JPH10247466A - Astigmatic correction method and astigmatic correction device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a charged particle optical lens barrel used in SEM, and to easily correct astigmatism with high accuracy. SOLUTION: A charged particle optical lens barrel has stigmata for correcting astigmatism. An astigmatic correction method of the charged particle optical lens barrel includes a first step S1 of focusing an objective lens into a secondary particle signal image obtained by extracting a secondary particle signal obtained by two-dimensionally scanning a sample with an electron beam, and also includes a second step S2 of correcting astigmatism by adjusting the stigmata to reduce the astigmatism of the electron beam with respect to the secondary particle signal image obtained in an adjusted focusing position.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for correcting astigmatism in a charged particle optical column such as a scanning electron microscope used for observing a fine pattern of a semiconductor device.

[0002]

2. Description of the Related Art In recent years, a scanning electron microscope (SEM) has been used to observe the surface state of a semiconductor device, and its resolution has been reduced to less than 1 nm. FIG. 14 shows a general configuration of this type of SEM, which will be briefly described.

A cathode 1 and an anode 2 constitute a field emission type electron gun 3 from which a high brightness electron beam 4 is generated. The electron beam 4 is reduced through a condenser lens 5, an aperture 7 and an objective lens 6, and is irradiated on the surface of the sample 8. At this time, the detector 11 detects secondary electrons emitted from the sample surface. The electron beam 4 is two-dimensionally scanned on a sample 8 placed on a sample table 9 by a deflector 10 controlled by a scanning control circuit 12. By synchronizing with this and displaying the signal of the detector 11 on the image display 13,
Information on the sample surface can be obtained.

Incidentally, in a focusing system using such an objective lens 6, there is an aberration generally called "astigmatism". This occurs because the position where the electron beam is focused in two directions perpendicular to the beam axis is shifted in the axial direction. When such aberrations exist, the beam cross-sectional shape on the sample surface changes according to the focal length of the objective lens 6 as shown in FIG.
It changes as shown in (a) to (c). Note that FIG.
The beam diameter at which the circular beam shown in (b) is obtained is
This is larger than the case where such astigmatism does not exist. Such aberration occurs because the focal lengths in the two directions are different.

Therefore, a mechanism 15 for generating a quadrupole electric or magnetic field called "stigmata" is conventionally provided in a charged particle optical lens barrel comprising an electron gun 3, a condenser lens 5, an objective lens 6, an aperture 7, and the like. It is provided to correct astigmatism. Specifically, 4 below the objective lens 6
A quadrupole stigmat consisting of individual electromagnets or electrodes is provided so that it can be rotated 45 degrees around the optical axis, or two sets of quadrupole stigmaters are provided that are shifted from each other by 45 degrees and the strength of each is adjusted. , X-direction and y-direction.

Hereinafter, a method of actually adjusting astigmatism will be described with reference to FIG. When the sample as shown in FIG. 16A is observed, if astigmatism is present, the image on the display 13 has, for example, a cross-sectional shape of the beam shown in FIG.
In the case of (a), as shown in FIGS. 16 (b) and (c), a direction in which the edge resolution is high and a direction in which the edge resolution is low appear. The beam shape is changed by changing the focal length of the objective lens 6 as shown in FIG.
In the case of (c), the direction in which the resolution of the edge is high is switched. Therefore, the stigmator 15 is adjusted so that the image has no direction even if the focal length of the objective lens 6 is changed. At this time, the beam cross-sectional shape is as shown in FIG.
As shown in (a) to (c), only the diameter of the circular beam changes.

Here, in the case of automatically correcting astigmatism, the following operation is required. First, the contrast of images obtained at different focal lengths of the lenses is obtained, and the position where the contrast is maximum is obtained. When the astigmatism is large, two maximal positions may appear, but in such a case, the position between them is taken. At the focal position determined in this way, the stigmata in the x direction is adjusted to maximize the contrast. Then adjust the y-direction stigmata to maximize contrast. The above operation is repeated, and the converged position is determined as the optimum position.

However, this type of astigmatism correction method has the following problems. That is, since the adjustment of the stigmata depends on the visual sense of the operator, for example, the operator adjusts the stigmata while watching the image on the display device, the adjustment varies depending on the skill of the operator. If the adjustment of the stigmata is not perfect, the observation accuracy will be reduced. In addition, it takes time for adjustment to be performed by humans,
The sample is irradiated with a beam more than necessary. This is not preferable in many cases because it may damage the sample or change the resist shape. Furthermore, correction is difficult when the sample has directionality.

In the method of automating the correction of astigmatism, the range of correction that can be performed on an actual sample is smaller than that performed by a human, and in most cases, the correction of astigmatism is performed by hand. Inferior. For example, when observing a sample as shown in FIG. 16A when the cross-sectional shape of the beam is as shown in FIG. 18, the resolution is substantially equal to the direction perpendicular to the two sides, and the astigmatism Difficult to correct. This also applies to the case where the astigmatism correction is automated.

[0010]

As described above, conventionally, correction of astigmatism in a charged particle optical lens barrel has to be adjusted by the skill of an operator because the adjustment of the stigmata is performed by the operator's vision. There are problems such as a difference, a long time for adjustment by a human for adjusting the stigmata, and difficulty in correcting when the sample has directionality. Further, the above problem is not limited to an observation apparatus such as an SEM, but can be similarly applied to a charged particle beam drawing apparatus using an electron beam or an ion beam.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has as its object to correct astigmatism in a charged particle optical column quickly and with high accuracy. It is an object of the present invention to provide an astigmatism correction method and an astigmatism correction device for a charged particle optical lens barrel that can prevent a decrease in observation accuracy due to a beam cross-sectional shape.

[0012]

[Means for Solving the Problems]

(Structure) In order to solve the above problem, the present invention employs the following structure. That is, the present invention provides a method for correcting astigmatism in a charged particle optical column, in which a charged particle beam having a stigmata for correcting astigmatism is used, and a charged particle beam is two-dimensionally placed on a sample. For the secondary particle signal image created by extracting the secondary particle signal obtained by scanning, the focal point of the lens system of the charged particle optical column is adjusted based on the secondary particle signal image information. First to match
And a second step of correcting the astigmatism of the secondary particle signal image obtained at the in-focus position by adjusting the stigmata so that the astigmatism of the charged beam is reduced. It is characterized by the following.

Here, preferred embodiments of the present invention include the following. (1) The first step is a third step of generating astigmatism by stigmata, a fourth step of inputting a secondary particle signal image in a state where the stigma is generated, and the input secondary particle signal A fifth step of obtaining the direction of astigmatism from the image, a sixth step of setting a focal position at a preset position, and a seventh step of inputting a secondary particle signal image at the set focal position Step, an eighth step of rotating the input secondary particle signal image so that the direction of astigmatism is horizontal or vertical in the image, and calculating image sharpness from the rotated secondary particle signal image A ninth step, an eleventh step of calculating a focus position at which the sharpness becomes a peak from among the sharpnesses obtained for the preset focus positions, and a calculated sharpness becomes a peak Focus on focus position And a twelfth step. (2) The second step is a thirteenth step of setting a stigmata value in the x direction, a fourteenth step of inputting a secondary particle signal image, and setting the input secondary particle signal image in advance. A fifteenth step of rotating the image by a predetermined angle, a sixteenth step of calculating the sharpness of the image from the rotated secondary particle signal image, and a step of setting a previously set x-direction stigmata value. An eighteenth step of calculating an x-direction stigmata value at which the sharpness peaks from the obtained sharpness, and adjusting the x-direction stigmata value to the calculated x-direction stigmata value at which the sharpness peaks. A nineteenth step and a twentieth step of setting a stigmata value in the y direction while holding the stigmata value in the x direction
, A twenty-first step of inputting a secondary particle signal image, a twenty-second step of rotating the input secondary particle signal image by a preset angle, and a rotated secondary particle signal image 23rd which calculates the sharpness of the image from
And a twenty-fifth step of calculating a y-direction stigmata value at which the sharpness peaks from the sharpness obtained for each of the preset y-direction stigmata values. The twenty-sixth adjustment of the y-direction stigmata to the y-direction stigmata at which the sharpness reaches a peak.
Steps. (3) The fifth step is to perform a Fourier transform in a secondary space of the secondary particle signal image to calculate a power spectrum.
7, a 28th step for obtaining an image obtained by binarizing the power spectrum, a 29th step for obtaining a main axis of the binarized image and an axis in a direction orthogonal to the main axis,
A thirtieth step of determining the strength and direction of astigmatism by determining the distance of each point in the digitized image to the principal axis and the distance to an axis perpendicular to the principal axis. (4) The ninth, sixteenth, and twenty-third steps divide the plurality of search regions set around each point in the secondary particle signal image into two regions, and calculate the variance between these two regions. Determined in each search area, the maximum or average value of the inter-area variance determined in each search area as the variance of each point in the image, and the variance of the variance determined at each point in the area set in the image The sum or average value is calculated as the sharpness of the secondary particle signal image. (5) The ninth, sixteenth, and twenty-third steps set two vertical and horizontal search areas in the secondary particle signal image with each point as the center of gravity in the image, and set these search areas in the image. two regions a centered on each point, each divided into B, the average luminance of pixels included the m _a and m _B in regions a and B, m _T region a, the average of the entire search area combined B when the _{brightness, σ = (m a -m T} ) × (m a -m T) + (m B -m T)
× (m _B −m _T ), the variance between the areas A and B in the horizontal search area and the vertical search area is determined, and the sum or average of these variances is calculated as a secondary particle. It is calculated as the sharpness of the signal image. (6) In an eleventh step, a focus position indicating the maximum value is selected from among the sharpness values calculated at a plurality of preset focus positions. (7) In the eleventh step, a sharpness curve showing a relationship between the focus position and the sharpness is created for the sharpness obtained at a plurality of preset focus positions, and a Gaussian curve is added to the sharpness curve. A curve such as a distribution curve or a quadratic curve is applied, and a focal position at a position where this curve becomes a peak is calculated. (8) In the eighteenth step, a plurality of preset x
From among the sharpness values calculated for the stigmata values in the direction, the x-direction stigmata value indicating the maximum value is selected. (9) In the eighteenth step, a plurality of preset x
For the sharpness determined at the stigmata value in the direction, x
Creates a sharpness curve showing the relationship between the stigmata value in the direction and the sharpness, applies a curve such as a Gaussian distribution curve or a quadratic curve to the sharpness curve, and calculates the focal position at the position where this curve becomes a peak I do. (10) The twenty-fifth step includes a plurality of preset y
From the sharpness calculated for the stigmata value in the direction, a stigmata value in the y direction that indicates the maximum value is selected. (11) The twenty-fifth step includes a plurality of preset y
For the sharpness determined at the stigmata value in the direction, y
Creates a sharpness curve showing the relationship between the stigmata value in the direction and the sharpness, applies a curve such as a Gaussian distribution curve or a quadratic curve to the sharpness curve, and calculates the focal position at the position where this curve becomes a peak I do.

According to the present invention, in a device for correcting astigmatism in a charged particle optical column having a stigmator for correcting astigmatism, a charged particle beam is two-dimensionally scanned on a sample. Secondary particle signal image input means for inputting a secondary particle signal image created by extracting a secondary particle signal; focusing position detecting means for focusing the secondary particle signal image; and the charged particle optical mirror A focus adjusting means for adjusting the focus position by changing the objective lens of the cylinder; and a non-charged beam non-charged beam with respect to the secondary particle signal image inputted by focusing on the focus position obtained by the focus position detection means. Astigmatism correction means for correcting the astigmatism by adjusting the stigmata so as to reduce the astigmatism is provided.

Here, preferred embodiments of the present invention include the following. (1) The stigmata corrects astigmatism in the x direction and the y direction, and the x direction stigmata adjuster adjusts the x direction stigmata by the astigmatism corrector and the astigmatism corrector adjusts the y direction by the astigmatism corrector. Y direction stigmata adjusting means for adjusting the direction stigmata. (2) Focusing point detection means, astigmatism direction calculation means for obtaining the direction of astigmatism from the secondary particle signal image input from the secondary particle signal image input means,
Image rotating means for rotating the secondary particle signal image input from the secondary particle signal image input means so that the direction of astigmatism is horizontal or vertical in the image; and sharpness of the image from the rotated secondary particle signal image And a sharpness peak focal position calculating means for calculating a focus position at which the sharpness becomes a peak from among the sharpnesses determined for the preset focus positions. thing. (3) astigmatism correction means for rotating the focused secondary particle signal image input by the in-focus position detection means by a preset angle, and rotated secondary particles A sharpness calculating means for calculating the sharpness of the image from the signal image; and a sharpness value from the sharpness values respectively determined for the preset stigmata value in the x direction and the stigmata value in the y direction. A sharpness peak stigmata value calculating means for calculating a stigmata value in the x direction and a stigmata value in the y direction to become a peak is provided. (4) The astigmatism direction calculation means obtains an image obtained by performing a Fourier transform of the secondary particle signal image in a two-dimensional space to obtain a binarized power spectrum, and a main axis of the binarized image and a direction orthogonal to the main axis. Is determined, and the strength and direction of astigmatism are determined by determining the distance of each point in the binarized image to the main axis and the distance to an axis perpendicular to the main axis. (5) The sharpness calculating means divides each of the plurality of search areas set around each point in the secondary particle signal image into two areas, and obtains a variance between these two areas in each search area, The maximum value or the average value of the inter-region variance values determined in each of these search regions is used as the variance value of each point in the image, and the sum or average value of the variance values obtained at each point in the region set in the image is calculated as
Calculate as the sharpness of the secondary particle signal image. (6) The sharpness calculation means sets two vertical and horizontal search regions in the secondary particle signal image with each point as the center of gravity in the image, and sets these search regions around each point. When the area is divided into two areas A and B, and m _A and m _B are the average luminance of the pixels included in the areas A and B, and m _T is the average luminance of the entire search area including the areas A and B, _{σ = (m A -m T)} × (m A -m T) + (m B -m T)
The variance values between the regions A and B in the search region in the horizontal direction and the search region in the vertical direction are obtained by the formula of × (m _B −m _T ), and the sum or average value of these variance values is calculated as Calculate as the sharpness of the signal image. (7) The sharpness peak focus position calculating means selects a focus position indicating the maximum value from the sharpness calculated at a plurality of preset focus positions. (8) The sharpness peak focal position calculating means creates a sharpness curve indicating the relationship between the focus position and the sharpness with respect to the sharpness obtained at a plurality of preset focus positions, and this sharpness Applying a curve such as a Gaussian distribution curve or a quadratic curve to the curve, and calculating a focal position at a position where this curve becomes a peak. (9) The sharpness peak stigmata value calculating means selects an x-direction stigmata value indicating a maximum value from among sharpness values calculated in a plurality of preset x-direction stigmata values, Further, while holding the x-direction stigmata value at the peak position, the y-direction stigmata value indicating the maximum value is selected from the sharpness calculated for a plurality of preset y-direction stigmata values. thing. (10) The sharpness peak stigmata value calculating means indicates a relationship between the stigmata value in the x direction and the sharpness with respect to the sharpness obtained in a plurality of preset stigmata values in the x direction. A curve such as a Gaussian distribution curve or a quadratic curve is applied to the sharpness curve, a focal position at a position where the curve becomes a peak is calculated, and the x-direction stigmata value is held at the peak position. Meanwhile, a sharpness curve indicating the relationship between the stigmata value in the y direction and the sharpness is created for the sharpness determined in a plurality of previously set stigmata values in the y direction. Applying a curve such as a Gaussian distribution curve or a quadratic curve, and calculating a focal position at a position where this curve becomes a peak.

(Operation) According to the present invention, a secondary particle signal image is created by extracting a secondary particle signal obtained by two-dimensionally scanning a charged particle beam on a sample.
Focus the lens system of the charged particle optical column on the secondary particle signal image, and then adjust the stigmata on the secondary particle signal image obtained at the in-focus position so that the astigmatism of the charged beam is reduced. This makes it possible to correct astigmatism. In this case, since the adjustment of the stigmata is not performed by relying on the visual sense of the operator, there is no inconvenience such as a difference in the adjustment or a long time required for the adjustment due to the skill of the operator. Therefore, correction of astigmatism can be easily performed with high accuracy, and it is possible to prevent a decrease in observation accuracy due to the cross-sectional shape of the beam due to astigmatism.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The details of the present invention will be described below with reference to the illustrated embodiments. FIG. 1 shows a processing flow of a method for correcting astigmatism in the charged particle optical column according to the embodiment. Here, the charged particle optical column forms an optical system of a scanning electron microscope (SEM), and its basic configuration is the same as that of FIG. First, a secondary particle signal image (SEM image) is captured by the SEM, and the in-focus position is detected and focused (first step: S1). Next, the stigmata is adjusted while focusing on the in-focus position to correct astigmatism (second step: S
2).

FIG. 2 shows a configuration diagram of a device for correcting astigmatism in a charged particle optical column according to the present embodiment. This device receives a detection signal of the detector 11 of the charged particle optical column in the SEM as shown in FIG. 14 and extracts a secondary particle signal by two-dimensional scanning of an electron beam. 21, a focus position detection unit 22 that obtains a focus position using the input secondary particle signal image, a focus adjustment unit 23 that controls an objective lens 24 to change the focus position, and astigmatism after focusing The astigmatism correction unit 25 for correcting aberrations, the x-direction stigmata adjustment unit 26 for adjusting the drive of the x-direction stigmata 27 under the control of the astigmatism correction unit 25, and y under the control of the astigmatism correction unit 25 It comprises a y-direction stigmata adjustment unit 28 for adjusting the drive of the direction stigmata 29.

Here, the objective lens 24 corresponds to the objective lens 6 in FIG. 14 described above, and is moved from the x-direction stigmata 27 and the y-direction stigmata 29 to the stigmata 15 in FIG.
Is configured.

A sample having a circular pattern as shown in FIG. 3 is prepared as a standard sample for adjusting astigmatism. Such a circular pattern includes a semiconductor contact hole, gold particles, and the like. Therefore, a sample having these is set in an SEM at the time of correcting astigmatism, and correction processing is performed.

FIG. 4 shows a processing flow of the technique (S1) of detecting and focusing on the in-focus position using the SEM image, which will be described below. First, the x-direction stigmata 27 and the y-direction stigmata 29 are largely displaced from their normal positions, and large astigmatism is intentionally generated (S3).
A large astigmatism can be easily generated by adjusting the x-direction stigmata 27 and the y-direction stigmata 29 to numerical values set in advance.

Next, a secondary particle signal image I whose focus position is slightly shifted from the focus position is input (S4). Then, the direction of astigmatism is obtained based on the input secondary particle signal image I (S5).

Here, a processing flow for obtaining the direction of astigmatism in S5 is shown in FIG. 5 and will be described. First, the secondary particle signal image I whose focus position is slightly shifted from the in-focus position is Fourier-transformed to obtain a power spectrum image If
Is created (S27). This image If is created such that a low-frequency component comes at the center of the image. Then, as shown in FIG. 6, a rectangular band B is set at a position at a fixed distance d from the center of the image If, and the luminance average a and the standard deviation σ inside the band B are set.
Ask for. The shape of the band B may be a ring instead of a rectangle. The value of the distance d and the width of the band B are experimentally set. Since the sample used for the adjustment is always the same, these values can be set once and used fixed thereafter.

Next, the power spectrum image If is converted to a +
Binarization is performed using 2σ (S28). Since the low frequency component near the center of the image If of the contact hole is extremely large, only the central portion of the image If can be extracted by this binarization. The image Ib obtained by binarization is subjected to labeling and histogram processing to remove a small area portion as noise. When astigmatism exists, the SEM image I is blurred in a certain direction, so that low-frequency components increase in the blurring direction. For this reason, as shown in FIG. 7, the region R (hatched part in the figure) in the image If has an elliptical elongated shape with the direction of the blur being the short axis.

However, actually, since the region R is not a true ellipse but has irregularities in the outline, it is not easy to detect the ellipse from the image Ib and determine the major axis length and the minor axis length. Therefore, the main axis a1 of the region R and the axis a2 orthogonal to the main axis are
It is obtained by calculating the following second moment (S
29).

Tan ² θ + [{M (2,0) -M (0,2)} / M (1,1)]
tanθ-1 = 0 M (p , q) = Σ (i, j) (i-xc) p (i-yc) q fij Here, (i, j) is a coordinate on the image Ib, ( x
c, yc) are the barycentric coordinates of the region R in Ib. Also,
fij is fij when the value of point (i, j) is 0 in Ib.
= 0, and fij = 1 when it is not 0.

The two solutions of θ1 and θ2 obtained by solving the above equation indicate the main axis direction and the axial direction orthogonal to the main axis.
In order to determine which of the axis a1 in the θ1 direction and the axis a2 in the θ2 direction is the main axis, the distance between each point included in the region R and a1 and a2 is obtained, and the average D1,
Find D2. If D1> D2, a1 is set as the main axis. Because the direction of blur is perpendicular to the main axis,
The direction perpendicular to the main axis is defined as the direction of astigmatism (S30).

In the above procedure for obtaining the direction of astigmatism, instead of calculating the second moment, the image Ib
The distance from the center of the image to the elliptical contour may be determined around the center of the image, and the vertical direction indicating the maximum distance may be set as the direction of astigmatism. Further, an average direction between the vertical direction indicating the maximum distance and the direction indicating the shortest distance may be set as the direction of astigmatism. Further, the direction indicating the shortest distance may be the direction of astigmatism.

After the step of determining the direction of astigmatism in S5 is completed, a secondary particle signal image is input while changing the focal position, and the focal point is determined. The range in which the focal point is changed is set in advance, and the focal point is shifted in predetermined steps within this range (S6). Then, a secondary particle signal image is input at the set focal position (S7). An affine transformation with the image center as the rotation center is applied to the input secondary particle signal image to obtain an image I _R obtained by rotating the image by the astigmatism direction θs (S8). By applying the conversion formula x ′ = x cos (θs) −y sin (θs) y ′ = x sin (θs) + y cos (θs) to the point (x, y) in the input image, Position (x ',
y ') is obtained. The sharpness is calculated for the rotated image I _R (S9).

If astigmatism is present, the above-mentioned FIG.
As shown in FIG. 18 (c) or FIGS. 18 (a) and 18 (c), since the beam shape changes by 90 degrees from the focal point position, the direction of blur (the direction of astigmatism) also varies from the focal point position. Changes 90 degrees. That is, in the rotated image I _R , the direction of the blur is the x direction from the y direction or the x direction from the y direction at the in-focus position.
Since it changes in the direction, the sharpness of the image is calculated from the degree of blur in the x direction and the y direction.

Then, the focus position setting step (S
Steps from 6) to the sharpness calculation step (S9) are repeated within a preset focus position range (S10). An in-focus image has more high-frequency components in the image than an out-of-focus image. For this reason, the focus position where the high-frequency component becomes maximum is usually calculated by obtaining the power spectrum of the frequency by Fourier transform or applying a high-frequency emphasizing filter such as a differential filter. However, since the SEM image contains many high-frequency components due to noise, it is difficult to examine the change in the high-frequency component on the sample in detail by using a Fourier transform or an ordinary differential filter.

Therefore, without being affected by random noise,
By applying a filter capable of enhancing the edge portion on the sample to the SEM image and using the sum of the brightness of the obtained edge-enhanced image as a measure of sharpness, the focus position can be stably set even for a noisy image. We have developed a method to find it. Figure 9 for details
This will be described with reference to FIG.

A search area (height H and width W, respectively) is set in the vertical and horizontal directions with each point P in the image as the center of gravity.
Then, the search area is divided into two areas A and B centered on the point P. Variance between the regions A and _{B σ = (m A -n T} ) × (m A -n T) + (m B -n T)
× (m _B −n _T ) is obtained in each of the horizontal search area and the vertical search area.
Referred to as _A and sigma _B. Here, σ _A and σ _B are in the region A
And B, the average luminance of the pixels included in B, and m _T is the average luminance of the entire search area including areas A and B. The inter-region variance shows a large value when a step-shaped edge exists between the region A and the region B, but a small value when the height or width of the search region is large to some extent, in a region containing only random noise. Become. Therefore, even in a noisy image, it is possible to emphasize only a true edge portion existing on the sample. At each point P in the image, the variance value σh of the horizontal search area and the variance value σv of the vertical search area are obtained,
If σh <σv, then σv, if σh> σv, σh
And a variance value σ.

The variance value σh of the horizontal search area has a large value when a vertical edge exists near the point P. Conversely, the variance value σv of the vertical search area has a large value when a horizontal edge exists near the point P. If the focus position is off,
Rotated image I _R is the horizontal direction (x-direction) or vertical direction (y-direction)
, One of the vertical edge and the horizontal edge is blurred, and at each point P, σh or σv
Either one of them shows a lower value than the in-focus position. Therefore, the sum S of the variance value σ for the entire image has the highest value for the image input at the in-focus position. For this reason, the sum S of the variance values at each point in the image is used as the sharpness that measures the degree of blur of the image.

The focal position is sequentially shifted at regular intervals from a preset start point to an end point, and a secondary particle signal image is input. Then, the sharpness Si at each focal position Fi is obtained. FIG.
, The horizontal axis represents the focal position Fi, and the vertical axis represents the sharpness S.
The sharpness curve in which i is plotted has an upwardly convex shape at which the focus point becomes a peak. Therefore, a sharpness curve is created, and the focus position Fmax at which the sharpness becomes a peak is obtained to determine the focus position. It can be detected (S11). Then, the objective lens 24 is adjusted to adjust the focal position to Fmax (S12).

As a method of calculating the peak position, in addition to a method of simply finding the focal position at which the sharpness becomes maximum, a curve such as a Gaussian distribution curve or a quadratic curve is applied to the sharpness curve, and the peak on this curve is obtained. A method for obtaining the position or the like can be applied. Further, the points used for obtaining the sharpness are not limited to all points in the image, but may be in a partial area. Further, a value obtained by normalizing the variance value σ of each point by the maximum value σmax of the variance value of each point, for example, the sum of values converted into an 8-bit image by an equation of (σ / σmax) × 255 is obtained, and It is also possible. It is also possible to use the variance between the areas on one line in which the height of the horizontal search area is set to H = 1 and the width of the vertical search area is set to W = 1.

In the sharpness calculation step (S9),
Although the sum S of the variance values at each point in the image is used as the sharpness for measuring the degree of blur of the image, the average value of the variance values at each point in the image may be used as the sharpness. Although the maximum value of the inter-region variance determined in each search region is set as the variance of each point in the image, the average value of the inter-region variance may be set as the variance of each point in the image.

FIG. 11 is a block diagram of the in-focus position detecting section 22 of the astigmatism correction device according to the present embodiment. A detection signal of the detector 11 of the charged particle optical column in the SEM as shown in FIG. 14 is input, and a secondary particle signal image input unit 21 for inputting a secondary particle signal image by two-dimensional scanning of an electron beam is obtained. An astigmatism direction calculation unit 30 for calculating the direction of astigmatism from the obtained image, and an astigmatism obtained by the astigmatism direction calculation unit 30 from the image input from the secondary particle signal image input unit 21. An image rotator 31 that rotates only in the direction of?
A sharpness calculation unit 32 that calculates the sharpness of the rotated image by the above-described method, and a focus position at which the sharpness reaches a peak is calculated from a sharpness curve created using the sharpness calculated at a plurality of focus positions. , Focus adjustment unit 23 and astigmatism correction unit 25
Peak focus position calculation unit 33 that outputs the focus position to
It is composed of Note that the sharpness calculation unit 32 outputs a preset focus position to the focus adjustment unit 23 to change the focus position.

FIG. 12 shows a processing flow of a method (S2) for correcting an astigmatism by inputting an in-focus image, which will be described below. First, the x-direction stigmata 27 is changed from a preset position for each preset step (S13), and a secondary particle signal image is input (S14). Next, the input secondary particle signal image is rotated by a preset angle θ1 (S
15). Usually, θ1 is 45 degrees. As the image rotation method, the same method as the in-focus position detection method is used. Then
The sharpness of the rotated image is obtained by the same method as the in-focus position detection method (S16). And. From the x-direction stigmata setting step (S13) to the sharpness calculation step (S1)
6) is repeated within a preset range (S1).
7).

Next, a stigmata value XSmax in the x direction at which the sharpness reaches a peak is obtained in the same manner as the in-focus position detecting method (S18). Then, the stigmator 27 in the x direction is adjusted to XSmax (S19). Next, the stigmata 29 in the y direction is changed from the preset position at every preset step (S2).
0) A secondary particle signal image is input (S21). Then, the input secondary particle signal image is rotated by the preset angle θ2 (S22). Usually, θ2 = 0 degrees. As the image rotation method, the same method as the in-focus position detection method is used. Next, the sharpness of the rotated image is obtained by the same method as the in-focus position detection method (S23). And. The steps from the y-direction stigmata setting step (S20) to the sharpness calculation step (S23) are repeated within a preset range (S24).

Next, a stigmata value YSmax in the y direction at which the sharpness reaches a peak is obtained in the same manner as the in-focus position detecting method (S25). Then, the stigmator 29 in the y direction is adjusted to YSmax (S26). With the above method, it is possible to correct the astigmatism of the SEM.

FIG. 13 shows the configuration of the astigmatism correction unit 25 of the astigmatism correction device. An image rotator 3 that rotates the focused secondary particle signal image input from the in-focus position detector 22 by a preset rotation angle.
4 and calculate the sharpness of the rotated image,
A sharpness calculation unit 35 that outputs a set value of the stigmata to the x-direction stigmata adjustment unit 26 and the y-direction stigmata adjustment unit 28, and a sharpness calculated by a plurality of x-direction or y-direction stigmata set values. A sharpness peak stigmata value calculation unit that calculates an x-direction or y-direction stigmata value at which sharpness peaks from the created sharpness curve and outputs it to the x-direction stigmata adjustment unit 26 or the y-direction stigmata adjustment unit 28 36.

As described above, according to the present embodiment, a secondary particle signal image is created by extracting a secondary particle signal obtained when the electron beam is two-dimensionally scanned on the sample, and this secondary particle signal image is created. , The focus of the objective lens 24 is adjusted. At this time, the stigmata 27 in the x direction and the y direction
29, an astigmatism is intentionally generated, the direction of astigmatism is obtained from the secondary particle signal image input in a state where the astigmatism is generated, and the secondary particle signal image input at a preset focal position is converted to the astigmatism. The image is rotated so that the direction of the aberration is horizontal or vertical in the image, the sharpness of the image is calculated from the rotated secondary particle signal image, and the focal position where the sharpness reaches a peak is calculated and adjusted to this focal position.

Next, the astigmatism is corrected by adjusting the stigmata of the secondary particle signal image obtained at the in-focus position so that the astigmatism of the electron beam is reduced. At this time, the x-direction stigmata value was set, the sharpness of the image obtained by rotating the input secondary particle signal image was calculated, and the x-direction stigmata value at which the sharpness reached a peak was calculated and calculated. The stigmata 27 in the x direction is matched with the stigmata value, the stigmata value in the y direction is set while the stigmata value in the x direction is maintained, and the sharpness of an image obtained by rotating the input secondary particle signal image is calculated. Then, a y-direction stigmata value at which the sharpness reaches a peak is calculated, and the y-direction stigmata 29 is matched with the y-direction stigmata value at which the sharpness reaches a peak.

Thus, S for observing the sample surface is
In EM, stigmata can be automatically adjusted without relying on the operator's vision, astigmatism can be corrected with high accuracy and easily, and the cross-sectional shape of the electron beam caused by astigmatism It is possible to prevent a decrease in observation accuracy due to the above.

The present invention is not limited to the above embodiment. In the embodiment, in the second step for correcting astigmatism, the stigmata value in the y direction is adjusted after the stigmata value in the x direction is adjusted, but the stigma value in the x direction is adjusted after the stigmata value in the y direction is adjusted. The data values may be matched. Furthermore, the stigmata values in the x direction and the y direction may be changed simultaneously to find the position where the sharpness reaches a peak. In the embodiment, a sample having a circular pattern is used as a standard sample. However, the sample is not limited to this, and a main axis exists when Fourier transform is performed to extract a low-frequency component by binarization. Any sample can be used as a standard sample.

The present invention is not limited to an observation apparatus such as an SEM, but can be applied to a charged particle writing apparatus for writing an LSI pattern using an electron beam, an ion beam, or the like. In addition, various modifications can be made without departing from the scope of the present invention.

[0048]

As described above in detail, according to the present invention, a charged particle beam obtained by two-dimensionally scanning a sample on a sample is obtained.
A secondary particle signal image is created by extracting a secondary particle signal, the lens system of the charged particle optical column is focused on this secondary particle signal image, and then the secondary particle signal obtained at the in-focus position The astigmatism can be corrected for the image by adjusting the stigmata so that the astigmatism of the charged beam is reduced. Therefore, correction of astigmatism can be easily performed with high accuracy, and it is possible to prevent a decrease in observation accuracy due to the cross-sectional shape of the beam due to astigmatism.

[Brief description of the drawings]

FIG. 1 is a view showing a processing flow of an astigmatism correction method according to an embodiment of the present invention.

FIG. 2 is a diagram showing a schematic configuration of an astigmatism correction device according to an embodiment of the present invention.

FIG. 3 is a diagram showing an example of a standard sample used in the embodiment.

FIG. 4 is a diagram showing a processing flow of a focusing method.

FIG. 5 is a diagram showing a processing flow of a method for obtaining a direction of astigmatism.

FIG. 6 is a view for explaining a band-like area for setting a binarization threshold;

FIG. 7 is a view for explaining a main axis and an axis in a direction orthogonal to the main axis.

FIG. 8 is a diagram showing a beam cross-sectional shape when astigmatism does not exist.

FIG. 9 is a diagram for explaining a search area for obtaining inter-area variance.

FIG. 10 is a diagram for explaining a sharpness curve.

FIG. 11 is a diagram showing a schematic configuration of a focus position detection unit.

FIG. 12 is a diagram showing a processing flow of a technique for correcting astigmatism.

FIG. 13 is a diagram illustrating a schematic configuration of an astigmatism correction unit.

FIG. 14 is a diagram showing a basic configuration of an SEM according to the related art.

FIG. 15 is a diagram showing a change in a cross-sectional shape of a beam according to a focal length.

FIG. 16 is a diagram showing distributions of a sample surface shape and a beam cross-sectional shape in the x and y directions.

FIG. 17 is a diagram showing a change in a beam cross-sectional shape when there is no astigmatism.

FIG. 18 is a diagram illustrating a change in the beam cross-sectional shape when there is astigmatism that deforms the beam cross-section in a direction inclined at 45 degrees to the x and y axes.

[Explanation of symbols]

 S1 to S30 Various steps 1 Cathode 2 Anode 3 Electron gun 4 Electron beam 5 Condenser lens 6, 24 Objective lens 7 Aperture mask 8 Sample 9 Sample table 10 Deflector 11 Detector 12 ... scanning control circuit 13 ... image display device 15 ... stigmata 21 ... secondary particle signal image input unit 22 ... focus point position detection unit 23 ... focus adjustment unit 25 ... astigmatism correction unit 26 ... x-direction stigmata adjustment unit 27 ... x-direction stigmata 28 ... y-direction stigmata adjustment unit 29 ... y-direction stigmata 30 ... astigmatism direction calculation unit 31 ... image rotation unit 32 ... clarity calculation unit 33 ... clarity peak focal position calculation unit 34 ... image Rotating unit 35: Clarity calculation unit 36: Clarity peak stigmata value calculation unit

Claims (13)

[Claims] 1. A method for extracting a secondary particle signal obtained by two-dimensionally scanning a charged particle beam on a sample using a charged particle optical column having a stigmator for correcting astigmatism. A first step of focusing the lens system of the charged particle optical column on the secondary particle signal image created by A second step of correcting the astigmatism by adjusting the stigmata so as to reduce the astigmatism of the charged beam based on the signal image information. Aberration correction method. 2. The method according to claim 1, wherein the first step includes a third step of generating astigmatism by the stigmata, and a fourth step of inputting a secondary particle signal image in a state where the astigmatism is generated. A fifth step of obtaining the direction of astigmatism from the secondary particle signal image, a sixth step of setting the focal position to a preset position, and inputting the secondary particle signal image at the set focal position A seventh step,
Eighth step of rotating the input secondary particle signal image so that the direction of astigmatism is horizontal or vertical in the image, and Ninth step of calculating the sharpness of the image from the rotated secondary particle signal image Step, an eleventh step of calculating a focus position at which the sharpness peaks from among the sharpness values determined for the preset focus positions, and a step of calculating the focus position at which the calculated sharpness peaks. 2. A method for correcting astigmatism in a charged particle optical lens barrel according to claim 1, comprising a twelfth step of adjusting. 3. The method according to claim 1, wherein the second step is a thirteenth step of setting a stigmata value in the x direction, a fourteenth step of inputting a secondary particle signal image, and a step of converting the input secondary particle signal image. First to rotate by a preset angle
Step 5, a sixteenth step of calculating the sharpness of the image from the rotated secondary particle signal image, and sharpness from among the sharpnesses respectively determined for the preset x-direction stigmata value. An eighteenth step of calculating an x-direction stigmata value at which the degree peaks, a nineteenth step of adjusting the x-direction stigmata value to the calculated sharpness x-direction stigmata value, and an x-direction A twentieth step of setting a stigmata value in the y direction while holding the stigmata value of the first step, a twenty-first step of inputting a secondary particle signal image, and setting of the input secondary particle signal image in advance. A twenty-second step of rotating the image by a predetermined angle, a twenty-third step of calculating the sharpness of the image from the rotated secondary particle signal image, and a previously set y-direction stigmata value. A twenty-fifth step of calculating a y-direction stigmata value at which sharpness peaks from the respective sharpness values obtained; and a y-direction stigma value at the calculated y-direction stigmata value at which the sharpness peaks. 26. The method for correcting astigmatism in a charged particle optical column according to claim 1, comprising a twenty-sixth step of adjusting the astigmatism. 4. The fifth step is a twenty-seventh step of performing a Fourier transform on the secondary space of the secondary particle signal image to calculate a power spectrum, and the twenty-eighth step of obtaining a binary image of the power spectrum. Step 29, a main axis of the binarized image and an 29th step of obtaining an axis in a direction orthogonal to the main axis, a distance between each point in the binarized image to the main axis and an axis in a direction orthogonal to the main axis 3. The astigmatism correction method for a charged particle optical column according to claim 2, comprising a thirtieth step of determining a strength and a direction of astigmatism by calculating a distance to the astigmatism. 5. The ninth step according to claim 2 or the sixteenth and twenty-third steps according to claim 3, wherein a plurality of search regions set around each point in the secondary particle signal image are respectively divided into two regions. And the variance between these two areas is determined in each search area, and the maximum or average value of the variance between the areas determined in each search area is set as the variance of each point in the image, and set in the image. And calculating the sum or average of the variance values obtained at each point in the region as a sharpness of the secondary particle signal image. 6. The ninth step according to claim 2, or the sixteenth and twenty-third steps according to claim 3, wherein two search areas in the vertical direction and the horizontal direction with each point being the center of gravity in the secondary particle signal image. Is set in the image, and these search areas are divided into two areas A and B centered on each point, respectively, and m _A
And m _B are the average luminance of the pixels included in the regions A and B, m
_T region A, when the average brightness of the entire search area combined _{B, σ = (m A -m} T) × (m A -m T) + (m B -m T)
The variance values between the regions A and B in the search region in the horizontal direction and the search region in the vertical direction are obtained by the formula of × (m _B −m _T ), and the sum or average value of these variance values is calculated as A method for correcting astigmatism in a charged particle optical column, wherein the method calculates the sharpness of a signal image. 7. The charging method according to claim 2, wherein in the eleventh step, a focus position showing a maximum value is selected from sharpness calculated at a plurality of preset focus positions. A method for correcting astigmatism in a particle optical column. 8. The eleventh step is to create a sharpness curve showing the relationship between the focus position and the sharpness with respect to the sharpness obtained at a plurality of preset focus positions, 3. The method for correcting astigmatism in a charged particle optical column according to claim 2, wherein a curve such as a Gaussian distribution curve or a quadratic curve is applied to the curve, and a focal position at a position where the curve has a peak is calculated. 9. The eighteenth step is to select a maximum x-direction stigmata value from among sharpness values calculated in a plurality of preset x-direction stigmata values. 4. The method for correcting astigmatism in a charged particle optical column according to claim 3, wherein 10. The sharpness indicating the relationship between the stigmata value in the x direction and the sharpness with respect to the sharpness obtained in a plurality of preset stigmata values in the x direction in the eighteenth step. 4. The charged particle optical mirror according to claim 3, wherein a curve is created, a curve such as a Gaussian distribution curve or a quadratic curve is applied to the sharpness curve, and a focal position at a position where the curve becomes a peak is calculated. A method for correcting astigmatism in a cylinder. 11. The twenty-fifth step is a step of selecting a y-direction stigmata value indicating a maximum value from among sharpness values calculated for a plurality of y-direction stigmata values set in advance. 4. The method for correcting astigmatism in a charged particle optical column according to claim 3, wherein 12. The sharpness indicating the relationship between the stigmata value in the y direction and the sharpness in the twenty-fifth step with respect to the sharpness determined in a plurality of preset stigmata values in the y direction. 4. The charged particle optical mirror according to claim 3, wherein a curve is created, a curve such as a Gaussian distribution curve or a quadratic curve is applied to the sharpness curve, and a focal position at a position where the curve becomes a peak is calculated. A method for correcting astigmatism in a cylinder. 13. A charged particle optical column having a stigmator for correcting astigmatism, and a secondary particle signal obtained by two-dimensionally scanning a charged particle beam on a sample is extracted. Secondary particle signal image input means for inputting a secondary particle signal image to be created, in-focus position detecting means for focusing the secondary particle signal image, and focusing by changing the objective lens of the charged particle optical column Focus adjusting means for adjusting the position; and the stigmata for reducing the astigmatism of the charged beam with respect to the secondary particle signal image input by focusing on the in-focus position obtained by the in-focus position detecting means. Astigmatism correction means for correcting astigmatism by adjusting the astigmatism of the charged particle optical column.