JP2011023248A - Strain evaluation method of image by charged particle beam device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of evaluating stain on an observed image which is difficult to check visually in observing by a charged particle beam device. <P>SOLUTION: Boundaries in the observed image and a reference line in a reference image are specified. The reference image is moved on an X-axis of the observed image and distances between the respective boundaries and the reference line on the X-axis with respect to a movement amount are obtained. Changes in the distances are approximated by a linear function and intervals of the respective boundaries are calculated. Respective strain amounts on the observed image are calculated as differences between the respective intervals and an average interval. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

The present invention relates to a distortion evaluation method for images acquired by a charged particle beam apparatus.

In recent years, semiconductor devices are highly miniaturized from the viewpoints of integration degree and power consumption. The machining process for that purpose is complicated, and evaluation by shape observation and length measurement is very important. A charged particle beam apparatus such as a scanning electron microscope (SEM) having a length measurement function is well known as a means for performing this evaluation.

In observation with a charged particle beam apparatus, distortion due to optical factors such as aberrations and disturbances such as mechanical and electrical noises becomes a problem. Such distortion cannot be ignored when high-precision length measurement is required. For example, in observation with a scanning electron microscope, there is a concern that the distortion in the peripheral portion of the image becomes larger than that in the central portion. This is because, as one factor, the electron beam irradiation position tends to move away from the position where the electron beam should be irradiated as the distance from the optical axis increases. That means it ’s bigger.

As a distortion correction method, a technique using a standard sample is disclosed in Patent Document 1. In this technique, a standard sample having a periodic structure with a known size is observed at a first magnification, and the distribution of the strain amount in the observed image is defined as an absolute strain. Next, a sample different from the standard sample is observed at the first magnification and the second magnification, and the observation image at the first magnification is expanded and contracted isotropically and compared with the observation image at the second magnification. From this comparison, the relative distortion at the second magnification is calculated, and the absolute distortion at the second magnification is obtained from the relative distortion at the second magnification and the absolute distortion at the first magnification. Finally, the observed image is corrected using these absolute distortions.

JP 2008-14850 A

As described above, in the shape observation, the distortion on the image does not matter so much. However, when high-accuracy measurement is required from low magnification to medium magnification (for example, up to about 20,000 times), if an existing standard sample with a pitch of several hundred nm is used as the calibration sample, The number of stripes that appear is huge. In particular, since charged particle beam apparatuses in recent years have been highly adjusted, the amount of distortion is several pixels or less on an image, and visual confirmation is practically difficult.

  Therefore, an object of the present invention is to provide a method for evaluating distortion on an observation image, which is difficult to visually confirm in observation with a charged particle beam apparatus.

One aspect of the present invention is an image distortion evaluation method using a charged particle beam apparatus, wherein (a) a boundary line is specified in an observation image of a standard sample, and (b) an observation image or an observation image is obtained at the same magnification. The reference image is extracted from the other images of the standard sample, (c) the reference line is specified from the boundary line in the reference image, the reference image is moved on the X axis of the observation image, and each boundary line with respect to the movement amount And (d) the distance between each boundary line is calculated from the change in the distance between the boundary line and the reference line, and (e) the difference between each distance and the average distance is the amount of distortion. Each step is calculated as follows. The average interval is an average value of at least two of the intervals.

The change in distance may be approximated by a linear function.

  In the image distortion evaluation method using the charged particle beam apparatus according to the present invention, a change in the distance between the boundary line and the reference line is approximated by a linear function, and the interval between the boundary lines is obtained with an accuracy of less than one pixel. Therefore, the distortion amount can be calculated with the same accuracy. Since the amount of distortion of several pixels, which is difficult to see visually, can be confirmed, length measurement can be performed with the same accuracy.

It is a schematic diagram which shows an example of the distortion of the observation image by a scanning electron microscope. It is an example of the observation image of a standard sample in one Embodiment of this invention. It is a flowchart which shows the distortion evaluation method of the image which concerns on one Embodiment of this invention. It is a schematic diagram which shows an example of a standard sample. It is a schematic diagram which shows the change of the distance of a boundary line and a reference line in one Embodiment of this invention. It is a figure showing the calculation process of the distortion amount in one Embodiment of this invention, (a) is a figure showing the boundary line of an observation image, (b) is the change of the distance calculated | required about each boundary line of (a), and its The figure which shows a linear function, (c) is a figure showing the distortion amount of the space | interval shown in (b). It is a schematic diagram which shows a mode that the reference | standard image which concerns on one Embodiment of this invention is moved on an observation image. It is a figure which shows an example of the observed image and the image extracted in one Embodiment of this invention. It is a block diagram of the scanning electron microscope which concerns on one Embodiment of this invention. It is an example of the observation image of a standard sample in one Embodiment of this invention. It is a schematic diagram which shows only the boundary line of the observation image shown in FIG. It is a figure which shows the calculation process of the rotation angle which concerns on one Embodiment of this invention, (a) is a schematic diagram which shows a mode that a reference line is moved to a Y-axis direction, (b) is the reference | standard calculated | required by the movement of a reference line It is a schematic diagram which shows a line, a reference boundary line, and distance. In one Embodiment of this invention, it is a schematic diagram which shows distribution of the distortion amount on each axis | shaft obtained before considering a rotation angle. In one Embodiment of this invention, it is a schematic diagram which shows distribution of the distortion amount on each axis | shaft obtained in consideration of the rotation angle. It is an example which applied the evaluation method of distortion concerning one embodiment of the present invention to an actual observation picture. It is an example of distribution of the distortion amount in the observation image of FIG.

An embodiment of the present invention will be described with reference to the drawings. Although the present embodiment will be described using observation images acquired by a scanning electron microscope, the present invention is not limited to other charged particle beam devices such as a transmission electron microscope (TEM), a scanning transmission electron microscope. (STEM: Scanning Transmission Electron Microscope), focused ion beam processing apparatus (FIB), etc. are also applicable. The observation image here is a so-called digital image composed of a plurality of pixels.

(principle)
FIG. 1 shows an example of distortion of an observation image by SEM. In this figure, the distortion on the observed image is increased by an arrow with increasing outward as the distance from the origin O increases. In the observation image 10 of FIG. 1, dotted lines 12 indicate the positions of the sample observed when the image is not distorted, and solid lines 13 indicate the positions of the sample observed when the dotted line 12 is distorted. Show.

The amount of distortion on the image (that is, the amount of movement of each point on the image caused by the presence or absence of distortion) can be decomposed into two components in the X-axis direction and the Y-axis direction. Therefore, a component in the X-axis direction can be calculated using a striped pattern parallel to the Y-axis as shown in FIG. 2 (hereinafter simply referred to as a pattern). If a striped pattern parallel to the X-axis is used, the Y-axis direction can be calculated. Can be calculated. Note that the X-axis is usually taken in the scanning line direction when the charged particle beam is two-dimensionally scanned.

Therefore, a method for evaluating distortion in the X-axis direction will be described below. As will be understood from the following description, the evaluation of distortion in the Y-axis direction may be performed by replacing the following X-axis and Y-axis with each other.

In the evaluation of the distortion amount according to the present invention, a sample in which a boundary line 23 between the bright part 21 and the dark part 22 as shown in the observation image 20 in FIG. 2 appears is used as a standard sample. An example of such a sample is a sample 13 in which convex portions 14 and concave portions 15 having known dimensions are periodically formed, such as a diffraction grating (see FIG. 4). Further, in the sample 13, it is desirable that the upper surface 14a and the side surface 14b of the convex portion 14 are formed at right angles or acute angles so that the boundary line 23 of the observation image 20 appears clearly. That is, a sample in which both side surfaces 14b and 14b of the convex portion 14 are overhang is most suitable. Such a sample can be manufactured by etching using a photolithography method or the like. As another standard sample, a sample (not shown) in which substances having different secondary electron emission rates are alternately formed on the surface can be considered. The above-described standard sample is merely an example, and it is only necessary to have a periodic structure in which a boundary line necessary for calculating the amount of distortion appears on the axis of the observation image for moving the reference image.

The procedure for carrying out the present invention will be described based on the flowchart shown in FIG.
First, an observation image 20 is acquired by a scanning electron microscope (see FIG. 9), and a boundary line 23 in the image is specified (step S1). The boundary line 23 is defined as a line formed by connecting pixels in a predetermined intensity range in the bright part 21 or the dark part 22 in series. This line is ideally a line in which pixels with uniform intensity as shown in FIG. 2 are arranged in a line, that is, a line formed by connecting pixels having the same intensity and a constant X coordinate. . However, in an actual image, even if the bright part 21 and the dark part 22 are arranged in parallel to the Y axis, the intensity for each pixel changes, or the X coordinate of the pixels on these boundaries also changes by about one pixel. There are many cases. The boundary line defined in the present embodiment not only means a geometrically ideal straight line, but also includes a line composed of pixels having variations in each coordinate.

Next, the reference image 24 is extracted from the observation image 20 and one boundary line in the reference image 24 is selected, and this is used as the reference line 25 (step S2). The image size of the reference image 24 selected at this time is smaller than that of the observation image 20. When the image size of the observation image 20 is, for example, about 1200 × 1400 pixels, the image size of the reference image 24 is set to, for example, about 256 × 256 pixels. It is desirable that the reference image 24 is extracted from a region with the least distortion in the observation image 20. In general, such a region is a region including a position corresponding to the optical axis P (see the drawing) of the charged particle beam apparatus. In the example shown in FIG. 2, the origin O is a position corresponding to the optical axis. The reference image 24 may be extracted from another observation image obtained by observing the sample 13 at the same magnification as when the observation image 20 was acquired.

Next, the reference image 24 is moved on the X axis of the observation image 20, and the distance between the boundary line 23 of the observation image 20 and the reference line 25 of the reference image 24 is calculated according to the amount of movement (step S3). Here, the distance between the boundary line 23 and the reference line 25 is an intersection point (hereinafter referred to as a reference point) 23a between the boundary line 23 and the X axis and an intersection point (hereinafter referred to as a reference point) 25a between the reference line 25 and the X axis. And distance. When the reference point 23a is in the positive direction of the X axis relative to the reference point 25a, the distance takes a positive value. When the reference point 23a is in the negative direction of the X axis relative to the reference point 25a, the distance is Define to take a negative value.

FIGS. 5A to 5G schematically show changes in the distance between the boundary line 23 and the reference line 25 when the reference image 24 is moved in the positive direction of the X axis. Each left figure shows the relative arrangement of the boundary line 23 and the reference line 25, and each right figure shows cumulatively the distance between the intersections 23a, 25a obtained from the arrangement of the left figure. As shown in these figures, the distance corresponding to the movement amount of the reference image 25 changes to the right (that is, the distance decreases as the movement amount increases).

In this way, a plurality of distances between the reference point 23a and the reference point 25a are obtained for one boundary line 23, and a plurality of similar distances are obtained for other boundary lines. As long as the processing speed is not hindered, it is better that the number of calculated distances is larger. Therefore, the reference image 24 is moved pixel by pixel on the X axis to determine the distance for each movement amount, or when the movement amount at one time is not one pixel, one boundary line is a distance calculation target. It is preferable to set a wide area. For example, in FIGS. 5A to 5G, the area set as the calculation target is set to the same size as the reference image.

Next, the change in distance obtained for each boundary line 23 is approximated by a linear function (step S4). FIG. 6A shows a part of the observation image 20. For convenience of explanation, the bright portion 21 and the dark portion 22 are not shown, and only the boundary line 23 is shown. FIG. 6B shows a change in distance obtained for each boundary line 23 in FIG. 6A and its linear function 27. FIG. 6B differs from FIG. 5 in that the general method of this approximation on the boundary line 23 is the least square method, but other approximation methods may be used.

Then, the movement amount (X coordinate) of the reference image 24 when these linear functions 33a to 33j become 0 is obtained (step S5), and the interval between adjacent boundary lines is calculated from this movement amount (step S6). For example, the interval G1 shown in FIG. 6B is calculated from the positions on the X axis of the linear function 27a and the linear function 27b. The same applies to the intervals G2 to G9, which are calculated from the positions on the X axis of adjacent linear functions. For convenience of the following description, it is assumed that the intervals G1 to G7 have the same value, the interval G8 is larger than the interval G7, and the interval G9 is larger than the interval G8.

As described above, the interval between the boundary lines 23 is not obtained from a single line profile (for example, a line profile on the X axis) relating to the intensity, but the boundary line on the X axis is determined from the change in the distance between the boundary line and the reference line. The positions of the boundary lines adjacent to each other are obtained from these positions. Since the position of the boundary line on the X axis is determined using a linear function, the interval between the boundary lines can be accurately calculated.

Next, an average value is obtained using the intervals G1 to G9 of each boundary line 23, and a difference between the intervals G1 to G9 and the average value is set as a distortion amount (step S7). The average value is calculated using at least two of the total intervals. For example, a plurality of intervals may be selected in the central portion of the image with the least distortion, that is, a region including a position corresponding to the optical axis, and an average value may be obtained from these. Further, in the case of an observation image acquired at a medium magnification or less, when it is considered that the distortion of the image is very small, an average value for all intervals may be obtained.

Each point of FIG.6 (c) has shown the distortion amount 28a-28i of the space | gap G1-G9 shown in FIG.6 (b), respectively. In this example, it is assumed that the average value is equal to the interval G1. Therefore, the distortion amounts 28a to 28g from the interval G1 to the interval G7 are zero. Further, the distortion amount 28h at the interval G8 is larger than the distortion amounts 28a to 28g, and the distortion amount 28i at the interval G9 is larger than the distortion amount 28h, and the calculated distortion amounts 28a to 28i and the distortion amount occur. The position (coordinates) on the image is stored in the storage means or the like of the charged particle device (step S8), and is output to the display unit (see FIG. 9) of the charged particle device as a calculation result (step S9). As the position of the distortion amount, one of the positions on the X axis of the two linear functions used for calculating the distortion amount or a position between them is stored.

Thus, in the distortion evaluation method according to the present invention, as described above, the interval between each boundary line is accurately calculated, so that the distortion amount using this can also be accurately calculated. In other words, since the change in distance is approximated by a linear function and the position of the boundary line is determined from the linear function, the position of the boundary line can be calculated with an accuracy of less than one pixel. Therefore, the amount of distortion can also be calculated with an accuracy of less than one pixel.

Further, as shown in FIG. 7, a plurality of axes X1, X2, and X3 that are parallel to each other are defined on the observation image 20, and the reference image 24 is moved on these axes, and the amount of distortion on each axis. May be calculated. When a plurality of axes parallel to the X axis are defined, a two-dimensional distribution is obtained for the amount of distortion in the X axis direction.

On the other hand, when calculating the amount of distortion in the Y-axis direction, an observation image in which a striped pattern is parallel to the X-axis is acquired, and the calculation of the amount of distortion described above is performed by replacing the X-axis and the Y-axis. Good. That is, the reference image is moved in the Y-axis direction, a change in the distance between each boundary line on the observation image and the reference line on the reference image is obtained, the change is approximated by a linear function, and adjacent to the obtained linear function. The distance of the boundary line is obtained, and the distortion amount in the Y-axis direction is calculated using the distance. Then, the calculated strain amount and the position (coordinates) on the image where the strain amount is generated are stored in the storage means of the charged particle device. Regarding the calculation of the amount of distortion in the Y-axis direction, a plurality of axes parallel to the Y-axis may be defined, and the reference image may be moved on these axes to calculate the amount of distortion on each axis. When a plurality of axes parallel to the Y axis are defined, a two-dimensional distribution is obtained for the amount of distortion in the Y axis direction.

If the strain amount in the X-axis direction and the Y-axis direction is obtained, a two-dimensional distribution of the strain amount can be obtained from the position. Therefore, a predetermined threshold is set, an area on the image where the distortion amount is equal to or less than the threshold is specified, and only that area is observed. That is, an area that at least one side of a straight line defined by the position (coordinates) of the distortion amount in the X-axis direction or the Y-axis direction is observed.

For example, FIG. 8 shows an image 30, and a grid-like line 31 represents distortion on the image. In this case, only the region 32 having a distortion amount that is equal to or less than a predetermined threshold is observed. This observation can be performed by scanning the electron beam only in the region 32 or extracting only the specific region as an observation image without changing the scanning region of the electron beam. In this way, an image with less distortion can be obtained by eliminating a portion with a large amount of distortion, and the accuracy at the time of side length is increased. In other words, an image in which the upper limit of the measurement error is defined can be acquired.

In the above distortion amount calculation method, the reference image is moved together with the reference line. However, in principle, only the reference point is moved on the X axis or the Y axis, and the distance from the reference point on the boundary line is calculated. Also good.

(Application 1)
An example in which the distortion evaluation method according to this embodiment is applied to a scanning electron microscope will be described. FIG. 9 is a block diagram of a scanning electron microscope according to the present invention.

As shown in FIG. 9, a scanning electron microscope 40 includes an electron gun 42 that emits an electron beam 41 along an optical axis P, a focusing lens 43, a beam deflector (scanning coil) 44, an objective lens 45, a sample, and the like. The stage 46, the detector 47, the power supply part 48, and the control apparatus 50 are provided.

The power supply unit 48 applies a current or voltage to the lenses 43 and 45 and the beam deflector 44. The power supply unit 48 is controlled by the control device 50.

After being accelerated, the electron beam 41 emitted from the electron gun 42 is focused by the focusing lens 43 and the objective lens 45, and is irradiated on the sample 13 while being periodically deflected by the beam deflector 44. Secondary electrons and reflected electrons emitted from the sample 13 by this irradiation are detected by the detector 47.

The control device 50 includes a CPU (central processing unit) 51 that constitutes a computer, storage means 52 such as a ROM, a RAM, and a hard disk, an input unit 53 such as a mouse and a keyboard as an interface, an observation image and an operation image. And a control unit 55 for controlling the power supply unit 48, the detector 47, and the like.

The CPU 51 performs various processes based on an image signal from the detector 47, an input signal from the input unit 53, a program and data stored in the storage unit 52, and the like. For example, the CPU 51 outputs an observation image to the display unit 54 based on the image signal from the detector 47.

Specifically, each process in the control device will be described.

First, the observation image 20 of the sample 13 is stored in the storage unit 52. At this time, the observation image 20 may be output to the display unit 54. CPU51 reads the observation image 30 from the memory | storage means 52, and extracts a reference | standard image (refer FIG. 2) from it. Alternatively, an image of the sample 13 that is different from the observation image is read, and the reference image 24 is extracted therefrom. The reference image 24 is stored in the storage unit 52.

The CPU 51 specifies the boundary line 23 and the reference line 25 from the intensity distribution of the observation image 20 and the reference image 24. Furthermore, the X axis and the Y axis of the observation image 20 are defined, and the reference image 24 is arranged on the X axis. This is achieved by associating each coordinate in the reference image 24 with each coordinate in the observation image. Then, the reference image 24 is moved in the positive direction of the X axis, and the distance between the boundary line 23 and the reference line 25 with respect to the movement amount is calculated. Specifically, the difference between the coordinates of the intersection of the boundary line 23 and the X axis and the coordinates of the intersection of the reference line 25 and the X axis is calculated. That is, the difference between the Y coordinates of both intersections may be obtained. Each distance with respect to the movement amount is stored in the storage means 52 as a table (not shown).

The CPU 51 refers to the table to approximate the change in the distance between each boundary line 23 and the reference line 25 with a linear function (see FIG. 6B), and obtains the intersection of the linear function and the X axis. The coordinates of each intersection are stored in the storage means 52.

Next, the CPU 51 selects each intersection of the two adjacent boundary lines 23 and calculates the interval. When all the intervals are calculated, at least two of these intervals are selected, and the average value is stored in the storage means 52 as the average interval.

Further, the CPU 11 subtracts the average value from the interval between the boundary lines 23 and stores the value in the storage means 52 as the distortion amount. As the position on the image where the amount of distortion occurs, the coordinates of the intersection of adjacent boundary lines or the coordinates between them are stored together.

The above processing is also executed for the Y-axis direction distortion amount calculation. Thereby, a two-dimensional distribution of the distortion amount is obtained. When this distribution is output to the display unit 54, a method that is easy for the operator to recognize, such as a numerical table, a color tone, a contour line, and a vector diagram, is applied.

The distortion amount threshold value is stored in advance in the storage means or is set by an input from the input unit 53 by the operator. Based on this threshold, the CPU 11 selects a region that is equal to or less than the threshold in the stored strain amount distribution, and stores the coordinates of the region in the storage unit 52. When the operator wants to observe a region having a distortion amount that is less than or equal to the threshold, the CPU 11 instructs the control unit 55 to generate a control signal necessary for observation based on the stored coordinates of the region. The control unit 55 controls the beam deflection unit 44 based on this instruction. The obtained image is stored in the storage unit 52 and output to the display unit 54. Alternatively, the observation image is stored without changing the irradiation region, only the region within the threshold value is extracted, and the extracted image is stored in the storage unit 52 again. In any case, the distortion amount of the image used for length measurement is equal to or less than the set threshold value.

(Application example 2)
As an application example of the present embodiment, a method for calculating the distortion amount of the interval between the adjacent convex portions 14 of the sample 13 will be described. As described above, the striped pattern shown in FIG. 2 appears in the observation of the sample 13, but in actual observation, this pattern is rotated on the image, that is, tilted with respect to both the X and Y axes. There are many. This is because it is difficult to accurately position the sample 13 so that an image of the observation image 20 is obtained, and depending on the type of the charged particle beam apparatus, the image rotates depending on the observation magnification from the characteristics of the electron optical system. Because it ends up.

Therefore, when the pattern is rotated on the observation image, the interval between the boundary lines and the amount of distortion are obtained as follows. FIG. 10 is an example of an observation image of the sample 13 and shows a state where the pattern is rotated. FIG. 11 shows a boundary line 63 between the bright part 61 and the dark part 62 in FIG. In the image 60 of FIG. 11, the rotation angle of the pattern, that is, the angle formed by each boundary line 63 with the X axis is defined as θ. Note that the angle θ is unknown at this time, and is calculated by a process described later.

First, a reference image is extracted from the image 60 or from an image obtained by observing the sample 13 at the same magnification as the image 60. The extraction of the reference image 50 is the same as that of the reference image 24 described above.

Next, one of the boundary lines 63 in the image 60 is specified as the reference boundary line 66, and one of the boundary lines in the standard image 64 is specified as the reference line 65. Then, the reference image 64 is moved on the Y axis of the image 60. For example, as indicated by an arrow in FIG. 12A, the reference image 64 is moved in the positive direction of the Y axis. Then, for each movement amount of the standard image 64, the distance in the X-axis direction between the standard line 65 and the reference boundary line 66 in the standard image 64 is obtained. Specifically, this distance is a distance between an intersection (reference point) 65a between the reference line 65 and the Y axis, and an intersection (reference point) 66a between a straight line parallel to the X axis including the intersection and the boundary line (reference point) 66a. At this time, when the reference point 66a is in the positive direction of the X axis relative to the reference point 65a, the distance takes a positive value, and when the reference point 66a is in the negative direction of the X axis relative to the reference point 65a, The distance is defined to be negative.

When the reference image 64 is moved in the positive direction of the Y axis under this definition, when the angle θ is less than 90 degrees, the distance corresponding to the movement amount of the reference image 64 changes to the lower right (that is, The distance decreases as the amount of movement increases). As shown in FIG. 12B, this change in distance is approximated by a linear function 67 using a least square method or the like. In other words, the distances shown in FIGS. 12A and 12B are simply proportional to the amount of movement of the reference line 65, but in many cases, each intersection 66a is not positioned on a straight line. A straight line is approximately obtained as a linear function. The obtained linear function is stored in the storage means, and the rotation angle θ of the pattern is obtained from the inclination of the linear function 67.

Next, a plurality of axes parallel to the X axis are defined in the image 60, and the distance between the reference line 65 and each boundary line 63 on these axes is calculated for each movement amount of the reference image 64. The method for calculating the distance is the same as that described in the principle section. For example, as shown in FIG. 11, three axes of axes X1, X2, and X3 are defined. Further, the reference image 64 including the reference line 65 is moved on these axes X1, X2, and X3, and the distance between the reference line and each boundary line on each axis X1, X2, and X3 is set for each movement amount. calculate. Then, as shown in FIG. 6B, these distance changes are approximated by a linear function, and the intersections of the linear functions and the axes X1, X2, and X3 are obtained.

Next, the intervals between the intersections on the axes X1, X2, and X3 are calculated. In the subsequent processing, the distortion amount is calculated on each of the axes X1, X2, and X3 in the same manner as the method described in the principle section. However, at this stage, as shown in FIG. 13, it is not specified which boundary line the calculated distortion amounts correspond to.

Therefore, using the linear function 67 obtained with respect to the distance between the reference boundary line 66 and the reference line 65, the distortion amount and each boundary line 63 are associated with each other on the axes X1, X2, and X3. Since the position (Y coordinate) of each axis X1, X2, X3 on the Y axis is known, by substituting these positions into a linear function, the amount of movement (X coordinate) where the distortion amount on each axis is located ) Is obtained. Using this deviation amount, the movement amount (X coordinate) of each distortion amount is corrected. For example, if the values obtained when the Y coordinates of the axes X1, X2, and X3 are substituted into the linear function 66 are d1, d2, and d3, respectively, these values are the above-described deviation amounts (FIG. 12 (where d2 is 0). Next, when the shift amounts d1, d2, and d3 are subtracted from the movement amounts obtained for the respective intervals in the respective axes X1, X2, and X3, as shown in FIG. 14, they are obtained for the respective axes X1, X2, and X3. The amount of movement (X coordinate) of each distortion amount matches. That is, the boundary line when the respective distortion amounts on the respective axes X1, X2, and X3 are obtained is specified.

An example in which the distortion amount and the adjacent boundary line are specified by performing the above processing is shown in FIGS. These drawings show the results of obtaining the amount of distortion at the intervals between the boundary lines using the actually acquired observation image 68. In FIG. 15, the white frame at the center of the observation image 68 represents the reference image 69 to be extracted, and the other white frames move the reference image 69 on the respective axes X1, X2, and X3 in the directions of the arrows. It shows the state when making it. FIGS. 16A to 16C respectively show the amounts of distortion on the axes X1, X2, and X3 obtained from the observation image 68 and the positions of the distortion amounts corrected in consideration of the rotation angle in the X-axis direction. ing. As can be understood from these figures, it can be confirmed that the amount of distortion of the observed image 68 varies little by little within a range of one pixel. From this result, for example, the distortion amount of the observation image 68 tends to increase as it goes to the right, and in particular, the distortion amount is 1.2 pixels near the right end on the axis X3, indicating the maximum value. Can be determined.

As described above, according to the present invention, the amount of distortion that includes a large number of boundary lines and is difficult to confirm visually can be calculated with an accuracy of less than one pixel, so that observation with the same accuracy is possible. Therefore, length measurement with the same accuracy is possible.

13 Standard sample 20, 30, 60, 68 Observation image 23, 63 Boundary line 24, 64, 69 Reference image 25, 65 Reference line 27, 67 Linear function 40 Scanning electron microscope

Claims (9)

An image distortion evaluation method using a charged particle beam device,
Identify the boundary line in the observation image of the standard sample,
Extracting a reference image from the observation image or another image of the standard sample acquired at the same magnification as the observation image;
Identifying a reference line from a boundary line in the reference image;
Moving the reference image on the X-axis of the observed image, and determining a distance on the X-axis between each boundary line and the reference line with respect to the movement amount;
Approximating each change in the distance with a linear function, and calculating the interval between the boundary lines from the linear function,
The difference between each said interval and the average interval is calculated as the amount of distortion,
The method for evaluating distortion of an image by a charged particle beam apparatus, wherein the average interval is an average value of at least two of the intervals. The interval for obtaining the average interval is selected from an area in the observation image that includes a position corresponding to the optical axis of the charged particle beam device and is smaller than the size of the observation image. Item 6. An image distortion evaluation method using the charged particle beam device according to Item 1. 3. The image distortion evaluation method using a charged particle beam device according to claim 1, wherein the average interval is an average value of all the intervals. 4. The image distortion evaluation method using a charged particle beam apparatus according to claim 2, wherein the X axis is a plurality of axes parallel to each other.   5. The image distortion evaluation method using a charged particle beam apparatus according to claim 1, wherein the reference image is moved pixel by pixel on the X axis. 6. The image distortion evaluation method according to claim 5, wherein the standard sample has a periodic structure in at least one predetermined direction so that the boundary line is obtained. Identifying one of the boundary lines as a reference boundary line in the observed image;
A rotation angle of the reference boundary with respect to the X axis is obtained;
Associating each interval with each boundary based on the rotation angle;
An image distortion evaluation method using the charged particle beam apparatus according to claim 6. 8. A distortion amount equal to or less than a predetermined threshold value is selected from each of the distortion amounts, and an area having at least one side of a straight line defined by the position of the distortion amount is observed. The distortion evaluation method of the image by the charged particle beam apparatus in any one. 9. The image distortion evaluation method using a charged particle beam apparatus according to claim 8, wherein the evaluation image is an image acquired by observation at a magnification of 20,000 times or less.

Images