Crystal orientation search system, crystal orientation search method, and crystal orientation search program

Abstract

To provide a crystal orientation search system capable of efficiently screening a visual field with respect to a polycrystalline or pulverized powder sample.SOLUTION: The crystal orientation search system 1 is a search system using a computer including a scanning transmission electron microscope 2 that can scan a sample with an electron beam and acquires a STEM image and a diffraction pattern in a STEM mode, a reference information storage unit 20 that stores reference information on a crystal in advance, a diffraction pattern collection control unit 21 that causes an electron beam to stay for a predetermined time in response to designation of a sample position assigned to a pixel on the STEM image and collects and stores a diffraction pattern, and an estimated crystal orientation selection unit 22 that selects an estimated crystal orientation at the sample position based on the designated crystal orientation, the reference information, and the diffraction pattern.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a technique for screening the field of view of a sample using a scanning transmission electron microscope. [Background technology]

[0002] Most of the materials used in industry are polycrystalline, including sintered bodies. In particular, in ceramics, the strength and properties of the material change depending on the structure and composition of the grain boundaries, making them extremely important for evaluating the properties of materials. Recently, the structure and composition of these grain boundaries have been observed and analyzed at atomic resolution using a scanning transmission electron microscope. However, to obtain atomic resolution images and atomic column elemental maps, it is necessary to align the crystal orientation with a specific (low-order) zone axis. Of course, if the crystal orientation is not aligned, neither a crystal structure image nor an atomic column elemental map can be obtained. A conventional method for searching for crystal orientation is described in Patent Document 1. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 3672728 Summary of the Invention [Problem to be solved by the invention]

[0004] Aligning the crystal orientation (aligning the crystal's low-order zone axis with the direction of incidence of the electron beam) is essential, and electron microscope users spend a lot of time on this task. For large grains or single-crystal samples, even if the field of view shifts slightly due to specimen tilt, as long as the shift is within the grain, the crystal orientation remains unchanged, so adjustments are not a major problem. However, for small grains, tilting the specimen to align the crystal orientation often causes the grain being observed to escape the field of view. Even for large grains, if the crystal orientation of the grain is significantly off and the tilt amount exceeds the tilt range of the goniometer, alignment is impossible, and it may be necessary to search for several grains or remake the specimen.

[0005] On the other hand, when the grain size is small, many grains can be seen in one field of view, so the user must first find grains oriented in the desired direction (preliminary screening). Although several methods exist for achieving this, currently no specific, user-friendly means are provided, and most users select grains to observe based on prior knowledge gained from experience of grain shape or non-quantitative empirical knowledge such as the brightness of dark-field and bright-field images. When many grains are in the field of view, the chances of finding grains close to the desired orientation increase.

[0006] Conventional techniques for this screening include orientation detection using precession diffraction (PED) and orientation detection using electron backscatter diffraction (EBSD), which is primarily performed with a scanning electron microscope (SEM). However, with EBSD, for example, the sample is essentially bulk, so even if the target grain is found, it cannot be immediately observed using a transmission electron microscope (TEM). The sample at that location must be removed using a focused ion beam (FIB) to create a thin section sample, which requires moving the sample between several pieces of equipment. Furthermore, when the grain size is submicron, EBSD requires precision in identifying the location and FIB processing, making it difficult to fabricate.

[0007] In addition, the precession diffraction (PED) method uses a TEM sample from the beginning, but requires an electron diffraction detector such as a dedicated camera to obtain a 2D map of the diffraction pattern. Furthermore, it takes a long time (sometimes several hours) to find the necessary orientation from the data cube containing the 2D diffraction pattern for each pixel (orientation analysis). As a result, this method requires a large financial and time investment in the equipment, so it is not widely used. Furthermore, precession diffraction requires special illumination lens systems and electron beam deflection system settings. Switching from such a setup to STEM mode using a high-resolution scanning transmission electron microscope requires significant changes to the lens settings and may require readjustment of the microscope's alignment.

[0008] The present invention has been made to solve the above-mentioned problems of the prior art. The object of the present invention is to provide a crystal orientation searching system, method, and program that can efficiently perform field screening for polycrystalline or crushed powder samples, sintered samples made by sintering powder, and metal or semiconductor samples made of polycrystalline materials. [Means for solving the problem]

[0009] The crystal orientation searching system according to the first aspect of the present invention comprises: a scanning transmission electron microscope capable of scanning an electron beam on a sample and acquiring STEM images and diffraction patterns in STEM mode; a reference information storage unit that stores reference information related to crystals in advance; a diffraction pattern collection control unit that, in response to the designation of a sample position assigned to a pixel on the STEM image, causes the electron beam to dwell for a predetermined time, and collects and stores the diffraction pattern; an estimated crystal orientation selection unit that selects an estimated crystal orientation at the sample position based on a designated crystal orientation, the reference information, and the diffraction pattern; The present invention is characterized in that it is a crystal orientation searching system using a computer having the above. Since several companies sell systems for collecting diffraction patterns, the diffraction pattern collection control unit and estimated crystal orientation selection unit can be combined as separate units. In this way, if a user already owns an electron diffraction collection system, it is possible to reduce costs by eliminating the collection unit.

[0010] A second aspect of the present invention is the crystal orientation searching system according to the first aspect, the estimated crystal orientation selection unit includes a direct spot detection unit that detects a direct spot having the maximum intensity from among the diffraction spots of the diffraction pattern; a distance / angle calculation unit that selects the detected direct spot and any two of the diffraction spots and calculates an angle formed by the two distances to the diffraction spots of the any two points; a crystal orientation candidate search unit that searches for a crystal orientation candidate based on the two distances, the formed angle, and the reference information; a crystal orientation candidate selection unit that selects the crystal orientation candidate to be an estimated crystal orientation based on the regularity of the diffraction spots; The present invention is characterized by having the following:

[0011] A third aspect of the present invention is the crystal orientation searching system according to the first aspect, the reference information comprises a reference diffraction pattern obtained by simulation or experiment at a predetermined camera length, a predetermined orientation, and an incident angle of the electron beam to the sample; a conversion unit in which the estimated crystal orientation selection unit converts the diffraction pattern into a rotation angle and distance axis or the rotation angle and logarithmic distance axis to generate a converted diffraction pattern; a coincidence calculation unit that calculates a coincidence between the reference diffraction pattern and the transformed diffraction pattern using a cross-correlation function; The estimated crystal orientation is selected based on the degree of coincidence.

[0012] A fourth aspect of the present invention is the crystal orientation searching system according to the first aspect, characterized in that a convolution operation or an autocorrelation operation is performed on the diffraction pattern and used to select the estimated crystal orientation. A fifth aspect of the present invention is the crystal orientation searching system according to the first aspect, the estimated crystal orientation selection unit has a direct spot detection unit that detects a direct spot with maximum intensity from among the diffraction spots of the diffraction pattern, The direct spot detection unit divides the diffraction pattern into a plurality of regions, compares the integrated intensity for each divided region, and detects the direct spot in a predetermined range where the integrated intensity is high.

[0013] A sixth aspect of the present invention is the crystal orientation searching system according to the first aspect, the estimated crystal orientation selection unit has a direct spot detection unit that detects a direct spot with maximum intensity from among the diffraction spots of the diffraction pattern, The estimated crystal orientation selection unit multiplies the diffraction pattern by a function whose value decreases as the distance from the direct spot increases, as a pre-processing of the diffraction pattern. A seventh aspect of the present invention is the crystal orientation searching system according to the fifth aspect, characterized in that it includes a process of comparing the integrated intensities in re-divided regions shifted by a distance smaller than the length of one of the divisions.

[0014] An eighth aspect of the present invention is the crystal orientation searching system according to the first aspect, characterized in that the reference information includes positions of diffraction spots corresponding to indices of crystal orientations, intensities of the diffraction spots, interplanar spacings corresponding to distances of the diffraction spots, and angles between the diffraction spots. A ninth aspect of the present invention is the crystal orientation searching system according to the first aspect, characterized in that it comprises a display control unit that displays a predicted diffraction pattern based on the estimated crystal orientation superimposed on the diffraction pattern. A tenth aspect of the present invention is the crystal orientation searching system according to the first aspect, characterized in that it further comprises a display control unit that displays the estimated crystal orientation and the sample position based on the estimated crystal orientation on the STEM image.

[0015] A crystal orientation searching method according to an eleventh aspect of the present invention comprises: A method for searching for a crystal orientation using a scanning transmission electron microscope that can scan an electron beam on a sample and acquire a STEM image and a diffraction pattern in a STEM mode, comprising: an acquisition and storage step of suspending the electron beam for a predetermined time in response to the designation of a sample position assigned to a pixel on the STEM image, and acquiring and storing the diffraction pattern; an estimated crystal orientation selection step of selecting an estimated crystal orientation at the sample position based on the designated crystal orientation, reference information previously stored about the crystal, and the diffraction pattern; The present invention is characterized in that the crystal orientation searching method is executed by a computer, comprising:

[0016] A crystal orientation searching program according to a twelfth aspect of the present invention comprises: A crystal orientation search program for searching for a crystal orientation using a scanning transmission electron microscope that can scan an electron beam on a sample and acquire a STEM image and a diffraction pattern in a STEM mode, an acquisition and storage process for dwelling the electron beam for a predetermined time in response to the designation of a sample position assigned to a pixel on the STEM image, and acquiring and storing the diffraction pattern; an estimated crystal orientation selection process for selecting an estimated crystal orientation at the sample position based on a designated crystal orientation, pre-stored reference information about the crystal, and the diffraction pattern; The present invention is characterized in that the present invention is a crystal orientation search program that causes a computer linked to the scanning transmission electron microscope to execute the above. [Effects of the Invention]

[0017] The present invention provides a crystal orientation search system, method, and program that can efficiently screen the field of view for polycrystalline or crushed powder samples, sintered samples made by sintering powder, and metal or semiconductor samples made of polycrystalline materials. [Brief explanation of the drawings]

[0018] [Figure 1] 1 is a schematic block diagram showing one configuration of a crystal orientation searching system according to an embodiment of the present invention. [Figure 2] 10(a) to 10(d) are diagrams for explaining how to specify the sample position. [Figure 3] 1 is a flowchart showing main processes performed in the present search system. [Figure 4] FIG. 10 is a diagram illustrating an example of a detection process for diffraction spots. [Figure 5] 10(a) to 10(i) are diagrams for explaining the filtering process performed on the spot-shaped diffraction spots. [Figure 6] 10(a) to 10(i) are diagrams for explaining the filtering process performed on the disk-shaped diffraction spot. [Figure 7] 10 is a flowchart showing an example of processing performed by an estimated crystal orientation selection unit. [Figure 8] 10(a) and 10(b) are diagrams for explaining how to calculate the distances from a direct spot to any two diffraction spots and the angles between them. [Figure 9] 10(a) and 10(b) are diagrams for explaining a process of selecting crystal orientation candidates based on regularity. [Figure 10] 10(a) and 10(b) are diagrams showing examples of application of the present search system when silicon is used as a sample. [Figure 11] FIG. 1 is a diagram showing a disk-shaped diffraction spot as an example of a diffraction pattern. [Figure 12] (a) to (h) are diagrams illustrating a method for simulating a diffraction pattern from crystal information using a method similar to PED. [Figure 13] 1 is an image showing the extracted area of ​​polystyrene. [Figure 14] 1 is an image showing the sample position. [Figure 15] FIG. 10 is a diagram showing diffraction spots detected from an image of an actual diffraction pattern. [Figure 16] FIG. 10 is a diagram showing an example of an image in which detected diffraction spot information and diffraction spot information predicted from estimated crystal orientation are displayed together with an image of an actual diffraction pattern. DETAILED DESCRIPTION OF THE INVENTION

[0019] A crystal orientation searching system (hereinafter, abbreviated as a searching system as appropriate) according to one embodiment of the present invention will be described below with reference to the drawings. [First embodiment] FIG. 1 is a schematic block diagram showing an example of the configuration of a search system. The search system 1 includes a scanning transmission electron microscope 2 that acquires STEM images and electron diffraction patterns (hereinafter referred to as diffraction patterns) in STEM mode, a linked computer 4, a display unit 5 such as an LCD display device, and an input unit 6 such as a keyboard and mouse. As shown in FIG. 1, the search system 1 of this embodiment includes a functional unit and a reference information storage unit 20 realized by cooperation between hardware such as a personal computer or a server computer and software such as a crystal orientation search program. The information stored in the reference information storage unit 20 is not particularly limited as long as it is information that can acquire information related to crystal orientation from a diffraction pattern. Generally, for a crystal, the positions of diffraction spots corresponding to the index of crystal orientation, the intensity of the diffraction spots, the interplanar spacing corresponding to the distance between the diffraction spots, the angle between the diffraction spots, etc. are stored in advance, as will be described later. The reference information includes crystal information, and can include not only data such as physical constants and material property values, but also diffraction patterns obtained in advance by simulation or experiment.

[0020] The functional units include a diffraction pattern collection control unit 21, an estimated crystal orientation selection unit 22, and a display control unit 23 that displays a predicted diffraction pattern superimposed on the diffraction pattern and also displays the sample position and estimated crystal orientation on the STEM image. The estimated crystal orientation selection unit 22 includes a direct spot detection unit 24, a distance / angle calculation unit 25, a crystal orientation candidate search unit 26, a crystal orientation candidate selection unit 27, etc. 1 has an electron gun 10, accelerating voltage devices 11a and 11b, a sample stage 13, an objective lens 14, a detection system plane 15, and a camera 3 for detecting STEM images and diffraction patterns. The detection system plane 15 is provided with associated detectors such as a bright-field detector and a dark-field detector.

[0021] <Pretreatment> This search system simulates electron diffraction patterns obtained at specific orientations based on given crystal information (e.g., the crystal's a, b, and c axes, α, β, and γ angles, and crystal system) and stores them as reference information in a database. Specifically, given crystal information, the positions and intensities of diffraction spots corresponding to the indexes can be calculated using kinematic diffraction theory. Furthermore, the interplanar spacing corresponding to the distance between each diffraction spot and the angle between each index spot are also calculated. Since this search system is intended for screening and specializes in low-order orientation estimation, it targets diffraction spots belonging to the zeroth-order Laue zone. The diffraction vector (h, k, l) represented by a diffraction spot belonging to the zeroth-order Laue zone and the incident orientation vector or crystal zone axis vector [u, v, w] are orthogonal to each other, and the inner product is zero, so the incident orientation [u, v, w] is estimated by utilizing this fact. This information is stored as reference information. This reference information will be discussed later.

[0022] <Specimen position specified by the user> First, the diameter of the illumination lens aperture is reduced to reduce the illumination angle onto the sample surface, and then the lens setting is changed to reduce the reduction ratio of the probe, thereby reducing the illumination angle. The standard for the size of the irradiation angle in this case is to ensure that the closest diffraction spot does not overlap with the direct beam, which is empirically a requirement for accurate detection. A STEM image is acquired using this set beam, and the user or the like specifies the positions assigned to the pixels of the STEM image. In this state, the reduction ratio is not so small, and the angular distribution of the illumination angle is a step function, so the diffraction spot often becomes a diffraction disk. Therefore, the user changes the aperture diameter and the lens system settings so that the diffraction disk and the direct disk do not overlap. Note that in this specification, the term "diffraction spot" includes both point-like diffraction spots and disk-like diffraction disks.

[0023] FIG. 2 is a diagram showing an example of a method for specifying a sample position. Figure 2(a) is a diagram showing the crystal grains (grains 29) of the sample, Figure 2(b) is a diagram showing the pattern of specified positions created by intersections 31 of a grid pattern 30, Figure 2(c) is a diagram showing the pattern created by line points 33 on a line pattern 32, and Figure 2(d) is a diagram showing the pattern created by multiple specified points 34 freely selected by the user. This type of method allows specifying multiple points over a wide range, but any method can be used to specify points for acquiring diffraction patterns as long as it covers all grains observable on the STEM image. The diffraction pattern collection control unit 21 then drives the scanning transmission electron microscope 2, and the specified positions and other information are recorded in the computer 4 attached to the microscope. Then, the collection software is started using a starting means such as a start button, and the beam is moved to the specified points in sequence, and the diffraction pattern at each point is recorded with a camera attached to the electron microscope, thereby collecting a group of diffraction patterns.

[0024] <User-specified crystal orientation> This search system assumes that the crystal orientation analysis of each diffraction pattern is not performed for all patterns, but rather the user specifies the desired crystal orientation. The system then selects diffraction patterns with the crystal orientation specified by the user. Furthermore, to find the crystal orientation, crystal information about the crystal of the sample is required. In this embodiment, the sample information does not include determining the orientation of an unknown sample.

[0025] FIG. 3 is a flowchart showing an example of processing performed in this search system. In SP1, in response to the user individually specifying a sample position or the position specification program specifying a sample position, the diffraction pattern collection control unit 21 dwells the electron beam for a predetermined time, collects and stores a group of diffraction patterns, and obtains a diffraction pattern for each sample position. Next, in SP2, the estimated crystal orientation selection unit 22 selects an estimated crystal orientation for each diffraction pattern by pattern diffraction based on the crystal orientation desired by the user. Then, in SP3, the display control unit 23 displays information related to the estimated crystal orientation together with a specified image such as a diffraction pattern image or STEM image. Figures 10, 14, 15, and 16 are examples of such display formats. Next, SP2 "Selection of estimated crystal orientation by pattern analysis" will be explained. <Selection of estimated crystal orientation by pattern analysis> The following formula (1) can be used to determine the crystal orientation.

[0026]

number

[0027] In equation (1), v is the orientation vector of the crystal, g1, g2, and g3 are vectors of non-parallel diffraction points detected on the diffraction pattern, x is the cross product (vector product), and / / indicates the parallel relationship. Here, the diffraction vectors g1, g2, and g3 in the zeroth-order Laue band are estimated from the diffraction pattern. These diffraction vectors are the intersections of the Ewald sphere and the zeroth-order reciprocal lattice. Considering the excitation error and the fact that many diffracted beams are excited within the range of the incident beam's illumination angle, three diffraction vectors are not necessarily required; even with two diffraction vectors, the crystal orientation can be determined from their cross product.

[0028] (Diffraction vector estimation) The index of the diffraction vector is estimated from the lattice spacing d value calculated from crystal information. The lattice spacing of each elementary lattice can be calculated from crystal structure information based on the following formulas (2) to (8), as described in publicly known literature.

[0029]

number

[0030] Formula (2) shows the interplanar spacing of a cubic crystal, formula (3) shows the interplanar spacing of a tetragonal crystal, formula (4) shows the interplanar spacing of a hexagonal crystal, formula (5) shows the interplanar spacing of a trigonal crystal, formula (6) shows the interplanar spacing of an orthorhombic crystal, formula (7) shows the interplanar spacing of a monoclinic crystal, and formula (8) shows the interplanar spacing of a triclinic crystal.

[0031] All that is needed for the calculation are the a, b, c and α, β, and γ of the unit cell. hkl are the plane indices. The calculated d value or its reciprocal 1 / d is compared with the measured value to assign an index. The position and intensity of the diffraction spot corresponding to each index can be obtained by calculating the crystal structure factor. The crystal structure factor F can be calculated using the following formula (9). T in this formula is the temperature factor of element n in the crystal, fn is the atomic scattering factor corresponding to the scattering angle, and xn, yn, and zn are the Wyckoff positions representing the positions of the elements in the unit cell. The temperature factor and atomic scattering factors use the reported values. The Wyckoff positions can be calculated from the basic atomic positions and space group contained in the crystal information.

[0032]

number

[0033] In addition, the user-specified zone axis vector Z(u,v,w) or the incident vector representing the incident direction and the diffraction vector contained in the zeroth-order Laue band that appears around the direct line are orthogonal to each other, so the dot product of the two vectors is zero. Furthermore, the face angle between two crystal lattice planes can also be calculated and compared with the measured value. These relationships are also used to determine candidate observable diffraction spots, and the measured diffraction spots are selected based on the d value, and the hkl of those spots is determined. For example, in a cubic system, the angle φ (face angle) between crystal faces can be expressed by the following formula (10).

[0034]

number

[0035] The condition indicating the orthogonal relationship between the incident vector and the diffraction vector is expressed by the following formula (11).

[0036]

number

[0037] (Detection of diffraction spots) It is necessary to efficiently detect the diffraction spots obtained from the measured diffraction pattern at the zone axis incidence of a specific crystal orientation. The diffraction spots at the zone axis incidence have the following characteristics. (i) The detected spots are not isolated but show a net pattern. (ii) The diffraction spots contained in the zeroth-order Laue band to be detected are distributed on the low-angle side from the direct spot.

[0038] In particular, to take advantage of property (ii), correction processes are performed such as multiplying the direct spot of the diffraction pattern obtained so as to ignore distant spots by a function such as a Gaussian distribution, which has a high value in the center and decreases in intensity toward the periphery. These correction processes have the effect of reducing the number of cases where distant spots contain large hkl indices, which increases the possibility of 1 / d combinations and increases the number of hkl index candidates. This method allows us to search only for diffraction spots (including the diffraction disk) near the direct spot (including the direct disk).

[0039] In addition, the increased possibility of a 1 / d combination means that, specifically, if the measured distance from zero is 1 / d, this 1 / d value is basically the vector sum of multiples of 1 / a, 1 / b, and 1 / c, which are the reciprocals of the unit cell sizes a, b, and c. Therefore, when a 1 / d with a large exponent is measured, if a and b are different values, various combinations are possible, such as the measured value being 10*1 / a or 9*1 / b. This means that the number of combinations of hkl values ​​that can be fitted increases, and the number of possible crystal orientation candidates also increases, making it difficult to estimate the crystal orientation.

[0040] 4 shows an example of the process of detecting diffraction spots (including disk-shaped diffraction disks) from a diffraction pattern obtained by the STEM method. The figure shows how an image with width W and height H is divided into boxes of a predetermined size S, and a sliding window 46 is slid across them. When detecting diffraction spots, a filter process is performed to remove noise from the image and make it easier to extract the peaks of the diffraction spots. This filtering process uses a convolution or autocorrelation operation. This filtering process can simultaneously remove noise and enhance diffraction spots. Next, the image is divided into small blocks, and the peak position, which is the local maximum value, is found in each block. Diffraction spots are detected from each block using the sliding window 46 method. Sliding allows overlapping areas to detect diffraction spots in detail. From the detected diffraction spots, diffraction spots that are close to each other or have low intensity are removed.

[0041] The images in Figures 5 and 6 show the results of filtering the diffraction patterns: Figure 5 shows the results of filtering the electron diffraction pattern containing spot-like diffraction spots, and Figure 6 shows the results of filtering the electron diffraction pattern containing disk-shaped diffraction spots.

[0042] In Figures 5 and 6, (a) to (c) are images used to explain the convolution operation, and (d) to (f) are images used to explain the autocorrelation operation. Also, (a) and (d) are the input image, (b) and (e) are the filtered images, and (c) and (f) are the aperture images. The size of the aperture image is determined in advance from the size of the direct spot. The aperture image is artificially generated from parameters such as diameter. The convolution operation is calculated using the input image and aperture image. In the case of autocorrelation operation, the operation is performed using only the input image.

[0043] 5 and 6, (g) is an image showing the profile width 48 of the input image, and (h) is an image showing the profile width 48 of the image after filtering by autocorrelation calculation. (i) is a diagram showing the improvement in detection accuracy due to noise removal after filtering, where the horizontal axis is position and the vertical axis is intensity, and shows the input image profile 49 and the autocorrelation calculation profile 50 after filtering. It can be seen from FIGS. 5 and 6 that the autocorrelation calculation profile 50 has improved detection position accuracy compared to the input image profile 49. In particular, in Figure 6(i), which shows the case of a disk, the input image profile 49 has a depression in the center, whereas the filtered autocorrelation calculation profile 50 has noise removed and peaks at the center of the disk, making it easier to detect the spot position.

[0044] 7 is a flowchart showing an example of the processing performed by the computer 4 from the time when the diffraction spot is detected to the time when the estimated crystal orientation is selected. These processing steps are performed by the estimated crystal orientation selection unit 22. In Figure 7, in step SP101, the direct spot with the maximum intensity is detected from the detected diffraction spots. In step SP102, the detected direct spot and any two diffraction spots are selected in sequence, and the angle formed by the two distances between the direct spot and the two diffraction spots is calculated. In step SP103, crystal orientation candidates are searched for and saved based on the angle formed by the two calculated distances and crystal information. Then, in step SP104, an estimated crystal orientation is selected and determined from the saved crystal orientation candidates based on the regularity of the diffraction spots.

[0045] (Detection of direct spots (including direct discs)) The processing related to the detection of direct spots performed by the direct spot detection unit 24 and / or the estimated crystal orientation selection unit 22 will be described. When searching for the direct spot (direct disc), the location of the highest intensity is taken as the direct spot position as a first approximation. To increase accuracy, the location of the highest convolution with the diffraction spot or disc shape on the diffraction pattern estimated by the user is taken as the direct spot position. In addition, as described above, the pattern based on the spot (disc) shape function and the direct spot obtained experimentally are fitted, the coefficient of the function with the smallest error is estimated, and the location of the center of that function is taken as the direct spot (disc) position.

[0046] In order to improve efficiency, a method for searching for direct spots involves dividing the diffraction pattern into sections, calculating the integrated intensity for each section, and then searching for the position of the direct spot after reducing the search range before examining the entire position of the diffraction pattern. For example, a method is used in which the collected diffraction pattern is divided into multiple regions, the integrated intensity for each region is compared, and direct spots are detected in a predetermined range of high integrated intensity. The predetermined range of high intensity can be determined in various ways, such as arranging the divided regions in descending order of intensity and limiting it to a certain number from the top, or searching for a range above a predetermined intensity threshold. In this case, to prevent the large direct spot position from being divided into multiple sections and falling outside the search range, the integrated intensities can be compared in re-divided regions shifted by a distance smaller than the length of one section. In this case, the integrated intensities can be calculated in sections shifted by an integer fraction (e.g., 1 / 2) of the divided range.

[0047] (Angle and distance calculation process) Fig. 8 is a diagram for explaining the calculation process of angles and distances. For example, distance-angle calculation unit 25 selects a detected direct spot and any two detected diffraction spots from a diffraction pattern such as that shown in Fig. 8(b), and calculates the distances and angles from the direct spot to the any two diffraction spots. To explain this with reference to Figure 8(a), any two points around the direct spot 36 are selected, and the first line segment 37 between the direct spot 36 and one diffraction point 38 and its distance d1, the second line segment 39 between the direct spot 36 and another diffraction point 40 and its distance d2, and the angle φ between the first line segment 37 and the second line segment 39 with the direct spot 36 as the vertex are calculated.

[0048] (Save crystal orientation candidates) The crystal orientation candidate search unit 26 searches for a matching simulated diffraction spot in a database for any crystal orientation based on the two calculated distances and angles, and if such a diffraction spot is found, saves the crystal orientation at that time as a candidate. This process is performed for all detected diffraction spots that are the search target, and the candidates are saved.

[0049] (Selection process of crystal orientation candidates based on regularity) Next, the stored candidates for crystal orientation are selected based on the crystal orientation specified by the user. Patterns in which the distances and angles between diffraction spots match may potentially match a large number of patterns in the database. The candidates for crystal orientation are checked using the regularity of the crystal, in this case symmetry. Crystals have the property of satisfying specific symmetry operations (point symmetry such as rotation, inversion, and reflection, and translational symmetry, etc.). For example, referring to Figure 8(a), it is checked whether there is a pair of diffraction spots on a line connecting another detected diffraction spot 40 and the direct spot 36, at an angle of 180 degrees and equidistant from each other. Furthermore, if these two spots appear, the cross product direction of the vectors of these two diffraction points can be identified as a candidate for the crystal orientation.

[0050] To explain this in Figure 9(a), the black circle in the center is the direct spot, the white circle is the location where reflection is expected, the black star is a diffraction spot where a pair was detected, and the white star is a diffraction spot where no pair was detected. In this case, diffraction spots where pairs were detected at equal distances are selected by symmetry, as in [P1,-P1], [P2,-P2], [P3,-P3], ..., [Pn,-Pn]. Furthermore, in addition to the two selected diffraction spots, other diffraction spots are also checked to see if there are diffraction spots on the opposite side. For example, the crystal orientation candidate selection unit 27 checks the number of such pairs and removes them from the crystal orientation candidates if the number is less than a threshold value. In this way, orientations estimated by improperly detected diffraction spots are removed, and the crystal orientations are selected. In this case, if it is determined that the diffraction spots satisfying the regularity are within a predetermined error range, it can be said that the candidate crystal orientation is selected from the candidates. One example of a condition for determining whether or not a given crystal orientation is within the predetermined error range is the number of pairs Pn detected at the location where reflection is expected.

[0051] In Figure 9(b), two large white circles indicate paired diffraction spots on a line connecting a detected direct spot with another direct spot, and at an angle of 180 degrees, at equidistant positions. The number of such pairs is then checked using a similar process, and diffraction spots with a large number are surrounded by thin squares. This method searches based on angles, distances, and crystal symmetry, so the amount of calculation is small and the crystal orientation can be searched for quickly.

[0052] Figures 10(a) and 10(b) show the results of applying the search system algorithm described in Figures 7 to 9 to actual data when silicon is used as the sample. Silicon has a cubic crystal system, and a, b, and c are all 5.43, and α, β, and γ are all 90°. In Figure 10(a), the diffraction spots detected from the diffraction pattern are indicated by circles 58, and autocorrelation calculations were used for the filter calculations. It can be seen that the diffraction spots can be successfully detected by using autocorrelation calculations. In the image of Figure 10(b), the diffraction spot predicted from the estimated crystal orientation is shown by the small circle 59. The circle 58 and the small circle 59 are in the same place, confirming that the estimation was correct. The estimated crystal orientation in this case was [1,1,1].

[0053] The crystal orientations that this embodiment primarily seeks to select are not orientations that include higher-order indices, but rather crystal orientations necessary for obtaining high-resolution images. These crystal orientations generally have indices of at most ±3 (between -3 and +3). The reason for this is that the orientations for obtaining high-resolution images are such that atoms of the same element type or atoms at equivalent positions in the crystal are visible overlapping. These conditions are advantageous for the following observations and analyses. (i) The positions of local atomic rows, such as the interface with other particles (grain boundaries), can be clearly observed. (ii) When only the same element is present on an atomic column, the element species in each column can be observed separately during useful atomic column analysis. This means that the 2D projection position of each atomic species can be determined, thereby obtaining important structural information about the sample. Furthermore, in a crystal orientation where multiple atoms are projected at the same position, for example, in a mixed crystal where oxides of element A and element B have the same structure, the abundance ratio of these elements at each atomic site can be directly observed.

[0054] The case where the diffraction pattern is disk-shaped will be further described below. As mentioned above, the diffraction spots in the diffraction pattern obtained in STEM mode are often disk-shaped, rather than the spot-like shapes obtained in normal diffraction mode. In normal STEM image acquisition mode, the diameter of these diffraction disks is large and they overlap with adjacent diffraction disks, but the diffraction pattern measured here should preferably be one in which the diffraction disks do not overlap. FIG. 11 shows an example of a diffraction pattern. In FIG. 11, the disk area 44 is an area with an error of n%, the diffraction pattern 43 indicated by the black circle is a diffraction pattern of the u1v1w1 crystal orientation, and the diffraction pattern 42 indicated by the gray oval is a diffraction pattern of the u2v2w2 crystal orientation. If the diffraction spots included in the diffraction pattern 43 are defined with the same error, there will be more distant spots, and the number of 1 / d value options will increase, making it difficult to determine which crystal orientation they belong to. Therefore, this confusion can be avoided by intentionally reducing the intensity of spots far from the direct spot. In the present invention, this attenuation rate is also added as a condition for orientation determination.

[0055] One method for identifying the position of the diffraction disk is to use the cross-correlation function with the diffraction pattern of the direct diffraction disk alone, and determine the point where the peak is located as the position of the diffraction disk. However, with this method, if there is intensity non-uniformity within the measured diffraction disk, the correlation function will be pulled to areas with high intensity, which will result in poor positioning accuracy. Therefore, as described above, one feature of this embodiment is that the autocorrelation function of the diffraction pattern can be selected as a detection target as a preprocessing step for detecting the diffraction pattern. The autocorrelation function is literally a function that superimposes a pattern and its copy pattern and calculates the sum of the products of the two while shifting the copy pattern. Therefore, if the diffraction pattern has a net-like, periodic repeating pattern, it is easy to imagine that each diffraction point will have a peak. In this case, even if there is a bias in intensity within each diffraction disk, the autocorrelation function is the sum of the entire products and is therefore insensitive to this bias, resulting in improved accuracy in determining the position of the diffraction point. Then, the crystal orientation is determined from the indices of the multiple diffraction vectors selected as described above based on the hkl of the diffraction points detected by the above method, and if specified by the user, is mapped onto the STEM image.

[0056] The advantages of screening using this first embodiment will be described below. (A) For polycrystalline or powdered samples containing multiple grains oriented in different directions, a STEM image of the sample position and a diffraction pattern corresponding to each pixel are acquired simultaneously or at intervals, and the above-mentioned selection process is performed. Since the pixels in the field of view correspond to the diffraction pattern, the position of the grain oriented in the desired crystal orientation can be found efficiently. (B) When an estimated crystal orientation that matches the crystal orientation set by the user within a specified error range is selected, the system has a display control unit that displays information related to the estimated crystal orientation in the STEM image and diffraction pattern, allowing for intuitive and easy-to-understand screening. In this case, image display control elements such as color, pattern, shape, brightness, and blinking can be changed for each orientation to make the display easier to understand. (C) A unique method is used to identify the direct beam position in the obtained diffraction pattern, allowing for high accuracy. (C1) Furthermore, since the above specific conditions can be relaxed appropriately, it is possible to extract regions that are close to the specific conditions, and it is also possible to extract regions that are close to the conditions even in samples with poor crystallinity. (D) By filtering the diffraction pattern with characteristics before performing an analysis to identify the crystal orientation from the obtained diffraction pattern, it is possible to accurately detect detection targets such as diffraction spots. (E) Normally, STEM images and diffraction patterns are acquired using devices such as an electron diffraction detector that is equipped in the STEM, and the method can be implemented simply by running a specified program on a computer linked to the STEM. This method does not require any additional hardware devices and can be implemented at low cost. (F) When a search is performed based on angles, distances, and crystal symmetry, the amount of calculation can be reduced, enabling high-speed screening.

[0057] <Options> The options in the first embodiment will be described below. As an operational option, it is also possible to calculate the tilt angle from the crystal orientation obtained from the diffraction pattern to the orientation specified by the user from tilt axis information of a pre-set sample tilt device (goniometer), and automatically tilt the sample to match the orientation specified by the user. Similar operations are partially described in Patent Document 1. Furthermore, since tilting the specimen inevitably involves a shift in the field of view, it is possible to calculate the amount of specimen movement from the correlation function with a previously acquired STEM image and feed this information back to the specimen movement device to ensure that the field of view is not missed. In addition, in order to effectively utilize the obtained orientation information as sample information, it is also possible to have a means for outputting map information so that it can be merged with other information such as element maps. In this way, information such as grain orientation can also be compiled as sample information.

[0058] In the basic form of this first embodiment, the sample is not tilted (X tilt: 0, Y tilt: 0), but by tilting the sample and obtaining a similar map, it is possible to increase the probability of finding particles that match the orientation specified by the user. In the simplest implementation, after completing a map at a tilt angle of 0 degrees, tilt the sample to an X tilt angle of 1 degree and obtain a map in the same way. By sequentially changing the tilt angle and obtaining maps, the map information is two-dimensional and the tilt information is also two-dimensional, so particles with a specific crystal orientation and tilt conditions can be obtained from a total of four-dimensional data cube. For example, by adjusting the sample position to X1 and Y1 with an X tilt of A degrees and a Y tilt of B degrees, it is possible to search for particles with the desired crystal orientation in the four-dimensional data cube. These operations can actually be performed even when the user is not present, and can be performed unattended at night based on a preset search program, resulting in significant labor savings.

[0059] In addition, to obtain a diffraction pattern, the electron beam is moved in STEM mode to obtain a diffraction pattern of a local area. However, it is also possible to use a small selected area aperture in TEM mode and, instead of moving the beam, move the transmitted beam over the selected area aperture to change the location where diffraction is obtained. This method is particularly effective for obtaining high-resolution TEM images, as it does not require a significant change in the lens setting current value compared to TEM mode. Even in this case, if a low-magnification image is taken in TEM mode beforehand, it is possible to map the desired orientation location on that image.

[0060] [Second embodiment] A second embodiment of the present invention will now be described. The operation of the second embodiment is similar to that of the first embodiment in terms of obtaining a diffraction pattern and searching for the center, but the method of searching for the crystal orientation is different. While the mathematical relationship between the diffraction vector and the incident axis vector was used in the first embodiment, the diffraction pattern is simulated from crystal information, as in precession diffraction (PED), and then a search is performed based on this reference diffraction pattern. The search function is a cross-correlation function.

[0061] Although a block diagram is omitted for this embodiment, in one form, the reference information includes a reference diffraction pattern obtained by simulation or experiment at a predetermined camera length, a predetermined orientation, and an incident angle of the electron beam to the sample, and the estimated crystal orientation selection unit 22 shown in Figure 1 includes a conversion unit (not shown, as in Figure 1) that converts the collected diffraction pattern into a rotation angle vs. distance axis or a rotation angle vs. logarithmic distance axis to generate a converted diffraction pattern, and a coincidence calculation unit (not shown, as in Figure 1) that calculates the degree of coincidence from the reference diffraction pattern and the converted diffraction pattern using a cross-correlation function, and the estimated crystal orientation selection unit 22 selects an estimated crystal orientation based on the degree of coincidence. The predetermined orientation in the reference information includes the orientation of a crystal axis, the orientation of a crystal plane, and the like.

[0062] FIG. 12 is a diagram for explaining a method for simulating a diffraction pattern from crystal information. Figures 12(a), (b), and (c) show the rotation angles of tens to hundreds of diffraction patterns P1, P2, and P3, respectively, as used in PED. Figures 12(d) and (e) plot the intensities of diffraction patterns P1 and P2, with the rotation angle of diffraction patterns P1 and P2 on the horizontal axis and distance on the vertical axis. Figure 12(g) shows the correlation function between P1 and P2 and the rotation angle. θ in Figure 12(g) indicates the rotation angle. First, the diffraction pattern is transformed into angle and distance space before the search is performed. To transform into angle / distance space, a radial line is drawn outward from the center position found by the search, and the intensity is plotted on the vertical and horizontal axes, with the angle of the normal and the distance between the pixel of the diffraction pattern and the center pixel.

[0063] In the diffraction pattern obtained in this way, even if the pattern is rotated, the entire pattern simply shifts in the direction of the angular axis, and therefore the cross-correlation function simply produces peaks at different positions, the intensity of which represents the degree of pattern coincidence. The patterns obtained in simulations or experiments depend on the rotation angle of the sample, and since the rotation angle is indefinite, searching with a pattern converted to angle / distance eliminates the need to calculate cross-correlation functions with many rotated patterns.

[0064] (Logarithmic transformation of distance axis) Figure 12(f) plots the intensity of diffraction pattern P3, with the horizontal axis representing the rotation angle of diffraction pattern P3 and the vertical axis representing logarithmic distance. Figure 12(h) plots the correlation function between P1 and P3 and the rotation angle θ. X in Figure 12(h) represents the expansion / contraction ratio due to camera length fluctuations. In actual experiments, it is desirable to eliminate the instability of the camera length of the diffraction pattern, which is an unspecified parameter and corresponds to the magnification of the image. Since the ratio of the diffraction spot's position from the center remains constant even if the camera length of the diffraction pattern changes, the logarithmic converted pattern will show the same pattern even if the camera length is different. Since it is the position that differs, by comparing the cross-correlation function after angle / logarithmic distance axis conversion, it is possible to measure the degree of pattern agreement from the peak intensity of the cross-correlation function, even if the camera length is different. As a by-product, the relative camera length can be measured from the deviation of the logarithmic distance axis. This relationship is Log(ka)=Log(k)+Log(a) This can be understood by recalling the formula:

[0065] This method is effective if you can specify a central pattern or the same point on both images. Once the transformation is complete, the rotation angle and magnification can be measured using the cross-correlation function. Furthermore, even if the magnification is different, the degree of match between the images or patterns can be immediately determined. Alternatively, after measuring the rotation angle and magnification, you can rotate and scale the images to the measured or reference values, and then compare the two images using the normal cross-correlation function on a pixel-by-pixel basis. The processing after searching for a matching pattern is performed in the same manner as in the first embodiment.

[0066] The crystal orientation searching systems according to the first and second embodiments can also be realized as a crystal orientation searching program that causes a computer linked to (including attached to) a scanning transmission electron microscope to execute predetermined processing. The crystal orientation searching system can also be understood as a crystal orientation searching method performed by a computer and a scanning transmission electron microscope. The functional units described in the crystal orientation searching system and the processing performed by each functional unit can be similarly realized as processing in a crystal orientation searching program and steps in a crystal orientation searching method, respectively. [Example]

[0067] A characteristic search example of the present invention will be described below. It should be noted that the present invention is not limited to the following examples. The results of applying the method of the first embodiment to polystyrene (isotactic polystyrene), which is a type of crystalline polymer material and has high crystallinity, will be described. Crystalline polymers are used in a variety of situations, and observation of their crystal structure is essential for the development of polymer materials. Polymer crystals form complex higher-order structures that are a mixture of crystalline and amorphous phases, so analyzing the spatial distribution of crystals is considered important.

[0068] The crystalline information for the polystyrene used this time was trigonal, with the aforementioned a being 2.19, b being 2.19, c being 0.665, α being 90°, β being 90°, and γ being 120°. The imaging conditions for acquiring the data were a JEM-F200 scanning transmission electron microscope manufactured by JEOL Ltd., with an accelerating voltage of 200 kV, a convergence angle of 0.65 mrad, and a Gatan K2 IS detector. The sample temperature was -177°C, and the exposure time was 0.013 seconds.

[0069] Figure 13 is an image showing the extracted area of ​​polystyrene. The photographed area was a 2.1 μm square area. With 300 pixels x 300 pixels, pixel information for a total of 90,000 points was obtained. Figure 14 shows an image in which pixels belonging to a diffraction pattern having a specific diffraction spot are extracted from a four-dimensional data cube (recording two-dimensional diffraction patterns corresponding to each two-dimensional pixel) acquired simultaneously with the image shown in Figure 13, and the sample position 52 associated with the estimated crystal orientation selected at the position on this image (pixel information) is highlighted for easy understanding by the user. Note that the dashed line 51 in Figure 14 is a leader line serving as display correspondence information showing the correspondence between the sample position 52 and Figures 15, 16, etc.

[0070] FIG. 15 is a diagram showing diffraction spots detected from an image of an actual diffraction pattern, and large circles 55 are diffraction spots detected from the image of the actual diffraction pattern. Figure 16 shows an image in which the large circle 55 and the small circle 56 indicating the position of the diffraction spot predicted from the estimated crystal orientation are superimposed on an image of the actual diffraction pattern. Referring to Figure 16, it can be seen that the large circle 55 and the small circle 56 are in almost the same place, so the crystal orientation was correctly estimated. The estimated crystal orientation in this case was [-1,1,1].

[0071] Although the present invention has been described above by way of example of an embodiment, the technical scope of the present invention is not limited to the configurations described in the above embodiment. The technical scope of the present invention should be determined based on the claims, and it goes without saying that various modifications, additions to the configuration, and improvements can be made within that scope. [Explanation of symbols]

[0072] 1: Crystal orientation search system 2: Scanning transmission electron microscope 20: Reference information storage section 21: Diffraction pattern collection control unit 22: Estimated crystal orientation selection section 23: Display control unit 24: Direct spot detection unit 25: Distance and angle calculation unit 26:Crystal orientation candidate search section 27: Crystal orientation candidate selection section

Claims

1. a scanning transmission electron microscope capable of scanning an electron beam on a sample and acquiring a STEM image and a diffraction pattern in STEM mode; a reference information storage unit that stores reference information related to crystals in advance; a diffraction pattern collection control unit that, in response to a designation of a sample position assigned to a pixel on the STEM image, causes the electron beam to dwell for a predetermined time, and collects and stores the diffraction pattern; an estimated crystal orientation selection unit that selects an estimated crystal orientation at the sample position based on a designated crystal orientation, the reference information, and the diffraction pattern; A crystal orientation search system using a computer having the above.

2. the estimated crystal orientation selection unit includes a direct spot detection unit that detects a direct spot having the maximum intensity from among the diffraction spots of the diffraction pattern; a distance / angle calculation unit that selects the detected direct spot and any two of the diffraction spots and calculates an angle formed by two distances to the diffraction spots from the selected two points; a crystal orientation candidate search unit that searches for a crystal orientation candidate based on the two distances, the formed angle, and the reference information; a crystal orientation candidate selection unit that selects the crystal orientation candidate to be an estimated crystal orientation based on the regularity of the diffraction spots; 2. The crystal orientation finding system according to claim 1, comprising:

3. the reference information comprises a reference diffraction pattern obtained by simulation or experiment at a predetermined camera length, a predetermined orientation, and an incident angle of the electron beam to the sample; a conversion unit in which the estimated crystal orientation selection unit converts the diffraction pattern into a rotation angle and distance axis or the rotation angle and logarithmic distance axis to generate a converted diffraction pattern; a coincidence calculation unit that calculates a coincidence between the reference diffraction pattern and the transformed diffraction pattern using a cross-correlation function; The crystal orientation searching system according to claim 1 , wherein the estimated crystal orientation is selected based on the degree of coincidence.

4. 2. The crystal orientation searching system according to claim 1, wherein a convolution operation or an autocorrelation operation is performed on the diffraction pattern to be used for selecting the estimated crystal orientation.

5. the estimated crystal orientation selection unit has a direct spot detection unit that detects a direct spot with maximum intensity from among the diffraction spots of the diffraction pattern, 2. The crystal orientation searching system according to claim 1, wherein the direct spot detection unit divides the diffraction pattern into a plurality of regions, compares the integrated intensities of the divided regions, and detects the direct spot in a predetermined range where the integrated intensities are high.

6. the estimated crystal orientation selection unit has a direct spot detection unit that detects a direct spot with maximum intensity from among the diffraction spots of the diffraction pattern, 2. The crystal orientation searching system according to claim 1, wherein the estimated crystal orientation selection unit multiplies the diffraction pattern by a function whose value decreases as the distance from the direct spot increases as the diffraction pattern pre-processes the diffraction pattern.

7. 6. The crystal orientation searching system according to claim 5, further comprising a process for comparing the integrated intensities in re-divided regions that are shifted by a distance smaller than the length of one of the divisions.

8. 2. The crystal orientation searching system according to claim 1, wherein the reference information includes positions of diffraction spots corresponding to indices of crystal orientations, intensities of the diffraction spots, interplanar spacings corresponding to distances of the diffraction spots, and angles between the diffraction spots.

9. 2. The crystal orientation searching system according to claim 1, further comprising a display control unit that displays a predicted diffraction pattern based on the estimated crystal orientation superimposed on the diffraction pattern.

10. 2. The crystal orientation searching system according to claim 1, further comprising a display control unit that displays the estimated crystal orientation and the sample position based on the estimated crystal orientation on the STEM image.

11. A method for searching for a crystal orientation using a scanning transmission electron microscope that can scan an electron beam on a sample and can acquire a STEM image and a diffraction pattern in a STEM mode, comprising: an acquisition and storage step of dwelling the electron beam for a predetermined time in response to the designation of a sample position assigned to a pixel on the STEM image, and acquiring and storing the diffraction pattern; an estimated crystal orientation selection step of selecting an estimated crystal orientation at the sample position based on the designated crystal orientation, reference information previously stored about the crystal, and the diffraction pattern; A computer-implemented method for searching for a crystal orientation, comprising:

12. A crystal orientation search program for searching for a crystal orientation using a scanning transmission electron microscope that can scan an electron beam on a sample and acquire a STEM image and a diffraction pattern in a STEM mode, an acquisition and storage process for dwelling the electron beam for a predetermined time in response to the designation of a sample position assigned to a pixel on the STEM image, and acquiring and storing the diffraction pattern; an estimated crystal orientation selection process for selecting an estimated crystal orientation at the sample position based on a designated crystal orientation, pre-stored reference information about the crystal, and the diffraction pattern; A crystal orientation search program that causes a computer linked to the scanning transmission electron microscope to execute the above.

Images