Charged particle beam device

Abstract

The present invention improves measurement precision by correcting distortion due to charging influence. To this end, in order to obtain a scan image, an imaging unit scans a sample with a charged particle beam a plurality of times under a prescribed optical condition. In obtaining the scan image, a computer integrates, by using an i-th transform matrix (i is an integer of 2 or more) calculated in advance, a corrected frame image obtained by correcting a frame image generated by an i-th scan for obtaining a scan image by using the imaging unit. The i-th transform matrix is calculated as a matrix for transforming a first coordinate group in a reference image including an image of a pattern into a second coordinate group corresponding to the first coordinate group in a frame image generated by an i-th scan for calculating a transform matrix by using the imaging unit.

Description

Charged particle beam device

[0001] This invention relates to a charged particle beam apparatus.

[0002] Depending on the material of the specific wafer and the optical conditions during imaging, distortion may occur in the SEM image due to charging when scanning the sample with an electron beam. As the pattern becomes smaller, the impact of distortion on length measurement increases, and a decrease in length measurement accuracy may lead to a decrease in yield. Patent Document 1 discloses a method for calculating the amount of stretching, rotation, and orthogonality by statistically calculating the amount of displacement of the shape corresponding to the position, in order to correct positional displacement due to wafer stretching.

[0003] Japanese Patent Publication No. 2002-251974

[0004] Patent Document 1 does not mention correction of image distortion due to static charge. The effect of static charge largely depends on the imaging conditions when acquiring SEM images, the pattern shape of the sample, and the material, and it can be said that a method for correcting the effect of static charge has not been established. Therefore, measurements are currently taken with the static charge affecting the image.

[0005] The system comprises an imaging unit that scans a sample with a charged particle beam under predetermined optical conditions, detects signal electrons emitted from the sample, and outputs a detection signal; a signal processing unit that associates the scanning position of the charged particle beam with the detection signal to generate a frame image; and a computer that integrates multiple frame images to obtain a scan image. A pattern is formed on the surface of the sample. To obtain a scan image, the imaging unit scans the sample multiple times with a charged particle beam under predetermined optical conditions. The computer, in obtaining a scan image, integrates a corrected frame image obtained by correcting the frame image generated by the i-th scan for obtaining the scan image by the imaging unit using a pre-calculated i-th transformation matrix (where i is an integer of 2 or more). To calculate the transformation matrix, the imaging unit scans the region of the sample containing the pattern multiple times with a charged particle beam under predetermined optical conditions. The i-th transformation matrix is ​​calculated as a matrix that transforms the first group of coordinates in a reference image containing the pattern image to a second group of coordinates corresponding to the first group of coordinates in the frame image generated by the i-th scan for calculating the transformation matrix by the imaging unit.

[0006] The technology disclosed herein aims to improve measurement accuracy by correcting distortion caused by electrostatic charge. Other challenges and novel features will become apparent from the description herein and the accompanying drawings.

[0007] This is a schematic diagram of the length-measuring SEM. This diagram illustrates the distortion that occurs in the frame image due to the charging of the sample surface. This is an example of a reference image and a frame image to be corrected. This is a flowchart of Example 1. This diagram illustrates the method of correcting the frame image in Example 1. This is a flowchart of Example 2. This diagram illustrates the method of calculating the transformation matrix p in Example 2. This is a flowchart of Example 3. This diagram illustrates the method of calculating the transformation matrix p in Example 3. This is a flowchart of Example 4. Transformation matrices p, p in Example 4. ＊ This is a diagram to explain the calculation method.

[0008] This embodiment will be described below with reference to the attached drawings. In the attached drawings, functionally identical elements may be indicated by the same number or corresponding number. The attached drawings show embodiments and implementation examples in accordance with the principles of this disclosure, but these are for the purpose of understanding this disclosure, and it is possible to change the configuration and structure and replace various elements without departing from the scope and spirit of the technical idea of ​​this disclosure. The description in this specification is merely a typical example and is not limited in any sense.

[0009] In the following description of embodiments, the disclosure shows an example of its application to a scanning electron microscope (SEM) using an electron beam. The disclosure may also be applied to charged particle beam apparatuses and systems using other microscopes, such as transmission electron microscopes (TEM), projection electron microscopes, and surface-illuminated electron microscopes, instead of a scanning electron microscope. Furthermore, the disclosure may also be applied to charged particle beam apparatuses and systems configured using multiple electron beams (multibeams) in the aforementioned electron microscopes, or to general observation systems.

[0010] Furthermore, in the functions, operations, processes, and flows of the embodiments described below, the explanation of each element and step will primarily focus on the computer or processor as the operating entity; however, the explanation may also focus on the "various programs" executed by the computer as the operating entity. Parts or all of the programs may be implemented using dedicated hardware, or they may be modularized. The various programs may be installed on the computer via a program distribution server or storage media.

[0011] Figure 1 is a schematic diagram of a Critical Dimension-Scanning Electron Microscope (CD-SEM), which comprises a computer 100, an imaging unit 101, and a signal processing unit 102. The computer 100 further includes a processor 103, an input / output device 104, and storage 105.

[0012] The imaging unit 101 is equipped with optical elements (collectively referred to as the electron optical system) involved in the irradiation and scanning of the electron beam. Specifically, it includes an electron gun 106, focusing lenses 108 and 109 for focusing the electron beam 107 emitted from the electron gun 106, a deflector 110 for deflecting the electron beam 107, and an objective lens 111 for controlling the height at which the electron beam 107 is focused. It is also provided with an aperture 130 that partially restricts the passage of the electron beam 107, a blanking deflector 131 that deflects the electron beam 107 out of the optical axis to limit the arrival of the electron beam 107 to the sample 112, and a blanking electrode 132 that receives the electron beam 107 deflected by the blanking deflector 131.

[0013] The electron beam 107 that has passed through the electron optical system described above is irradiated onto the sample 112 placed on the stage 113. Signal electrons 114, such as secondary electrons (SE) and backscattered electrons (BSE), emitted from the sample by the irradiation of the electron beam 107 are guided in a predetermined direction by a deflector 115 (first secondary electron aligner) for deflecting signal electrons. The deflector 115 is a so-called Wien filter, and it selectively deflects the signal electrons 114 in a predetermined direction without deflecting the electron beam 107 from the electron gun 106.

[0014] The signal electrons 114 that have passed through the detection aperture 116, which is provided to angularly discriminate the signal electrons 114, are guided by a deflector 123 (second secondary electron aligner) to a detector 119 positioned off-axis. A detector 121 is also provided to detect secondary electrons 120 generated by the collision of the signal electrons 114 with the detection aperture 116. An energy filter 122 is provided directly in front of the detector 119, and by energy discrimination, it is possible to selectively detect secondary electrons that are emitted vertically upward from the bottom of the semiconductor pattern formed on the sample 112 and have a trajectory that passes near the optical axis. The optical elements involved in the detection of signal electrons 114 as described above are collectively called the detection system.

[0015] The signal processing unit 102 generates an SEM image based on the output from the detection system. The signal processing unit 102 generates image data by storing the detection signal in a frame memory or the like, in synchronization with the scanning of the electron beam 107 by a scanning deflector (not shown). When storing the detection signal in the frame memory, the signal is stored at a position corresponding to the scanning position in the frame memory, thereby generating a signal profile (one-dimensional information) and a frame image (two-dimensional information).

[0016] The processor 103 integrates image data obtained based on multiple two-dimensional scans of the field of view (FOV). The image obtained in one two-dimensional scan is called a frame image, and the method of acquiring multiple frame images by scanning the FOV with the electron beam 107 multiple times and integrating them is called image integration. With this image integration, noise can be reduced by the cancellation effect of integrating each frame image, and the amount of detected signal can be increased by simple integration according to the number of frames to be integrated, thereby improving the signal-to-noise ratio of the SEM image (scanned image). In the following embodiment, an example of performing image integration on an FOV basis will be described, but it is not limited to this, and for example, image data of a specific part or region of interest (ROI) within the FOV may be selectively integrated to generate an SEM image.

[0017] First, the principle of distortion correction according to this disclosure will be explained. Since charge accumulates on the sample each time the sample surface is scanned by the electron beam, the magnitude of distortion due to charge differs for each frame image. Therefore, in this disclosure, when integrating frame images, a reference image is set, and the distortion due to charge in each frame image is corrected by comparing the frame image with the reference image, and an SEM image is generated by integrating the frame images for which the charge distortion has been corrected.

[0018] Figure 2 illustrates the distortion in frame images caused by charging of the sample surface. A raster scan is performed on a sample with a cross-shaped pattern formed on its surface, repeatedly scanning the electron beam in the X direction while sequentially shifting it in the Y direction. In the top row, "No Charge Effect," a frame image without pattern distortion is obtained. However, when charging occurs on the sample surface, this effect causes a shift in the electron beam's irradiation position on the sample. Distortion due to charging can be modeled as a combination of one or more changes from orthogonality change, rotation change, magnification change, and drift (position shift). Figure 2 schematically shows distorted frame images caused by orthogonality change, rotation change, and magnification change, respectively.

[0019] Figure 3 shows an example of a reference image 300 and a frame image 301 to be corrected. In this example, a hole pattern is regularly arranged on the sample surface. A collection of regularly arranged hole patterns like this is called a grid pattern. The reference image 300 is the reference image used to correct the frame image 301, and is, for example, the first frame image acquired. In the following explanation, a grid pattern is used as an example, but the type and shape of the pattern on the sample surface are not limited. For example, if the pattern is cross-shaped as in Figure 2, it is possible to perform the same processing using any point on the contour line of the pattern.

[0020] The reference image 300 contains a grid pattern with N (N=9 in the example shown) hole patterns. Here, the centroid position (X) of the hole pattern in the reference image 300 is... ｉ , Y ｉ ) (i=1 to N) is affected by the distortion caused by static charge, and the centroid position of the hole pattern in frame image 301 (x ｉ , y ｉ It can be considered that the displacement is (i = 1 to N). This is expressed in terms of a determinant as (Equation 1).

[0021]

[0022] A is called the design matrix and is a 2N x 6 matrix. The components of the matrix include the coordinates of the reference image. P is called the transformation matrix and is a 6 x 1 matrix, where P represents the effect of distortion due to charging. y is called the measurement matrix and is a 2N x 1 matrix. The components of the matrix include the coordinates of the frame image to be corrected.

[0023]

[0024] (Equation 2) shows how to calculate each component of the transformation matrix P. As described above, the strain due to charging can be modeled as a result of a combination of one or more changes from among the changes in orthogonality, rotation, magnification, and offset (position shift). The right-hand side of (Equation 2) shows that the transformation matrix P is a matrix representing the change in orthogonality (φ: deformation angle, see Figure 2), a matrix representing the change in rotation (θ: rotation angle, see Figure 2), and a matrix representing the change in magnification (S ｘ: Magnification in X direction, S ｙ : Magnification in Y direction), drift (t ｘ : Offset in X direction, t ｙ : Offset in Y direction), which is obtained as the product of matrices.

[0025] In order to obtain the transformation matrix p, (Equation 1) is converted into (Equation 3).

[0026]

[0027]

[0028] (Equation 3) is obtained by multiplying the pseudoinverse matrix of the design matrix A by the measurement matrix y, and the transformation matrix p is obtained so as to minimize the correction residual δ shown in (Equation 4). Therefore, by inversely transforming each pixel coordinate of the frame image 301 based on the transformation matrix p, a frame image (corrected frame image) from which the influence of distortion caused by charging has been removed can be obtained. Although the transformation matrix p has been described herein based on the least square method, the maximum likelihood method or the like can also be used.

[0029] As described above, an SEM image is obtained by accumulating corrected frame images in which distortion caused by charging has been reduced from the frame images, and length measurement is performed from the SEM image, thereby enabling more accurate length measurement.

[0030] Depending on the measurement conditions or the wafer being measured, distortions other than those caused by static charge cannot be ignored as factors contributing to the distortion of the frame image. For example, in overlay measurements where SEM images for length measurement are acquired using BSE, a large dose of electron beam causes sample damage such as resist shrinkage. Therefore, when frames are stacked, the sample damage progresses and the pattern becomes distorted. Such distortions caused by sample damage can also be suppressed by correction based on the transformation matrix p. Sample damage includes not only shrinkage but also pattern thickening due to contamination, and in either case, the length measurement accuracy can be improved by the method disclosed herein. The reason why the correction residual δ is not zero is that the frame image contains distortions caused by factors other than static charge. The upper and lower stage ratio of the deflector 123 and the deflector 115 is adjusted to eliminate distortions of magnification, rotation, and orthogonality with respect to the imaging unit 101, but the correction residual δ is proportional to the deflection distortion that occurs because these distortions do not become completely zero even after adjustment.

[0031] Furthermore, to perform corrections with higher accuracy, it is advisable to incorporate perturbations into the design matrix A and transformation matrix p shown in (Equation 1), and to perform higher-order corrections instead of linear corrections. For example, if you want to incorporate quadratic functions, add X to the design matrix A. ２ XY, Y ２ The term is added. Alternatively, when correcting the frame image based on the transformation matrix p, higher-order terms can be incorporated by correcting the distortion of the frame image using spline interpolation.

[0032] When measuring under conditions where static charge is easily generated, or when measuring wafers that are prone to static charge, drift, magnification, rotation, and orthogonality deviations occur during scanning due to the effects of static charge. In Example 1, length measurement with suppressed static charge effects is achieved by correcting these distortions that occur during electron beam scanning.

[0033] Figure 4 shows a flowchart of the measurement procedure in which the length-measuring SEM shown in Figure 1 performs charge correction. This assumes a scenario where the length-measuring SEM performs in-line measurement of a pattern formed on a mass-produced wafer.

[0034] A mass production line typically has multiple length measuring SEMs installed. On any one of these SEMs, necessary information such as the layout of the measurement recipe (operation program) used for mass-produced wafers, the coordinates of the measurement pattern, and the measurement conditions are input from the input / output device 104 to create a measurement recipe (operation program) and store it in the storage 105. At this time, the charge correction function is set in the measurement recipe. The details of the charge correction function can be set for each wafer and condition to be corrected. The created recipe can be deployed and stored on other length measuring SEMs connected via the network (S01).

[0035] Next, the processor 103 of the length measuring SEM executes the measurement recipe created in step S01 and starts measuring the target wafer (S02). First, the processor 103 moves the FOV to the coordinates of the pattern to be measured and acquires the number of frame images specified in the measurement recipe (S03). In Example 1, the frame image of the first frame captured is set as the reference image (S04), a transformation matrix p is calculated for each frame image from the second frame onward (S05), and correction is performed using the calculated transformation matrix p (S06).

[0036] This is schematically shown in Figure 5. Using the first frame image 401 as the reference image, the transformation matrix p is applied to each frame image from the second frame onward. ｋ We find (k=2 to n) and, for the k-th frame image, the transformation matrix p ｋ Corrected frame images are obtained by performing an inverse transform using the method described above. The corrected frame image of the second frame, 402, is corrected frame image 412, and the corrected frame image of the nth frame, 403, is corrected frame image 413. In this way, corrected frame images can be generated for each frame image that are fixed to the magnification of the reference image and have minimal distortion and drift such as rotation and orthogonality.

[0037] In this way, by integrating the corrected frame images, a SEM image for length measurement is obtained (S07), and the length measurement value is obtained (S08). This makes it possible to perform measurements with minimal distortion of magnification, rotation, orthogonality, and drift caused by the effect of static charge. In the flowchart of Figure 4, the correction of the frame image is explained after the frame image is acquired, but the frame image acquisition and frame image correction may be performed in parallel.

[0038] In Example 2, variations in measurement due to different optical conditions are suppressed by correcting the difference in the magnitude of charge-induced distortion caused by different optical conditions. As an example of different optical conditions, we will explain using the difference in the number of integrated images. Figure 6 shows a flowchart of the measurement procedure in which the measurement SEM shown in Figure 1 suppresses variations in charge-induced distortion caused by optical conditions.

[0039] The creation of the measurement recipe is the same as in Example 1, but this time, multiple condition correction recipes are created in which only one of the optical conditions, the number of integrated images, is different. For example, condition correction recipes are created with the number of integrated images set to 4, 8, 16, and 32 (S11).

[0040] The condition correction recipe is executed to obtain a condition correction SEM image (S12). The flow described in Example 1 is also performed for the condition correction recipe, but the SEM image obtained by the condition correction recipe is called the condition correction SEM image.

[0041] One of the condition correction SEM images is set as the reference image (S13), and a transformation matrix p is calculated for each of the other condition correction SEM images (S14). Figure 7 shows the transformation matrix p for each of the condition correction SEM images with image integration counts of 8, 16, and 32, with the condition correction SEM image with image integration counts of 4 as the reference image (p Ａ , p Ｂ , p Ｃ A schematic example of calculating the transformation matrix p(p) is shown below. Ａ , p Ｂ , p Ｃ The offset is saved as an offset file (S15), and the offset file is linked to the measurement recipe (S16).

[0042] The measurement recipe is executed and the target wafer is measured (S17). The step of executing the measurement recipe and obtaining a length-measuring SEM image is as described in Example 1. The obtained length-measuring SEM image is corrected according to the reference optical conditions (S18). For example, if the measurement recipe has 16 integrated images, the coordinates of each pixel in the length-measuring SEM image are corrected using a transformation matrix p Ｂ By performing an inverse transformation based on this, correction can be made so that the charge effect is equivalent to the charge effect when the number of integrated images is 4 (reference optical conditions). The measured length value is obtained from the corrected measuring SEM image (corrected measuring SEM image) (S19). This makes it possible to match the measured length results of measuring SEM images obtained under different optical conditions with the measured length results of measuring SEM images obtained under reference optical conditions.

[0043] Furthermore, the transformation matrix p stored in the offset file allows for the matching of measurement results when the same length-measuring SEM measures wafers of the same process. When measuring wafers of different length-measuring SEMs or processes, it is necessary to recalculate the transformation matrix p. In addition, although the image integration count is used as an example of optical conditions here, differences in charge effects caused by differences in optical conditions such as acceleration voltage, probe current, and electron beam scanning speed can be similarly corrected.

[0044] In in-line measurement of mass-produced wafers, multiple length measuring SEMs are used, but even when measuring the same wafer, the charge effect differs depending on the potential reference value, the amount of landing electrons, or the in-plane distribution for each device. In Example 3, the variation in length measurement due to device differences is suppressed by correcting for the difference in the magnitude of charge-induced distortion which differs for each length measuring SEM. Figure 8 shows a flowchart of the length measurement procedure in which the length measuring SEM shown in Figure 1 suppresses the variation in charge-induced distortion due to device differences.

[0045] From among the multiple length-measuring SEMs, one unit is selected as the reference unit (referred to as the reference unit), and a measurement recipe is created in the reference unit in the same manner as in Example 1, and this is used as the machine error correction recipe. The machine error correction recipe created in the reference unit is then deployed to the other length-measuring SEMs and stored in the storage 105 (S11).

[0046] Each measuring SEM executes a recipe for instrument error correction and acquires an SEM image for instrument error correction (S22). The flow described in Example 1 is also performed for the instrument error correction recipe, but the SEM image acquired by the instrument error correction recipe is called an SEM image for instrument error correction.

[0047] The SEM image used for instrument error correction of the reference machine is set as the reference image (S23), and a transformation matrix p is calculated for each of the SEM images used for instrument error correction of the other measuring SEMs (S24). Figure 9 shows the transformation matrix p for each of the SEM images used for instrument error correction of the first to third measuring SEMs, with the SEM image used for instrument error correction of the reference machine as the reference image (p Ｄ , p Ｅ , p Ｆ A schematic example of calculating the transformation matrix p(p) is shown below. Ｄ , p Ｅ , p Ｆ The result is saved as an offset file (S25), and the offset file is linked to the measurement recipe (S26).

[0048] The measurement recipe is executed and the target wafer is measured (S27). The step of executing the measurement recipe and obtaining a length-measuring SEM image is as described in Example 1. The obtained length-measuring SEM image is corrected to match the reference instrument (S28). For example, if the length-measuring SEM on which the measurement recipe was executed is the first length-measuring SEM, the coordinates of each pixel in the length-measuring SEM image are converted to a transformation matrix p Ｄ By performing an inverse transformation based on this, correction can be made so that the charge effect is equivalent to the charge effect of the reference instrument. The measured length value is obtained from the corrected measuring SEM image (corrected measuring SEM image) (S19). This makes it possible to match the measured length results of measuring SEM images obtained from multiple measuring SEMs with the measured length results of measuring SEM images obtained from the reference instrument.

[0049] Furthermore, the transformation matrix p stored in the offset file allows for the matching of measurement results when the same length-measuring SEM measures wafers of the same process. When measuring wafers of different length-measuring SEMs or processes, it is necessary to recalculate the transformation matrix p.

[0050] In SEM, images are generated by irradiating the sample with an electron beam. Therefore, it is unavoidable that SEM images will include the effects of electric charge. In Example 4, an ideal SEM image without the effects of electron beam irradiation is generated, and the measured length values ​​are estimated.

[0051] The creation of the measurement recipe is the same as in Example 1, but this time, multiple ideal state estimation recipes are created with only the number of integrated images differing. For example, ideal state estimation recipes with image integrated numbers of 4, 8, 16, and 32 are created (S31).

[0052] The recipe for estimating the ideal state is executed, and an SEM image for estimating the ideal state is obtained (S32). The flow described in Example 1 is also performed for the recipe for estimating the ideal state, but the SEM image obtained by the recipe for estimating the ideal state is called the SEM image for estimating the ideal state.

[0053] One of the SEM images used for estimating the ideal state is set as the reference image (S33), and a transformation matrix p is calculated for each of the other SEM images used for estimating the ideal state (S34). Figure 11 shows the SEM image used for estimating the ideal state with 32 integrated images as the reference image, and the transformation matrix p for each of the SEM images used for estimating the ideal state with 4, 8, and 16 integrated images (p G , p H , p I A schematic example of how to calculate ( ) is shown below.

[0054] Next, the transformation matrix p(p G , p H , p I ) and an approximate formula representing the relationship between the number of integrated images is calculated (S15). That is, an approximate formula is created with the number of integrated images as the explanatory variable and each component of the transformation matrix p as the objective variable, and by calculating each component of the transformation matrix p when the number of integrated images is 0, the transformation matrix p when the number of integrated images is 0 is calculated. ＊ We estimate this.

[0055] The measurement recipe is executed and the target wafer is measured (S37). At this time, the number of integrated images in the measurement recipe is the number of integrated images in the SEM image set as the reference image in step S33. The process of executing the measurement recipe and obtaining the SEM image for length measurement is as described in Example 1. The obtained SEM image for length measurement is transformed into a transformation matrix p ＊ Correction is performed using (S38). The measured length is obtained from the corrected SEM image for length measurement (corrected SEM image for length measurement) (S39). This makes it possible to estimate the measured length of the SEM image for length measurement under ideal conditions, without electron beam irradiation.

[0056] Furthermore, the parameters set stepwise in step S31 are not limited to the number of integrated images, but may also be the probe current or the electron beam scanning speed. Both are parameters that affect the charge state of the sample, making it possible to estimate the measured length value in an ideal state where the charge effect is suppressed. In particular, it is possible to improve measurement accuracy in patterns where shrinkage occurs, such as resists and low-K films.

[0057] The present invention is not limited to the embodiments described above, and includes various modifications. For example, the embodiments and modifications 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 replace parts of the configuration of one embodiment or modification with the configuration of another embodiment or modification, and it is also possible to add the configuration of another embodiment or modification to the configuration of one embodiment or modification. In addition, it is possible to add, delete, or replace parts of the configuration of each embodiment or modification with other configurations.

[0058] For example, by obtaining SEM images from different measurement points and using one of the SEM images from a single measurement point as the reference image to determine a transformation matrix p, it is possible to measure the in-shot distortion and inter-shot distortion of the exposure machine from the differences in the transformation matrix p. Furthermore, it is also possible to measure the in-plane uniformity of etching, film deposition, and CMP. Alternatively, for groove patterns and hole patterns, by obtaining SEM images for each pattern and using one of the SEM images as the reference image to determine a transformation matrix p, it is possible to quantitatively evaluate the processing accuracy of etching from the variation in the transformation matrix p.

[0059] 100: Computer, 101: Imaging unit, 102: Signal processing unit, 103: Processor, 104: Input / Output device, 105: Storage, 106: Electron gun, 107: Electron beam, 108, 109: Focusing lens, 110: Deflector, 111: Objective lens, 112: Sample, 113: Stage, 114: Signal electrons, 115: Deflector, 116: Detection aperture, 119: Detector, 120: Secondary electrons, 121: Detector, 122: Energy filter, 123: Deflector, 130: Aperture, 131: Blanking deflector, 132: Blanking electrode, 300: Reference image, 301: Frame image, 401, 402, 403: Frame images, 412, 413: Corrected frame images.

Claims

1. The system comprises: an imaging unit that scans a sample with a charged particle beam under predetermined optical conditions, detects signal electrons emitted from the sample and outputs a detection signal; a signal processing unit that associates the scanning position of the charged particle beam with the detection signal and generates a frame image; and a computer that integrates a plurality of the frame images to obtain a scan image, wherein a pattern is formed on the surface of the sample; the imaging unit scans the sample multiple times with a charged particle beam under the predetermined optical conditions in order to obtain the scan image; the computer, in order to obtain the scan image, integrates a corrected frame image obtained by correcting the frame image generated by the i-th scan by the imaging unit to obtain the scan image using a pre-calculated i-th transformation matrix (i is an integer of 2 or more); and the imaging unit scans the region of the sample including the pattern multiple times with a charged particle beam under the predetermined optical conditions in order to calculate the transformation matrix. The i-th transformation matrix is ​​calculated as a matrix that transforms a first group of coordinates in a reference image containing the pattern image to a second group of coordinates corresponding to the first group of coordinates in a frame image generated by the i-th scan by the imaging unit for calculating the transformation matrix, in a charged particle beam apparatus.

2. The charged particle beam apparatus according to claim 1, wherein the reference image is a frame image generated by the first scan performed by the imaging unit to calculate the transformation matrix.

3. The computer in claim 1 is a charged particle beam apparatus that acquires length values ​​from the scanning image.

4. The charged particle beam apparatus according to claim 1, wherein the components of the transformation matrix are obtained as the product of a matrix representing the orthogonality change of the image, a matrix representing the rotation change of the image, a matrix representing the magnification change of the image, and a matrix representing the drift of the image.

5. In claim 1, the first coordinate group is (X ｋ , Y ｋ ) (k=1 to N), the second coordinate group is (x ｋ , y ｋ ), (X ｋ , Y ｋ ) of said first coordinate group corresponds to (x ｋ , y ｋ ) of said second coordinate group, and when a component of the conversion matrix is P ｍｎ (m=1 or 2, n=1, 2 or 3), the relationship of (Formula 1) is satisfied, A charged particle beam device, wherein the conversion matrix is calculated so as to minimize a correction residual δ calculated by (Formula 2).

6. The charged particle beam apparatus according to claim 1, wherein the imaging unit obtains a first scan image under first optical conditions, corrects the first scan image to a second scan image using a condition correction transformation matrix, the computer scans a region of the sample including the pattern multiple times under first optical conditions and second optical conditions to calculate the condition correction transformation matrix, and obtains a first condition correction scan image and a second condition correction scan image, and the condition correction transformation matrix is ​​calculated as a matrix that transforms the first coordinate group in the first condition correction scan image to the second coordinate group corresponding to the first coordinate group in the second condition correction scan image.

7. Charged particle beam apparatus according to claim 1, wherein the imaging unit obtains a first scan image under predetermined optical conditions, corrects the first scan image to a second scan image using a machine error correction transformation matrix, the computer scans a region of the sample including the pattern multiple times under predetermined optical conditions to calculate the machine error correction transformation matrix, and obtains a first machine error correction scan image, the machine error correction transformation matrix is ​​calculated as a matrix that converts the first coordinate group in the first machine error correction scan image to the second coordinate group corresponding to the first coordinate group in the second machine error correction scan image acquired by the reference machine, and the second machine error correction scan image is a scan image obtained by the reference machine scanning a region of the sample including the pattern multiple times under predetermined optical conditions to calculate the machine error correction transformation matrix.

8. The charged particle beam apparatus according to claim 1, wherein the imaging unit obtains a first scan image under first optical conditions, corrects the first scan image to a second scan image using an ideal state estimation transformation matrix, the computer scans a region of the sample including the pattern multiple times under multiple optical conditions to obtain multiple ideal state estimation scan images, and the ideal state estimation transformation matrix is ​​estimated based on a plurality of transformation matrices that convert one of the plurality of ideal state estimation scan images to any other of the plurality of ideal state estimation scan images.

Images