JPS62239116A - Method and device for focus position detection - Google Patents

Abstract

PURPOSE:To find an optimum position with a high precision by detecting a ratio of the area of a thin film pattern to the foundation area of an object to be measured in the area where a stripe pattern is projected and correcting the focus position decision result detected in accordance with a contrast curve of the stripe pattern. CONSTITUTION:Pattern masks 3a and 3b are irradiated through a lens 2, and light having a wavelength lambda1 or longer is intercepted. The light of stripes which are projected to a sample 6 has the wavelength lambda1 or longer cut off by a filter 14 and is focused on an image sensor 9, and meanwhile, the light of said stripes has the wavelength lambda1 or shorter cut off by a filter 15 to eliminate influences of masks 3a and 3b and is focused on an image sensor 16. The output of the image sensor 16 is binarized, and a pattern density is operated by a pattern density calculating circuit 17 and is inputted to a contrast detecting circuit 18. The circuit 18 calculates the contrast of projected stripe patterns 3a and 3b and uses the calculated pattern density to correct the extent of deviation of the focus, and a stage 12 is finely moved in accordance with the direction of deviation. By this constitution, the optimum focus position is detected with a high precision.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a method and device for detecting a focal point position in a microscope device for observing semiconductor wafers, etc., and particularly relates to a method and device for detecting a focal point position in a microscope device for observing semiconductor wafers, etc., and particularly for detecting a thin film pattern having steps such as a resist pattern on a wafer surface. The present invention relates to a focus position detection method and apparatus that are capable of always detecting the optimum focus position even when the focus position is

[Conventional technology]

Highly integrated LS 1. When inspecting the appearance of bubble memories, image pickup tube face plates, etc., a high-magnification microscope is used because these items have fine patterns of 2 to 37 J.

The depth of focus of a high-magnification microscope is 1/711 or less, and a precise automatic focusing mechanism is required when visual inspection is performed automatically.

Conventional focal position detection devices for this purpose include a contact type device using a stylus with a sharp tip, a device that determines the gap between the nozzle end and the surface of the object based on the flow resistance of the air ejected from the thin nozzle tip, and a device that uses capacitance to determine the gap between the nozzle end and the surface of the object. There were non-contact devices that measure gaps. However, in contact type devices, there is a risk that the sample may be damaged by the stylus. The non-contact type device for determining the gap requires a nozzle and a pickup in addition to the objective lens of the microscope, so there is a problem that it cannot be used when there is no space.

In response to this, the applicant previously proposed a device that detects the focal position by projecting a striped pattern onto the sample surface and detecting its contrast (Japanese Patent Laid-Open No. 58-705
No. 40, JP-A-58-120106).

Figure 13 shows its basic configuration. Light from light source 1 is transmitted through lens 2.
The two pattern masks 3a and 3b are illuminated through the light.

As shown in FIG. 14, the pattern masks 3a and 3b are striped patterns in which bright and dark areas with clear contrast are arranged alternately. are arranged side by side within. The pattern of this mask is projected onto a sample 6 via a semi-transparent mirror 4 and an objective lens 5.

The pattern projected onto the sample surface is formed by an objective lens 5 and a semi-transparent mirror 4.
．． An image is formed on the light-receiving surface of the image sensor 9 via the image sensor 7 .

An example of the detection signal of the image sensor 9 is shown in FIG. The left side is the projection pattern of the mask 3a, and the right side is the detection signal of the projection pattern of the mask 3b. Here, the contrasts Ca and Cb on the left and right sides are the projection pattern 3a,
Since the imaging plane 3b is shifted in the optical axis direction, it changes depending on the position of the sample 6 in the optical axis direction as shown in FIG. Therefore, the focusing position of the image sensor 8 for sample observation shown in FIG. 13 is set to a point Z where the contrasts Ca and Cb are equal.
. If set to , the direction of defocus can be determined by comparing both contrasts. In the example of FIG. 13, the contrast detection circuit 10 includes the projection pattern 3a,
The direction of defocus is determined from the contrast of 3b, and a signal corresponding to the direction is output.

The drive circuit 11 drives a stage fine movement mechanism 12 composed of, for example, a piezo element. This allows the image sensor 8 to be automatically focused.

[Problem that the invention seeks to solve]

As described above, according to the apparatus shown in FIG. 13, it is possible to project a striped pattern onto the sample and detect the focal point position by the contrast of the obtained reflected light. For example, if a semiconductor wafer has a thickness d of 1 as shown in
When using a sample in which a thin film pattern 21 such as a resist pattern of a size similar to that of a capacitor is formed on a base 20 with a step difference, if an objective lens with a high magnification of several tens of times or more is used, the surface Zd of the base 20, the thin film Surface Z of pattern 21
There is a problem in that no matter which side of u is focused, the pattern edges are observed to be blurred. For such samples,
In order to maximize the contrast of the pattern edges, it is necessary to focus on a point zo approximately in the middle of Zd and Zu.

However, the automatic focus detection device having the above configuration has a problem in that the stripe pattern is projected onto the thin film pattern 21, but the focal position varies depending on whether it is projected onto the base 20. As shown in FIG. 18, the thin film pattern 21
When the striped patterns 3a and 3b are projected onto the surface Zu of the thin film pattern 21, the focus is on the surface Zu of the thin film pattern 21, and when the striped patterns 3a and 3b are projected onto the base 20 as shown in FIG. The focus is on the surface Zd. In other words, as shown in Figs. 20 and 21, the intersection of the contrast curves Ca and Cb shifts back and forth with respect to the optimum focus position @Zll, so the optimum focus position is always determined for the image sensor for sample observation. There was a problem that the sample could no longer be moved, and the detected sample image would be out of focus6.The purpose of the present invention is to:
It is an object of the present invention to provide a highly accurate focus position detection method and apparatus that solves the problems of the prior art described above and can always detect the optimum focus position of a sample to be observed even when observing a sample with a large step difference at high magnification. be.

[Means for solving problems]

The above objective is the area ratio of the thin film pattern to the area of the base part of the object to be measured in the region on which the striped pattern is projected, that is, the pattern density (where the area of the base part is a and the area of the thin film pattern part is b), the pattern density is ρ is p= (b/
(a+b)) given by X100) and corrects the focus position determination result detected from the contrast curve of the striped pattern.

[Operation] When the pattern density in the stripe pattern projection area is 100%, that is, when the stripe pattern is projected onto the pattern part of the object to be measured as shown in FIG. 18, the focus judgment position determined from the intersection of the contrast curves is Zu, and as shown in FIG. 19, when the pattern density in the striped pattern area is 0%, that is, when the striped pattern is projected onto the underlying part of the object to be measured, the focus judgment position determined from the intersection of the contrast curves is Z
d. That is, the focus determination position determined from the intersection of the contrast curves changes linearly with the pattern density within the striped pattern projection area, as shown in FIG.

Therefore, the pattern density within the stripe pattern projection area is detected with respect to the focus determination position determined from the contrast curve obtained by stripe pattern projection, and this is used to ensure that the focus position is always Z. Highly accurate focusing becomes possible by applying the correction.

[Embodiments of the invention]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a configuration diagram illustrating one embodiment of the present invention.
Components that are the same as in the figures are given the same reference numerals. As the light source 1, one having substantially uniform spectral characteristics in the visible light region, such as a xenon lamp, is used. Light from light source 1 passes through lens 2
The striped pattern 3 is illuminated on the sample 6 through the semi-transparent mirror 4 and the objective lens 5.
Project a and 3b. Here, pattern masks 3a and 3b
The bright part transmits light of all wavelengths, and the dark part has the characteristic of blocking light of wavelength λ1 or less, as shown in FIG. The bright and dark striped pattern projected onto the sample 6 forms an image on the light-receiving surface of the image sensor 9 via the objective lens 5, the semi-transparent mirrors 4.7 and 13, and further via the optical filter 14. The optical filter 14 is provided with a second filter to enhance the contrast of the striped pattern.
As shown in the figure, it has the property of blocking light with a wavelength of λ or more. The light reflected by the semi-transparent mirror 13 enters the image sensor 16 via the optical filter 15. The optical filter 15 has the characteristics shown in FIG. 3 similarly to the dark areas of the pattern masks 3a and 3b. By using the optical filter 15, the image sensor 16 can image the area on which the striped pattern of the sample 6 is projected, almost without being affected by the pattern masks 3a, 3b. A pattern density calculation circuit 17 calculates the pattern density of the sample 6 from the output signal of the image sensor 16.

FIG. 4 shows the configuration of an embodiment of the pattern density calculation circuit, in which the output from the image sensor 16 is converted into a binarization circuit 4.
Binarization is performed using a certain threshold VTR using 0. 41 is a counter that counts the number of binary signals "1", 11017, and when the brightness of the pattern part formed on the base of the sample is brighter than the base part, the number of "1" is counted. The number of "0" corresponds to the area of the pattern part, the number of II OII corresponds to the area of the base part, and conversely, when the brightness of the pattern part is darker than the base part, the number of "0" corresponds to the area of the pattern part, ′
The number of 1" corresponds to the area of the base portion. Therefore, the pattern density of the sample in the striped pattern projection area can be calculated from the counting result of the counter 41.

FIG. 5 is a block diagram showing the configuration of another embodiment of the pattern density calculation circuit, in which difference circuits 42, 43, integration circuits 44,
It consists of a division circuit 45. After subtracting the brightness of the base portion measured in advance from the output of the image sensor 16, the difference is integrated by an integration circuit 44.

The pattern density can be calculated by further dividing the output of the integrating circuit 44 by the difference between the brightness of the pattern portion and the brightness of the underlying portion.

The pattern density thus determined is input to the striped pattern contrast detection circuit 18 (FIG. 1). The contrast detection circuit 18 calculates the contrasts Ca and Cb of the projected fringe patterns 3a and 3b imaged by the image sensor 9, but when determining the defocus direction from the calculated contrast, the pattern density calculated by the pattern density calculation circuit 17 is used. to correct the amount of defocus. That is, as shown in FIG. 6, the pattern density is ρ. If it is larger than ρ, use a negative correction amount. If it is smaller, a positive correction amount is added to the amount of defocus determined from the contrast curve, and then the direction of defocus is determined. The drive circuit 11 is a contrast detection circuit 18
The stage fine movement mechanism 12 is driven in accordance with the defocus direction output from the stage.

When the sample 6 includes a transparent pattern such as a resist pattern formed on a Si wafer, the difference in brightness between the base part and the pattern part is not clear, and it is difficult to determine the pattern density with the configuration of the embodiment shown in FIG. It can be difficult. 7th
The figure shows the configuration of another embodiment of the invention for such a sample. The dark portion of the striped pattern mask has a characteristic of blocking light having a wavelength of λ1 or more, as shown in FIG. 9, as described in the embodiment of FIG. In this embodiment, the light reflected by the semi-transparent mirror 13 is further divided into two optical paths by the semi-transparent mirror 30, one of which is transmitted to the image sensor 16 via an optical filter 34, and the other is transmitted to the image sensor 36 via an optical filter 35. Make it incident. Here, the optical filter 34 transmits light in a wavelength range of λ1 to λ2 as shown in FIG. 10, and the optical filter 35 transmits light in a wavelength range of λ2 to λ as shown in FIG. Keep it.

If the pattern formed on the sample 6 is a transparent pattern such as a resist pattern, the interference intensity will vary depending on the wavelength of the illumination light, so the transmission wavelengths λ1 to λ and the characteristics of the optical filters 34 and 35 should be appropriately selected. For example, image sensor 1
The amounts of light received by the transparent pattern portions 6 and 36 are different. On the other hand, if the spectral reflectance of the base is uniform to some extent, by comparing the output signals of the image sensors 16 and 36, it is possible to distinguish between the transparent pattern portion and the base portion and calculate the pattern density. If the pattern density can be calculated, the focal position can be detected by the same procedure as explained in the embodiment of FIG. 1. FIG. 8 shows the configuration of an embodiment of the pattern density calculation circuit in the embodiment of FIG. 7. A difference circuit 46 calculates the difference between the output signal of the image sensor 16 and the output signal of the image sensor 36.
Calculate with. As described above, since the interference intensity of the transparent pattern portion varies depending on the wavelength of the illumination light, the output signal of the difference circuit 46 has a value of zero for the base portion and a non-zero value for the pattern portion. Therefore, after the output signal of the difference circuit 46 is binarized by the binarization circuit 47 at an appropriate threshold value V'TH, the counter 47 counts the number of "0" or "1".
Pattern density can be calculated.

FIG. 12 shows the configuration of still another embodiment of the present invention, and the difference from the embodiment in FIG. 1 is that in FIG. 1, images of two striped patterns are captured by one image sensor. On the other hand, in this embodiment, the image of one striped pattern 3 is captured by two image sensors 9 and 9' which are shifted back and forth in the optical axis direction. In this embodiment, since the same striped pattern is imaged by two image sensors, it is not affected by the reflectance of the sample. For this reason, it becomes possible to improve the precision of focal position detection. It goes without saying that a similar method can be applied to the embodiment shown in FIG.

〔Effect of the invention〕

As explained above, according to the present invention, the focal position can be detected with high accuracy even when observing a sample with a large step difference at high magnification, and when applied to semi-dedicated wafer appearance inspection equipment, etc. , which has a remarkable effect on improving reliability.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention, FIGS. 2 and 3 are characteristic diagrams of the optical filter in FIG. 1, and FIGS. 4 and 5 are pattern densities as shown in FIG. 1. A block configuration diagram of the calculation circuit, FIG. 6 is an explanatory diagram of defocus amount correction in the embodiment, FIG. 7 is a configuration diagram of another embodiment of the present invention, and FIG. 8 is a pattern density calculation circuit in FIG. 7. FIG. 9, FIG. 10, and FIG. 11 are characteristic diagrams of the optical filter in FIG. 7, FIG. 12 is a configuration diagram of still another embodiment of the present invention, and FIG. 13 is a conventional FIG. 14 is a diagram showing the arrangement of the pattern mask in FIG. 13, and FIG. 15 is an explanatory diagram of the contrast signal.
Figure 6 is an explanatory diagram of the change in contrast signal depending on the sample position.
Figure 17 is a cross-sectional view of the wt sample, Figures 18 to 2
Figure 2 is an explanatory diagram of erroneous detection of the focal position in the conventional method.
In FIG. 18, the pattern mask 3 is projected onto the thin film pattern 21, and in FIG. 19, the pattern mask 3 is projected onto the base 21.
When projected on top, FIGS. 20 and 21 are diagrams showing contrast curves for FIGS. 18 and 19, respectively,
FIG. 22 is a diagram illustrating changes in focus determination position with respect to pattern density. Explanation of symbols> 1...Light source 2...Lens 3.3a,
3b...Pattern mask 4.7.13.30.31...Semi-transparent mirror 5...Objective lens

Claims (1)

[Claims] 1. A light pattern formed by a striped pattern mask that periodically combines bright and dark areas is projected onto an object to be measured, and the reflected light pattern from the object to be measured is imaged by an image sensor to generate light. In a focus position detection method that obtains brightness information corresponding to a pattern and detects the positional relationship between the surface of the object to be measured and the focal plane of the microscope from the contrast of this brightness information, the reflected light pattern from the object to be measured is detected by the image sensor. An image is captured by a second image sensor arranged separately from the second image sensor, and the area ratio between the base part and the pattern part in the area on which the light pattern of the object to be measured is projected is calculated from the second image sensor output signal. A method for detecting a focal point position, comprising: calculating and correcting a focal position of an object to be measured obtained from the brightness information according to the calculated area ratio. 2. A first optical system that projects and images a light pattern formed by a striped pattern mask that periodically combines bright and dark areas onto an object to be measured, and a second optical system that causes an image sensor to image a reflected light pattern from the object to be measured. and a contrast detection circuit that obtains brightness information corresponding to the light pattern from the image sensor output signal, processes the contrast of this brightness information, and detects the focal position of the object to be measured. a second image sensor that images the reflected light pattern from the object to be measured separately from the image sensor; and a second image sensor that captures an image of the reflected light pattern from the object to be measured; A pattern density calculation circuit for calculating the area ratio between the base portion and the pattern portion, and a correction means for correcting the focal position of the object to be measured obtained from the brightness information in accordance with the output of the pattern density calculation circuit. A focal position detection device characterized by: 3. The bright part of the striped pattern mask transmits light of all wavelengths, and the dark part has a characteristic of blocking only light of a certain wavelength band λ_u, and the light receiving surface of the second image sensor has the property of transmitting light of all wavelengths. 3. The focal position detection device according to claim 2, further comprising an optical filter that blocks only the light of λ_u. 4. The bright part of the striped pattern mask transmits light of all wavelengths, and the dark part has a characteristic of blocking only light of a certain wavelength band λ_u, and the second image sensor is composed of a plurality of image sensors. , wavelength band λ on each light-receiving surface.
3. The focal position detection device according to claim 2, further comprising an optical filter that transmits only light in wavelength bands other than _u that do not overlap with each other. 5. The focal position according to claim 3, wherein the pattern density calculation circuit is a circuit that calculates the pattern density using the average brightness of the output of the second image sensor. Detection device. 6. A patent characterized in that the pattern density calculation circuit is a circuit that binarizes the output of the second image sensor using a certain threshold and calculates the pattern density from the area ratio of the binarized image. A focal position detection device according to claim 3. 7. The pattern density calculation circuit is a circuit that calculates the pattern density by comparing outputs of a plurality of image sensors in the second image sensor with each other, as set forth in claim 4. Focus position detection device.