Charged particle beam device and three-dimensional shape estimation method

Abstract

A plurality of observation images are acquired while varying, for each, the focal position of a charged particle optical system and / or the retention height of a sample 130 by a stage 119, an evaluation value is calculated for each of a plurality of local region images that constitute the observation images, said evaluation value varying in accordance with the deviation between the focal position of the charged particle optical system and the sample in the optical axis direction of the charged particle optical system, a three-dimensional distribution of the evaluation values is calculated, said three-dimensional distribution being in a three-dimensional space that is defined by an axis corresponding to coordinates of the observation images and an axis corresponding to the focal position of the charged particle optical system or the retention height of the sample by the stage, and the three-dimensional shape of the sample is estimated from the three-dimensional distribution of the evaluation values.

Description

Charged particle beam apparatus and method for estimating three-dimensional shape

[0001] The present invention relates to a charged particle beam apparatus and a method for estimating three-dimensional shapes.

[0002] When observing samples using a transmission electron microscope (TEM) or scanning transmission electron microscope (STEM), it is necessary to appropriately adjust the orientation of the sample with respect to the direction of the charged particle beam. For example, the surface of the sample needs to be approximately perpendicular or parallel to the direction of the charged particle beam. The following explanation will use STEM as an example.

[0003] For example, as shown in Figure 8A, when the surface of the sample is parallel to a plane S perpendicular to the direction of propagation of the charged particle beam indicated by the arrow, a clear observation image like the one shown in Figure 8B is obtained. In contrast, as shown in Figure 8C, when the surface of the sample is inclined with respect to plane S, a blurry observation image like the one shown in Figure 8D is obtained. This difference arises because the interfaces of the structure within the sample (crystal structure in this example) are parallel to the direction of propagation of the charged particle beam. This is often true for observation targets such as semiconductor devices, where the orientation of the structure and interfaces within the observation target is appropriately controlled.

[0004] For samples containing single-crystal regions, the plane inclination of the sample can be determined using the diffraction pattern obtained by irradiating the single-crystal region with a charged particle beam. However, this method cannot be applied to samples that do not contain single-crystal regions.

[0005] A technique for determining the surface shape of a sample is known, as described in Patent Document 1 (see, for example, the abstract).

[0006] Japanese Patent Publication No. 2022-185757

[0007] The technology described in Patent Document 1 relates to scanning electron microscopy, and requires multiple detection units to grasp the shape of the sample surface. Furthermore, since the observation image of a typical STEM does not change according to the angular relationship of the detector to the sample, as is the information obtained using the multiple detection units described in Patent Document 1, it is difficult to directly apply the technology described in Patent Document 1.

[0008] This invention provides a technique for generating information to adjust the tilt of a sample with respect to the direction of irradiation by a charged particle beam.

[0009] A typical example of the invention disclosed in this application is as follows: A charged particle beam apparatus comprising a stage for holding a sample, a charged particle optical system for irradiating the sample with a charged particle beam, a detector for detecting signal electrons emitted by the interaction between the sample and the charged particle beam, and a control system for controlling the stage, the charged particle optical system, and the detector. The stage is configured to allow adjustment of the sample holding height at least in the direction of the optical axis of the charged particle optical system, and the charged particle optical system is configured to allow adjustment of the focal position in the direction of the optical axis of the charged particle optical system. The control system acquires multiple observation images based on the signal electrons detected by the detector, each while changing at least one of the focal position of the charged particle optical system and the sample holding height by the stage. For each of the multiple local region images constituting the observation image, it calculates an evaluation value that changes in accordance with the deviation of the charged particle optical system in the direction of the optical axis of the charged particle optical system from the sample in the direction of the optical axis of the charged particle optical system, obtains a three-dimensional distribution of the evaluation value in a three-dimensional space defined by an axis corresponding to the coordinates of the observation image and an axis corresponding to the focal position of the charged particle optical system or the sample holding height by the stage, and estimates the three-dimensional shape of the sample from the three-dimensional distribution of the evaluation value.

[0010] According to one aspect of the present invention, information regarding the three-dimensional shape of a sample can be generated. Other problems and novel features will become apparent from the description herein and the accompanying drawings.

[0011] This is a diagram showing an example of the configuration of a scanning transmission electron microscope. This is a flowchart illustrating an example of the correction process performed by the control system. This is a diagram showing how a through-focus image is acquired. This is a diagram illustrating an example of an observed image. This is a graph showing the evaluation value S for each focal position in a certain local region. This is a graph showing an example of the three-dimensional distribution of the evaluation value S. This is a diagram for explaining the discrete wavelet transform. This is a diagram illustrating the method for calculating the evaluation value S. This is a diagram illustrating the method for calculating the evaluation value S. This is a diagram illustrating the method for calculating the evaluation value S. This is a diagram illustrating the problem. This is a diagram illustrating the problem. This is a diagram illustrating the problem.

[0012] The embodiments of the present invention will be described below with reference to the drawings. However, the present invention is not to be construed as being limited to the embodiments described below. It will be readily apparent to those skilled in the art that the specific configuration can be modified without departing from the spirit or intent of the present invention.

[0013] In the configuration of the invention described below, identical or similar components or functions are denoted by the same reference numerals, and redundant descriptions are omitted.

[0014] The designations "First," "Second," "Third," etc., used in this specification are for the purpose of identifying constituent elements and do not necessarily limit their number or order.

[0015] The positions, sizes, shapes, and ranges of each component shown in the drawings, etc., may not represent the actual positions, sizes, shapes, and ranges, etc., in order to facilitate understanding of the invention. Therefore, the present invention is not limited to the positions, sizes, shapes, and ranges, etc., disclosed in the drawings, etc.

[0016] Figure 1 shows an example of the configuration of a STEM. The STEM 100 includes a stage for holding a sample, a charged particle optical system for irradiating the sample (thin film sample) held on the stage with a charged particle beam (electron beam in this example), an electron optical system tube 101 that incorporates a detector for detecting signal electrons emitted as a result of the interaction between the sample and the charged particle beam, and a control system 102.

[0017] The electron optical system barrel 101 includes an electron source 111, an electrode 112, a condenser lens 113, a condenser diaphragm 114, an axial adjustment deflector 115, an aberration correction device 116, a scan deflector 117, an objective lens 118, a stage 119, an objective diaphragm 120, an axial adjustment deflector 121, a field-limiting diaphragm 122, an imaging lens 123, and a detector 124.

[0018] The control system 102 includes a control device 141 and a computer 142. The control system 102 controls the electron optical system tube 101 using a plurality of control circuits. More specifically, the control system 102 creates arbitrary electron-optical conditions by acquiring values ​​from a target device via control circuits and inputting values ​​to the target device via control circuits. Electro-optical conditions include, for example, the focal position, the diameter and shape of the charged particle beam (probe), and aberrations. The focal position can be controlled by adjusting the current of the objective lens 118. Astigmatism can be controlled by adjusting the current of the aberration correction device 116.

[0019] The control system 102 generates an image from the signal detected by the detector 124. In this embodiment, the control system 102 analyzes the image to calculate the tilt of the sample with respect to a plane perpendicular to the direction of the charged particle beam. Furthermore, the control system 102 in this embodiment controls the electron optical system tube 101 to reduce the tilt of the sample with respect to the aforementioned plane.

[0020] Figure 2 is a flowchart illustrating an example of a correction process performed by the control system 102.

[0021] The control system 102 adjusts the adjustment parameters corresponding to the electron-optical conditions and acquires multiple observation images (through-focus images) with different focal positions of the charged particle optical system (step S101). The control system 102 stores the adjustment parameters in association with the acquired observation images.

[0022] Figure 3A shows how a through-focus image is acquired. Here, we consider the case where the sample is inclined with respect to a plane S perpendicular to the optical axis direction (Z direction) of the electron optical system. The charged particle beam 301 is scanned over a two-dimensional region (XY plane) parallel to plane S. Furthermore, let's assume that F(-2), F(-1), F(0), F(+1), and F(+2) are set as focal positions for acquiring the observed image. Although not particularly limited, F(0) is the position where the center of the field of view is in just focus, while F(-2) and F(-1) are in the overfocus position, and F(+1) and F(+2) are in the underfocus position. For example, using the height of focal position F(0) as a reference, the heights of focal positions F(-2) and F(-1) are -900 nm and -450 nm, respectively, and the heights of focal positions F(+2) and F(+1) are 900 nm and 450 nm, respectively.

[0023] Figure 3A shows the scanning of a charged particle beam 301 with its focal point at focal point F(-2). Because the sample 302 is inclined with respect to the surface S, a clear image is obtained in the region of the observed image where the focal point of the charged particle beam 301 is on the sample 302, but the image becomes blurry as the focal point of the charged particle beam 301 moves away from the sample 302. The observed images acquired at each focal point are shown as observed images 311 to 315 in Figure 3B.

[0024] To obtain a through-focus image, it is necessary to make the distance between the charged particle optical system's focal position and the sample along the optical axis of the charged particle optical system different. For this reason, the focal position in the optical axis direction of the charged particle optical system may be adjusted by adjusting the objective lens 118, or the sample holding height may be adjusted in the optical axis direction of the charged particle optical system using the stage 119. This embodiment describes the former as an example, but the latter may also be adopted, in which case the evaluation value is calculated in accordance with the sample holding height by the stage 119. Note that the focal position of the charged particle optical system may be adjusted by an electromagnetic lens provided in the charged particle optical system, in addition to or instead of the objective lens 118.

[0025] The control system 102 uses these observation images to generate sample position information regarding the position of the sample in the optical axis direction of the charged particle optical system (step S102). For example, sample position information is generated by the following procedure.

[0026] (S102-1) The control system 102 selects one observation image and divides it into multiple regions. Each of the divided regions is called a local region. For example, each observation image is divided into M × N (M = N) grid regions. Figure 3B shows local region images 321 to 325 of local region LA(m, n) (1 ≤ m ≤ M, 1 ≤ n ≤ N) located at the same position in each observation image. Here, when dividing the observation image into local regions, the division may be made such that the periphery of adjacent local region images overlaps with each other.

[0027] (S102-2) The control system 102 calculates an evaluation value for each local region included in the observation image to evaluate the degree of focus shift of the local region image. In this embodiment, an evaluation value S corresponding to sharpness is calculated. For example, the evaluation values ​​S for local region images 321 to 325 are calculated as S1 to S5, respectively (see Figure 3B). The evaluation value only needs to change in accordance with the shift in the optical axis direction of the charged particle optical system between the focal position of the charged particle optical system and the sample, and in addition to sharpness, statistical values ​​such as the dispersion of image intensity can also be used. An example of how to calculate the evaluation value will be described later.

[0028] (S102-3) The control system 102 determines whether the processing in (S102-1) and (S102-2) has been completed for all observed images. If the processing is not completed for all observed images, the control system 102 returns to the process in (S102-1).

[0029] (S102-4) When the processing for all the observation images is completed, the control system 102 obtains the three-dimensional distribution of the evaluation value S in the three-dimensional space defined by the axes (X, Y) corresponding to the in-field coordinates and the axis (Z) corresponding to the focal position. FIG. 3C is a graph showing the evaluation value S for each focal position with respect to the local region LA(m, n) shown in FIG. 3B, where the horizontal axis represents the focal position and the vertical axis represents the evaluation value S. FIG. 4 is a graph representing the evaluation value distribution for each local region in the three-dimensional space with the axis corresponding to the focal position as the Z-axis, and the axes corresponding to the coordinates of the local region LA as the X-axis and Y-axis. That is, FIG. 4 corresponds to the graph obtained by rotating the graph of FIG. 3C counterclockwise by 90° and arranging it according to the arrangement of the local regions LA.

[0030] Note that in the graph shown in FIG. 4, the magnitude of the evaluation value S is originally plotted in color. Although the color information is lost due to grayscale, the low evaluation values and high evaluation values are plotted in complementary colors, and the evaluation values in between are plotted by the gradation of these two colors. In the dark gray region 401 near the center of the columnar evaluation value distribution 400, the evaluation value S of the local region LA is high. On the other hand, in the dark gray regions 401 above and below the region 401 in the columnar evaluation value distribution 400, the evaluation value S of the local region LA is low.

[0031] Subsequently, the control system 102 estimates the three-dimensional shape of the sample based on the sample position information (step S103). Here, the three-dimensional shape of the sample refers to the region occupied by the surface of the sample, which is a thin film sample, the central plane in the thickness direction of the sample, or the plane corresponding to a specific depth position in the thickness direction of the sample determined according to the observation conditions, in the three-dimensional space. The sample located in this three-dimensional space is modeled and called the sample surface.

[0032] It can be estimated that the sample surface exists in the region 401 with a high evaluation value S in the evaluation value distribution 400 in the three-dimensional space obtained in step S102. Here, a method for the control system 102 to estimate the sample surface 402 from the evaluation value distribution 400 in the three-dimensional space will be described. When the sample surface 402 is estimated as a plane, (Equation 1) holds.

[0033] Here, S is the evaluation value S, d is the scale of the evaluation value S in the three-dimensional space, e is the offset of the evaluation value S in the three-dimensional space, x is the X coordinate in the observation image, y is the Y coordinate in the observation image, a is the X-direction inclination amount of the sample surface 402, b is the Y-direction inclination amount of the sample surface 402, c is the height-direction (Z-direction) offset of the sample surface 402, and σ is the variance of the evaluation value S in the height direction (Z-direction). By fitting equation (1) to the evaluation value distribution 400 in the three-dimensional space, the values of each coefficient are estimated. The inclination of the sample is calculated as the inclination amount between the plane (XY plane) perpendicular to the optical axis direction of the charged particle optical system and the sample. The coefficients a and b respectively correspond to the two inclination angles (X-direction inclination amount, Y-direction inclination amount) shown in FIG. 4.

[0034] The control system 102 calculates an optimal inclination control angle based on the three-dimensional shape of the sample (step S104). For example, when the three-dimensional shape of the sample is substantially flat, the control system 102 calculates the inclination control angle by converting the two inclination angles calculated in step S103 into the inclination control amount of the stage 119. This is because the X-axis and Y-axis in the observation image are determined by the deflection direction of the charged particle beam by the scan deflector 117, while the inclination adjustment axis of the stage 119 does not necessarily coincide with the deflection direction of the scan deflector 117. Also, when the three-dimensional shape of the sample is not substantially flat, the control system 102 may calculate the inclination angle between the plane perpendicular to the direction of the charged particle beam and the plane formed by the local region specified by the user, and calculate the inclination control angle by converting it into the inclination control amount of the stage 119.

[0035] The control system 102 controls the stage 119 based on the inclination control angle calculated in step S104 to hold the sample perpendicular to the optical axis direction of the charged particle optical system, and adjusts the focus position (step S105).

[0036] The control system 102 determines whether to end the correction (step S106). For example, when the number of corrections is greater than a predetermined value, the control system 102 ends the correction. Also, the control system 102 calculates the inclination amount of the sample with respect to the plane perpendicular to the direction of the charged particle beam using the observation image acquired after the correction, and ends the correction when the inclination amount is small.

[0037] Furthermore, the control system 102 may calculate evaluation values ​​for each region of an observed image after acquiring that image. In other words, it may calculate evaluation values ​​in parallel with the acquisition of the observed image. This enables efficient calculation.

[0038] Furthermore, although this example shows the control system 102 controlling the stage 119 based on the three-dimensional shape of the sample, in addition to or instead of controlling the tilt of the stage, the deflection of the charged particle beam may be controlled by the axial adjustment deflector 115.

[0039] Furthermore, in step S103, when the surface function used to estimate the three-dimensional shape of the sample is expressed using two components, X and Y, to represent the coordinates corresponding to the position within the field of view, it is also possible to use a form that includes not only a term composed of first-order coefficients of X and Y, but also a term composed of second-order or higher coefficients of X and Y. In this case, the first-order coefficients of X and Y correspond to the linear change of the sample surface with respect to each coordinate axis direction, i.e., the amount of inclination, while the second-order or higher coefficients correspond to the warp or deflection of the sample surface.

[0040] If the sample is warped or deflected, the inclination of the sample may have different values ​​in different regions within the observation field. In this case, the average inclination of the entire sample can be determined from the aforementioned first-order coefficient, and if it is desired to accurately determine the inclination of a local region on the sample, a second-order or higher coefficient can be used to obtain the local inclination. This allows for accurate adjustment of the inclination of the sample relative to the surface of the sample being observed, even for samples that are warped or deflected, by controlling the inclination of the stage 119 or the deflection of the axial adjustment deflector 115. Alternatively, the inclination of the sample can be adjusted from the aforementioned first-order coefficient, and then the axial adjustment deflector 115 can be controlled so that the direction of propagation of the charged particle beam is approximately perpendicular to the surface of the sample being observed, corresponding to the local inclination obtained from the second-order or higher coefficient.

[0041] In this case, methods for representing the warp or deflection of the sample include not only methods using higher-order polynomials, but also techniques such as interpolation to a portion of the distribution of focal points that come into focus on the sample in each measured region.

[0042] Furthermore, while the reference plane for calculating the tilt angle was described here as being perpendicular to the optical axis direction of the charged particle optical system, equivalent effects can be obtained by using a different plane as the reference plane. For example, the tilt angle calculated using a plane tilted at a known fixed angle relative to the plane perpendicular to the optical axis direction of the charged particle optical system as the reference plane can be used as the tilt angle relative to the plane perpendicular to the optical axis direction of the charged particle optical system by subtracting the known fixed angle from the calculated tilt angle. As the method of determining the angle reference is arbitrary, the reference plane for calculating the tilt angle is not limited to the plane perpendicular to the optical axis direction of the charged particle optical system, but can be set arbitrarily by the user. Examples of arbitrarily set reference planes include, in addition to the plane perpendicular to the optical axis direction of the charged particle optical system, the reference horizontal plane in the design of the stage, and the central axis in the design of the electron optical system tube.

[0043] The method for calculating the evaluation value S is described below. In this example, the evaluation value S is calculated to evaluate the sharpness of the image by applying a discrete wavelet transform to the observed image. The discrete wavelet transform transforms the image into wavelets (localized small waves / bases) that are superimposed with varying scales and positions. Therefore, it is possible to evaluate local frequency information while preserving the positional information of the image.

[0044] The discrete wavelet transform is conceptually explained using Figure 5. For simplicity, an example of performing a discrete wavelet transform on a one-dimensional signal (original signal) 501 is shown. The original signal 501 is decomposed into an approximate signal 502 and a detailed signal 503. Here, the approximate signal 502 is the average value of two consecutive scale divisions of the original signal 501, and the detailed signal 503 is the difference between the original signal 501 and the approximate signal 502. By expanding the approximate signal 502 with a scaling function 504 and the detailed signal 503 with a wavelet function 505, a scaling coefficient sequence 506 and a wavelet coefficient sequence 507 are obtained. Here, the scaling function 504 is a function that takes a value of 1 for two consecutive scale divisions, and the wavelet function 505 is a function that takes values ​​of -1 and 1 for two consecutive scale divisions. The wavelet coefficient sequence 507 represents local intensity changes. For example, analysis of the original signal 501 reveals that the steepest intensity change occurred in region 508 where the absolute value of the wavelet coefficient sequence 507 is at its maximum (this corresponds to the interval between scales 4 and 6 of the original signal 501).

[0045] The discrete wavelet transform of the original signal is called a first-order discrete wavelet transform. The scaling coefficient sequence 506 can be subjected to a further wavelet transform, which is called a second-order discrete wavelet transform, and similarly, higher-order discrete wavelet transforms can be performed. In low-order discrete wavelet transforms, information on local intensity changes can be obtained for the high-frequency components of the original signal, and as the order of the discrete wavelet transform increases, information on local intensity changes can be obtained for the lower-frequency components of the original signal.

[0046] In this embodiment, the same process is performed on a two-dimensional local region image. The method for calculating the evaluation value S by the control system 102 will be explained using Figures 6A, 6B, and 6C.

[0047] Figure 6A is a local region image, where the local region image 600 shows a schematic and simplified representation of the pattern. Figure 6B shows the first-order transverse wavelet transform image (1HL) 601 of the local region image 600, and Figure 6C shows the first-order longitudinal wavelet transform image (1LH) 602 of the local region image 600.

[0048] If the local region image 600 is an image composed of pixel values ​​I(x,y), then the transverse wavelet transform image 601 is a coefficient value w １ｈ This is an image composed of (x, y) and shows the high-frequency components (image changes) of the local region image 600 in the lateral direction. The longitudinal wavelet transform image 602 has coefficient values ​​w １ｖ This image is composed of (x, y) and shows the high-frequency components (image changes) of the local region image 600 in the vertical direction. In the transverse wavelet transform image (1HL) 601 and the longitudinal wavelet transform image (1LH) 602, the change from a light region to a dark region is shown by a bright line, and the greater the change, the brighter the line (closer to white). The change from a dark region to a light region is shown by a dark line, and the greater the change, the darker the line (closer to black).

[0049] As explained in the example of a one-dimensional signal, a discrete wavelet transform of a (j+1)th-order (j≧0) low-frequency component image ((j+1)LL, corresponding to the scaling coefficient sequence in a one-dimensional signal) can yield a (j+2)th-order transverse wavelet transform image ((j+2)HL) and a longitudinal wavelet transform image ((j+2)LH). Higher-order discrete wavelet transform images show information at lower spatial frequencies (resolution).

[0050] Within a sample, especially in regions containing structures such as edges, the intensity change of the image becomes steep when the image is in focus, and the absolute value of the discrete wavelet transform coefficients becomes large. Conversely, when the image is out of focus, information from a wider area than when the image is in focus is used to obtain the image intensity, resulting in a relatively gentler intensity change in that region, and the absolute value of the discrete wavelet transform coefficients becomes small. Thus, the magnitude of the absolute value of the discrete wavelet transform coefficients indicates the degree of deviation from the focal point.

[0051] Therefore, in this embodiment, a multi-resolution analysis is performed on the local region image 600 using discrete wavelet transform, and the coefficient values ​​w of the transverse wavelet transform images obtained by discrete wavelet transforms of multiple orders are obtained. １ｈ The maximum value C of the absolute value of (x, y) １ｈ and coefficient values ​​of the vertical wavelet transform image lol １ｖ The maximum value C of the absolute value of (x, y)１ｖ By multiplying it with, the evaluation value S is calculated. That is, the evaluation value S can be expressed as follows. Evaluation value S = (C １ｈ · C １ｖ ) · (C ２ｈ · C ２ｖ ) ··· (C Ｊｈ · C Ｊｖ ) Here, when j is an integer from 1 to J, C ｊｈ = |w ｊｈ (x, y)|max, C ｊｖ = |w ｊｖ (x, y)|max The maximum value C １ｈ and the maximum value C １ｖ are multiplied together to ensure that both the x-direction edge and the y-direction edge included in the local region image 600 are captured. Also, by multiplying the results of higher-order discrete wavelet transforms from one-dimensional discrete wavelet transforms, even when there is a lot of noise in the local region image 600, the degree of deviation from the focal position can be robustly reflected in the evaluation value S.

[0052] Note that the absolute value of the wavelet transform coefficient may be determined in an image where noise has been removed from the data, or in an image to which information restoration processing has been applied using technologies such as neural networks and compressive sensing. Also, although multiple algorithms are known as methods for obtaining the wavelet transform coefficient, the type thereof is not limited in the present invention.

[0053] Note that although an example of calculating the evaluation value S using discrete wavelet transform has been described, it may also be calculated using a Sobel filter, a Prewitt filter, or a Laplacian filter.

[0054] Also, when calculating the evaluation value S, the maximum value of the absolute value of the coefficient value of the wavelet transform image was used, but a specific percentile value (for example, 99%) may be used instead. This is to prevent the situation where there is a lot of noise in the local region image and an outlier caused by the noise is adopted as the maximum value of the absolute value of the wavelet coefficient.

[0055] Alternatively, the same processing may be performed using observation images obtained with a charged particle beam having astigmatism. The aberration correction device 116 corrects the astigmatism of the charged particle beam and then generates a predetermined amount of astigmatism. When the optical system has astigmatism, the shape of the charged particle beam differs in the region before and after the circle of least confusion. Therefore, the evaluation values ​​will differ in the direction of image blur between the image obtained when the charged particle beam is irradiated onto the sample upstream of the circle of least confusion and the image obtained when the charged particle beam is irradiated downstream of the circle of least confusion. By using this characteristic, observation images with different image blur characteristics above and below the focal point can be obtained, making it possible to capture the deviation from the focal point with greater accuracy.

[0056] In the above explanation, the evaluation value S of the coordinates included in the local region image is a common value, and the spatial resolution in the X and Y directions depends on the size of the local region that divides the observed image. For this reason, it may be possible to calculate the coefficients by multiplying the wavelet coefficients of multiple orders obtained by performing multi-resolution analysis using discrete wavelet transform on the local region image, according to the scale of the local region image. As shown in Figure 5, the sequence of higher-order wavelet coefficients is shorter than the sequence of lower-order wavelet coefficients. That is, as shown in Figure 7, the third-order wavelet transform image 703 is smaller than the second-order wavelet transform image 702, and the second-order wavelet transform image 702 is smaller than the first-order wavelet transform image 701. Therefore, these wavelet transform images are scaled up to match the local region image 700, and the coefficient values ​​w of the transverse wavelet transform images obtained by discrete wavelet transforms of multiple orders for each pixel are calculated. １ｈ The absolute value of (x, y) and the coefficient values ​​of the longitudinal wavelet transform image w １ｖ The evaluation value S is calculated by multiplying the product of the absolute value of (x, y) by the evaluation value S. That is, the evaluation value S can be expressed as follows: Evaluation value S = (|w １ｈ (x, y) |・| w １ｖ (x, y) |)・(|w ２ｈ (x, y) |・| w ２ｖ (x, y) |)...(|w Ｊｈ (x, y) |・| w Ｊｖ(x, y)|) This makes it possible to calculate the evaluation value S with a spatial resolution of a few pixels.

[0057] According to this embodiment, even for samples that do not have single-crystal regions, the three-dimensional shape of the sample can be calculated, and furthermore, the inclination of the sample can be calculated. In addition, automatic correction to reduce the inclination of the sample with respect to the aforementioned plane can be performed at high speed.

[0058] The present invention is not limited to the embodiments described above, and various modifications are included. For example, the embodiments described above are explained in detail to make the present invention easier to understand, and are not necessarily limited to those having all the configurations described. Furthermore, it is possible to add, delete, or replace some of the configurations of the embodiments with other configurations. For example, although this embodiment was explained using STEM as an example, the same applies to TEM. Also, the signal electrons detected by the detector are signal electrons emitted as a result of interaction between the sample and the charged particle beam, but are not limited to electrons that have passed through the sample, and may be electrons scattered from the sample or secondary electrons emitted from the sample.

[0059] 100: Scanning transmission electron microscope, 101: Electron optics tube, 102: Control system, 111: Electron source, 112: Electrode, 113: Condenser lens, 114: Condenser diaphragm, 115: Axis adjustment deflector, 116: Aberration correction device, 117: Scan deflector, 118: Objective lens, 119: Stage, 120: Objective diaphragm, 121: Axis adjustment deflector, 122: Field limiting diaphragm, 123: Imaging lens, 124: Detector, 130: Sample, 141: Control device, 142: Computer, 301: Charged particle beam, 302: Sample, 311: Observation image, 312: Observation image, 313: Observation image, 314: Observation image, 315: Observation image, 321: Station Local region image, 322: Local region image, 323: Local region image, 324: Local region image, 325: Local region image, 400: Evaluation value distribution, 401: Region, 402: Sample surface, 501: Original signal, 502: Approximate signal, 503: Detailed signal, 504: Scaling function, 505: Wavelet function, 506: Scaling coefficient sequence, 507: Wavelet coefficient sequence, 508: Region, 600: Local region image, 601: Transverse wavelet transform image, 602: Longitudinal wavelet transform image, 700: Local region image, 701: 1st order wavelet transform image, 702: 2nd order wavelet transform image, 703: 3rd order wavelet transform image.

Claims

1. The system comprises: a stage for holding a sample; a charged particle optical system for irradiating the sample with a charged particle beam; a detector for detecting signal electrons emitted as a result of interaction between the sample and the charged particle beam; and a control system for controlling the stage, the charged particle optical system, and the detector. The stage is configured to adjust the height of the sample held at least in the direction of the optical axis of the charged particle optical system; the charged particle optical system is configured to adjust the focal position in the direction of the optical axis of the charged particle optical system; the control system acquires a plurality of observation images based on the signal electrons detected by the detector, each while changing at least one of the focal position of the charged particle optical system and the height of the sample held by the stage; for each of the plurality of local region images constituting the observation image, it calculates an evaluation value that changes in accordance with the offset of the focal position of the charged particle optical system and the sample in the direction of the optical axis of the charged particle optical system; and it determines the three-dimensional distribution of the evaluation value in a three-dimensional space defined by an axis corresponding to the coordinates of the observation image and an axis corresponding to the focal position of the charged particle optical system or the height of the sample held by the stage. A charged particle beam apparatus for estimating the three-dimensional shape of a sample from the three-dimensional distribution of the aforementioned evaluation values.

2. The charged particle beam apparatus according to claim 1, wherein the local region image is an image obtained by dividing the observation image into a plurality of regions.

3. The charged particle beam apparatus according to claim 1, wherein the control system controls the stage based on the estimated three-dimensional shape of the sample.

4. The charged particle optics according to claim 1, comprising a deflector for adjusting the direction of propagation of the charged particle beam, wherein the control system controls the deflector based on the estimated three-dimensional shape of the sample.

5. The charged particle beam apparatus according to claim 1, wherein the control system estimates the amount of inclination of the sample surface, which is a model of the sample, with respect to a reference plane, as the three-dimensional shape of the sample.

6. The charged particle beam apparatus according to claim 5, wherein the control system models the sample surface including a term composed of a coefficient of order two or higher of the coordinates of the observation image, and estimates the local inclination of the sample surface.

7. The charged particle beam apparatus according to claim 1, wherein the control system calculates the evaluation value based on a plurality of order wavelet coefficients obtained by performing a multi-resolution analysis of the local region image using discrete wavelet transform.

8. The charged particle beam apparatus according to claim 7, wherein the control system calculates the evaluation value of coordinates included in the local region image as the product of the maximum value or a specific percentile value of the wavelet coefficients of each order calculated for the local region image.

9. The charged particle beam apparatus according to claim 7, wherein the control system calculates the evaluation value of coordinates included in the local region image by multiplying the wavelet coefficients of each order calculated for the local region image in accordance with the scale of the local region image.

10. The charged particle optical system comprising an aberration corrector for correcting astigmatism of the charged particle beam, and the control system comprising a charged particle beam apparatus for acquiring the observation image while the charged particle beam has astigmatism.

11. The charged particle beam apparatus according to claim 7, wherein the evaluation value is calculated based on coefficients obtained by applying one of the Sobel filter, Previtt filter, or Laplacian filter instead of the discrete wavelet transform.

12. A method for estimating the three-dimensional shape of a sample held in a charged particle beam apparatus, wherein the charged particle beam apparatus comprises a stage for holding the sample, a charged particle optical system for irradiating the sample with a charged particle beam, a detector for detecting signal electrons emitted as a result of interaction between the sample and the charged particle beam, and a control system for controlling the stage, the charged particle optical system, and the detector, wherein the stage is configured to adjust the holding height of the sample in at least the direction of the optical axis of the charged particle optical system, the charged particle optical system is configured to adjust the focal position in the direction of the optical axis of the charged particle optical system, the control system acquires a plurality of observation images based on the signal electrons detected by the detector, each changing at least one of the focal position of the charged particle optical system and the holding height of the sample by the stage, and calculates an evaluation value that changes in accordance with the deviation of the focal position of the charged particle optical system from the sample in the direction of the optical axis of the charged particle optical system for each of the plurality of local region images constituting the observation image, A method for estimating a three-dimensional shape, comprising determining the three-dimensional distribution of the evaluation value in a three-dimensional space defined by an axis corresponding to the coordinates of the observation image and an axis corresponding to the focal position of the charged particle optical system or the holding height of the sample by the stage, and estimating the three-dimensional shape of the sample from the three-dimensional distribution of the evaluation value.

13. The control system in claim 12 is a method for estimating a three-dimensional shape of a sample, which estimates the amount of inclination of the sample surface, which is a model of the surface of the sample, with respect to a reference plane, as the three-dimensional shape of the sample.

14. The control system according to claim 13, a method for estimating a three-dimensional shape, wherein the control system converts the amount of inclination into an amount of inclination control for controlling the stage.

15. The control system according to claim 12, a method for estimating a three-dimensional shape in which the control system calculates the evaluation value based on a plurality of order wavelet coefficients obtained by performing a multi-resolution analysis by discrete wavelet transform on the local region image.

Images