Method for analyzing quasicrystal

Abstract

To provide a method for analyzing quasicrystals capable of specifying an orientation of quasicrystals in addition to easily identifying which portions of a TEM observation surface are quasicrystals.SOLUTION: A FIB device is used to prepare a sample for TEM observation from a base material obtained through a quasicrystal production process (for example, a process in which a crystalline material is rapidly cooled from a high temperature). The sample is observed using a combination of bright-field and dark-field TEM observation to identify quasicrystals based on differences in the contrast of observation images. An orientation of the quasicrystals is also identified by the observation images obtained using LAADF-STEM and HAADF-STEM methods.SELECTED DRAWING: Figure 5

Description

[Technical field]

[0001] The present invention relates to a method for analyzing quasicrystals in a TEM observation sample cut out from a base material using an FIB device. [Background technology]

[0002] The electron diffraction pattern of quasicrystals was discovered by Shechtman et al. in Israel, and the arrangement had 10-fold rotational symmetry, which is impossible in crystals. By examining the diffraction patterns at various incident angles, it became clear that the quasicrystal had a three-dimensional structure of a regular icosahedron with a 5-fold axis of rotational symmetry. This became clear and was published in 1984 (see Non-Patent Document 1).

[0003] Quasicrystals have a unique atomic structure that does not have translational symmetry, unlike crystals, and many researchers are conducting basic and applied research on them. In order to conduct such research, it is necessary to distinguish, or identify, quasicrystals in the parent material obtained through the quasicrystal production process.

[0004] Conventionally, a method has been used to identify quasicrystals by cutting out a tiny thin film sample from a base material using a focused ion beam device (hereinafter referred to as a FIB device), randomly observing the observation surface with a transmission electron microscope (hereinafter referred to as a TEM), and examining the electron diffraction image. In this case, as shown in Figure 1, if a quasicrystal is found at least in one place on the observation surface, an electron diffraction image characteristic of the quasicrystal can be obtained by changing the observation angle with a two-axis sample holder (see Non-Patent Documents 1 and 4). [Prior art documents] [Non-patent literature]

[0005] [Non-Patent Document 1] Takeuchi, Edakawa, Crystals, Quasicrystals, Amorphous (2008) Uchida, Rokakuho. [Non-Patent Document 2] Hiraga, Mysterious Structure of Quasicrystals (2003), Agne Technology Center. [Non-Patent Document 3] JEOL, JEM-2100 Instruction Manual (2014). [Non-Patent Document 4] Y. Kaneko, Y. Arichika, T. Ishimasa, Phil. Mag. Lett., 81 (2001) 777. Summary of the Invention [Problem to be solved by the invention]

[0006] However, quasicrystals are quite different from single-crystal materials such as silicon, and even when performing TEM observation of thin-film samples, it was difficult to determine where the quasicrystals were located and to distinguish which ones were quasicrystals from the numerous crystal grains that appeared within the observation surface.

[0007] Therefore, the present invention aims to provide a method for easily identifying which parts of a TEM observation surface are quasicrystals, and further to provide a method for identifying the orientation of such quasicrystals, since the observation surface is likely to contain a large number of quasicrystals with different orientations. [Means for solving the problem]

[0008] In view of the above-mentioned problems, the method for analyzing quasicrystals of the present invention is characterized in that a sample for TEM observation is prepared from a base material using an FIB device, and the sample is observed using a combination of bright-field observation and dark-field observation of the TEM, thereby identifying the quasicrystals and their orientation based on the difference in contrast of the observed image. Effect of the Invention

[0009] According to the present invention, it is possible to easily identify which parts of the TEM observation surface are quasicrystals. Also, the orientation of the quasicrystals can be identified by the images observed by the LAADF-STEM method or the HAADF-STEM method. [Brief description of the drawings]

[0010] [Figure 1]FIG. 1 shows the observation directions of a regular icosahedron (five-fold, two-fold, and three-fold rotational symmetry). [Diagram 2] Electron diffraction pattern showing five-fold rotational symmetry [Diagram 3] Electron diffraction pattern showing two-fold rotational symmetry [Figure 4] Electron diffraction pattern showing three-fold rotational symmetry [Diagram 5] This is a bright-field TEM image. [Figure 6] STEM image (camera length 12 cm) [Figure 7] HAADF-2 image (acceptance angle 35~90mrad) [Figure 8] FIG. 1 is a schematic diagram of dark-field image detection. [Figure 9] TEM bright field image of aluminum electrode [Figure 10] STEM image of an aluminum electrode (camera length 12 cm) [Figure 11] HAADF-5 image of an aluminum electrode (take-off angle 80-220 mrad) [Figure 12] Electron diffraction pattern of an aluminum electrode (white area A in HAADF) [Figure 13] Electron diffraction pattern of an aluminum electrode (black area B in HAADF) [Figure 14] TEM bright field image [Figure 15] STEM image (camera length 120 cm) [Figure 16] STEM image (camera length 40 cm) [Figure 17] HAADF-1 image (acceptance angle 30~80mrad) [Figure 18] HAADF-3 image (acceptance angle 40~110mrad) [Figure 19] HAADF-4 image (acceptance angle 60~160mrad) [Figure 20] TEM bright field image [Figure 21] Electron diffraction pattern showing three-fold rotational symmetry [Figure 22] TEM bright-field image of a sample processed by FIB (1) [Diagram 23] TEM bright-field image of a sample processed by FIB (2) [Figure 24] TEM bright-field image of a sample processed by FIB (3) [Diagram 25] TEM bright-field image of a sample processed by FIB (4) DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, a method for identifying a quasicrystal according to an embodiment of the present invention will be described with reference to the drawings.

[0012] <Basic methods for identifying quasicrystals> To prove that a quasicrystal is a quasicrystal, for example, in the case of a regular icosahedral quasicrystal, its electron diffraction image must show 5-fold, 2-fold, or 3-fold rotational symmetry (see Figures 2 to 4). First, a sample for TEM observation is prepared using an FIB device from a base material obtained through a quasicrystal preparation process (for example, a process of rapidly cooling a crystalline material from a high temperature). As a result of examining electron diffraction images at multiple points on the TEM observation surface of the sample, it was found that a method for identifying quasicrystals based on differences in the contrast of the observed images is effective. In other words, by examining the parts that appear black (high atomic density) in the TEM bright-field image, quasicrystals could be found with a high probability (see Figures 2 and 5).

[0013] The dark areas in the bright-field image indicate high density, suggesting that quasicrystals are denser than amorphous metals or crystalline alloys of the same composition. Therefore, the way to find quasicrystals was to look for dark areas in the bright-field image.

[0014] Although it was found that there was a high probability that quasicrystals existed in the black areas, the contrast of bright-field images changes with slight deviations in the thickness of the sample or in the focus, making the reliability of identifying quasicrystals low. Therefore, the inventors considered using dark-field images. As one method, they attempted to visualize quasicrystals by selectively using bright-field and dark-field TEM images, but they determined that analysis using dark-field images requires advanced technology and time, and is not suitable for investigating a larger area.

[0015] Therefore, the present inventors have noticed that a TEM equipped with scanning transmission electron microscope (hereinafter referred to as STEM) functions can be used to observe bright and dark fields by changing the camera length, and have found that a method for identifying quasicrystals from dark-field images obtained by the STEM method is useful.

[0016] As mentioned above, quasicrystals were found with a high probability in the black areas of the TEM bright-field image, but in combination with this, by changing the STEM camera length and examining the change in black and white shade (change in contrast), it was found that the white areas in the dark field were highly likely to be quasicrystals, and analysis of the dark-field image (LAADF-STEM method) was effective in identifying quasicrystals. Here, LAADF-STEM is an abbreviation for Low Angle Annular Dark Field Scanning Transmission Electron Microscopy.

[0017] Therefore, we found that identifying high-density areas, i.e. areas that appear black in bright-field images and white in dark-field images, could be a way to find quasicrystals.

[0018] In short, it is effective to identify quasicrystals based on the difference in contrast of the observed images by combining bright-field and dark-field observations with a TEM. However, dark-field observation is easier to perform using an annular dark-field detector (ADF) in a STEM than by shifting the optical axis of the bright-field observation.

[0019] <Method for Identifying Quasicrystals and Specifying Orientations by HAADF-STEM Method> Furthermore, by using the function of HAADF-STEM, more precise image information can be captured than from the changes in the contrast (black-and-white shading) of the dark-field image of LAADF-STEM. Here, the HAADF-STEM method is also abbreviated as the High Angle Annular Dark Field Scanning Transmission Electron Microscopy.

[0020] This HAADF-STEM method is a type of STEM method. It is a method in which an electron beam focused on a micro-region is two-dimensionally scanned on a thin film sample, and the intensity of electrons scattered at high angles is recorded by an annular detector to obtain a two-dimensional intensity distribution image. Since the intensity of electron scattering is proportional to the square of the atomic number Z, the contrast of the HAADF image emphasizes the differences between atoms and is called Z contrast (see Non-Patent Document 1).

[0021] Note that the variable range of the camera length of the STEM used in this case is 2, 8, 12, 20, 40, 80, 120, 200 cm. Also, there are the following five types of HAADF settings. In this case, the larger the capture angle (θ2 - θ1) of scattered electrons, the narrower the detection area (see Fig. 8, Non-Patent Document 3). ·HAADF-1: Capture angle 30 - 80 mrad, detection area, large ·HAADF-2: Capture angle 35 - 90 mrad, (camera length, increasing↑) ·HAADF-3: Capture angle 40 - 110 mrad, detection area, medium ·HAADF-4: Capture angle 60 - 160 mrad, (camera length, decreasing↓) ·HAADF-5: Capture angle 80 - 220 mrad, detection area, small In the HAADF-STEM method, it was expected that the quasicrystal grains would appear white and that the quasicrystals could be identified by the dark-field image of the LAADF-STEM method and the observation image of the HAADF-STEM method. However, after examining many grains, most of them have diffraction images with 5-fold rotational symmetry and 2-fold rotational symmetry (see Figs. 2 and 3).

[0022] It is unnatural that quasicrystals with 2-fold and 5-fold orientations are concentrated within the observation surface, and it was strange that a diffraction image with 3-fold rotational symmetry (Figure 4) was not obtained. Figure 5 is a bright-field image of a TEM sample prepared with an FIB device, and the black parts are thought to be areas of high density. Normally, these appear white in a dark-field image, but we found a part that did not appear white (see Figures 6 and 7). When we examined this part, a diffraction image with 3-fold rotational symmetry appeared (see Figures 20 and 21).

[0023] The areas that appear black in the bright-field image but do not appear white in the dark-field image are quasicrystals with three-fold rotational symmetry. We performed a detailed analysis of the change in contrast in LAADF-STEM (when the camera length is changed) and when the take-off angle in HAADF-STEM is changed.

[0024] Although comparing bright-field images and LAADF-STEM dark-field images was also effective (see Figures 14 to 16), the images observed using the HAADF-STEM method were easier to understand, and as a result of observations under various conditions, subtle differences in contrast could be identified in the images observed using the HAADF-STEM method more easily than in the images observed using the LAADF-STEM method (see Figures 17, 18, and 19). When examining grain agglomerates identified using this method, it was found that three-fold rotational symmetry was easier to find.

[0025] More specifically, the following points have become clear through multiple analytical experiments. (1) If a certain part (grain mass) appears black in a TEM bright-field image, appears white in the LAADF STEM dark-field image, and appears white in the HAADF STEM dark-field image, the density of this part is higher than the surrounding area and represents a 5-fold or 2-fold rotationally symmetric plane (orientation) of a quasicrystal. (2) On the other hand, if a certain part (grain mass) appears black in the TEM bright-field image, in the LAADF STEM dark-field image, and in the HAADF image, it represents a three-fold rotationally symmetric plane (orientation) of the quasicrystal. The reason that this part appears black in the bright-field image is not due to high density, but is probably due to differences in sample thickness or focus deviation. (3) Furthermore, when a certain part (grain mass) appears black in a TEM bright-field image and gray (an intermediate color between black and white) in the LAADF and HAADF STEM dark-field images, this often represents a three-fold rotationally symmetric plane (orientation) of a quasicrystal.

[0026] In other words, this suggests that the whiter something appears in LAADF and HAADF, the denser it is compared to the surroundings (however, black and white is not an absolute difference but a relative difference with the surroundings). (4) At this stage, we do not have the detection capability to distinguish between 5-fold rotationally symmetric surfaces (orientations) and 2-fold rotationally symmetric surfaces (orientations). We needed to prove that we could distinguish the difference in density between 5-fold rotationally symmetric surfaces (orientations), 2-fold rotationally symmetric surfaces (orientations), and 3-fold rotationally symmetric surfaces (orientations), but this was not easy.

[0027] <Verification of the effectiveness of the HAADF-STEM method by observing semiconductor devices> The only difference between 5-fold, 2-fold, and 3-fold rotational symmetry in quasicrystals is the distance between diffraction points (the apparent distance between atoms), and it was presumed that the contrast between the images observed using LAADF-STEM and HAADF-STEM represents a difference in orientation.

[0028] It was possible to qualitatively (or intuitively) imagine from the electron diffraction images that the interatomic distances in planes (orientations) with 5-fold rotational symmetry and planes (orientations) with 2-fold rotational symmetry are shorter than the interatomic distances in planes (orientations) with 3-fold rotational symmetry.

[0029] However, because the atomic structure of quasicrystals that lack translational symmetry is complex, it is difficult to calculate interatomic distances from electron diffraction images alone (see Non-Patent Document 1). Therefore, we considered applying this method to other materials and tried LAADF and HAADF-STEM analysis on the polycrystalline silicon gate and aluminum wiring used in trench-type power MOSFETs.

[0030] Since the trench gate has a large area, it was expected that the contrast could be identified by the HAADF-STEM method. However, the grain clusters of polycrystalline silicon were small, and diffraction images that met the criteria for analysis could not be obtained, so the electron diffraction images could not be compared. However, the comparison of aluminum electrodes could be easily done (see Figs. 9, 10, and 11).

[0031] The difference in the orientation of the grain clusters can be easily identified even in the SIM observation image of the FIB-processed cross-section. However, the difference in contrast on the electrode also clearly appears in the observation images of LAADF-STEM and HAADF-STEM. When comparing the electron diffraction images of the white region (A) and the black region (B), the difference in orientation was obvious (see Figs. 12 and 13). This indirectly proved that the contrast (black and white shading) in LAADF and HAADF represents the density difference (the difference in interatomic distance).

[0032] Even for aluminum with a crystal structure, the difference in contrast can be explained by the difference in the distance between diffraction points. Therefore, it can be said that the difference in orientation is also represented in the case of quasicrystals. In this way, the effectiveness of the LAADF-STEM method and the HAADF-STEM method could be confirmed.

[0033] <FIB Processing of Quasicrystals> Observation examples of bright-field images of multiple TEM observation samples cut out from the same material using an FIB apparatus are shown in Figs. 22 to 25. Although there are some refinements in the FIB processing conditions, significant differences appear in the finished state. This is because the size of the grain clusters varies greatly depending on the sample. What is common to these samples is that the size of the grain clusters becomes smaller as it approaches the sample surface.

[0034] Initially, the electron diffraction images obtained from the FIB-processed samples did not show the patterns specific to quasicrystals. However, it was found that they were affected by FIB processing because crystal-like grain clusters were accidentally found. This is thought to be due to the influence of heat generation during FIB processing, which changes the state of the quasicrystal. The amorphization of crystalline materials by FIB processing has been pointed out before, but sufficient attention is also required for quasicrystals.

[0035] <Summary> (1) By observing using bright-field or dark-field TEM, quasicrystals can be easily identified based on the difference in contrast of the observed images. (2) Dark-field observation is effectively performed using the LAADF-STEM or HAADF-STEM techniques. In particular, the HAADF-STEM technique has the advantage of being able to distinguish subtle differences in contrast. (3) For aluminum with a crystalline structure, the images observed using the LAADF-STEM and HAADF-STEM methods show differences in crystal orientation. Even in the case of quasicrystals, the images observed using the LAADF-STEM and HAADF-STEM methods are thought to show differences in orientation, making them useful for identifying the orientation of quasicrystals. In addition, Z-contrast not only shows differences in atomic number Z, but also shows differences in atomic density of the same atom (differences in crystal orientation).

Claims

1. A method for analyzing quasicrystals, comprising the steps of: preparing a sample for TEM observation from a base material using an FIB device; observing the sample using a combination of bright-field and dark-field TEM observation; and identifying the quasicrystals and specifying their orientation based on differences in contrast in the observed image.

2. 2. The method for analyzing quasicrystals according to claim 1, wherein the dark-field observation is performed by a LAADF-STEM method.

3. 2. The method for analyzing quasicrystals according to claim 1, wherein the dark-field observation is performed by a High Accuracy Atomic Force-STEM (HAADF-STEM) method.

Images