Microscopy, 2014, Vol. 63, No. S1
i4 the microstructural tracking technique [5]. The technique provides information about individual grain orientations and 1-micron-level grain morphologies in 3D together with high-density local strain mapping. The application of this technique to the deformation behavior of a polycrystalline aluminium alloy will be demonstrated in the presentation [6]. The synchrotron-based microtomography has been mainly utilized to light materials due to their good X-ray transmission. In the present talk, the application of the synchrotron-based microtomography to steels will be also introduced. Degradation of contrast and spatial resolution due to forward scattering could be avoided by selecting appropriate experimental conditions in order to obtain superior spatial resolution close to the physical limit even in ferrous materials [7].
References
Downloaded from http://jmicro.oxfordjournals.org/ at The Chinese University of Hong Kong on November 14, 2015
1. Toda H., Maire E., Aoki Y., Kobayashi M.: J. Strain Anal. Eng., 46(2011), 549–561. 2. Kobayashi M., Toda H., Kawai Y., Ohgaki T., Uesugi K., et al. Acta Mater., 56 (2008), 2167–2181. 3. Toda H., Minami K., Koyama K., Ichitani K., et al. Acta Mater., 57 (2009), 4391–4403. 4. Toda H., Yamamoto S., Kobayashi M., Uesugi K., Zhang H.: Acta Mater., 56(2008), 6027–6039. 5. Toda H., Ohkawa Y., Kamiko T., Naganuma T., Uesugi K., et al. Acta Mater., 61(2013), 5535–5548. 6. LeClere D.J., Kamiko T., Mizuseki Y., Suzuki Y., Toda H., et al. Proc. of ICAA13 (2012), 9–14. 7. Seo D., Tomizato F., Toda H., Uesugi K., Takeuchi A., et al. Appl. Phys. Lett., 101 (2012) 261901. doi: 10.1093/jmicro/dfu038
High dimensional data driven statistical mechanics Yoshitaka Adachi and Sunao Sadamatsu Department of Mechanical Engineering, Graduate School of Science and Engineering, Kagoshima University, Korimoto 1-21-24, Kagoshima, 890-0065, Japan In “3D4D materials science”, there are five categories such as (a) Image acquisition, (b) Processing, (c) Analysis, (d) Modelling, and (e) Data sharing. This presentation highlights the core of these categories [1]. Analysis and modelling A three-dimensional (3D) microstructure image contains topological features such as connectivity in addition to metric features. Such more microstructural information seems to be useful for more precise property prediction. There are two ways for microstructure-based property prediction (Fig. 1A). One is 3D image data based modelling such as micromechanics or crystal plasticity finite element method. The other one is a numerical microstructural features driven machine learning approach such as artificial neural network or Bayesian estimation method. It is the key to convert the 3D image data into numerals in order to apply the dataset to property prediction. As a numerical feature of microstructures, grain size, number of density, of particles, connectivity of particles, grain boundary connectivity, stacking degree, clustering etc. should be taken into consideration. These microstructural features are so-called “materials genome”. Among those materials genome, we have to find out dominant factors to determine a focused property. The dominant factorzs are defined as “descriptor(s)” in high dimensional data driven statistical mechanics. Image acquisition It is important for researchers to choice a 3D microscope from various microscopes depending on a length-scale of a focused microstructure. There is a long term request to acquire a 3D microstructure image
Fig. 1. (a) A concept of 3D4D materials science. (b) Fully-automated serial sectioning 3D microscope “Genus_3D”. (c) Materials Genome Archive (JSPS).
Microscopy, 2014, Vol. 63, No. S1
i5
more conveniently. Therefore a fully automated serial sectioning 3D optical microscope “Genus_3D” (Fig. 1B) has been developed and nowadays it is commercially available. A user can get a good contrast image and more than one hundred sections in a day using Genus_3D, namely 3D image capturing is now a day or hour work rather than monthly work. Data sharing It is also another important issue to archive all materials genome data including process and properties for modelling-assisted high-throughput materials design. Japan Society for the Promotion of Science (JSPS) 176 working group has announced to commence “Materials Genome Archive” service (Fig. 1C). Reference
doi: 10.1093/jmicro/dfu086
Recent improvement of a FIB-SEM serial-sectioning method for precise 3D image reconstruction – application of the orthogonally-arranged FIB-SEM
Recent topics and Future prospects We have applied this instrument for wide area of microstructure analysis; Metals and Alloys, Semiconductor devices, Battery electrodes, Minerals, Biomaterials, and so on. In my presentation, I would like to introduce some of our application results and will discuss about future development of the methodology of a FIB-SEM serial sectioning. As the applied research field becomes wider, various requests for the method were arisen. However, most requests can be summarized as follows: observation of larger area, expansion of applicable sample, obtain many kind of information, linkage with other instruments. Acknowledgments The instrument introduced in this work was installed at NIMS by a part of “Low-carbon research network Japan” funded by the MEXT, Japan.
References Toru Hara Advanced Key Technology Division, National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 3050044, Japan Introduction We installed the first “orthogonally-arranged” FIB-SEM in 2011. The most characteristic point of this instrument is that the FIB and SEM columns are perpendicularly mounted; this is specially designed to obtain a serial-sectioning dataset more accurately and precisely with higher contrast and higher spatial resolution compare to other current FIB-SEMs [1]. Since the installation in 2011, we have developed the hardware and methodology of the serial-sectioning based on this orthogonal FIB-SEM. In order to develop this technique, we have widely opened this instrument to every researcher of all fields. In the presentation, I would like to introduce some of application results that are obtained by users of this instrument. The characteristic points of the orthogonal system Figure 1 shows a difference between the standard and the orthogonal FIB-SEM systems: In the standard system, shown in Fig.1(a), optical axes of a FIB and a SEM crosses around 60deg., while in the orthogonal system (Fig.1(b)), they are perpendicular to each other. The standard arrangement (a) is certainly suitable for TEM lamellae preparation etc. because the FIB and the SEM can see the same position simultaneously. However, for a serial-sectioning, it is not to say the best arrangement. One of the reasons is that the sliced plane by the FIB is not perpendicular to the electron beam so that the background contrast is not uniform and observed plane is distorted. On the other hand, in case of the orthogonally-arranged system,(b), these problems are resolved. In addition, spatial resolution can keep high enough even in a low accelerating voltage (e.g. 500V) because a working distance is set very small, 2mm. From these special design, we can obtain
Fig. 1. Schematic illustration described (a) a standard type arrangement, (b) an orthogonal type arrangement.
1. Hara T., Tsuchiya K., Tsuzaki K., Man X., Asahata T., Uemoto A. “Application of orthogonally arranged FIB-SEM for precise microstructure analysis of materials”, Journal of Alloys and Compounds 577S (2013), 717–721. 2. Hara T. “3D Microstructure observation of materials by means of FIB-SEM serial sectioning”: KENBIKYO 49–1, (2014), 53–58. doi: 10.1093/jmicro/dfu077
Nonlinear intensity attenuation with increasing thickness and quantitative TEM tomography of micron-sized materials Jun Yamasaki1,2, * Research Center for Ultra-High Voltage Electron Microscopy, Osaka University, Ibaraki, Osaka 567-0047, Japan, and 2EcoTopia Science Institute, Nagoya University, Nagoya 464-8603, Japan
1
*[email protected] Nowadays three-dimensional (3D) analyses of nanometer-sized and sub-micron-sized objects have been widely achieved by tomography in transmission electron microscopes (TEM). One of the next methodological targets should be quantitative 3D reconstructions in which not only the shape but also the internal density is correctly reproduced. This is, however, generally hindered by the nonlinearity between projection mass-thickness and image intensity. In the case of BF-TEM images, the ideal exponential attenuation with increasing thickness is disturbed by multiple scatterings. The nonlinearity in the tilt series should induce an inaccurate density distribution in the reconstructed volume. In the present study, the nonlinear attenuation effect has been analyzed using amorphous carbon microcoils (CMCs). Their welldefined shapes and compositional homogeneity are quite useful to estimate the mass-thickness. For measurements over a wide range of acceleration voltages (400, 600, 800 and 1000 kV), we used a highvoltage electron microscope (JEOL: JEM-1000KRS). The intensity attenuation with increasing thickness was measured in the BF-TEM images taken with various sizes of objective apertures (OAs). The results have shown that using a smaller OA and a lower voltage enhances the nonlinearity in the intensity attenuation. It is considered that such nonlinear attenuation should induce failures in conversion from intensity to thickness and thus inhibits
Downloaded from http://jmicro.oxfordjournals.org/ at The Chinese University of Hong Kong on November 14, 2015
1. “Basic to application of 3D materials science”, Adachi Y., Koyama T., Uchida Rokakuho Pub., (2014).
the serial-sectioning dataset from rather wide area (∼100um) with high spatial resolution (Max. 2×2×2nm). As this system has many kinds of detectors: SE, ET, Backscatter Electron(Energy-selective), EDS, EBSD, STEM(BF&ADF), with Ar+ ion-gun and a plasma cleaner, many kinds of signals can be obtained simultaneously.
i4 the microstructural tracking technique [5]. The technique provides information about individual grain orientations and 1-micron-level grain morphologies in 3D together with high-density local strain mapping. The application of this technique to the deformation behavior of a polycrystalline aluminium alloy will be demonstrated in the presentation [6]. The synchrotron-based microtomography has been mainly utilized to light materials due to their good X-ray transmission. In the present talk, the application of the synchrotron-based microtomography to steels will be also introduced. Degradation of contrast and spatial resolution due to forward scattering could be avoided by selecting appropriate experimental conditions in order to obtain superior spatial resolution close to the physical limit even in ferrous materials [7].
References
Downloaded from http://jmicro.oxfordjournals.org/ at The Chinese University of Hong Kong on November 14, 2015
1. Toda H., Maire E., Aoki Y., Kobayashi M.: J. Strain Anal. Eng., 46(2011), 549–561. 2. Kobayashi M., Toda H., Kawai Y., Ohgaki T., Uesugi K., et al. Acta Mater., 56 (2008), 2167–2181. 3. Toda H., Minami K., Koyama K., Ichitani K., et al. Acta Mater., 57 (2009), 4391–4403. 4. Toda H., Yamamoto S., Kobayashi M., Uesugi K., Zhang H.: Acta Mater., 56(2008), 6027–6039. 5. Toda H., Ohkawa Y., Kamiko T., Naganuma T., Uesugi K., et al. Acta Mater., 61(2013), 5535–5548. 6. LeClere D.J., Kamiko T., Mizuseki Y., Suzuki Y., Toda H., et al. Proc. of ICAA13 (2012), 9–14. 7. Seo D., Tomizato F., Toda H., Uesugi K., Takeuchi A., et al. Appl. Phys. Lett., 101 (2012) 261901. doi: 10.1093/jmicro/dfu038
High dimensional data driven statistical mechanics Yoshitaka Adachi and Sunao Sadamatsu Department of Mechanical Engineering, Graduate School of Science and Engineering, Kagoshima University, Korimoto 1-21-24, Kagoshima, 890-0065, Japan In “3D4D materials science”, there are five categories such as (a) Image acquisition, (b) Processing, (c) Analysis, (d) Modelling, and (e) Data sharing. This presentation highlights the core of these categories [1]. Analysis and modelling A three-dimensional (3D) microstructure image contains topological features such as connectivity in addition to metric features. Such more microstructural information seems to be useful for more precise property prediction. There are two ways for microstructure-based property prediction (Fig. 1A). One is 3D image data based modelling such as micromechanics or crystal plasticity finite element method. The other one is a numerical microstructural features driven machine learning approach such as artificial neural network or Bayesian estimation method. It is the key to convert the 3D image data into numerals in order to apply the dataset to property prediction. As a numerical feature of microstructures, grain size, number of density, of particles, connectivity of particles, grain boundary connectivity, stacking degree, clustering etc. should be taken into consideration. These microstructural features are so-called “materials genome”. Among those materials genome, we have to find out dominant factors to determine a focused property. The dominant factorzs are defined as “descriptor(s)” in high dimensional data driven statistical mechanics. Image acquisition It is important for researchers to choice a 3D microscope from various microscopes depending on a length-scale of a focused microstructure. There is a long term request to acquire a 3D microstructure image
Fig. 1. (a) A concept of 3D4D materials science. (b) Fully-automated serial sectioning 3D microscope “Genus_3D”. (c) Materials Genome Archive (JSPS).
Microscopy, 2014, Vol. 63, No. S1
i5
more conveniently. Therefore a fully automated serial sectioning 3D optical microscope “Genus_3D” (Fig. 1B) has been developed and nowadays it is commercially available. A user can get a good contrast image and more than one hundred sections in a day using Genus_3D, namely 3D image capturing is now a day or hour work rather than monthly work. Data sharing It is also another important issue to archive all materials genome data including process and properties for modelling-assisted high-throughput materials design. Japan Society for the Promotion of Science (JSPS) 176 working group has announced to commence “Materials Genome Archive” service (Fig. 1C). Reference
doi: 10.1093/jmicro/dfu086
Recent improvement of a FIB-SEM serial-sectioning method for precise 3D image reconstruction – application of the orthogonally-arranged FIB-SEM
Recent topics and Future prospects We have applied this instrument for wide area of microstructure analysis; Metals and Alloys, Semiconductor devices, Battery electrodes, Minerals, Biomaterials, and so on. In my presentation, I would like to introduce some of our application results and will discuss about future development of the methodology of a FIB-SEM serial sectioning. As the applied research field becomes wider, various requests for the method were arisen. However, most requests can be summarized as follows: observation of larger area, expansion of applicable sample, obtain many kind of information, linkage with other instruments. Acknowledgments The instrument introduced in this work was installed at NIMS by a part of “Low-carbon research network Japan” funded by the MEXT, Japan.
References Toru Hara Advanced Key Technology Division, National Institute for Materials Science, 1-1 Namiki, Tsukuba, Ibaraki 3050044, Japan Introduction We installed the first “orthogonally-arranged” FIB-SEM in 2011. The most characteristic point of this instrument is that the FIB and SEM columns are perpendicularly mounted; this is specially designed to obtain a serial-sectioning dataset more accurately and precisely with higher contrast and higher spatial resolution compare to other current FIB-SEMs [1]. Since the installation in 2011, we have developed the hardware and methodology of the serial-sectioning based on this orthogonal FIB-SEM. In order to develop this technique, we have widely opened this instrument to every researcher of all fields. In the presentation, I would like to introduce some of application results that are obtained by users of this instrument. The characteristic points of the orthogonal system Figure 1 shows a difference between the standard and the orthogonal FIB-SEM systems: In the standard system, shown in Fig.1(a), optical axes of a FIB and a SEM crosses around 60deg., while in the orthogonal system (Fig.1(b)), they are perpendicular to each other. The standard arrangement (a) is certainly suitable for TEM lamellae preparation etc. because the FIB and the SEM can see the same position simultaneously. However, for a serial-sectioning, it is not to say the best arrangement. One of the reasons is that the sliced plane by the FIB is not perpendicular to the electron beam so that the background contrast is not uniform and observed plane is distorted. On the other hand, in case of the orthogonally-arranged system,(b), these problems are resolved. In addition, spatial resolution can keep high enough even in a low accelerating voltage (e.g. 500V) because a working distance is set very small, 2mm. From these special design, we can obtain
Fig. 1. Schematic illustration described (a) a standard type arrangement, (b) an orthogonal type arrangement.
1. Hara T., Tsuchiya K., Tsuzaki K., Man X., Asahata T., Uemoto A. “Application of orthogonally arranged FIB-SEM for precise microstructure analysis of materials”, Journal of Alloys and Compounds 577S (2013), 717–721. 2. Hara T. “3D Microstructure observation of materials by means of FIB-SEM serial sectioning”: KENBIKYO 49–1, (2014), 53–58. doi: 10.1093/jmicro/dfu077
Nonlinear intensity attenuation with increasing thickness and quantitative TEM tomography of micron-sized materials Jun Yamasaki1,2, * Research Center for Ultra-High Voltage Electron Microscopy, Osaka University, Ibaraki, Osaka 567-0047, Japan, and 2EcoTopia Science Institute, Nagoya University, Nagoya 464-8603, Japan
1
*[email protected] Nowadays three-dimensional (3D) analyses of nanometer-sized and sub-micron-sized objects have been widely achieved by tomography in transmission electron microscopes (TEM). One of the next methodological targets should be quantitative 3D reconstructions in which not only the shape but also the internal density is correctly reproduced. This is, however, generally hindered by the nonlinearity between projection mass-thickness and image intensity. In the case of BF-TEM images, the ideal exponential attenuation with increasing thickness is disturbed by multiple scatterings. The nonlinearity in the tilt series should induce an inaccurate density distribution in the reconstructed volume. In the present study, the nonlinear attenuation effect has been analyzed using amorphous carbon microcoils (CMCs). Their welldefined shapes and compositional homogeneity are quite useful to estimate the mass-thickness. For measurements over a wide range of acceleration voltages (400, 600, 800 and 1000 kV), we used a highvoltage electron microscope (JEOL: JEM-1000KRS). The intensity attenuation with increasing thickness was measured in the BF-TEM images taken with various sizes of objective apertures (OAs). The results have shown that using a smaller OA and a lower voltage enhances the nonlinearity in the intensity attenuation. It is considered that such nonlinear attenuation should induce failures in conversion from intensity to thickness and thus inhibits
Downloaded from http://jmicro.oxfordjournals.org/ at The Chinese University of Hong Kong on November 14, 2015
1. “Basic to application of 3D materials science”, Adachi Y., Koyama T., Uchida Rokakuho Pub., (2014).
the serial-sectioning dataset from rather wide area (∼100um) with high spatial resolution (Max. 2×2×2nm). As this system has many kinds of detectors: SE, ET, Backscatter Electron(Energy-selective), EDS, EBSD, STEM(BF&ADF), with Ar+ ion-gun and a plasma cleaner, many kinds of signals can be obtained simultaneously.
