I have. FIGS. 8 and 9 are diagrams for explaining the detection principle based on the knife edge method. The knife edge method is conventionally known, and the basic principle is that, as shown in FIGS. 8 and 9, the reflected light from the wafer 306 which is the sample to be inspected is shielded by the shielding plates 420 and 422 which are knife edges. Through two optical sensors 428 and 4
30 and receives the signals on the wafer 3 based on the detection signals AD.
06 is to detect the height of the irregularities on the surface. 8 and 9, a quarter of the inspection optical system shown in FIGS. 3 and 4 is used.
The wavelength plate 412, the polarization beam splitter 410, and the half mirrors 416 and 418 are omitted, and the imaging lenses 424 and 42 are omitted.
6 illustrates an example in which the objective lens 414 and the objective lens 414 are integrally formed. Therefore, in FIG. 8 and FIG.
It is shown that there are shielding plates 420, 422 between the optical sensors 24, 426 and the split optical sensors 428, 430.

Normally, if only the height of the unevenness on the surface of the wafer 306 is to be detected, the half mirror 418 is omitted, and the reflected laser light from the half mirror 416 is received by one two-divided optical sensor 430, and the detection is performed. The height information ε can be obtained based on the signals A and B. 2 split optical sensor 4
The length L of the boundary between the 30 light receiving regions corresponds to the detection region 308 of the wafer 306. The height information ε is the detection signals A and B
Is applied to the arithmetic expression (AB) / (A + B). For example, as shown in FIG.
If the wafer 306 has no projections (projections) or depressions (dents) and its surface is at the focal point of the objective lens 414,
The image point of the reflected laser light is formed on the surface of the two-divided optical sensor 430, and near the boundary between the two divided light receiving regions. As a result, the detection signals A and B output from the two-segment optical sensor 430 become substantially equal, so that the height information ε =
(AB) / (A + B) becomes “0”.

As shown in FIG. 8C, when a projection (projection) is present on the wafer 306, the image forming point of the reflected laser beam is located on the rear side of the two-divided optical sensor 430. Therefore, the left side of the two-divided optical sensor 430 (the detection signal A
On the side), a semicircular beam spot of the reflected laser light having a diameter corresponding to the size (height) of the projection (projection) is formed. Height information ε = (AB) /
(A + B) is a positive value (ε> 0). The height information ε corresponds to the size (height) of the projection (projection). Conversely, as shown in FIG. 9A, when there is a concave portion (dent) in the wafer 306, the imaging point of the reflected laser light is located in front of the two-divided optical sensor 430. Right side of the two-segment optical sensor 430 (detection signal B side)
Is irradiated with a beam spot of a semicircular reflected laser beam having a diameter corresponding to the size (depth) of the concave portion (recess). Height information ε = (AB) / (A
+ B) is a negative value (ε <0). The height information ε corresponds to the size (depth) of the concave portion (depression).

As described above, although the reflected laser light from the half mirror 416 is received by only one split optical sensor 430, the height information ε can be obtained based on the detection signals A and B. In the micro-projection inspection apparatus according to this embodiment, the half mirror 418 is provided, and the reflected laser light is split into two parts by a shielding plate (knife edge) 420 having a complementary positional relationship with the shielding plate 422. The light is received by the sensor 428, and height information ε is obtained based on the detection signals A to D. The height information ε is obtained by calculating the values of the detection signals A to D by an arithmetic expression {(AB) + (CD)} /
It is calculated by applying (A + B + C + D). As shown in FIGS. 8A and 8B, when the wafer 306 has neither a convex portion (projection) nor a concave portion (dent), the values of the detection signals A to D output from the two-divided optical sensors 428 and 430 are changed. Since they are almost equal, the height information ε is “0”.

As shown in FIG. 8C and FIG.
If there is a convex portion (projection) on the left side (detection signal A side) of the two-part optical sensor 430 and the two-part optical sensor 4
On the right side of 28 (on the side of the detection signal C), a semicircular beam spot of the reflected laser light having a diameter corresponding to the size (height) of the projection (projection) is irradiated. Therefore, the height information ε = {(AB) + (CD)} / (A
+ B + C + D) is a positive value (ε> 0). FIG.
In the case where a concave portion (dent) exists in the wafer 306 as in (a) and (b), the right side of the two-divided optical sensor 430 (the detection signal B side) and the left side of the two-divided optical sensor 428 (the detection signal D side) ) Is irradiated with a semicircular beam spot of the reflected laser light having a diameter corresponding to the size (depth) of the concave portion (recess). Therefore, the height information ε in this case is
= {(AB) + (CD)} / (A + B + C + D)
Is a negative value (ε <0).

As described above, the two systems of the two-divided optical sensor 4
Since the height information ε is obtained based on the detection signals A to D from the signals 28 and 430, the following effects are obtained in addition to the effect that the detection accuracy is improved. That is, the semiconductor laser light source 400, the S-polarizing plate 401, the collimating lens 402, the acousto-optic deflector 404, the deflector driving circuit 405, the fθ lens 406, the lens 408, the polarizing beam splitter 410, and the quarter Thermal fluctuation or acousto-optic deflector 4 such as wave plate 412 and objective lens 414
When spot deviation (Δt) occurs as shown in FIGS. 9C and 9D due to the deflection error of the optical sensor 04 itself, the spot position on the surface of the two-divided optical sensors 428 and 430 moves to the left (detection signal A side). It shifts. The spot scanning deviation (Δt) refers to a case where the laser beam falls at a certain angle, although the laser beam must fall vertically from the objective lens 414 to the wafer 306.

In such a case, the half mirror 418 is omitted as in the related art, and the reflected laser light from the half mirror 416 is received by only one split optical sensor 430.
When height information ε is obtained based on the detection signals A and B,
8C and 9C, the left side (detection signal A side) of the two-part optical sensor 428 is irradiated even though the size (height) of the convex portion (projection) is the same. The spot area of the reflected laser light to be performed differs in the cases of FIGS. 8C and 9C. If the spot areas are different as described above, an error height is detected in the case of FIG. 9C, although the actual height of the projections (projections) is the same. . That is, since the magnitude of the detection signal A is larger in the case of FIG. 9C than in the case of FIG. 8C, the arithmetic expression (AB) /
This is because the value (AB) on the numerator side of (A + B) fluctuates according to the magnitude of the detection signal A.

However, as in the microprojection inspection apparatus according to the present embodiment, two systems of the two-part split optical sensor 42 are used.
8,430 based on the detection signals A to D from the height information ε.
, The magnitude of the detection signal A becomes
Even if the case of FIG. 9C is larger than that of FIG. 9C, the magnitude of the detection signal C is complementary in the case of FIG. 9D than in the case of FIG. 8D. Becomes smaller. That is,
Equation for calculating height information ε {(AB) + (CD)}
The value on the numerator side of / (A + B + C + D) ｛(AB) + (C
-D)｝ is a substantially constant value without fluctuating. Therefore, according to the micro-projection inspection apparatus according to this embodiment, the thermal fluctuation of the optical element of the light projecting optical system and the acousto-optic deflector 4
The spot scanning deviation (Δ
Even if t) occurs, there is an effect that highly accurate height measurement can be performed without being affected by the occurrence of t).

When detecting a defect such as the height or shape of a three-dimensionally structured inspection object as in the microprojection inspection apparatus according to this embodiment, the inspection is performed with a three-dimensional object such as a pattern existing around the inspection object. The object must be clearly identified. Until now, when a plurality of inspection objects are arranged in a complicated manner on a sample to be inspected, complicated coordinate setting management has been performed for each of the inspection objects. Therefore, in the micro-projection inspection apparatus according to the present embodiment, detection signals A to D output from two-divided optical sensors 428 and 430 are provided.
Based on the above, the coordinate position of the bump to be inspected is measured in advance with high accuracy. Hereinafter, a method of measuring the coordinate position of the inspection object will be described with reference to FIGS.

FIG. 10 is a diagram for explaining an outline of a method for measuring the coordinate position of a bump to be inspected by binarizing a detection signal. FIG. 10A is a diagram of a part of the wafer surface on which bumps are formed, as viewed from above.
FIG. 10B is a diagram showing a cross-sectional shape of the bump of FIG. As shown in FIG. 10A, the bumps 72 are formed on the pads 70 on the pads 70 as shown in FIG. 10B.
It is formed so that it becomes. When the detection area indicated by the arrow 101 is scanned by the beam spot of the laser beam applied to the surface of the wafer 306, the A / D converters 504, 5
A signal representing the reflected light luminance as shown in FIG. That is, the reflected light luminance signal is the detection signal A output from the two-divided optical sensors 428 and 430.
To D, that is, the above-described arithmetic expression ｛(AB) +
(CD)｝ / (A + B + C + D) on the denominator side (A +
B + C + D).

This reflected light luminance signal (A + B + C + D)
It is known that indicates a specific value depending on the surface shape, base shape, material, and the like of the bump to be inspected and the wafer 306 to be inspected. That is, it is known that the reflected light luminance signal of the bump 72 shows an extremely small value as compared with the reflected light luminance signal of the surrounding wafer 306 as shown in FIG. Therefore, in this embodiment, a threshold value is set at approximately the middle of the respective reflected light luminance signals of the bump 72 and the wafer 306, and when the threshold value is larger than this threshold value, the low level is set to “0”.
When the value is small, the high level is set to “1”, and the binarized sampling of the surface of the wafer 306 as the sample to be inspected is performed.

FIG. 10D shows the result of binarized sampling of the reflected light luminance signal of FIG. 10C.
FIG. 10E shows that the area around the bump 72 in FIG.
It is a figure showing the result in the case of being quantified. As apparent from these figures, by binarizing and sampling the surface of the wafer 306, it is possible to specify the position where the bump exists with high accuracy without performing complicated coordinate management as in the related art. become able to. FIG.
FIG. 9 is a diagram schematically illustrating an adjacent comparison performed based on the coordinate positions of the binarized and sampled bumps by the measurement method of FIG. FIG. 11A shows a state of binarization around a bump to be compared for adjacent chips.
(B) is a graph showing the value of the height information ε near the center of each bump. In the figure, regions 75 to 79 of high level “1” are portions where bumps of each chip exist, and the comparison pitch of the adjacent comparison target is specified based on these. Then, the adjacent comparison based on the comparison pitch is performed based on the height information ε as shown in FIG. The result of the adjacent comparison is as shown in FIG. That is, the regions 75, 76, and 78 are determined to be normal as a result of the adjacent comparison, the region 77 is determined to be a missing bump defect, and the region 79 is determined to be a bump height defect. It should be noted that it is possible to detect a bump defect based on the binarized bump coordinate position in FIG. That is, in the case of FIG. 11A, it is possible to detect that the lower right portion of the region 77 is missing in the binarized sampling data, and that a missing bump defect exists in this portion.

FIG. 12 is a diagram showing a method of calculating the final bump height information ε ₀₁ and ε ₀₂ obtained by the defect detection program in the image processing / control device 500. FIG.
FIG. 4 is a diagram showing the concept of the necessity and method of calculating bump height information ε ₀₁ and ε ₀₂ . FIG. 13A is a view of a part of the wafer surface when there is nothing between the bumps as viewed from above. FIG.
FIG. 3A is a diagram illustrating a cross-sectional shape of each bump. FIG.
FIG. 13C is a waveform diagram showing the value of the bump height information ε, which is the result of the height measurement calculation performed on the bumps as shown in FIGS. FIG. 13D is a top view of a part of the wafer surface when a circuit wiring pattern or the like exists between bumps. FIG.
FIG. 13E is a diagram illustrating a cross-sectional shape of each bump and wiring pattern in FIG. FIG.
It is a waveform diagram which shows the value of height information (epsilon) of a bump which is a result of height measurement calculation processing performed with respect to a bump and a wiring pattern like (d) and (e).

As shown in FIGS. 13 (a) and 13 (b), when there is nothing between bumps, the above equation {(AB) + (CD)} / ( The height information ε of the bump obtained according to (A + B + C + D) is substantially the same as the cross-sectional shape of the bump in FIG. 13B, as shown in FIG. Therefore, in such a case, there is no problem even if the defect detection is performed based on the height information ε.

However, if a circuit wiring pattern exists between bumps as shown in FIGS. 13D and 13E, the operation formula ｛(AB) + (CD) ｝ / (A
+ B + C + D) bump height information ε obtained according to
As shown in FIG. 13 (f), the portion where the bump exists is almost the same as the cross-sectional shape of the bump in FIG. 13 (b), and there is no problem. is information epsilon becomes high level difference (height displacement) signal as shown in FIG. 13 (f), the height information including the different false information (imaginary displacement) epsilon ₁ is detected from the actual wiring pattern.

As shown in FIG. 8, the false report (imaginary displacement amount) ε ₁ is a difference value between the light receiving areas of the two-divided optical sensors 428 and 430, which are not shielded by the shielding plates 420 and 422. It is equal to the sum of (AC) and the difference value (BD) between the one-half area to be shielded. Therefore, the detection signals A to D output from the A / D converters 504 and 505 are calculated by using the arithmetic expression {(AC) + (BD)} /
By applying the (A + B + C + D ), false information (imaginary displacement) obtaining the epsilon _1. Figure 13 (g) is a waveform diagram showing a value of false information (imaginary displacement) epsilon ₁ obtained in this manner. As can be seen, false information from the height information of the bump epsilon to remove (imaginary displacement) epsilon _1, as shown in FIG. 13 (f)
The waveform of FIG. 13G may be subtracted from the waveform of FIG.

FIG. 12 (a) shows the bump height information ε _{01 obtained} by subtracting the waveform of FIG. 13 (g) from the waveform of FIG. 13 (f).
Image processing and control device 500 configured to output
FIG. 8 is a diagram showing a part of a processing block generated by a defect detection program in the inside. Processing block 121
/ D converters 504 and 505 output detection signals A to
D (A−B) + (C−D)} /
The height information ε of the bump is calculated by executing the height measurement calculation processing according to (A + B + C + D), and the pattern edge according to the calculation formula {(AC) + (BD)} / (A + B + C + D). calculating a false report (imaginary displacement) epsilon ₁ by performing the extraction processing. Processing block 12
2, by subtracting the false information (imaginary displacement) epsilon ₁ from the height information of the bumps calculated epsilon by process block 121, a false alarm, as shown in FIG. 13 (h) (imaginary displacement) epsilon
_The final bump height information ε01 not including ₁ is output.

On the other hand, as another method for removing false information (imaginary displacement) epsilon ₁ from the height information of the bumps epsilon, 13
To false alarm (imaginary displacement) epsilon ₁ of the waveform of (g), previously setting a predetermined threshold value, if the detection error (imaginary displacement) epsilon ₁ value exceeds this threshold, Processing is performed so as to mask the output of the bump height information ε. FIG. 12B compares the waveform of FIG. 13G with a predetermined threshold, masks the output of the bump height information ε of FIG. FIG. 11 is a diagram showing a part of a processing block generated by a defect detection program in the image processing / control device 500 configured to output height information ε ₀₂ . The processing block 123 calculates the arithmetic expression ｛(AB) + (C−) based on the detection signals A to D output from the A / D converters 504 and 505, similarly to the processing block 121.
D) Calculate bump height information ε by executing height measurement calculation processing according to｝ / (A + B + C + D),
Arithmetic expression {(AC) + (BD)} / (A + B + C +
Calculating a false report (imaginary displacement) epsilon ₁ by performing the pattern edge extraction processing in accordance with D). Comparator 1
24, the processing false information calculated in the block 123 (imaginary displacement) compares the epsilon ₁ and threshold value, a signal of a false alarm high when (imaginary displacement) epsilon ₁ is below the threshold value "1", If false alarm (imaginary displacement) epsilon ₁ is greater than the threshold the low level "0" signal, and outputs to the gate circuit 125. The gate circuit 125 inputs the bump height information ε calculated in the processing block 123 to one terminal, and inputs the high-level “1” or low-level “0” signal output from the comparator 124 to the other terminal. By opening the gate only when a high-level “1” signal is input, and closing the gate when a low-level “0” signal is input,
And outputs height information epsilon ₀₂ bumps removed the false information from the height information of the bump epsilon (imaginary displacement) epsilon _1. The image processing / control device 500 obtains the height information ε measured as described above.
_By inspecting whether there is a defect in a ball, a bump, or the like based on ₀₁ or ε ₀₂ , the height inspection of the minute projection can be increased without being affected by a false report (imaginary displacement amount) ε _1. Can be done with precision.

FIG. 14 is a view showing a modification of the micro-projection inspection apparatus shown in FIG. 3, in which only the components corresponding to the light projecting optical system and the bump defect detecting optical system in FIG. 3 are shown. Has been omitted. In FIG. 14, the same components as those in FIG.
The description is omitted. 14 is different from that of FIG. 3 in that a parallel beam of S-polarized light components transmitted through a collimator lens using a polygon mirror 403 instead of the acousto-optic deflector 404 of FIG. Is converted into deflection light reciprocating in the Y direction. Since the polygon mirror 403 is used, the semiconductor laser light source 400 (not shown), the S-polarizing plate 401 and the collimating lens 402 are arranged on the fθ lens 406 side, and the S-polarized component laser light is
03 will be irradiated. The polygon mirror 403
The rotational drive is controlled by the deflector drive circuit 405.

The bump defect detecting optical system shown in FIG. 14 differs from that shown in FIG. 4 in that a knife edge prism 423 is used instead of the half mirror 418 and the shielding plates 420 and 422.
Is used to reflect the reflected laser light from the half mirror 416 to the two-divided optical sensors 428 and 430. The knife-edge prism 423 is formed of a right-angle prism, and is arranged such that the top side of the tip coincides with the deflection direction. Therefore, the reflected laser light is branched by the knife edge prism 423 in two directions, up and down, and each of the branched laser lights is formed into an image forming lens 424 or 424 in a state where the lower half or upper half area of the reflected laser light is shielded. 426
Through the optical sensors 428 and 430.
When the deflected light is generated by using the polygon mirror 403 in this manner, the deflection of the polygon mirror causes
Spot scanning deviation (Δt) as shown in (c) and (d)
However, since the height information ε is obtained based on the detection signals A to D from the two systems of the two-divided optical sensors 428 and 430 as described above, a highly accurate height is obtained without being affected by the height information ε. Measurements can be made. Further, by using the knife edge prism 423, the number of parts can be reduced, and adjustment such as optical axis alignment can be easily performed.

FIG. 15 is a view showing a modification of the wafer surface defect detection optical system of the micro-projection inspection apparatus shown in FIG. 15, the same components as those in FIG. 4 are denoted by the same reference numerals, and the description thereof will be omitted. The optical system for detecting a defect on a wafer surface shown in FIG. 15 includes an imaging lens 440, an intermediate lens 442, a cross-shaped spatial filter 444, an imaging lens 446, and a photomultiplier 438.
The imaging lens 440 is arranged on the optical axis of the transmitted laser light from the half mirror 416, and forms an image of the transmitted laser light. The intermediate lens 442 forms the Fourier surface of the objective lens on the spatial filter 444. Cross-shaped spatial filter 4
44 is the X direction (main scanning direction) and the Y direction (deflection direction)
And a cross-shaped shielding member formed along. The cross-shaped spatial filter 444 includes a half mirror 416
Is disposed on the optical axis of the laser light that has passed through, and cuts the zero-order light component of the transmitted laser light, and forms a wiring pattern extending in the X and Y directions formed on the wafer surface by a cross-shaped shielding member. , And only high-order reflected and scattered light other than the zero-order light component is passed. The imaging lens 446 forms an image of the laser beam that has passed through the cross-shaped spatial filter 444 on the photomultiplier 438. The photomultiplier 438 receives the laser light passing through the cross-shaped spatial filter 444 and condensed by the imaging lens 446, and outputs a current corresponding to the intensity of the laser light.
As described above, by using the cross-shaped spatial filter 444, diffracted light such as a wiring pattern can be cut, and defects on the wafer surface can be detected with high accuracy. Note that the wafer surface defect detection optical system of FIG. 15 and the bump defect detection optical system of FIG. 14 may be combined.

In the above-described embodiment, the coordinate positions of the prismatic bumps to be inspected are determined with high precision in advance based on the detection signals (reflected light luminance signals) A to D output from the two-split optical sensors 428 and 430. Next, a method for measuring the coordinate position of the inspection object based on the current (scattered light luminance signal) output from the photomultiplier 438 will be described. FIG. 16 is a diagram schematically illustrating a method for measuring the coordinate position of a prismatic bump as an inspection object by binarizing a scattered light luminance signal. FIG. 16A is a diagram of a part of the wafer surface on which the prismatic bumps and the wiring patterns are formed, as viewed from above. FIG. 16B is a diagram showing a cross-sectional shape of the prismatic bump and the wiring pattern of FIG. 16A. As shown in FIG. 16A, the prismatic bumps 161 to 163 are
6 are formed so as to be convex portions (projections) as shown in FIG. When the detection area indicated by the arrow 168 is scanned by the beam spot of the laser light applied to the wafer surface, the A / D converter 510 outputs the signal shown in FIG.
A signal representing the scattered light luminance as shown in (c) is output.

The scattered light luminance signal has a specific value depending on the surface shape, base shape, material, wiring pattern, foreign matter 167, etc. of the prismatic bumps 161 to 163 to be inspected and the wafer to be inspected. It is known to show. That is, it is known that the scattered light luminance signals of the prismatic bumps 161 to 163 exhibit an extremely large value as compared with the scattered light luminance signals of the surrounding wafers as shown in FIG. Further, the scattered light luminance signal of the wiring pattern and the foreign matter 167 has a pulse-like waveform. Therefore, in this embodiment, a threshold value is set at almost the middle of the scattered light luminance signals of the prismatic bumps 161 to 163, and when the threshold value is larger than the threshold value, the high level is set to “1”. When the value is set to “0”, binarized sampling of the surface of the wafer to be inspected is performed.

FIG. 16D shows a binarized signal obtained as a result of binarized sampling of the scattered light luminance signal shown in FIG. 16C, and FIG. FIG. 7B is a diagram illustrating a result when the periphery of the prism bumps 161 to 163 is binarized. As is apparent from these figures, by binarizing and sampling the surface of the wafer, the position where the prismatic bumps exist can be specified with high accuracy without performing complicated coordinate management as in the related art. become able to. Note that the scattered light luminance signal of the foreign matter 167 is
Since the pulse waveform is larger than the threshold value, FIG.
As shown in FIG. 16D, a pulse-shaped waveform 167a having a sufficiently smaller width than the pulse-shaped waveforms 161a to 163a of the prismatic bumps 161 to 163 is detected.
It appears as a pattern corresponding to the foreign matter 167 as shown in FIG. Thus, the foreign matter 167 can be easily detected.

FIG. 17 is a diagram schematically illustrating a method of measuring the coordinate position of a ball bump to be inspected by binarizing a reflected light luminance signal. FIG. 17 (a)
FIG. 5 is a diagram of a part of the wafer surface on which ball bumps are formed, as viewed from above. FIG. 17B is a diagram showing a cross-sectional shape of the ball bump of FIG. As shown in FIG. 17A, the ball bumps 171 to 173 are
The lower end portions are formed on 4 to 176 by an annealing process as shown in FIG. When the detection area indicated by the arrow 178 is scanned by the beam spot of the laser beam applied to the wafer surface, the A / D converters 504 and 505 output the signals shown in FIG.
A signal representing the reflected light luminance as shown in (c) is output. That is, this reflected light luminance signal is
8,430, that is, the sum of the detection signals A to D, that is, the above-described arithmetic expression {(AB) + (CD)} / (A
+ B + C + D) on the denominator side (A + B + C + D).

It has been found that this reflected light luminance signal shows a specific value depending on the surface shape, base shape and material of the ball bumps 171 to 173 to be inspected and the wafer to be inspected. I have. That is, the reflected light luminance signals of the ball bumps 171 to 173 are as shown in FIG.
As shown in the figure, it is known that the vicinity of the top of the ball bumps 171 to 173 shows a reflected light luminance signal of a large value, but the vicinity of the outer periphery other than the top shows an extremely small value of the reflected light luminance signal. That is, when the beam spot of the laser beam irradiates the ball bumps 171 to 713, the reflected light luminance signal becomes extremely small. However, when the beam spot irradiates the vicinity of the top of the ball bumps 171 to 173, the reflected light luminance is reduced. When the signal increases and the beam spot irradiates the vicinity of the outer periphery of the ball bumps 171 to 173 again, the reflected light luminance signal decreases. Therefore,
In this embodiment, a threshold value is set at approximately the middle of the reflected light luminance signals near the ball bumps 171 to 173 and near the top and the outer periphery. The case is set to the low level “0”, and the binarized sampling of the wafer surface as the sample to be inspected is performed.

FIG. 17D shows a binarized signal as a result of binarized sampling of the reflected light luminance signal of FIG. 17C, and FIG. 17E shows FIG. FIG. 9B is a diagram illustrating a result when the area around the ball bumps 171 to 173 is binarized. As is clear from these figures, by binarizing and sampling the surface of the wafer including the ball bumps 171 to 173, the position where the ball bump exists can be specified with high accuracy. In the case of FIG. 17, since the beam spot of the laser beam scans near the center of the ball bumps 171 to 173 as indicated by an arrow 178, the width W1 of the binarized signal in FIG. 171 indicates the size (diameter), and the width W2 indicates position information near the top.

FIG. 18 is a diagram showing an example of a method of detecting the center of gravity and the height of the ball bump based on the binarized signal of FIG. 17D. Ball bump 17 on pad 174
In the case where a plurality of scans as shown in the figure are performed by using the beam spot of the laser light with respect to 1, the binarized signal has waveforms L1 to LF as shown on the right. Waveform L
The waveforms 1 to L5 and LB to LF are signals indicating a small concave portion of the reflected light luminance signal only in that portion because the beam spot passes only near the outer periphery of the ball bump 171. The widths W11 to W15 of the concave portions at this time
Is that the beam spot gradually increases in accordance with the distance over which the ball bump 171 is scanned, and conversely, the widths W1B to W1F of the concave portions gradually decrease in accordance with the scanning distance. In the waveforms L6 to LA, the reflected light luminance signal is a signal indicating a convex portion in a concave portion because the beam spot passes in the order of the outer periphery of the ball bump 171, the vicinity of the top portion, and the vicinity of the outer periphery. . The width W16 to W of the concave portion at this time
1A shows that the beam spot gradually increases in accordance with the distance over which the ball bump 171 is scanned, and the width W18 of the waveform L8
Shows the maximum, and then gradually decreases.
On the other hand, the widths W21 to W25 of the protrusions gradually increase in accordance with the distance over which the beam spot scans the vicinity of the top 171a of the ball bump 171; I have. Therefore, based on the waveforms L1 to LF, the ball bump 171 is formed.
Can be detected. That is, the largest value among the widths W21 to W25 is the width W.
23, the center position of the ball bump 171 in the X direction can be obtained by obtaining the center of the width W23, and the center position of the ball bump 171 in the Y direction can be obtained based on the detected waveform L8 of the width W23. You can ask. Since the center positions of the ball bump 171 in the X and Y directions are obtained in this manner, the waveform L8 at this time is obtained.
By obtaining the height information ε of the ball bump using the height measurement waveform corresponding to the above, the accurate height of the ball bump 171 can be measured. Also, width W11 to width W
The one with the largest value in 1F is the width W18,
By determining the width W18, the size of the ball 171 in the X direction can be determined. Further, the size of the ball 171 in the Y direction can be accurately measured by obtaining the detected waveform L1 having the width W11 to the detected waveform LF having the width W1F. A defect in the height and size of the ball bump can be detected based on the height information and size information of the ball bump thus obtained.

FIG. 19 is a diagram schematically illustrating a method of measuring a defect of a ball bump to be inspected by binarizing a scattered light luminance signal. FIG. 19 (a)
FIG. 4 is a view of a part of the wafer surface on which a defective ball bump and a wiring pattern are formed as viewed from above. FIG.
(B) is a figure which shows the cross-sectional shape of the ball bump and wiring pattern of FIG.19 (a). As shown in FIG. 19A, the ball bumps 191 to 193 are
6 are formed so as to be convex portions (projections) as shown in FIG. The ball bump 191 has a chipping defect on the right side surface. The ball bump 192 has a concave defect on the entire upper side. The foreign matter 197 adheres to the left side surface of the ball bump 193. When the detection area indicated by the arrow 198 is scanned by the beam spot of the laser light applied to the wafer surface, the A / D converter 510 outputs the signal shown in FIG.
A signal representing the scattered light luminance as shown in (c) is output.

The scattered light luminance signal depends on the defect shape of the ball bumps 191 to 193 to be inspected, the surface shape and the underlying shape of the wafer to be inspected, the material thereof, the wiring pattern, the foreign matter 197 and the like. It has been found to exhibit certain values. That is, as shown in FIG. 19C, the scattered light luminance signals of the chipped defect of the ball bump 191, the concave defect of the ball bump 192, and the adhered foreign matter 197 of the ball bump 193 are, as shown in FIG. It is known that the value shows an extremely large value as compared with. The wiring pattern has a pulse-like waveform.
Therefore, in this embodiment, the ball bumps 191 to 1
A threshold value is set almost at the center of the scattered light luminance signal of 93, and when it is larger than this threshold value, a high level “1” is set.
When the value is smaller, the low level is set to “0”, and the binarized sampling of the surface of the wafer to be inspected is performed.

FIG. 19D shows a binarized signal as a result of binarized sampling of the scattered light luminance signal of FIG. 19C, and FIG. 19E shows a binarized signal. FIG. 9B is a diagram showing a result when the area around the ball bumps 191 to 193 is binarized. In the case of the ball bump 191, a waveform 191A corresponding to the vicinity of the top 191a is provided.
The waveform 191B corresponding to the missing defect 191b appears as a binary signal. In the case of the ball bump 192, only the waveform 192B corresponding to the concave defect 192b appears. In the case of the ball bump 193, a waveform 193A corresponding to the vicinity of the top 193a and a waveform 193A corresponding to the adhered foreign matter 197 are provided.
7B appear as a binary signal. By comparing the shapes of the defects thus appearing based on the coordinate positions of the pole bumps obtained in FIG. 17, each defect can be extracted.

In the above embodiment, bumps have been described by way of example. However, the present invention relates to bumps, balls, projections for soldering, spacers for LCD filters, and the like.
It goes without saying that the present invention can be applied to the detection of various kinds of minute projections. Further, in the above-described embodiment, the optical sensor of the focus error detection optical system has been described as an example in which the optical sensor is divided into two. However, it is needless to say that the number of divisions may be larger than this. It is needless to say that the CCD light receiving element may be used. In the above-described embodiment, an example has been described in which the wafer surface defect detection optical system is provided so that the wafer surface can be inspected at the same time. However, these may be omitted. In the above-described embodiment, the semiconductor laser light source has been described as an example of the light projecting optical system. However, it goes without saying that other light sources such as white light may be used. In the above embodiment, the wafer mounting table is moved by the XY stage moving mechanism. However, it goes without saying that the optical system may be moved. In the above-described embodiment, an acousto-optic deflector and a polygon mirror have been described as examples in which the light beam is deflected so as to reciprocally scan at a predetermined cycle. However, a galvano mirror, a digital micromirror device (DMD), etc. It goes without saying that other deflectors may be used.

[0069]

According to the minute projection inspection apparatus of the present invention,
There is an effect that the height of the minute protrusion can be inspected with high accuracy and at high speed.

[Brief description of the drawings]

FIG. 1 is a diagram showing a conventional technique of an inspection apparatus that optically inspects whether there is a defect in the appearance of a ball or a bump provided for connecting a wiring such as an IC package or the like.

FIG. 2 is a diagram showing a conventional technique of an inspection apparatus that can detect whether a laser beam is reflected on a plus position (+) or a minus position (−) of a sample to be inspected.

FIG. 3 is a diagram showing a schematic configuration of a micro-projection inspection apparatus of the present invention.

FIG. 4 is a diagram three-dimensionally showing a part of the inspection optical system in FIG. 3;

FIG. 5 is a diagram illustrating a state of a beam spot that relatively scans on a wafer in XY directions.

FIG. 6 is a diagram showing a state in which a plurality of bumps are regularly arranged on a bump formation line provided along the outer periphery of a chip.

FIG. 7 is a diagram showing a scanning state of a laser beam on a bump on a bump formation line.

FIG. 8 is a diagram for explaining a detection principle based on a knife edge method employed in the minute projection inspection apparatus according to the embodiment.

FIG. 9 is another diagram showing an outline of a detection principle based on the knife edge method employed in the minute projection inspection apparatus according to the embodiment.

FIG. 10 is a diagram schematically illustrating a method of measuring a coordinate position of a bump as an inspection object by binarizing a detection signal.

11 is a diagram schematically illustrating an adjacent comparison performed based on the coordinate positions of the bumps binarized and sampled by the measurement method of FIG. 10;

FIG. 12 shows final bump height information ε obtained by a defect detection program in the image processing / control device 500.
FIG. 6 is a diagram showing a method of calculating ₀₁ and ε ₀₂ .

FIG. 13 is a diagram showing the necessity of calculating bump height information ε ₀₁ and ε ₀₂ and a concept of a calculation method.

FIG. 14 is a view showing a modification of the minute projection inspection apparatus shown in FIG. 4;

FIG. 15 is a view showing a modification of the wafer surface defect detection optical system of the micro projection inspection apparatus shown in FIG. 4;

FIG. 16 is a diagram schematically illustrating a method for measuring the coordinate position of a prismatic bump as an inspection object by binarizing a scattered light luminance signal.

FIG. 17 is a diagram schematically illustrating a method of measuring a coordinate position of a ball bump as an inspection object by binarizing a reflected light luminance signal.

FIG. 18 is a diagram illustrating an example of a method of detecting the center of gravity and height of a ball bump based on the binarized signal of FIG. 17D.

FIG. 19 is a diagram schematically illustrating a method of measuring a defect of a ball bump as an inspection object by binarizing a scattered light luminance signal.

[Explanation of symbols]

306: Wafer 400 ... Semiconductor laser light source 401 ... S polarizing plate 402 ... Collimate lens 403 ... Polygon mirror 404 ... Acousto-optic deflector 405 ... Deflector drive circuit 406 ... Fθ lens 408 ... Lens 410 ... Polarization beam splitter 412 ... Quarter 1 wavelength plate 414 ... objective lens 416, 418 ... half mirror 420, 422 ... shielding plate 423 ... knife edge prism 424, 426 ... imaging lens 428, 430 ... 2 split optical sensor 432 ... spatial filter 434, 440, 446 ... Image lens 436 Slit 438 Photomultiplier 442 Intermediate lens 444 Cross-shaped spatial filter 500 Image processing / control device 502A to 502D 506 Amplifier 504, 505, 510 Analog-digital (A /
D) Converter 508: Defect detection circuit 161-163: prismatic bumps 171-173, 191-193 ... ball bumps 164-166, 174-176, 194-196 ... pads 167, 197: foreign matter

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) H01L 21/60 H01L 21/66 J 21/66 21/92 604 (72) Inventor Hideo Ishimori Shibuya-ku, Tokyo 3-16-3 Higashi Hitachi Electronics Engineering Co., Ltd. (72) Inventor Takahito Tabata 3-16-3 Higashi 3 Chome, Shibuya-ku, Tokyo F-term in Hitachi Electronics Engineering Co., Ltd. 2F065 AA01 AA17 AA24 BB03 CC19 CC26 FF01 FF10 GG04 HH04 HH09 HH13 JJ01 JJ05 JJ17 LL04 LL14 LL28 LL33 LL36 LL57 PP12 PP22 QQ31 RR06 UU07 2G051 AA61 AB02 BA10 CA01 CA07 CB01 DA07 DA08 EA11 EA12 2H088 FA02 FA11 DB18 AFA18 FA11

Claims (13)

[Claims] 1. A projection system for irradiating a light beam so that the surface of a sample to be inspected is located at a focal point, and deflecting a predetermined inspection area on the sample to be linearly reciprocally scanned at a predetermined cycle. An optical optical system means, and two semi-circular light beams whose diameter is a straight line drawn by the optical axis of the light beam to be reciprocally scanned, wherein different light beams are shielded from each other, A detection optical system means for generating an image from the light beam reflected from the surface of the light beam, and provided corresponding to one of the two light beams generated by the detection optical system means; A first optical sensor unit that has a light receiving area divided at least into two so that the diameter of the light beam becomes a boundary, and outputs a detection signal corresponding to each light receiving area; To the other of the two light beams A second optical sensor means provided correspondingly and having at least two light receiving areas divided so that the diameter of the semicircular light beam is a boundary, and outputting a detection signal corresponding to each light receiving area; Stage means for mounting the sample to be inspected; drive means for relatively moving the stage means, the light projecting optical system means and the detection optical system means; and Control means for calculating height information of the minute projections formed on the sample to be inspected based on the detection signal output from the sensor means. 2. The optical system according to claim 1, wherein the detection optical system means generates the two light beams by reflecting a part of the light beam reflected from the surface of the sample to be inspected and transmitting the rest. A half-mirror unit, a first shielding plate unit that generates the semicircular light beam by shielding one side of one of the light beams generated by the half mirror unit, and a half-mirror unit. A second shielding plate means for generating a shielded semicircular light beam in an area different from the area shielded by the first shielding plate means by shielding one side of the other light beam; First imaging lens means for imaging the semicircular light beam generated by the first shielding plate means, and the semicircular light beam generated by the second shielding plate means Microprojections Inspection apparatus characterized by being configured to include a second imaging lens means for imaging. 3. The detection optical system according to claim 1, wherein the detection optical system means is provided such that a straight line drawn by an optical axis of the light beam reflected from the surface of the sample to be inspected coincides with a top side of a tip portion, Knife edge prism means for reflecting the light beam reflected from the surface of the sample to be inspected on respective slopes to generate the two semicircular light beams in which different regions are shielded, and the knife edge prism First imaging lens means for imaging one of the semicircular light beams reflected by the means, and second imaging means for imaging the other of the semicircular light beam reflected by the knife edge prism means. An apparatus for inspecting minute projections, comprising: an imaging lens means. 4. The microprojection inspection apparatus according to claim 1, wherein an acousto-optic deflector is used to deflect the light beam so as to linearly scan back and forth at the predetermined cycle. 5. The microprojection inspection apparatus according to claim 1, wherein a polygon mirror is used to deflect the light beam so as to linearly scan back and forth at the predetermined cycle. 6. The light-receiving element according to claim 1, wherein a zero-order light component of the light beam reflected from the surface of the sample to be inspected is cut and high-order reflected scattered light other than the zero-order light component is received. A microprojection inspection apparatus characterized by comprising: 7. The method according to claim 1, wherein the control unit sets a sum of the detection signals output from the optical sensor unit as a reflected light luminance signal from the surface of the test sample, and a value of the reflected light luminance signal. And a predetermined value obtained by comparing the position information of the minute protrusions formed on the sample to be inspected, and controlling the stage means based on the position information. . 8. The device according to claim 1, wherein the control unit is the detection signal output from the two divided light receiving regions of the first optical sensor unit, and is a region where the light beam is not blocked. The signal output from the light receiving area corresponding to the side of the light beam is referred to as a first detection signal, and the signal output from the light receiving area corresponding to the light blocking area of the light beam is referred to as a second detection signal. The third detection signal is the detection signal output from the light receiving area divided into two of the optical sensor means and output from the light receiving area corresponding to the unshielded area of the light beam, When a signal output from the light receiving area corresponding to the shielded area of the light beam is a fourth detection signal, a value obtained by subtracting the second detection signal from the first detection signal and the 3 detection signal The first value obtained by dividing the sum of the value obtained by subtracting the fourth detection signal from the first detection signal and the sum of the first to fourth detection signals is the height information of the microprojections formed on the sample to be inspected. A microprojection inspection apparatus characterized by calculating as: 9. The control device according to claim 8, wherein the control unit calculates a value obtained by subtracting the third detection signal from the first detection signal and a value obtained by subtracting the fourth detection signal from the second detection signal. And a value obtained by subtracting a value obtained by dividing the sum of the first and fourth detection signals by the total sum of the first to fourth detection signals from the first value, the height of the microprojections formed on the test sample. A microprojection inspection device, which is calculated as information. 10. The control unit according to claim 8, wherein the control unit calculates the fourth detection value from a value obtained by subtracting the third detection signal from the first detection signal and the second detection signal.
The value obtained by masking the first value when a value obtained by dividing the sum of the detection signal and the sum of the detection signals by the total sum of the first to fourth detection signals is larger than a predetermined value. Is calculated as height information of the minute projections formed on the sample to be inspected. 11. The apparatus according to claim 1, wherein the control unit performs a defect inspection of a minute projection formed on the inspection sample based on the calculated height information. Characteristic micro-projection inspection device. 12. The method according to claim 6, wherein the control unit compares the value of the detection signal output from the light receiving element unit with a predetermined value to obtain positional information of the minute projection formed on the sample to be inspected. A microprojection inspection apparatus characterized in that it is obtained. 13. The control unit according to claim 7, wherein the control unit compares a value of the reflected light luminance signal output from the optical sensor unit with a predetermined value when the minute projection has a spherical shape. The micro-projection inspection apparatus according to claim 1, wherein at least one of the center position information and the size information of the spherical micro-projection is obtained.