JP2003279311A - Optical elevation measuring device and method therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To quickly and accurately measure elevation of a structure formed dispersively. <P>SOLUTION: A device comprises a moving means (21) for moving a focus position of an objective lens (8) to the light axis direction; a first focusing optical system which constitutes a common focus optical system irradiating an object to be measured (9) with light through the objective lens, relatively moving the object to be measured and the objective lens to the light axis direction by means of the moving means, and gaining the reflection light from the object to be measured as a sectioning image; and a second focusing optical system for focusing the sectioning image by the first focusing optical system on an imaging means (12). From the relation between the focal point and the brightness value by integrating the outputs of a plurality of pixels of the imaging means including a bright point center where the reflection light from near the top of the structure or the predetermined center position of the structure, the elevation of the object to be measured is obtained. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical height measuring apparatus and method for measuring the height of a structure discretely formed on a substrate such as a wafer or glass epoxy.

[0002]

2. Description of the Related Art In recent years, the number of electrodes of an LSI chip has increased and the packaging density has increased due to high integration of LSI. From such a background, bump electrodes are being rapidly adopted as electrodes of LSI chips.

FIG. 10 is a schematic diagram of an LSI chip having bump electrodes as a sample. As shown in FIG. 10, a plurality of hemispherical bumps 100 are formed on the LSI chip 101. The size of the bumps and the pitch between the bumps are various, but for example, the bump radius is 50 μm and the pitch is 200 μm. LSI chip is 10mm x 1
If it is 0 mm, a huge number of bump electrodes, which number up to several thousand, are formed on this LSI chip.

FIG. 11 shows the LSI chip 1 shown in FIG.
It is a figure which shows the state which 01 and the board | substrate 102 were connected. Figure 1
As shown in FIG. 1, the LSI chip 101 on which a plurality of bumps are formed is a substrate 102 for connecting the LSI chip 101.
The electrodes of the substrate 102 and the bump electrodes of the LSI chip 101 are connected to each other. This is called flip-chip connection, but in order to make a normal connection, the bump shape, especially its height, must be accurately controlled as described later.

FIG. 12 is an enlarged view of the AA 'cross section of FIG. As shown in FIG. 12, the heights of the respective bumps 121 need to be aligned with the height level indicated by the dotted line in design, but in reality, like the black-painted bumps 122, the design height is allowed. There may be higher or lower ranges. If the flip-chip connection shown in FIG. 11 is performed in such a state, a defective connection between the bump and the substrate 102 will occur. Therefore, it is necessary to make the heights of the bumps formed on the LSI chip uniform within the allowable range.

From this background, there is a demand for in-line inspection of the height of all bumps with an accuracy of several μm before flip-chip connection. Moreover, since it is an in-line inspection, the inspection time is directly reflected in the cost of the LSI chip, so that high-speed inspection is essential.

An apparatus for the above-mentioned demand for high-speed and high-precision bump height inspection is disclosed in Japanese Patent Laid-Open No. 9-12673.
No. 9 publication. The three-dimensional shape measuring device disclosed in JP-A-9-126739 is a height measuring device using a confocal optical system. Laser scanning type and disc type (Nipkow disc, etc.) are known as confocal optical systems.
Both are optical systems having a function of converting the distribution in the height direction (optical axis direction) into the detected light amount.

FIG. 13 is a diagram showing a schematic configuration of a confocal optical system. The light emitted from the light source 211 is the pinhole 2
The light is focused on the sample 215 through the beam splitter 12, the beam splitter 213, and the objective lens 214. The light reflected by the sample 215 is reflected by the objective lens 214 and the beam splitter 213.
The light is focused on the detection pinhole 216 through the photodetector 21.
Light is received at 7. Here, assuming that the sample 215 is displaced downward by ΔZ in the optical axis direction, the light reflected by the sample 215 passes through the path shown by the dotted line in FIG. 13 and spreads greatly on the detection pinhole 216. Therefore, the amount of light that can pass through the detection pinhole 216 is extremely small,
The passing light amount can be regarded as substantially zero.

FIG. 14 is a diagram showing the relationship (I-Z characteristic) between the position of the sample in the Z direction and the amount of light I passing through the detection pinhole 216. FIG. 14 shows the relationship between the position Z of the sample 215 and the light amount I based on the focal position and the light amount I when the numerical aperture (NA) of the objective lens 214 shown in FIG. It is a figure, and when the sample is in a focus position (Z = 0), the light amount I is the largest (I = 1), and the light amount I increases as it deviates from the focus position.
Is decreasing. Therefore, when the sample is observed with the confocal optical system, only the vicinity of the focus position height looks bright.
This is called the sectioning effect of the confocal optical system.

In an ordinary optical microscope, a blurred image of a portion deviated from a focus position and an image at a focus position are observed in an overlapping manner.
In the confocal optical system, a sectioning image only at the in-focus position is observed due to the sectioning effect, which is a big difference from an ordinary optical microscope. The sectioning effect is more remarkable as the NA of the objective lens is larger. For example, when NA = 0.3, the sectioning effect is ± 10 μ from the focus position.
Only the sectioning image of the sample within m can be observed.

The sample 215 is scanned in the XY directions by a stage (not shown), or a galvanomirror is used as is well known, so that the pinhole image on the sample surface is, for example, X-axis.
By scanning in the Y direction and scanning the stage in the Y direction, one slice image is obtained. Next, the stage and the objective lens 214 are moved in the Z direction, and I-Z of each XY coordinate position is obtained from the sectioning images at a plurality of focal positions.
The height of the entire sample surface can be measured by, for example, interpolating a curve.

However, this method requires movement of the stage in at least two axes, so that it is impossible to obtain the required high-speed measurement. Therefore, JP-A-9-1
Japanese Patent No. 26739 discloses a bump height measurement which utilizes the IZ characteristic of this confocal optical system and can realize high speed. In order to utilize the I-Z characteristic of the confocal optical system, a rotary plate equipped with a plurality of parallel flat glass plates placed behind the objective lens is rotated at high speed to determine the relative position between the focus position and the sample. Move in the Z direction at high speed. The focal position is discretely moved according to the thickness of the parallel flat glass plate, and a sectioning image is obtained by the CCD camera. The number of sectioning images is the same as the number of parallel flat glass plates. The measurement range in the Z direction is determined by the thickness difference between the thickest parallel flat glass plate and the thinnest parallel flat glass plate.
The sampling interval in the Z direction can be made fine, or the Z measurement range can be expanded.

[0013]

The realization of the bump measuring device by configuring the principle confocal system shown in FIG. 13 as the scanning type is extremely effective in realizing the high speed required as described above. Have difficulty.

Further, in JP-A-9-126739,
This is effective when the bump electrode is large to a certain extent, the number of bumps is small and the arrangement is fixed, but when the bump becomes small and various bump chips of different sizes are to be dealt with, it is effective. Problem occurs.

In recent years, the bump electrodes have been further miniaturized, and the pitch is 50 μm and the bump height is 10 to 20 μm.
Things are about to be realized. As miniaturization of bumps progresses in this way, it becomes extremely difficult to align the production positions of the CCD light receiving portion and the pinhole plate. In addition, different size samples would have to be replaced with different size pinhole plates. To avoid this inconvenience, it is necessary to increase the magnification of the pinhole projected on the CCD. Then, it is necessary to divide one chip into a plurality of areas for measurement, or to use a huge CCD and an advanced optical system capable of coping with it. However, in the former case, it is expected that the manufacturing position adjustment of the CCD light receiving part and the pinhole plate and the exchange of the pinhole plate will be complicated, and the measurement time will be significantly increased.

In the case of JP-A-9-126739,
When the glass plate is rotated, the relative positions of the pixels of the CCD, which is the image sensor, and the image of the object to be measured are displaced due to changes in the aberrations of the glass plates of different thicknesses. This is because the position of the measurement object on which the measurement light strikes is different for each sectioning image at the timing of measuring the intensity.

It is considered that these are caused by vibrations that occur when scanning the measurement beam, when the sample stage or the objective lens moves in the Z direction, or where the adjustment of the optical system cannot be ideally performed. Therefore, it is not only a defect that occurs only in the above two examples, but also a problem that is common to confocal measurement of the confocal disk rotation method.

It is an object of the present invention to provide an optical height measuring apparatus and method for measuring the height of discretely formed structures at high speed and with high accuracy.

[0019]

In order to solve the above problems and achieve the object, the optical height measuring apparatus and method of the present invention are configured as follows.

(1) In the optical height measuring apparatus of the present invention, the structures which are discretely arranged on the substrate surface are used as the measurement object, and the height of the measurement object from the substrate surface is optically measured. In the optical height measuring device for measuring the distance, the moving means for moving the focal position of the objective lens in the optical axis direction, and the moving means for irradiating the object to be measured with light through the objective lens. The first imaging optical system that constitutes a confocal optical system that acquires the reflected light from the measurement target as a sectioning image while moving the measurement target and the objective lens relatively in the optical axis direction by And the first
A second image forming optical system for forming a sectioning image by the image forming optical system on the image pickup means, and the center of the bright spot on which the reflected light from the vicinity of the vertex of the structure is formed or is determined in advance. The height of the object to be measured is obtained from the relationship between the brightness value and the focus position obtained by integrating the outputs from the plurality of pixels of the imaging unit including the center position of the structure.

(2) The optical height measuring device of the present invention is the device described in (1) above, and the first imaging optical system has a confocal disc.

(3) In the optical height measuring method of the present invention, the structures which are discretely arranged on the substrate surface are used as the measuring object, and the height of the measuring object from the substrate surface is optically measured. In the optical height measuring method for measuring the object, while irradiating the measuring object with light through an objective lens, while relatively moving the measuring object and the objective lens in the optical axis direction, , The reflected light from the measurement object is acquired as a sectioning image, and the sectioning image is displayed as C
An image is formed on a CD, and outputs from a plurality of pixels of the image pickup unit including a bright spot center where reflected light from the vicinity of the vertex of the structure is imaged or a predetermined center position of the structure are integrated. The height of the measurement target is obtained from the relationship between the brightness value and the focus position at this time.

[0023]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a diagram showing a schematic configuration of an optical height measuring device according to an embodiment of the present invention. The device of FIG. 1 is roughly divided into three functions. That is, the illumination optical system including the light source 1, the lens (L) 2 and the beam splitter (BS) 3, the confocal disc 4, the imaging lens (T
L) 6, λ / 4 wave plate 7, diaphragm 13, objective lens (O
B) a first imaging optical system composed of 8 and a lens (L) 10,
2nd which consists of diaphragm 14, lens (L) 11, and CCD 12
It is divided into imaging optics.

Sample stage 15 below the objective lens 8
A sample 9 is placed on. Further, the CCD 20 and the focus moving device 21 are connected to the computer 20. The confocal disc 4 may have any pattern as long as it produces a sectioning effect.

The light emitted from the light source 1 is collimated by the lens 2, reflected by the beam splitter 3, and then incident on the confocal disk 4. The confocal disc 4 is, for example, a disc having a plurality of pinholes (for example, Nipk).
ow disk), which is rotated by the motor 5. The light that has passed through the plurality of pinholes of the confocal disk 4 reaches the sample 9 through the imaging lens 6, the λ / 4 wavelength plate 7, the diaphragm 13, and the objective lens 8.

In FIG. 1, for convenience, the light passing through two of the pinholes on the confocal disk 4 is shown. The pinhole on the confocal disk 4 and the focal plane of the objective lens 8 are conjugate with each other, and the imaging lens 6, the objective lens 8 and the telecentric diaphragm 13 are arranged in a bilateral telecentric system.
Further, the light source 1 and the telecentric diaphragm 13 have a conjugate relationship. Therefore, the first imaging optical system including the confocal disc 4, the imaging lens 6, the objective lens 8 and the like has intensity information indicating the height distribution in the optical axis direction of the sample 9 in the IZ characteristic of the confocal optical system. With the function to convert to.

The confocal disk 4 and the CCD 12 are in a conjugate relationship by the lenses 10 and 11. Lenses 10, 11,
The second imaging optical system including the CCD 12 and the like also includes the lens 10,
11 and a telecentric diaphragm 14 are arranged to form a double-sided telecentric system. The second image-forming optical system does not have to be a telecentric system, but if the length of the second image-forming optical system does not matter, a telecentric system that does not easily cause a decrease in peripheral light amount is more preferable.

By the first image forming optical system and the second image forming optical system,
The CCD 12 captures a sectioning image only near the focal plane of the objective lens 8. When the sectioning image captured by the CCD 12 is displayed on the monitor, only the vicinity of the focal plane looks bright and the portion displaced from the focal plane in the optical axis direction looks dark. Therefore, with the focus moving device 21,
If the sample stage 15 on which the sample 9 is placed or the objective lens 8 is moved in the optical axis direction to acquire a plurality of sectioning images, the three-dimensional information of the sample 9 can be obtained. At this time, the XY measurement range is the imaging field of view of the CCD 12, and the Z measurement range is the range in which the focus is moved to capture the sectioning image.

The images obtained when observing the spherical bumps formed on the LSI chip with the confocal optical system will be considered below with reference to FIGS. 2 and 3A and 3B. FIG. 2 is an enlarged view of a portion from the sample 9 to the diaphragm 13 in FIG. In FIG. 2, for the sake of convenience, the sample 9 under the objective lens 8 has one bump 91 and a wafer 92.
And the luminous flux passing through the diaphragm 13 is
It shows a light beam that has passed through a pinhole formed on the confocal disk 4. In reality, a plurality of light beams from a plurality of pinholes will pass through the diaphragm 13.

The light flux that has passed through the diaphragm 13 in FIG. 2 passes through the objective lens 8 and is condensed at a point P near the apex of the bump 91. By the way, since the bump has a hemispherical shape as shown in FIG. 2, a part of the reflected light reflected at the point P is eclipsed to the outside of the objective lens 8. The farther the point P is from the apex, the more the objective lens 8 out of the reflected light thereat.
The rate of re-injection into is decreasing.

A condition for resolving this phenomenon is that the principal ray L passing through the center of the diaphragm 13 is reflected at the point P on the bump and the principal ray L reflected at the point P is re-incident on the objective lens 8. In addition,
The numerical aperture (NA) of the objective lens 8 is NA = sin θ from θ in FIG. It is assumed that the bump is a perfect hemisphere, and the tangent line of the bump at the point P is B1-B2 and the perpendicular line to the tangent line B1-B2 at the point P is C1-C2.

As is apparent from FIG. 2, the chief ray L is a point P.
In order to be reflected by and be re-incident on the objective lens 8, the angle formed by the chief ray L incident on the point P from the objective lens 8 and C1-C2 needs to be θ / 2 or less. Chief rays L and C1
When the angle formed by −C2 is θ / 2, the distance from the center of the bump 91 to the point P in the X-axis direction (horizontal direction in the drawing) is R · sin (θ / 2 when the radius of the bump 91 is R. ). Therefore, the full width area of the point P at which the principal ray L can re-enter the objective lens 8 is 2R · sin (θ / 2).
When the point P further moves away from the center of the bump 91, the reflected light that can re-enter the objective lens 8 sharply decreases.

Next, a state in which a large number of bumps formed on the LSI chip are observed will be described. FIG. 3A is a diagram showing a confocal image when the vicinity of the bump 91 vertex on the LSI chip is focused. FIG. 3 (a)
Then, if the width of the white and brightly visible region shown at the center of the bump 91 is φ, then φ = 2R · sin (θ / 2). Actually, the chief ray incident on the bump 91 is the bump 9
When the distance from the center of 1 is great, the tangent line B1-B2 at the point P has a strong inclination, so that the ray reflected at the point P largely deviates from the principal ray L and cannot be re-incident on the objective lens 8. appear.

Further, due to the sectioning effect of the confocal optical system shown in FIG. 14, when the objective lens 8 having NA 0.3 is used, only a sectioning image of several μm can be observed in the Z direction from the bump apex. The reflected light from the wafer surface of the LSI chip located several tens of μm below the bump apex cannot be detected. Therefore, in FIG.
As shown in, a bright image can be observed only near the apex of the bump 91. In FIG. 3A, the density of the wafer 92 surface of the LSI chip and the density of the mesh portions of the bumps 91 are different, but this is for the purpose of explanation, and in reality, the density of FIG. ), Only the vicinity of the apex of the bump 91 looks bright, and the other mesh portions are in a dark state.

As the in-focus position is moved closer to the wafer 92 in FIG. 2, the apex of the bump 91 gradually becomes darker due to the sectioning effect of the confocal optical system, and the surfaces of the bump 91 and the wafer 92 eventually become dark. When the focus position is aligned with the wafer 92 surface, the wafer 92 surface looks bright.
Then, in the state of being focused on the surface of the wafer 92, FIG.
As shown in FIG.
Only two sides look bright.

As an example for realizing a wide-field, high-resolution measuring device according to the principle of the present invention, the NA of the objective lens 8 of the first imaging optical system is about 0.3, and the sample 9 on the disk surface is If the imaging magnification of is about 3 times and the magnification of imaging the disk 4 of the second imaging optical system on the surface of the CCD 12 is about 1/3 times (the total magnification becomes 1 times), it is very general. Even if the cheap CCD used is used, 10mm ×
A field of view of 10 mm and a resolution of 1 μm or less can be obtained relatively easily.

The sectioning range is NA 0.3.
Was ± 10 μm, but this is due to the fact that a blurred image can be superimposed on Z
Note that it is the maximum range of directions, not the resolution itself. The resolution is the sectioning range, the number of sectioning images, the interpolation method, and the electrical S / N.
It depends on various factors.

From the plurality of sectioning images obtained in this way, an I-Z curve of a necessary portion (where bumps are present) is formed by using the above-described interpolation operation or the like, whereby a minute and high-density bump array is formed. The height can be measured at high speed. In order to perform a more accurate measurement, the sectioning interval may be narrowed to improve the interpolation accuracy. If it is desired to further reduce the measurement time, the measurement accuracy may be slightly deteriorated, but it is also easy to shorten the Z scanning time and the image processing time by increasing the sectioning interval.

Next, a method of removing the cause of deterioration of measurement accuracy common to the confocal height measuring method will be described in detail.

4 (a) to 4 (e) and FIG. 5 (a) to
(E) is a figure which shows a part of light-receiving part of the CCD camera which receives the bright spot reflected from the bump vicinity. In (a) to (e) of FIG. 4 and (a) to (e) of FIG.
Pixels for which a curve is to be calculated are shown by meshes and are set as target pixels. Also, the white circles indicate the bright spots reflected from the vicinity of the bump vertices, but the sizes are not necessarily as shown in the figure, and may be larger or smaller.
Therefore, the white circle may be considered as a mark indicating an area representative of the spread of the bright spots. The black circle indicates the center of the bright spot area.

First, description will be made from (a) to (e) of FIG. Z0 in FIG. 4A indicates the state of bright spots near the in-focus position. Z ± 11 and Z ± 21 of (b) to (e) of FIG.
Shows a spread area of the bright spots when the focal point is moved up and down (Z-axis direction) about the focus position of Z0 by the same distance. 4A to 4E each show a case where the center (black circle) of the bright spot does not move at all.

Therefore, when the IZ curve is fitted to the change in the light amount of five target pixels having different Z positions in the vicinity of the bump apex, the ideal symmetrical curve 61 (complementary curve 61) shown by black circles and solid lines in FIG. ) Is completed.
The Z position corresponding to the in-focus position (bump height) read from the curve 61 is shown by a solid vertical line 62 in FIG.

On the other hand, in (a) to (e) of FIG.
As the focus position moves above and below the focus position (Z-axis direction), the optical axis shifts due to the movement of the sample stage 15 or the objective lens 8 in the optical axis direction, causing the center of the bright spot (black circle) to move in the horizontal direction. Moving. It does not matter whether the phenomenon in which the center of the bright spot is displaced is caused by the vibration of the stage or the like, or by the running of the optical system, and the phenomenon is caused for some reason.

At this time, the pixels of interest in FIGS. 5A to 5E are shown in FIG. 4 at Z0 near the in-focus position shown in FIG.
Only in the case of (b) of FIG. 5 in which reflected light of equal intensity is received in the same state as in (a) of FIG. On the other hand, it receives light of higher intensity than in the case of FIG. In other cases (c) to (e) of FIG.
The center of the bright spot moves in the direction away from the center of the pixel of interest, and all the values are lower than those in (c) to (e) of FIG. This is because the light intensity in the spread range of the bright spot becomes lower toward the periphery, so that if the bright portion at the center of the bright spot deviates from the pixel of interest, the amount of light taken in by the pixel of interest will decrease.

When the IZ curve is fitted to the changes in the light amount of the five target pixels having different Z positions again, a slightly distorted curve 63 (complementary curve) as shown by a mesh circle and a broken line in FIG. 6 is formed. . When the Z position corresponding to the in-focus position is read from this curve 63, the peak position of the broken line in FIG. 6 becomes the vertical line 64, which is clearly the ideal I-Z.
It is the peak position of the curve (solid line in FIG. 6) and deviates from the vertical line 62.

As described above, if one pixel of the CCD is focused and an IZ curve is formed from the change in the received light intensity, the height measurement value is always changed if the image laterally shifts for each sectioning image. I understand that I will do it.

In the present invention, when measuring the height of one bump, (a) to (e) of FIG. 4 and (a) to (a) of FIG.
As shown in (e), the IZ curve is not formed from the intensity change of the bright spot center for only one pixel of interest, but the computer 20 integrates the outputs of a plurality of pixels of the CCD including the bright spot center. The I-Z curve is calculated from the luminance value and the height is measured. That is, the computer 20 uses the focus moving device 2
1 to control the sample stage 15 or the objective lens 8
While controlling the movement in the direction of the optical axis to change the focal position, the integral value of the amount of light received by a plurality of pixels centered on the brightest pixel in the CCD 12 on which the reflected light from the vicinity of the bump apex is imaged is obtained. , The height of the bump is determined from the relative movement amount of the sample 9 and the objective lens 8 and the IZ curve of the luminance value. Thereby, the above-mentioned error can be remarkably improved.

In the experiments conducted by the present inventors, as shown in FIG. 7, the center positions of the bright spots of a solder bump having a diameter of 100 μm and a height of 80 μm are within 6 pixels within one pixel of the CCD (2 μm when the CCD pixel is 10 μm × 10 μm). Every other time, the height repeatability was compared. As a result, 3
It was found that the value of σ was about 0.7 μm when there was only one pixel for integrating the light amount, but it was about 0.12 μm when measured with the integrated intensity of 5 × 5 pixels. By the way,
The CCD camera used in this experiment is MODEL made by Sony Corporation.
The XC-7500.

Further, in this experiment, the area of the bump apex which appears bright in FIG. 3A is about 3 × 3 pixels at a total magnification of 1 ×, which is a small area compared to the entire bump area and is a hemisphere. You are only looking at the area near the top.

However, the integration area of the light quantity may not be increased as much as possible. 3 (a), (b)
Structures other than bumps with high reflectance in the vicinity of the bump area shown in (ex. A1 wiring)
Therefore, it should be stopped in the bump area because it may cause an error.

The reason why the bump height measurement is improved by the light amount integration as described above is that the signal is completely cut off (isolation) between the pixels of the CCD as shown in FIGS. It is considered that this is not due to the fact that the charges of the adjacent pixels flow into the adjacent pixels.

FIG. 8A shows a state in which a light spot having a size equal to or smaller than the CCD pixel is continuously scanned within the range of the drawing on the three pixels a, b and c of interest, and FIG. (B) is pixel a,
The photoelectric conversion outputs of b and c are shown.

Further, FIG. 8C shows a change in the sum of the intensities of the three pixels a, b and c. From FIG. 8C, it can be seen that the intensity change depending on the location of the bright spot is unexpectedly small. This is I-Z for each pixel of CCD
The most reason for obtaining a reproducible value is to calculate from the intensity obtained by integrating the intensity of a plurality of pixels including the apex of the bump, rather than calculating the curve to measure the height of the bump.

The method of selecting the light amount integration region for obtaining the effect of the present invention is not particularly limited, and the strongest reflection during focusing in the bump region 91 as shown in FIG. 9A, for example. Pixel where light is observed (the brightest pixel) 92
A predetermined continuous area (for example, 5 × 5)
Pixel) is selected as the light amount integration area, or (b) in FIG.
As shown in, various methods such as selecting a plurality of bright pixels 93 (for example, 6 pixels) in order from the brightest pixel 92 in the bump region 91 are conceivable. Also, you can select multiple pixels arranged in a line centered on the brightest pixel,
A plurality of pixels arranged in a cross shape with the brightest pixel as the center may be selected and integrated. Alternatively, in addition to the method of integrating the outputs from a plurality of pixels around the center of the bright point of the bump apex imaged on the image sensor such as CCD, the center of each bump can be determined from the design data or the image captured by the image sensor. Find the position,
It is also possible to integrate the outputs from a plurality of pixels with the pixel corresponding to this center position as the center.

As a means for acquiring a plurality of sectioning images, the sample stage 1 is used as described above.
In addition to the method of moving the objective lens 5 or the objective lens 8 in the optical axis direction, the imaging lens 6 may be moved in some cases. In this case, the imaging lens 6 will generally be lighter in weight than the objective lens 8 and the sample stage 15, so that higher-speed Z scanning can be performed.

The present invention is not limited to the above-mentioned embodiments, but can be carried out by appropriately modifying it without departing from the scope of the invention.

[0058]

According to the present invention, it is possible to provide an optical height measuring apparatus and method for measuring the height of a structure formed discretely at high speed and with high accuracy.

[Brief description of drawings]

FIG. 1 is a diagram showing a schematic configuration of an optical height measuring device according to an embodiment of the present invention.

FIG. 2 is an enlarged view showing a schematic configuration of an optical height measuring device according to an embodiment of the present invention.

FIG. 3 is a diagram showing a confocal image when focused on a bump vertex on an LSI chip and a confocal image when focused on a wafer surface of an LSI chip according to an embodiment of the present invention.

FIG. 4 is a diagram showing a state in which a bright point near a bump vertex does not fluctuate on a CCD and is blurred with Z movement according to an embodiment of the present invention.

FIG. 5 is a diagram showing a state in which a bright spot near a bump apex is defocused while fluctuating on a CCD along with Z movement according to an embodiment of the present invention.

FIG. 6 is a diagram showing a difference in height measurement value by the complementary calculation according to the embodiment of the present invention.

FIG. 7 is a diagram regarding repeatability according to the embodiment of the present invention.

FIG. 8 is a diagram showing the reason for improving bump height measurement by light amount integration according to the embodiment of the present invention.

FIG. 9 is a diagram showing how to select a light amount integration region according to the embodiment of the present invention.

FIG. 10 is an LSI having a bump electrode according to a conventional example.
The schematic plan view of a chip.

FIG. 11 is a side view showing a state where an LSI chip and a substrate according to a conventional example are connected.

12 is an enlarged view of an AA ′ cross section of FIG. 10 according to a conventional example.

FIG. 13 is a diagram showing a schematic configuration of a confocal optical system according to a conventional example.

FIG. 14 is a diagram showing IZ characteristics according to a conventional example.

[Explanation of symbols]

1 ... Light source 2 ... Lens 3 ... Beam splitter 4 ... Confocal disc 5 ... Motor 6 ... Imaging lens 7 ... λ / 4 wave plate 8 ... Objective lens 9 ... Sample 10 ... Lens 11 ... Lens 12 ... CCD 13 ... Aperture 14 ... Aperture 15 ... Sample stage 20 ... Computer 21 ... Focus moving device 91 ... Bump 92 ... Wafer

   ─────────────────────────────────────────────────── ─── Continued front page    F-term (reference) 2F065 AA02 AA04 AA06 AA24 AA53                       BB01 CC19 CC25 DD06 FF04                       FF10 HH03 JJ01 JJ26 LL28                       LL30 LL36 LL46 LL59 MM13                       PP02 PP24 QQ14 QQ17 QQ24                       QQ25 QQ31 QQ36                 4M106 AA01 AB20 BA04 CA48 DB04                       DB07 DB13 DB30 DJ12

Claims (3)

[Claims] 1. An optical height measuring device for optically measuring the height of the measurement object from the surface of the substrate, wherein structures discretely arranged on the substrate surface are used as the measurement object. Moving means for moving the focal position of the objective lens in the optical axis direction, and irradiating the measuring object with light through the objective lens, and moving the measuring object and the objective lens by the moving means. While moving relatively in the axial direction,
A first imaging optical system that constitutes a confocal optical system that acquires reflected light from the measurement target as a sectioning image, and a second imaging image that forms a sectioning image by the first imaging optical system on an imaging unit. The image forming optical system of the above, and from a plurality of pixels of the image pickup means including a bright spot center where reflected light from the vicinity of the vertex of the structure is imaged or a center position of the structure determined in advance. An optical height measuring device, characterized in that the height of the object to be measured is obtained from the relationship between the brightness value obtained by integrating the output and the focus position. 2. The optical height measuring apparatus according to claim 1, wherein the first imaging optical system has a confocal disc. 3. An optical height measuring method for optically measuring the height of the measuring object from the surface of the substrate, wherein the structures discretely arranged on the substrate surface are used as the measuring object. While irradiating the measurement object with light through an objective lens, while relatively moving the measurement object and the objective lens in the optical axis direction, a sectioned image of reflected light from the measurement object is obtained. A plurality of the imaging means including the center position of the bright spot on which the reflected light from the vicinity of the vertex of the structure is imaged or the center position of the structure determined in advance. An optical height measuring method, characterized in that the height of the measuring object is obtained from the relationship between the luminance value obtained by integrating the output from the pixel and the focal position at this time.