Instructions for use: the html version works with all tools of your browser. For searching press Ctrl-F, Back brings you to previous view, etc. Images with high resolution will expand upon a mouse click.

ISBN 978-80-260-6721-4

Copyright© 2014 Czechoslovak Microscopy Society

No parts of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means without the written permission of the publisher, except as stated below. Single photocopies of abstracts may be made for private study or research. Illustrations and short extracts from the text of individual abstracts may be copied provided that the source is acknowledged and permission of the authors has been obtained.

The publishers have made every effort to reproduce the figures at optimum contrast and with optimum resolution from the original material provided by the authors. The publishers accept no liability for the loss of image quality resulting from supply of low quality material or improperly prepared digital records. The content of the individual abstracts remains the responsibility of the respective authors.

The abstracts should not be cited in bibliographies. Material contained herein should be treated as personal communication and should be cited as such only with consent of the author.

 

Table of content

Plenary lectures

Type of presentation: Plenary

IMC-PL-6095 Light Microscopy at the Nanoscale

Cremer C.1,2
1Superresolution Microscopy, Institute of Molecular Biology (IMB), Mainz, Germany, 2Kirchhoff Institute of Physics (KIP), and Institute of Pharmacy&Molecular Biotechnology (IPMB) University Heidelberg, Heidelberg, Germany
c.cremer@imb-mainz.de

Novel developments in optical technology and photophysics made it possible to radically overcome the diffraction limit (ca. 200 nm laterally, 600 nm along the optical axis) of conventional far-field fluorescence microscopy. Presently, three principal “nanoscopy” families have been established: “Nanoscopy” based on focused laser beams, like 4Pi-, STED- (STimulated Emission Depletion)-, and RESOLFT- (Reversible Saturable OpticaL Fluorescence depletion Transitions) microscopy; nanoscopy based on Structured Illumination Excitation (SIE), like SMI (Structured Modulated Illumination) microscopy, SIM (Structured Illumination Microscopy) and PEM (Patterned Excitation Microscopy); and nanoscopy based on various modes of Localization Microscopy, like PALM (PhotoActivated Localization Microscopy) and FPALM (Fluorescence Photoactivable Localization Microscopy), GSDIM (Ground State Depletion Imaging Microscopy), SPDM Spectral Precision Distance/Spatial Position Determination Microscopy), STORM (STochastic Optical Reconstruction Microscopy) and dSTORM (direct STORM). These and related far-field light microscopy methods have opened an avenue to image nanostructures down to single molecule resolution; they made possible to measure the size of molecule aggregates of few tens of nm diameter and to analyze the spatial distribution of individual molecules with a light optical resolution down to the few nanometer range, corresponding to ca. 1/100 of the exciting wavelength. Application examples obtained by focused, structured, and localization techniques cover a variety of biostructures, such as membrane complexes, neuronal synapses, cellular protein distribution, nuclear nanostructures, as well as the “nanoimaging” of individual viruses and lithographically generated nanostructures. Each of the nanoscopy methods described has its peculiar advantages; as a whole, they provide a tool set of light microscopy approaches to the nanoscale and open a wide range of perspectives in Biology, Medicine and the material sciences. Further improvements are expected to make possible a three-dimensional lightoptical resolution down to the 1 nm scale. The combination with Electron- and X-ray microscopy techniques is anticipated to provide further nanostructural insights.

C. Cremer, Optics far Beyond the Diffraction Limit: From Focused Nanoscopy to Spectrally Assigned Localization Microscopy (2012). In: Springer Handbook of Lasers and Optics, 2nd edition (F. Träger, Edit.), pp. 1351 – 1389.
C. Cremer, B.R. Masters (2013) Resolution enhancement techniques in microscopy. Eur. Phys. J. H 38: 281–344.


Type of presentation: Plenary

IMC-PL-6096 Bioimaging at the nanoscale -- Single-molecule and super-resolution fluorescence microscopy

Zhuang X.1
1Department of Chemistry and Chemical Biology, Department of Physics, Howard Hughes Medical Institute, Harvard University, Cambridge
zhuang@chemistry.harvard.edu

Dissecting the inner workings of a cell requires imaging methods with molecular specificity, single-molecule sensitivity, molecular-scale resolution, and dynamic imaging capability such that molecular interactions inside the cell can be directly visualized. Fluorescence microscopy is a powerful imaging modality for investigating cells largely owning to its molecular specificity and dynamic imaging capability. However, the spatial resolution of light microscopy, classically limited by the diffraction of light to a few hundred nanometers, is substantially larger than typical molecular length scales in cells. Hence many subcellular structures and dynamics cannot be resolved by conventional fluorescence microscopy. We developed a super-resolution fluorescence microscopy method, stochastic optical reconstruction microscopy (STORM), which breaks the diffraction limit. STORM uses single-molecule imaging and photo-switchable fluorescent probes to temporally separate the spatially overlapping images of individual molecules. This approach has allowed multicolor and three-dimensional imaging of living cells with nanometer-scale resolution and enabled discoveries of novel sub-cellular structures. In this talk, I will discuss the general concept, recent technological advances and biological applications of STORM.                                                                                                                                                                                                                                                                                                                                                                               


Type of presentation: Plenary

IMC-PL-6097 Imaging and Spectroscopy of Individual Atoms in Nanostructured Materials

Suenaga K.1
1National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Japan
suenaga-kazu@aist.go.jp

It has remained a challenge for scientists to image and discriminate individual atoms since Dalton first proposed distinct properties of atoms in his atomic theory. The requirements to analyze the atomic structures of matter with elemental information are nowadays increasing in importance of cutting-edge research. An elemental analysis down to the single atom limit was first demonstrated with the successful detection of a Gd dopant atom in carbon nano-peapods using a STEM-EELS technique at 100kV [1]. Specimen damage due to the high dose of the incident electron beam, which is required to isolate the signals from individual atoms, is an intrinsic problem for such a highly delicate analysis. Furthermore it is important to prevent the atoms from being kicked out during the observations. In order to reduce the atomic movements and also to enhance the EELS contrast, a lower accelerating voltage is preferred for single atom detection by STEM. Sawada et al. designed a new type of aberration corrector with triple dodecapole elements (the delta system) to reduce the higher-order geometric astigmatism [2, 3, 4], which is critical for the STEM performance operated at low accelerating voltages, i.e., 15 to 60 kV. Here, I demonstrate successful single-atom imaging and spectroscopy in nanostructured materials using STEM together with EELS and/or EDX.

Fig. 1 shows an example for chemical analysis of individual molecules. A carbon nanotube encapsulating two different metallofullerenes (La@C82 and Ce@C82) is examined at 30 kV operating voltage [5]. The annular dark-field (ADF) image clearly shows the molecular structures encapsulated inside the SWNTs (Fig. 1a). Each molecule carries one metal atom, appearing in brighter contrast, inside the cage. We can identify these atoms by simultaneous EELS. Fig. 1b shows two EELS spectra recorded from two atoms. The EELS spectrum shown in green corresponds to the atom indicated by the green arrow. This spectrum is the sum of four spectra, each of which had an acquisition time of 0.05 s. The resulting signal-to-noise ratio is high enough to isolate the La N-edge. On the other hand, the atom indicated by the blue arrow is assigned as Ce. Moreover, its peak position (≈122 eV) fits very well with that for Ce3+ [6]. Though the two edges of La N and Ce N overlap severely, we could identify the elements (La: Z = 57 and Ce: Z = 58) comprising the two encaged atoms. Fig. 1c shows the ADF image, and the elemental mappings for La, Ce, and carbon are shown in Figs. 1d, e, and f, respectively. A further comparison of simultaneous EELS and EDX measurement allows us to directly estimate the fluorescent yield of single atoms [7, 8].
The interrupted periodicities of 2D materials such as graphene, h-BN, and MX2 (dichalcogenides) are of great interest because they govern the physical/chemical properties. Atomic defects, such as a vacancy or impurity/dopant in single-layered materials are investigated with atomic precision. A single-layer of MoS2 exhibits interesting physical properties. The electrical conductivity of MoS2 can be further modulated by doping, such as Re (n-type) and Au (p-type). Typical ADF images of single-layered Re-doped and Au-doped MoS2 are presented in Fig. 2, respectively. The dopants, Re (Z = 75) and Au (Z = 79), appear in brighter contrast in the ADF images than both Mo (Z = 42) and S (Z = 16). Chemical analysis by means of EDX was also done to confirm the doping elements [9]. ADF image in the inset of Fig. 2(left) clearly shows that Re atoms sit at the Mo sites. The Re dopants are well dispersed in MoS2 layers and seldom form clusters on the host material. On the other hand, the Au dopants at similar concentration tend to aggregate on the MoS2 surface (Fig. 2 right). The Au atoms are indeed mobile under the electron beam [9].
A monovacancy in h-BN can be also examined by STEM-EELS (Fig. 3). Core-level spectroscopy on the nitrogen atoms in the vicinity of the boron vacancy was carried out [10]. As shown in Fig. 3a, a monovacancy is induced at the boron site by the knock-on effect, which can be proved by the fact that the darkest contrast appears in the middle of three nitrogen atoms showing brighter contrast. A line spectrum is recorded across the VB (boron monovacancy) along the yellow arrow. From the line spectrum, three typical spectra for the nitrogen K-edge were extracted, with probe positions corresponding to the yellow circles in Fig. 3b. While the first and third spectra are quite similar to the one for the sp2-bonded nitrogen atoms in h-BN with the known * peak at 401 eV, the second spectrum recorded near the VB indeed shows a sharp pre-peak around 392 eV. Although the spectra are rather noisy because of the minimized acquisition time, this pre-peak appears at the same energy level in many different experiments, and arises reproducibly at other VB sites and represents the lowered LUMO state [10].
Identification of individual atoms and examination of their electronic properties in materials are the ultimate goals of all microscopy-based analytical techniques. It is clear that the bonding/electronic states are now accessible from single atoms through EELS fine-structure analysis. For example the radical carbon atoms at the graphene edge have been successfully identified [11, 12, 13]. Moreover the active point defects in 2D materials can now be caught red-handed [14, 15, 16]. I will also show some of the atomic level observations of alloying behavior and phase transition phenomenon of 2D materials, that used to be investigated only by the macroscopic viewpoint [17, 18].                                                                                                                                                                              

References:
[1] K. Suenaga et al., Science, 290 (2000) 2280-2282
[2] H. Sawada, et al., J. Electron Microscopy, 58 (2009) 341-347
[3] H. Sawada, et al., Ultramicroscopy, 110 (2010) 958-961
[4] T. Sasaki, et al., J. Electron Microscopy, 59 (2010) S7-S13
[5] K. Suenaga, Y. Iizumi and T. Okazaki Eur. Phys. J. Appl. Phys., 54, 33508 (2011).
[6] K. Suenaga et al., Nature Chem., 1 (2010). 415-418
[7] K. Suenaga, et al., Nature Photonics, 6 (2012) 545-548
[8] L. Tizei et al., (in this conference)
[9] Y. C. Lin et al., Adv. Mater., (2014). DOI:10.1002/adma.201304985
[10] K. Suenaga, H. Kobayashi, and M. Koshino, Phys. Rev. Lett., 108 (2012). 075501
[11] K. Suenaga and M. Koshino, Nature 468 (2010) 1088-1090
[12] J. Warner et al., Nano Lett., 13 (2013) 4820-4826
[13] J. H. Warner et al., (unpublished)
[14] K. Suenaga et al., Nature Nanotech., 2, 358-360 (2007).
[15] Z. Liu et al., Nature Commun., 2, 213 (2011).
[16] O. Cretu, Y. C. Lin and K. Suenaga, Nano Lett., 14 (2014) 1064-1068
[17] D. O. Dumcenco et al., Nature Commun. 4 (2013) 1351 (5 pages)
[18] Y. C. Lin et al., Nature Nanotech., in press, (2014).


The present research is supported by a JST-CREST and Research Acceleration Programs. All my colleagues in AIST, Y.C. Lin, O. Cretu, L. Tizei, Z. Liu, M. Koshino, Y. Sato, and R. Senga, are gratefully acknowledged. Drs. H. Sawada, T. Sasaki, M. Mukai, Y. Kohno, M. Morishita and K. Kimoto are also acknowledged for the development of dedicated microscopes.

Fig. 1: Single molecular spectroscopy of mixed peapods (La@C82 and Ce@C82) at 30 kV[5]. (a) An ADF image with a rectangle showing where the spectrum image was taken. (b) Two EELS spectra recorded from two metal atoms. The atom indicated by the green arrow is assigned as La, and the other, indicated by the blue arrow, as Ce. (c) ADF image and (d, e, f) chemical maps for La, Ce, and carbon, respectively. Scale bar = 1 nm.

Fig. 2: Detection of single dopant atoms in single-layered MoS2[7]. (Left) An ADF image of Re-doped MoS2. The Re substitution at Mo site (Re@Mo) is pointed by a green arrow. (Right) An ADF images of Au-doped MoS2, where an Au adatom (indicated by a white arrow) located at the hollow-center (Au-HC). Scale bar = 0.3nm.

Fig. 3: Core-level spectroscopy of monovacancy in h-BN layer [8]. (a) ADF image shows a monovacancy in single-layer h-BN. Line spectrum was recorded along the yellow line. (b) Schematic presentation (red: nitrogen, blue: boron) of boron monovacancy. (c) Nitrogen K-edge fine structures extracted from the line-spectrum. Each of the three corresponds approximately to the probe positions marked in (b). A prominent pre-peak in the nitrogen K-edge can be found at 392 eV in the spectrum recorded at position 2, i.e., near the boron vacancy site.

Type of presentation: Plenary

IMC-PL-6098 Electron Tomography for Nanoscale Materials Science

Midgley P.1
1Department of Materials Science and Metallurgy, University of Cambridge, Cambridge, UK
pam33@hermes.cam.ac.uk

The nanoscale complexity of modern materials and devices, be they structural or functional in design, requires high spatial resolution characterisation in all 3 dimensions. The remarkable power and flexibility of a modern TEM makes it the ideal tool for such 3D nanoscale imaging and analysis. Over the past 15 years or so, electron tomography (3D imaging) has grown from a niche technique to one which is now firmly established as an almost routine tool for the 3D study of materials. Early electron tomography used many of the ideas and practices established first in the life sciences. Here, a tilt series of bright-field (BF) images are acquired by rotating the sample about a single axis and recording images every 1-2°. Typically, in the electron microscope, the range of sample tilt is limited either by the sample itself (becoming too thick) or by the objective lens pole piece gap. As such, it is therefore likely that the full tilt range is not accessible, this leads to a ‘missing wedge’ of information and the reconstructions suffer from artefacts, especially an elongation parallel to the optic axis. Dual axis tomography can help in this regard, reducing the missing information through a second tilt series about an axis mutually perpendicular to the first.

For many materials problems, however, BF images may not be ideal and the introduction of STEM HAADF tomography offered materials microscopists an imaging mode that, in many cases, is much more suited for tomography, providing images with greatly reduced diffraction contrast, with a signal that in most cases varies monotonically with thickness (satisfying the projection requirement) and providing compositional contrast through the atomic number (Z) dependence of the high angle (Rutherford-like) scattering [1]. STEM tomography has now become for many the technique of choice for 3D nanoscale imaging in materials science. Fig. 1 shows two examples of STEM HAADF tomography [2,3]. In Fig. 1(a) we see Ge precipitates within an Al-rich matrix revealing a wide variety of morphologies and clear orientation relationships and in (b) the 3D distribution of Ru-Pt catalyst nanoparticles (1-2nm in size) decorating the surface of a mesoporous silica support – here we see only the external surface. The colour of the support indicates the surface curvature with a strong preference of the nanoparticles to be anchored at the ‘saddle points’. STEM tomography (both BF and ADF) has also been developed for the study of defects (especially dislocations) where the reconstruction (or 3D representation) of the dislocation resembles a ‘string’ running through space.

Although determination of the 3D morphology of materials at the nanoscale is now essentially routine, to achieve a high fidelity reconstruction typically ca. 100 images are needed across the tilt range. For many specimens long acquisition times, and thus extended exposure, can lead to damage. However, the number of images required in the tilt series can be reduced if there is prior knowledge about the specimen being reconstructed. Such prior knowledge can be used within a discrete tomography reconstruction (using the physical discreteness of the sample) or, perhaps more generally, within a compressed sensing framework where the primary requirement is that the sample may be described as being ‘sparse’ in some transform domain [4,5]. This sparsity constraint turns out to be very powerful and high fidelity reconstructions can be achieved with remarkably few images (in some cases an order of magnitude reduction compared to conventional reconstructions), see Fig. 2.

Coupling tomography acquisition with analytical techniques, such as EDX and EELS, allows a more detailed exploration of the sample’s chemistry as well as its morphology. Early efforts in this direction included the use of EFTEM, especially using the low loss regime (where loss probability is relatively high), EDX and core-loss EELS. Inevitably, although the speed and efficiency of spectrometers has improved greatly over the past few years, the acquisition time needed for multi-dimensional ‘spectrum-images’ is considerably higher than a conventional image. To keep the total exposure to a reasonable level, fewer images are recorded in the tilt series – ideally perhaps only every 10 or 20°. The reduction in data must be compensated by an increase in prior knowledge to achieve a high fidelity reconstruction; for such ‘multi-dimensional microscopy’ [6], compressed sensing offers an important framework to achieve this. As an example, Fig. 3(a) shows a composite figure illustrating the localised surface plasmon resonances from a silver nanocube. The reconstruction was undertaken on a series of spectrum-images recorded about a single tilt axis every 15°. The 4mm symmetry of the cube-substrate system was imposed at the reconstruction stage as well as a constraint that the reconstruction could be considered as being sparse in a wavelet domain. That constraint provided a reconstruction relatively free of artefact even when using few images [7]. Interpretation of the reconstruction seen in Fig.3(a) can be made within a quasi-static approximation and related back to the potential induced by the electron beam acting back on the electron. Mapping electro-magnetic potentials is also possible using electron holography and coupled with tomography was able to yield 3D reconstructions of the built-in potential near a p-n junction in a silicon device, see Fig. 3(b) [8]. 3D magnetic fields require an enhanced approach using dual axis geometry to determine all the components of the magnetic potential A (or induction B). Here, physical constraints (e.g. in the form of Maxwell’s equations), perhaps again within a compressed sensing framework, could be used to improve a reconstruction of the electro-magnetic potential.

So, what of the future? The electron tomography community is pushing in many directions. Atomic resolution tomography has been demonstrated in some cases: by assuming periodicity within a nanocrystal, the position of an isolated atom in a matrix can be determined and even the location of atoms around a dislocation core. Synergisitc studies with atom probe tomography have been demonstrated already and this may, in the future, develop into an important correlative approach. Mapping physical properties in 3D at the nanoscale continues to be an exciting prospect. Whilst early work showed this to be feasible, further development is needed to improve reconstruction quality. Given the almost ubiquitous use now, in materials-based tomography at least, of iterative techniques (e.g. SIRT, ART, etc) the conventional projection / back-projection approach could evolve into a more model-based one incorporating a detailed description of the beam’s interaction with the sample along its trajectory (e.g. dynamical effects). By iteratively refining an initial model, increased detail about the sample may be obtainable (e.g. strain, fields, induced charges). Lastly, industry requires a robust nanoscale metrology technique that provides reliable 3D measurements of length, porosity, distributions etc. We are still some way in many cases of being able to provide such data with statistical confidence (i.e. error bars!) on our 3D measurements. Improved reconstructions, with fewer artefacts, incorporating prior knowledge, should allow improved and unbiassed segmentation and thus will go a long way to providing a true 3D nanometrology technique.                                                                                                                                                    

References:
[1] P.A. Midgley et al., Chem. Commun. (2001) 907
[2] K. Kaneko et al., Ultramicroscopy 108 (2008) 210
[3] E.P.W. Ward et al., J. Phys. Chem. C 111 (2007) 11501
[4] Z. Saghi et al., Nano Letters 11 (2011) 4666
[5] R. Leary et al., Ultramicroscopy 2013 131 70-91
[6] P.A. Midgley and J.M. Thomas, Angewandte Chemie (2014) DOI: 10.1002/anie.201400625
[7] O. Nicoletti et al., Nature 502 (2013) 80
[8] A. Twitchett-Harrison et al., Nano Letters 7 (2007) 2020


The author thanks his many colleagues, past and present, who have contributed to the work presented here including most recently J.M. Thomas, R. Leary, Z. Saghi, D. Holland, K. Kaneko, S. Hata, O. Nicoletti, F. de la Peña, C. Ducati. PAM acknowledges funding from the European Research Council under FP7/2007-2013 / ERC grant agreement 291522-3DIMAGE.

Fig. 1: (a) 3D reconstruction of Ge precipitates in an Al-rich matrix showing colour-coded to highlight theor different morpholgy [2]; (b) Ru-Pt catalyst particles (red) shown on a colour-coded silica surface where the blue regions indicate positive Gaussian curvature (saddle points) [3].

Fig. 2: (a) Comparison of reconstructions using SIRT and compressed sensing (CS) codes for an iron oxide nanoparticle with a concavity; (b) the apparent concavity volume as a function of projection number [4].

Fig. 3: (a) Colour composite figure indicating five surface plasmon modes on a silver nanoparticle, 100nm in size [7]; (b) Reconstructed electrostatic potential near a p-n junction in a silicon device showing sub-surface carrier depletion [8].

IFSM symposium

Type of presentation: Plenary

IFSM-PL-1670 From Atomic Structure to Properties of Oxides: Applications of Aberration-corrected Transmission Electron Microscopy

Jia C. L.1,2
1Ernst Ruska-Centre for Microscopy and Spectroscopy with Electrons and Peter Grünberg Institute, Forschungszentrum Jülich GmbH, Jülich, Germany, 2International Centre for Dielectric Research, Xi'an Jiaotong University, Xi'an, China
c.jia@fz-juelich.de

Functional oxides provide an important part of the material basis for multifunctional devices as a result of their exceptional range of physical properties. These properties, in turn, depend strongly on the crystal structures, chemical compositions and defect configurations of the materials, which can be characterized on the atomic scale.

In a high-resolution transmission electron microscope equipped with an aberration corrector, the spherical aberration coefficient CS of the objective lens can be tuned to either a positive or a negative value. The use of a negative value of CS combined with an overfocus setting of the objective lens is used in the negative CS imaging (NCSI) technique [1]. Images obtained using the NCSI technique show superior atomic column contrast and intensity than corresponding positive CS images [2], especially for weakly scattering oxygen columns that are in close proximity to strongly scattering cation columns.

Using the NCSI technique, we have investigated the atomic details near 180° domain walls in thin films of PbZr0.2Ti0.8O3 [3,4]. The relative displacements of ions have been measured and on this basis the local polarization across the wall has been calculated. Using this technique we have studied the atomic structure of LaO-TiO2-type interfaces in epitactic LaAlO3/SrTiO3 heterostructures [5]. The prominent result is the oxygen octahedron rotation and the TiO6 octahedra distortion induced by LaAlO3 in SrTiO3 at the interface. The cation-oxygen octahedra represent the prominent structural element of perovskites, which can be modified by distortions, rotations, and particular atomic shifts. Small atomic rearrangements as they are expected to occur at the interfaces between perovskites of different structure can change dramatically the electronic system.

We have recently used the NCSI technique to perform quantitative comparisons between experimental and simulated images on an absolute intensity basis after taking into account the effects of the modulation transfer function of the camera and additional image spread [6]. This absolute intensity matching approach not only allows atomic column positions and defect structures to be determined with picometer precision, but also allows the local chemistry and the three-dimensional morphology of a crystal to be determined on the atomic scale.

Figure 1 shows results obtained from a study of the relationship between the atomic structure and properties of BiFeO3, a room temperature multiferroic material. In the rhombohedrally-distorted perovskite unit cell of BiFeO3 (shown in Fig. 1a), characteristic structural features include relative shifts between the cations and the oxygen anions along the [111] axis and rotations of oxygen octahedra about the [111] axis, which are related to the ferroelectric polarization and the antiferromagnetic properties of the material, respectively. Both the atomic shifts and the rotations of the octahedra can be quantified using the NCSI and ACM techniques and used to understand the electrical and magnetic properties of the material. Figure 1b shows an atomic-resolution image of a 109° domain boundary (thick arrow) between two domains. The use of NCSI conditions and a particular specimen thickness result in the atomic columns appearing bright on a dark background. The domains in the material can then be distinguished by measuring the positions of the atomic columns inside individual unit cells.

In Fig. 1(b), the domain above the boundary is oriented along the [110] direction. The O column positions are shifted upward and downward (Fig. 1c), corresponding to alternating rotations of octahedra. A corresponding off-centre displacement of Fe with respect to the middle point of the line connecting two neighbouring (left and right) O positions is visible and oriented in a downward direction. In this orientation, the [001] component (red arrow) of the [111] polarization vector can be measured and the octahedron rotation can be revealed. Below the boundary (Fig. 1d), the domain is viewed along the [1 ̅10] direction. The octahedron rotation is now not visible due to the overlap of atoms (Fig. 1d). However, the full vector (red arrow) of the atomic column displacement is now revealed. In this way, the polarization of the domain can be determined unambiguously.                                                                                                                       

References:

C.L. Jia, M. Lentzen, K. Urban, Atomic-Resolution Imaging of Oxygen in Perovskite Ceramics. Science 299, 870 (2003).
C.L. Jia L. Houben, A. Thust,and J. Barthel, On the benefit of the negative-spherical-aberration imaging technique for quantitative HRTEM. Ultramicroscopy 110, 500 (2010).
C.L. Jia et al., Atomic-scale study of electric dipoles near charged and uncharged domain walls in ferroelectric films. Nature Mater. 7, 57 (2008).
C.L. Jia et al., Direct observation of continuous electric dipole rotation in flux-closure domains in ferroelectric Pb(Zr,Ti)O3, Science 331, 1421 (2011).
C.L. Jia et al., Oxygen octahedron reconstruction in the SrTiO3/LaAlO3 heterointerfaces, Phys. Rev. B 79, 081405(R) (2009).
C.L. Jia et al., Atomic-scale measurement of structure and chemistry of a single-unit-cell layer of LaAlO3 embedded in SrTiO3. Microsc. Microanal. 19, 310 (2013).


This work was carried out in collaboration with L. Jin, S.B. Mi, K. Urban, A. Thust, J. Barthel, L. Houben, M. Lentzen, D. Hesse and M. Alexe.

Fig. 1: (a) Schematic diagram of the pseudocubic unit cell of BiFeO3. (b) Atomic-resolution image of a 109° domain wall (thick arrow) separating two domains: the domain above the wall and the magnification in (c) correspond to a [110] viewing direction, while the domain below the wall and the magnification in (d) correspond to a [1 ̅10] viewing direction. The red arrows denote the polarization.

Type of presentation: Plenary

IFSM-PL-1780 Some surprises in electron diffraction physics and imaging.

Spence J. C.1
1Physics Department, ASU, Tempe , Az. USA. 85282, and LBNL USA.
spence@asu.edu

The multiple scattering theory on which modern electron microscopy (EM) is based had been fairly well worked out by about 1960, following work by Bethe, Sturkey, Heidenreich, Hirsch, Howie, Whelan, Cowley and Moodie and others. Nevertheless many surprises remained in the ensuing 50 years. For me the most important of these have been i) The finding that multiple energy-loss effects can be removed from EELS spectra, using earlier work on cosmic ray showers. ii) The richness of the "point-projection" geometry, championed by Gabor in 1949. In turn this has produced Ptychography, the theory of STEM lattice imaging for crystals and low-voltage field-emission point-projection imaging. It is remarkable that coherent overlapping convergent beam orders provide a solution to the phase problem, an atomic-resolution "shadow image", Talbot self-imaging, and in-line holography. iii) The discovery of "forbidden" termination reflections and their value for imaging surfaces and sub-surface dislocations and kinks. Their monolayer sensitivity is remarkable. iv) The detection of coherent bremstrahlung tunable X-ray emission lines in STEM EDX. It is remarkable that these lines can be indexed, and are absent when reflections are forbidden by symmetry. v) The explanation for dynamically forbidden reflections, which cancel due to symmetry-related paths for all thickness. vi) The usefulness of electron channelling effects (Alchemi) on EDX for locating foreign atoms in several fields (turbine blades, mineralogy), previously an academic curiosity. vii) The achievement of aberration correction. viii) The success of our TEM CCD camera, whose impact on cryo-em tomography we never anticipated. ix) The surprising sensitivity of low-angle scattering to atomic bonding, with the zero-order scattering the most sensitive quantity known. x) The finding that sufficiently short pulses of radiation can outrun radiation damage, thus breaking the nexus between damage, resolution and particle size if a large number of particles can be packed into a near delta-function pulse. xi) The information extracted from ELS spectra, with its unrivalled spatial resolution.

  The changing agenda of EM over this half-century, from the study of bulk defects such as dislocations, and atomic resolution imaging of interfaces, to nanoscience, cryo-electron and in-situ microscopy (liquid cells, catalysis) has been fascinating to watch. Recent developments - atomic resolution imaging with characteristic X-rays, direct injection detectors, sub-Angstrom resolution, high-resolution imaging in 3D, fast diffraction and imaging - continue to surprise. References in: High Resolution Electron Microscopy (Spence, 4th ed. 2014) and Electron Microdiffraction (Spence & Zuo, 1992).


To many colleagues and friends over half a century in many countries, and to the US funding agencies and Arizona State University.

Type of presentation: Plenary

IFSM-PL-6099 Macromolecules in Motion: Visualization by 4-Dimensional Cryo-Em

Steven A. C.1
1Laboratory of Structural Biology Research, National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda, USA
stevena@mail.nih.gov

Cryo-electron microscopy offers a unique capability to determine the 3-D structures of macromolecular complexes. However, insight into biological activity requires understanding the structural transitions that the complex of interest undergoes. It is not possible, even in principle, to visualize the same molecule in successive states as this would involve the prohibitively difficult task of thawing the specimen, inducing the conformational change, re-vitrifying and re-locating the molecule. However, dynamics may be addressed by a statistical approach in which classification techniques are applied to data sets imaging conformationally mixed populations. Then, provided that there is a basis for ordering the various conformers in a temporal sequence, the reaction dynamics of the complex may be described and movies made.

The process of virus maturation is amenable this approach. With many viruses, the precursor particle undergoes radical structural changes as it matures into an infectious virion. We have investigated the maturation of bacteriophage HK97 capsid, an icosahedrally symmetric shell composed of 420 protein subunits, which expands from 45 nm to 55 nm and angularizes as it matures. These changes in morphology reflect large rotations of the protein subunits and local remodeling (1), and the pathway proceeds via three metastable intermediates (2). The capsid of herpes simplex virus, an animal virus, follows a similar pathway, which may be traced to a capsid protein domain similar in structure to that of HK97, but it passes through many more intermediates (3). Recently, we have found that bacteriophage phi6, which has a RNA genome rather than a DNA genome and an entirely different capsid protein fold from HK97, also undergoes massive subunit rotations and matures via two intermediates (4, 5) – Figure 1.

For this approach to visualization of conformational dynamics, several conditions must be met (6). The differences between states must be large if differences in the images that arise from viewing geometry are to be separated from real structural differences. The number of distinct conformational states must be relatively small. To establish a time-line, one must be able to induce the reaction of interest on a time-scale of seconds to hours. Notwithstanding, recent technical advances in automated collection of large data sets, the improved resolution and signal-to-noise ratio of direct detection cameras, and sophisticated classification techniques promise to expand the range of applicability.                                                                                                                                            

References

1. J.F. Conway et al. Science 292: 744-748 (2001)
2. R. Lata et al., Cell 100: 253-263 (2000)
3. J.B. Heymann et al., Nature Struct. Biol. 10, 334-341 (2003)
4. D. Nemecek et al., J. Mol. Biol. 414, 260-271 (2011)
5. D. Nemecek et al., Structure 21, 1384-83 (2013)
6. J.B. Heymann et al., J. Struct. Biol. 147, 291-301 (2004)


I thank many colleagues, particularly Drs N. Cheng, J.F. Conway, J.B. Heymann and D. Nemecek. This research has been supported by the intramural research program of NIAMS/NIH.

Fig. 1: Maturation dynamics of bacteriophage phi6 capsid visualized by 4-dimensional cryo-EM. The pathway progresses through four states: the initially assembled procapsid with its deeply indented facets; two expansion intermediates with near-planar facets; and the spherical mature nucleocapsid. The capsid is built from 60 P1A/P1B dimers, where P1A and P1B are chemically identical but conformationally distinct (non-equivalent) protein subunits. The top row shows renderings of the outer surfaces viewed along a 5-fold symmetry axis. The middle row shows models of the respective capsids consisting of a pentamer of P1A subunits (blue, green) and surrounding P1B subunits (red, yellow), viewed from above. The bottom row shows slabs through a portion of the structures, passing through one vertex (P1A’s blue, P1B’s blue). The procapsid-to-intermediate 1 transition is achieved by rotations of P1B subunits about the line connecting two 3-fold icosahedral axes. Further expansion to intermediate 2 is achieved by outward movement of the P1A subunits. The final step to nucleocapsid involves primarily local changes affecting the P1A subunits. Adapted from reference (5), where movies of the transition are available, courtesy of J.B. Heymann. Scale bars: 100 Å.

Type of presentation: Plenary

IFSM-PL-6100 From the Prague Spring to a Spring in Electron Microscopy

Křivánek O. L.1
1Nion Co., Kirkland and Dept of Physics, ASU, Tempe, USA
krivanek@nion.com

Prague is my native city: I was born in Praha-Bubeneč, on the plateau behind the Prague Castle. I grew up in the era of the Czechoslovak Socialist Republic (ČSSR), when the Soviet Union and its satellites prided themselves on their space exploits and their education systems. Among the special efforts they made were competitions for talented youngsters in mathematics and physics, and I used to enjoy those. In my senior high school year, I qualified for the national round in both math and physics, and in physics I was invited onto the national team of three that represented Czechoslovakia at the 2nd International Physics Olympiad, held in Budapest in June 1968. Back then only the Soviet Union and its satellites participated – Western Europe, USA and other countries joined the Physics Olympiad later. Our team did well: we got a joint second place with the Hungarians and the East Germans, with the Soviets winning the first place. I have since then had the pleasure of working with one other former International Physics Olympian – Niklas Dellby, my partner at Nion.

That same summer I took the entrance exam to Charles University in Prague, to study physics. I passed and promptly took off on a trip I had planned: a vacation in the south of France, followed by a stay in London where I was planning to work in a summer job while improving my English. 1968 was the year of the famous Prague Spring, when “socialism with a human face,” which included many democratic measures, was introduced by a group of reformers led by Alexander Dubček (Fig. 1a), much to the displeasure of the old guard in the Kremlin. As I was boarding the train to France, my father told me: “If the Soviets invade, stay in the West.” I had not been following the political situation very closely, so this came as a surprise instruction to me. The Soviets invaded 4 weeks later (Fig. 1b), while my whole family happened to be in the West: my parents on vacation in Austria, my sister working in a summer job in France and me working as an office helper in London. We got together on the phone, and decided that none of us would go back to Prague, at least not for the time being. (See [1] for an especially lucid account of the Prague Spring.)

People were very sympathetic to citizens of a small country invaded by Soviet tanks, and the British National Union of Students had a special place in its London office for notices of available openings for prospective Czech and Slovak students. I was checking it daily while working in a new job, as a carpenter. Around the end of September, a small notice appeared, announcing that the University of Leeds was going to offer up to 5 scholarships to qualified Czechoslovak students. I called them up and caught the train to Leeds soon thereafter. There was an entrance interview during which it became clear that I knew my physics all right, and also that 3 years of high school English and a vocabulary of perhaps 3000 English words were not nearly enough for me to slot painlessly into the British university system.

Leeds took a chance on me, and at first they must have wondered how it would turn out. In my first year I got a First in math – understanding equations did not require much English – but only a Pass in physics, in which there were long textbook passages that I studied laboriously, with a dictionary in hand. I did better in later years, graduating with a First, at the top of my class. I was then accepted to do a physics Ph.D. in Cambridge, with Archie Howie as my inspiring supervisor. In my first year, our lab was not far from Ellis Cosslett’s, after whom the award I received is named, and who has been one of my heroes in electron microscopy, especially after I came to appreciate the pioneering nature of much of the work of his group.
I greatly enjoyed my time in Cambridge, both inside and outside the Physics Department. I learned a lot, made many friends, and made good use of Cambridge’s excellent extra-curricular facilities. I raced for Cambridge against Oxford in skiing and won the special and parallel slaloms at the 1975 Oxford-Cambridge ski race, in the Italian Dolomites. The 8-man boat crew I joined the previous spring (Fig. 1c) did three bumps and an overbump in the Cambridge May Races, and by Cambridge tradition, we got to keep our oars as souvenirs.

After Cambridge, I worked at Kyoto University for 3 months, and did post-docs at Bell Labs and UC Berkeley, where I joined the group of Gareth Thomas in the Materials Science Department. Being in Materials Science made me feel that I had to make a choice: I could concentrate on the materials we were studying and become a materials scientist, or on the instruments and techniques we were using and remain a physicist. I had done a little instrument design work and liked it, so the second option seemed more attractive. The technique I thought was especially interesting was a new one (to me) called Electron Energy Loss Spectroscopy (EELS). I got my first taste of it at the 1978 Cornell workshop, where I met people who became lifelong friends, such as Phil Batson, Christian Colliex, Ray Egerton, and Mike Isaacson. One was expected to build one’s own spectrometer in those days – there were no commercial models. When I got back to Berkeley, I climbed the stairs to Professor Thomas’s office and said: “I think I should build an energy loss spectrometer. It will allow us to study oxygen concentrations at grain boundaries in nitrogen ceramics.” – a subject the group was focusing on. Gareth asked just one question: “How much will it cost? ”, I replied “about $10k”, and I had my first OK to build a major instrument.

The spectrometer came together quickly and produced good results (Fig. 2a). In the summer and autumn of 1979, I was showing the results at various conferences. At one of them, at NBS in Washington, Nancy Tighe came up to me and said: “I think your spectrometer would interest Peter Swann of Gatan. You should give him a call.” This started my fruitful collaboration with Peter, from whom I learned on many fronts. Peter passed away in the summer of 2013, and many of us miss him very much.

Over the next year, Peter Swann, Joe Lebiedzik and I, with input from Mike Scheinfein, designed and built a second-generation serial EEL spectrometer. I also started in a new job, as Associate Director of the NSF-funded HREM facility at Arizona State University. With my collaborators at ASU, we applied the spectrometer to many interesting problems, and put together the EELS Atlas [2] that is used to this day. ASU was a great place to work. There were many good instruments, several leading researchers in electron microscopy, and stimulating annual schools and workshops (Fig. 1d), whose organization was my responsibility.
The pull of Gatan, however, proved irresistible when Peter moved its R&D facility from Pittsburgh to California, and in 1985 I became Director of Research at Gatan. A very productive period followed, during which I had the privilege of working with many talented researchers and designers: Dan Bui, Niklas Dellby, Garry Fan, Stuart Friedmann, Sander Gubbens (the current President of Gatan), Robert Keeney, Bernd Kraus, Mike Kundmann, Mike Leber, Chris Meyer, Paul Mooney, Ming Pan, Nils Swann, Peter Swann, Marcel Tence and Jacob Wilbrink, among others. We introduced a number of innovative products, including parallel EELS, imaging filters, CCD cameras, scanned image acquisition systems and DigitalMicrograph software. Gatan grew nearly 10x in size during this time, and I learned that developing instruments commercially can be a great way to fund instrumentation research, especially when working with like-minded researchers and lean and understanding administrations.
The next big change in my scientific life came when Peter decided to retire in 1992, and “professional managers” took over at Gatan. My freedom to do interesting projects was greatly restricted, and I started to look around. It had been clear to me since about 1990 that having managed to correct the second order aberrations of the quadrupole optics of imaging filters, I had a good chance of correcting third order aberrations – a classic problem in electron optics since Scherzer’s work on the subject in the 1930s and 40s. It seemed too speculative a project for Gatan, however, and so I explored doing it elsewhere. My first try for corrector funding was a chat with Uli Dahmen, the Berkeley NCEM director, who consulted with Bob Gottschall, his manager at DOE. Bob’s answer was apparently “over my dead body.” He had gotten burned funding Crewe’s corrector attempts, which never led to a working instrument.

I was more successful persuading Mick Brown of my Alma Mater, Cambridge University, who had a spare VG cold field emission (CFE) scanning transmission electron microscope (STEM), that we should jointly build a corrector for it. We applied for funding to the British Royal Society and secured the maximum allowed amount from the Paul Instrument Fund: £80k. I then moved to Cambridge with my family for two wonderful years. Niklas Dellby and others joined the project, and we had a working proof-of-principle STEM corrector about 2 years later [3], the same summer (1997) as the Heidelberg-Julich CTEM corrector started working.
The 100 kV VG STEM we built our corrector for was older than a research student who joined the project (Andy Lupini), and it had poor aberration coefficients (Cs~Cc~3.5 mm). We improved its resolution, but we did not beat any resolution records relative to the best uncorrected instruments. (The same was true for the Heidelberg effort – 1 MV microscopes were then giving higher resolution than their corrected 200 kV CTEM.) However, a corrector of an improved design we built for Phil Batson’s extensively modified VG at IBM Yorktown Heights achieved a double distinction: it led to the first STEM able to focus an electron probe to <1 Å diameter [4], and it was, as far as I know, the first commercial corrector (delivered in June 2000).

Aberration correction soon became a “hit”, with CEOS GmbH supplying correctors to all the regular manufacturers of electron microscopes, and the company Niklas Dellby and I started near Seattle, Nion, concentrating on correctors for CFE STEM and going it alone. Our idea was a somewhat crazy one: that we could extend our prowess in correctors by designing a whole new electron microscope, and that we would do it better than the regular manufacturers. Not many thought that we would succeed. But there were early believers to whom we owe a great deal, such as John Silcox, Andrew Bleloch, Steve Pennycook and Christian Colliex. Benchmarks established subsequently by Nion for resolution, stability, probe current, ultra-high vacuum, freedom from contamination and powerful software [5,6] have persuaded many others. 

Nion’s very capable team - Niklas Dellby, Neil Bacon, George Corbin, Peggy Cramer, Zeno Dellby, Russ Hayner, Petr Hrncirik, Tracy Lovejoy, Chris Meyer, Savath Phoungphidok, Michael Sarahan, Gwyn Skone, Zoltan Szilagyi, Janet Willis, Tad Yoo and myself for now, and growing, has done some amazing things. We first delivered 10 aberration correctors for VGs, then moved onto making whole electron microscopes. Currently we’re manufacturing Nion microscopes #10-13, and the interest in our instruments is on the rise. Building the instruments has been made easier by the close collaboration we enjoy with Czech Republic’s Delong Instruments, especially Vladimír Kolařík and Petr Homolka. Nion’s progress has also been helped by two simple facts: ordering an electron microscope from a small company is a gutsy thing to do, and gutsy scientists tend to be first-rate. (Figs 2b-3a,b) and references [7-13] show some of the revolutionary results they and their collaborators have obtained with Nion microscopes.

Aberration correction has ushered in an era of electron microscopy in which we can see the structure, composition and bonding of materials better than ever before. It amounts to a new spring in electron microscopy, best captured by the words of David Cockayne: “it is as though a veil of fog has lifted from our samples.” It is about to get better still, because of an exciting new development: studying energy losses with sub-20 meV energy resolution and sub-nm spatial resolution. This has been made possible by Nion’s new monochromator [14], which has been the subject of two separate talks at this congress [15], and which promises to make vibrational excitations in materials (phonons) readily observable (Fig. 4), at a high spatial resolution. It will probably also allow hydrogen to be mapped in the electron microscope, using energy losses that accompany high-angle scattering of fast electrons by hydrogen nuclei. 

My scientific instrumentation journey began with EELS and progressed onto aberration correction and high resolution STEM. It has now come back to EELS, with an energy resolution about 100x better than on my first try. My life’s journey began in Prague, and Prague is where this congress has been held. Both journeys are reminiscent of the famous lines by T.S. Eliot [16]:                                                                                

We shall not cease from exploration,
And the end of all our exploring
Will be to arrive where we started
And know the place for the first time.                                                                                                       

So let us celebrate exploration (also known as research) and knowing where we came from. And also congresses such as IMC, which enrich our knowledge of our field, and of ourselves.                                                                                                                                                        

[1] A. Levy, Rowboat to Prague (ISBN 0-670-60920-X), reprinted as So Many Heroes (ISBN 978-0933256125). See also http://en.wikipedia.org/wiki/Alan_Levy
[2] C.C. Ahn and O.L. Krivanek, EELS Atlas (1983) Gatan and the ASU HREM facility.
[3] O.L. Krivanek et al., Proc. EMAG 1997, IOP Conf. Ser. No 153 (J. Rodenburg, ed.) 35-40.
[4] P.E. Batson, N. Dellby and O.L. Krivanek, Nature 418 (2002) 617-620.
[5] O.L. Krivanek, et al., Ultramicroscopy 108 (2008) 179-195.
[6] N. Dellby et al., The European Physical Journal Applied Physics 54 (2011) 33505 (11 pages).
[7] D.A. Muller et al., Science 319 (2008) 1073-1076.
[8] O.L. Krivanek et al., Nature 464 (2010) 571-574.
[9] T.C. Lovejoy et al., Appl. Phys. Letts 100 (2012) 154101 to 154101-4.
[10] P.Y. Huang et al., Nano Letters 12 (2012) 1081-1086.
[11] W. Zhou et al., Microscopy and Microanalysis 18 (2012) 1342-1354.
[12] Q.M. Ramasse et al., Nano Letters 13 (2013), 4989–4995.
[13] J. Lee at al., Nature communications 4 (2013) 1650.
[14] O.L. Krivanek et al., Microscopy 62 (2013) 3-21.
[15] N. Dellby et al., these proceedings and O.L. Krivanek et al., these proceedings.
[16] T.S. Elliot, Four Quartets (1943) ISBN 978-0156332255.


Fig. 1: a) Alexander Dubček, who led the Prague Spring. b) Soviets tanks rolling through Prague’s Wenceslas square. c) Rowing on the river Cam. I am in seat #5 (bow = 1), holding the oar that does not quite match the others. d) 1981 ASU meeting. Front row: Mike Isaacson, Alan Craven, John Spence, John Venables, Albert Crewe, John Cowley, Bernard Jouffrey, Ian Wardell, Ondrej Krivanek, Colin Humphreys. Spot Ray Carpenter, Mark Disko, Murray Gibson, Sumio Iijima, Kazuo Ishizuka, Masashi Iwatsuki, Charlie Lyman, Peggy Mochel, Steve Pennycook, Jing Zhu and others in the photo.

Fig. 2: a) EEL spectrum of BaTiO3 recorded with the serial EEL spectrometer I built at Berkeley, at about 2 eV resolution. b) Reversible atomic motion in monolayer graphene: one of the 6 substitutional Si atoms moves right, left, right. Nion UltraSTEM100, 60 kV, 6 s between frames. (Ref. [13])

Fig. 3: a) EEL L2,3 spectrum from a single Si atom replacing a C atom in graphene (line) and theoretical fits (solid spectra). The right fit allowed the Si atom to “pop out” 0.65 Å from the graphene plane (inset) and gave better agreement. (Ref. [12]) b) Results from Nion microscopes. Nature vol. 464 (2010) issue 7288 cover: image of BN monolayer with impurities by Matt Chisholm, processing by the author and Tim Pennycook. Angewandte Chemie vol. 50 (2011) issue 43 cover: image of MoS2 by Quentin Ramasse. Nature Materials vol. 11 (2012) issue 10 cover: EELS elemental map by Julia Mundy.

Fig. 4: 60 kV results from the Nion High Energy Resolution Monochromated EELS-STEM (HERMES). a) Spectra obtained with the slit out and in, slit-in acquisition time 0.25 s, courtesy Niklas Dellby (Nion) and Philip Batson (Rutgers U.). b) spectrum from titanium hydride (acq. time 10 s), courtesy Peter Crozier and Jiangtao Zhu (ASU), and Tracy Lovejoy (Nion).

IT-1. Electron optics and optical elements

Type of presentation: Invited

IT-1-IN-2012 Advances in electron vortex experiments in the TEM

Verbeeck J.1, Béché A.1, Clark L.1, Guzzinati G.1, Juchtmans R.1, Van Boxem R.1, Van Tendeloo G.1
1EMAT, University of Antwerp, Groenenborgerlaan 171, 2020 Antwerp, Belgium
jo.verbeeck@ua.ac.be

Electrons in a transmission electron microscope are successfully described by linear combinations of plane waves. The sample and the magnetic lenses deform the wavefronts of these waves in a way that transfers information from the sample onto a detection plane. Alternatives to this plane wave basis are however possible and especially cylindrical harmonics are an interesting option. The plane waves are replaced by waves which have a typical azimuthal phase factor exp(i m φ) with φ the angle in the plane perpendicular to the optical axis and m the so-called topological charge. Such waves are orbital angular momentum (OAM) eigenstates in the sense that a normalized cylindrical wave carries exactly mħ angular momentum around the cylinder axis. These waves are often referred to as vortex waves and they attract considerable attention in many different fields of physics including optics, acoustics, radio communication and quantum information [1].
Electron vortices were theoretically predicted to possess also a quantised magnetic moment mµB on top of the common OAM of mħ due to their electrostatic charge [2]. It took untill 2010 before pure electron vortex modes were demonstrated in a transmission electron microscope [3,4]. Since then, many different ways of producing these waves have followed (see e.g. fig.1), each with different advantages and disadvantages. The latest addition, sketched in fig.2, is the production of electron vortex waves making use of a thin single domain magnetic needle approximating a magnetic monopole [5]. This method holds great promise as it offers pure vortex modes at full beam current. Apart from producing single vortex modes, we also focused on the detection of the OAM in an arbitrary wave. Several methods are possible and will be discussed. In terms of the interaction with a sample we observed magnetic dependence in EELS spectra of ferromagnetic samples relating to electron magnetic chiral dichroism and its X-ray counterpart X-ray magnetic chiral dichroism. On top of this, we will discuss the use of vortex beams in elastic diffraction and the transfer of angular momentum to rotate nanoparticles.

References

[1] J. F. Nye and M. V. Berry., Proc. of the R. Soc. of London. A. 336/1605 (1974) 165.

[2] K. Bliokh et al., Phys. Rev. Lett. 99 (2007) 190404.

[3] M. Uchida and A. Tonomura., Nature, 464/7289 (2010) 737.

[4] J. Verbeeck et al., Nature 467/7313 (2010) 301.

[5] A. Béché et al., Nat. Phys.10/1 (2013) 26.


This work was financially supported by the European Union: ERC grant 246791 COUNTATOMS, ERC Starting Grant 278510 VORTEX, Integrated Infrastructure Initiative grant 312483-ESTEEM2.

Fig. 1: Using a probe aberration corrector and an annular condensor aperture, an approximation to an azimuthally varying phase plate can be obtained. This is one of the alternative ways to create electron vortex beams.

Fig. 2: Sketch of the creation of an electron vortex beam by impinging a plane wave to the end of a long bar magnet approximating a magnetic monopole.

Type of presentation: Invited

IT-1-IN-2597 Electron Optics for High-brightness High-beam-current Column Design --- extracting micro-amperes from point cathodes ---

Fujita S.1, Takebe M.1, Wells T.2, El-Gomati M. M.3, Shimoyama H.4
1SHIMADZU Corporation, Kyoto, Japan, 2York Probe Sources Ltd., York, United Kingdom, 3University of York, York, United Kingdom, 4Meijo University, Nagoya, Japan
fujita@shimadzu.co.jp

A principal goal of designing electron probe forming system is to focus desired beam current into as small a spot on the target as possible. Increasing demand for analytical measurements is making desired beam current higher than ever (Ib>10nA). This article describes strategies to design high-brightness high-beam-current electron optical columns.

Figure 1 shows Probe Property that relates the beam current Ib to the probe size d. The dotted curve assumes as electron source a thermionic gun while the dashed curve is for conventional ZrO/W (100) Schottky emitter (SE) gun system. A higher brightness of the latter makes the probe size substantially smaller in the middle beam current regime. However, the probe blurs fast once the current exceeds a certain threshold.

Probe Property is limited by three different mechanisms with increasing beam current order:

Beam Current Regime   “low”                “middle”          “high”
Limiting Mechanisms     wavelength       brightness       angular intensity
                                        chromatic(OL)   spherical(OL)   spherical(Gun)

Attempts were made to improve “high” beam current performance by increasing the source angular intensity and suppressing the gun spherical aberration. Extended Paraxial Trajectory Method is used to analyze electron rays starting from cathode surface with large slopes [1]. The emission characteristic of SE gun is then given by optical parameters familiar in lens designs.

The first strategy is to adopt an emitter whose tip radius is significantly larger [2]. Figure 2 compares a scaled-up emitter (giant SE = GSE) with a conventional SE. The tip size effect is reflected in “electron gun focal length,” f. The angular intensity is given by JΩ = f2*js where js is the cathode current density. Since the focal length is roughly proportional to the tip size, a large tip leads to an improved angular intensity.

The second strategy is to immerse the emitter in the condenser lens field, which is known to result in a suppressed spherical aberration.

Figure 3 compares the source emittance diagrams of conventional SE and GSE. GSE’s wide and less-distorted diagram demonstrates its high-beam-current capability. It is expected GSE’s improved emittance extends the “middle” beam current regime to Ib ~ 1μA (see Fig.1).

A test column was constructed by combining the GSE gun with an objective lens designed for efficient X-ray detection. SEM image observations at Vacc = 10kV over beam current range 100pA <Ib< 3μA confirmed semi-quantitatively the predicted probe property given in Fig.1.

[1] S.Fujita, M.Takebe, W.Ushio and H.Shimoyama, J.Electron Microsc. 59, 3 (2010).
[2] S.Fujita, T.R.C.Wells, W.Ushio, H.Sato, and M.M.El-Gomati, J.Microsc. 239, 215 (2010).


The authors thank Shimadzu Corporation for the support of this work as well as for the permission of the publication.

Fig. 1: Probe Properties with three different electron sources. Expected probe size is plotted against the beam current.

Fig. 2: Comparison of tip geometries of conventional SE and scaled-up emitter (GSE). Angular intensity can be increased by a larger tip radius.

Fig. 3: Source emittance diagrams of conventional SE and GSE. GSE’s wide and less-distorted emittance demonstrates an improved high-beam-current capability of the emitter.

Type of presentation: Oral

IT-1-O-1664 Magnetic-Field-Superimposed Cold Field Emission Gun for 1.2-MV Transmission Electron Microscope

Kasuya K.1, Kawasaki T.1, Moriya N.1, Arai M.1, Furutsu T.1
1Central Research Laboratory, Hitachi, Ltd., Akanuma 2520, Hatoyama, Saitama, 350-0395, Japan
keigo.kasuya.bp@hitachi.com

     A magnetic-field-superimposed cold field emission gun (M-FEG) was developed for a 1.2-MV transmission electron microscope (TEM)[1]. This microscope is intended to have a point resolution of 40 pm and to take atomic-scale three-dimensional images by electron holography.

     Figure 1 shows the cross section of the developed M-FEG. The gun is designed to have a high brightness and stable emission current. The gun is equipped with a pre-accelerator magnetic lens placed close to the emitter [2]. The superimposed magnetic field causes the emitted electrons to converge so that the aberration-caused blurring with subsequent electrostatic lenses is minimized. As a result, the inherent high brightness of the cold field emitter can be obtained. The chambers of the gun are differentially evacuated with three non-evaporative getter (NEG) pumps and four ion pumps. The pressure of the first chamber, where the emitter is placed, was 3×10-10 Pa. This small pressure stabilizes time variations of the emission current [3].

     Figure 2 shows the measured time variations of the probe and total currents. After performing flashing of the emitter, the initial probe current of 1 nA was obtained at the total current of 1 µA. The probe current stayed almost constant for more than 10 hours during the initial period of the measurement. The 90% decrease time, at which the current falls to 90% of the initial value, was prolonged to 900 min in comparison with 3 min in a previous gun at 5×10-8 Pa [4]. The variation in the probe current over the course of the initial 8 hours was 5.2%.

     Another advantage of the pressure reduction is the increase in probe current. It increased two times higher than that of the conventional field emission gun operating at 10-8 Pa. This reason can be explained by the fact that the clean emitter surface has higher probe current density than the adsorbed surface. The gun provided large probe currents ranging from 1 to 170 nA for total currents ranging from 1 to 300 µA.

     The resulting current characteristics ensure that the 1.2-MV TEM will have fine resolution with a high S/N ratio. The illumination system of the microscope is discussed by Kawasaki in this conference.

[1] K. Kasuya et al., submitted to J. Vac. Sci. Technol. B.

[2] M. Troyon, Optik 57, 401 (1980).
[3] K. Kasuya et al., J. Vac. Sci. Technol. B 28, L55 (2010).
[4] T. Kawasaki et al., J. Elec. Microsc. 49, 711 (2000).

 


This research was supported by the Japan Society for the Promotion of Science through the FIRST Program, initiated by the Council for Science and Technology Policy.

Fig. 1: Cross-section of the developed magnetic-field-superimposed cold field emission gun (M-FEG). The pressure of the first chamber was 3 ×10-10 Pa.

Fig. 2: Measured time variations of probe and total currents. The 90% decrease time of the probe current was 900 min. The variation in the probe current over the initial 8 hours was 5.2%.

Type of presentation: Oral

IT-1-O-1982 Challenges in Phase Plate Development and Applications

Sader K.1, Buijsse B.1, van Duinen G.1, Danev R.2
1FEI, Achtseweg Noord 5, 5651 GG Eindhoven, The Netherlands, 2Max Planck Institute of Biochemistry, Am Klopferspitz 18, 82152 Martinsried, Germany
kasim.sader@fei.com

While there have been attempts to implement phase plates in transmission electron microscopes (TEMs) over a long period of time, a publication by Danev and Nagayama [1] renewed interest that functional phase plates could be produced. In particular in life sciences, the development of thin film vitrification techniques has enabled the examination of unstained macromolecules and thin cells in the electron microscope, but also created the need for phase contrast. Conventionally, contrast at low spatial resolutions has been generated by using a strong defocus, but with the added consequence of introducing oscillations in the contrast transfer function. A phase plate allows one to work in-focus, with a large increase in the contrast at low spatial resolutions.

Many types of phase plates have been proposed, but the most widespread implementation has been the original thin-film Zernike phase plate. This type of phase plate has shown practical performance, especially in life science applications. The most widely tested film type is amorphous carbon, but these suffer from aging problems, making frequent exchanges of the phase plate necessary. Alternatives to conventional amorphous carbon have been investigated and silicon-based films show promise in terms of longevity.

In close collaboration with the Max Planck Institute of Biochemistry in Martinsried, FEI have developed a new type of phase plate with properties that make it very suitable for implementing it as a user friendly device in our TEMs. It produces high-contrast images, providing excellent contrast transfer in the low resolution range which is particularly relevant for cryo-electron tomography and may provide benefits for single particle analysis in the case of small and heterogeneous particles. No fringing effects around high-contrast features are observed and CTF oscillations can be avoided up to better than 10Å while maintaining contrast transfer at low spatial frequencies. Transmission losses by the phase plate are very modest. Moreover, the phase plate shows consistent performance for at least half a year of usage.

To facilitate routine phase plate usage we have added extra alignments and control panels to the microscope software. In particular, accurate adjustment of beam deflection pivot points is included to ensure a stable beam position at the plane of the phase plate. Also, software has been developed to easily navigate the phase plate in the back focal plane. We are developing detailed phase plate workflows for our applications software that will provide a seamless integration of the phase plate in the (automated) applications. In this talk a selection of results will be shown from cryo electron tomography.

[1] R. Danev, K. Nagayama, Ultramicroscopy 88, 243-252 (2001)


Fig. 1: Tecnai F20 results from cryo electron tomography on doxorubicin using conventional TEM at 4 μm defocus. Experimental conditions: total dose of 85 e-/Å2, tilt range +/-60º. 

Fig. 2: Tecnai F20 results from cryo electron tomography on doxorubicin using phase plate TEM at 0.5 μm defocus. Experimental conditions: total dose of 85 e-/Å2, tilt range +/-60º.

Type of presentation: Oral

IT-1-O-2071 Sculpturing the electron wave function using nanoscale phase masks

Shiloh R.1, Lereah Y.1, Lilach Y.1, Arie A.1
1Department of Physical Electronics, Fleischman Faculty of Engineering, Tel Aviv University, Tel Aviv, Israel
royshilo@post.tau.ac.il

Electron beams are extensively used in lithography, microscopy, material studies and electronic chip inspection. Today, beams are mainly shaped using magnetic or electric forces, enabling only simple shaping tasks such as focusing or scanning. Recently, binary amplitude gratings achieved complex shapes. These, however, generate multiple diffraction orders, hence the desired shape, appearing only in one order, retains little of the beam energy. Here we demonstrate a method in electron-optics for arbitrarily shaping electron beams into a single desired shape, by precise patterning of a thin-membrane. It is conceptually similar to shaping light beams using refractive or diffractive glass elements such as lenses or holograms - rather than applying electromagnetic forces, the beam is controlled by spatially modulating its wavefront. Our method allows for nearly-maximal energy transference to the designed shape, and may avoid physical damage and charging effects that are the scorn of commonly-used (e.g. Zernike and Hilbert) phase-plates. The experimental demonstrations presented here – two solutions to the free-space wave equation: on-axis Hermite-Gauss and Laguerre-Gauss (vortex) beams, and computer-generated holograms – are a first example of nearly-arbitrary manipulation of electron beams. Our results herald exciting prospects for microscopic material studies, research in electron-matter interaction, enables electron lithography with fixed sample and beam and high resolution electronic chip inspection by structured electron illumination.


The work was supported by the Israel Science Foundation, grant no. 1310/13 and the German-Israeli Project cooperation.

Fig. 1: On-axis generation free-space modes: images taken at different effective distances near the diffraction plane. (A) Unmodulated beam passing through the membrane, (B) Hermite-Gauss11-like, (C) Laguerre-Gauss01-like (vortex), (D) Bragg diffraction pattern used as metric, (E) Bragg grating, (F) HG11-generating mask, (G) vortex-generating mask.

Fig. 2: On-axis holograms: (A) “TAU” hologram produced by the mask in (B); inset: magnification showing ~60nm holes composing the pixels. (C) Electrons orbiting a nucleus hologram produced by the mask in (D); inset: magnification showing the centre of the mask. Note: contrast and brightness levels in (C) were altered for visibility.

Type of presentation: Oral

IT-1-O-2914 Tuning and Operation of a sub-20 meV Monochromator

Dellby N.1, Lovejoy T. C.1, Křivánek L. O.1
1Nion Co., 1102 Eighth St., Kirkland, WA 98033, USA
dellby@nion.com

When aiming for simultaneous high energy resolution and high spatial resolution in a monochromated scanning transmission electron microscope (STEM), three locations in the microscope are critical:
1) the monochromator’s (MC’s) energy-selecting slit, where the pass-band of energies admitted into the rest of the column is determined,
2) the sample, where the tuning determines the spatial resolution, and
3) the detector of the electron energy loss spectrometer (EELS).
To optimize the performance of the entire system, aberrations in all three locations must be accurately and repeatably tuned, so as to produce the smallest possible beam crossover at each place. In typical operation, all three crossovers are images of the field emission source, and upstream crossovers are re-imaged in subsequent stages. A mistuned monochromator can be largely compensated by a pre-sample aberration corrector that is mistuned in the opposite direction, or by a mistuned EELS.
The ideal method for monochromator tuning should therefore measure the actual aberrations at the plane of the energy selecting slit and not be affected by post-monochromator optics. We use a variation of the method developed by Foucault[1]: we image the far-field shadow of the energy-selecting slit near which the beam crossover is formed.
With a monochromatic beam coming into the monochromator, the aberrations would be tuned when the far-field image of the slit fades out uniformly as the slit is closed up. Non-zero focus and astigmatism would produce a stripe across the image of the beam-defining aperture, and one would focus and stigmate to make the stripe wider until it fills the aperture.
In practice, however, the slit is illuminated with an energy-dispersed beam some 300 meV wide, i.e. about 20 times larger than the energy width of our usual monochromated beam. This means that electrons with different incoming energies fill different parts of the aperture with stripes of different energies (Fig. 1), and the total beam after the slit is an incoherent superposition of a distribution of slit positions.
Fortunately, the tuning information is imprinted on the coherence properties of the beam exiting the MC slit, and we use this to determine the tuning at the slit to first and higher orders (Fig. 2). The end result is a repeatable tuning of the MC to 15 meV and better in Nion’s High Energy Resolution Monochromated EELS STEM (HERMESTM) (Fig. 3), as well as an ability to refocus the beam at the sample to sub-nm dimensions [2].
[1] L. Foucault, Comptes Rendus Academie des Sciences 47 (1858) 958-959.
[2] OL Krivanek et al, Microscopy 62 (2013) 3-21


Fig. 1:  Idealized far-field image (Ronchigram) for three electrons beams with slightly different energies, with astigmatism present at the MC slit.

Fig. 2:  Fourier transform of a Ronchigram obtained with the beam tuned at the MC slit.

Fig. 3:  Zero loss peaks before and after MC tuning

Type of presentation: Poster

IT-1-P-1486 Magnetic monopole like fields and electron vortices

Béché A.1, Van Boxem R.1, Van Tendeloo G.1, Verbeeck J.1
1EMAT, University of Antwerp, Groenenborgerlaan 171, 2020 Antwerp, Belgium
armand.beche@uantwerpen.be

The search for magnetic monopole particles has been in vain so far. However, an approximation to a magnetic monopole field can be obtained at the tip of a long, thin, nanoscopic magnetic needle [1,2]. We demonstrate that the interaction of an electron beam with such a field produces an electron vortex beam just like was predicted for a true magnetic monopole [3]. The total orbital angular momentum (OAM) produced by the magnetic needle can be precisely tuned by carefully selecting the amount of magnetic flux via the needle cross section.

The magnetic needle is extracted from a 60 nm thick nickel film using focused ion beam (FIB) milling. It is then deposited on top of a gold plated silicon-nitride grid with one end suspended over a pre-cut aperture hole (Fig.1 A). This aperture allows the impinging electron beam to interact with only one end of the needle. The magnetic field at the tip causes the fast electrons to obtain a spiral phase shift via the Aharanov-Bohm effect as revealed by holography in field free conditions in a transmission electron microscope (TEM) (Fig. 1B). The width of the needle is reduced in the FIB until the flux approaches one fluxon (total phase shift of to 2pi). Comparing the experimental results with simulations (Fig. 1C), an OAM of 0.8 was estimated.

In order to confirm the existence of a vortex after letting an electron beam interact with the magnetic needle aperture, a focal series was acquired in the far field plane of the needle (Fig.2 A). The presence of a dark center which does not disappear upon focusing is typical for a vortex beam, as demonstrated in simulated images (Fig. 2B). A second confirmation of the vortex character was made by cutting the slightly defocused far field images with the sharp edge of an objective aperture and noting the configuration of the Fresnel fringes [4]. Close to the vortex core, the phase dislocation pattern appears in the Fresnel fringes (Fig. 3A). The number of non-connected lines gives an approximation of the total OAM, close to 1 in the present case, confirming the holography result (Fig. 1B). The Fresnel fringes agree remarkably well with simulations (Fig. 3B).

An aperture containing such a monopole-like field provides a unique way of creating electron vortex beams with a pure OAM value, independent of the electron energy. As almost all the incoming electrons transforms into a specific OAM state, a high intensity vortex beam is created, greatly improving the potential for atomic scale magnetic measurements at much improved signal to noise ratios.

1. Béché A. et al., Nature Physics (2014), 10, p. 26-29.
2. Kasama T. et al., MRS Proc. (2004), 839, p. 107-118.
3. Aharonov Y. and Bohm D., Phys. Rev. (1959), 115, p. 485-491.
4. Verbeeck J. et al., Nature (2010), 467, p. 301-304.


This work was financially supported by the European Union: ERC grant 246791 COUNTATOMS, ERC Starting Grant 278510 VORTEX, Integrated Infrastructure Initiative grant 312483-ESTEEM2.

Fig. 1: A: Overview of needle surrounded by an aperture. B: Experimental phase map at the tip of the needle, figured by the dash square in B. The phase rosacea is scaled from 0 to 2pi. C: Simulated phase map for a total phase shift of 0.8x2pi over the full aperture.

Fig. 2: A: Experimental focus series of the aperture in far field conditions. The destructive interference center is typical of a vortex beam. B: Simulation of the focal series using the phase profile displayed in Fig. 1C.

Fig. 3: A: Cut of the defocused far field image of the needle aperture by a sharp edge, revealing a dislocation like feature in the Fresnel fringes close to the vortex center. As only one branch cannot connect, the total OAM is close to 1. B: Simulation of the cut aperture in the far field using the phase profile displayed in Fig 1C.

Type of presentation: Poster

IT-1-P-1500 Measuring the Orbital Angular Momentum of Electron Vortex Beams in the TEM

Guzzinati G.1, Clark L.1, Béché A.1, Verbeeck J.1
1EMAT, University of Antwerp, Groenenborgerlaan 171, 2020 Antwerp, Belgium
giulio.guzzinati@uantwerpen.be

The exchange of orbital angular momentum (OAM) in the interaction between an electron beam and a sample is determined by the properties of the sample and the beam [1,2]. Studying this interaction could enable a new class of OAM based microscopy techniques if convenient measurement of OAM exchange would exist. These techniques could then be used to study in the TEM, among others, the magnetic state of atoms and the transfer of OAM nanoparticles.

Electron beams possessing intrinsic orbital angular momentum have recently risen to attention after the prediction and demonstration of electron vortex beams[3-5]. This discovery has led to the rapid development in the field of singular electron optics [1-7].
In order to employ electron vortices as a probe to study the OAM exchange between a beam and a sample, methods to manipulate or measure the OAM of the beams are fundamentally important. While several methods have been designed to produce vortex beams, there has not been an equal progress in the detection and measurement of intrinsic OAM in the electron microscope.

Aiming to bridge this gap, we have implemented several diffraction based OAM measurement methods: using a forked grating hologram, a triangular geometrical aperture, a knife-edge and an astigmatic phase plate. Fig.1 shows an overview of the experimental results of the different methods when different incoming vortex beams are used as input.
In particular the triangular aperture and the astigmatic phase allow to recognize high order vortex beams easily , but they require to record and analyze a full 2D diffraction pattern. Intentional astigmatic aberration is easier to implement but the OAM is revealed by observing the beam waist rather than the far field pattern which may be a disadvantage in scanned electron probe setups.
On the other hand the hologram and the knife-edge are only appropriate for the measurement of lower values of OAM, but they allow the measurement to be reduced to a simple electron counting process which makes them ideally suited for automated OAM measurement [7].

[1] P. Schattschneider et al., Phys. Rev. B 85, 134422 (2012).
[2] A. Béché et al., Nat. Phys.10/1 26 (2013).
[3] K. Bliokh et al., Phys. Rev. Lett. 99 190404 (2007).
[4] M. Uchida and A. Tonomura., Nature, 464/7289 737 (2010).
[5] J. Verbeeck et al. Nature 494, 331–335 (2013).
[6] V. Grillo et al., Phys. Rev. X 4, 011013 (2014).
[7] G. Guzzinati et al., Phys. Rev. A 89, 025803 (2014).


We acknowledge funding from the European Union under the FP7 program, ERC Starting Grant No. 278510 VORTEX and Integrated Infrastructure Initiative No. 312483 ESTEEM2.

Fig. 1: (a) Schematic representation of the experiment, depending on the OAM of the input beam and type of aperture used different patterns are produced. Experimental data are show for (b) forked hologram, (c) triangular aperture (d) knife-edge (e) astigmatic aberration.

Type of presentation: Poster

IT-1-P-1541 Investigation of physical and chemical method to produce Möllenstedt electrostatic biprism for off-axis electron holography experiment

Cours R.1, Houdellier F.1
1CEMES-CNRS, Université de Toulouse, 29 Rue Jeanne Marvig, 31055 TOULOUSE FRANCE EU
robin.cours@cemes.fr

Denis Gabor has developed electron holography in 1948, as a method used to quantitatively retrieve the phase of the electron wave. D. Gabor proposed a configuration where the perturbed wave (object wave) and the reference unperturbed wave are observed in a common optical plane below the sample. In this plane a superimposition of the two waves can occur. This superimposition will induce an interference phenomenon and create the so-called in-line electron hologram, used to retrieve the phase difference between the two waves. In this configuration the sample is then out of focus. In 1955 G. Möllenstedt and H.Düker invented the biprism for electrons, a metallic wire biased relatively to the earth. The biprism effectively splits the electron beam into an object wave and a reference wave, which by electrostatic fields are brought to overlap onto one another. An interference pattern will be observed below the wire plane while the sample can still be in focus. This configuration, known as off-axis electron holography, is the one commonly used in all the major holography studies from dopant profiling to strain mapping through studies of nanomaterials magnetic configurations. Biprisms in common use today are constructed by coating ultrasmall quartz fibers with noble metals. The resulting biprisms, although they are quite small by most fabrication standards (approximately 700 nm in diameter), can have various mechanical, electrical, structural … properties. Depending on the quality of the biprism, the properties of the off axis hologram can be strongly modified. As an example, to avoid vibration, which drastically decrease the interference fringes contrast, the wire should be very taut; to minimize charge effect, which induce Fresnel fringes phenomena, the wire should be extremely clean; to increase the phase coherence of the beam across the biprism the wire should be the smaller possible, …

Regarding all these drastic requests that the wire should fulfilled to be a suitable biprism, the question of reproducibility become deeply problematic using standard biprism fabrication method. This question become even more crucial regarding our new microscope, the In situ interferometry TEM (I2TEM), a HF3300 TEM that fits with 4 biprisms wire used for various electron holography developments. In order to choose the most reproducible way which will give the best wire properties (size, vibrations, cleanliness, …), we have investigated several methods to produce them from chemical method to FIB (Focused Ion Beam) approach. The combination of these methods allowed us to make numbers of high quality biprism wire with a higher reproducibility rate.


This work has been supported by the French National Research Agency under the "Investissement d'Avenir" program reference No. ANR-10-EQPX-38-01.

Fig. 1: A: SEM image of a Wollaston wire thinned using a FEI Helios FIB B: The same wire installed inside a Hitachi HF2000 TEM

Fig. 2: C: Special and ultrafast chemical etching method using nanowetting of hot HNO3 onto a Wollaston wire

Type of presentation: Poster

IT-1-P-1672 Development of illumination system of a 1.2 MV Field Emission Transmission Electron Microscope

Kawasaki T.1, Kasuya K.1, Furutsu T.1, Ono S.1, Arai M.1, Moriya N.1
1Central Research Laboratory, Hitachi, Ltd., Hatoyama 2520,Saitama 350-03, Japan
takeshi.kawasaki.qb@hitachi.com

       In the FIRST Tonomura project, we have been developing a 1.2 MV field-emission transmission electron microscope (FE-TEM) for the atomic resolution three-dimensional reconstruction of electro-magnetic fields by electron holography. Here FIRST stands for funding program for world-leading innovative R&D on science and technology. In this paper we report its illumination system with the following requirements:
       (1) high brightness beam for electron holography
       (2) current fluctuation less than 10 % over 8 hours for stable observation
The requirement (1) is discussed in this presentation and the requirement (2) is discussed by Kasuya in this conference.
       Figure 1 shows schematic view of the illumination system and three ray paths. Separate valves are placed between the FE gun and the accelerator tube so that conditioning of emission and high-voltage can be performed separately. The pre-accelerating magnetic lens focuses the beam near the first electrode of the accelerator tube where the Butler lens is formed (Case A), and then the spherical aberration of the accelerator tube can be suppressed. When the magnetic lens excitation becomes stronger, the electron trajectory focuses twice in the accelerator tube (Case C). Between Case A and Case C, beams focus near the condenser lens and cannot focus on the specimen position (Case B). To obtain high brightness beam, total aberration of the illumination system has to be minimized. The optimum condition of the pre-accelerating magnetic lens was obtained by calculating mean brightness and probe current of the spot focused on the specimen position as a function of the lens excitation using WR5 software (MEBS Ltd.). The FE-cathode source diameter, the angular current density, and the energy spread are assumed to be 5 nm, 30 μA/sr, and 0.3 eV, respectively. Figure 2 shows the results. Two peaks of the brightness exist: The left peak corresponds to Case A, the right peak corresponds to Case C, and the bottom region D between two peaks corresponds to the Case B. Preliminary experimental results using the 1 MV FE-TEM showed the following:
       (1) existence of two brightness peaks
       (2) the maximum brightness of 1.8×1010 A/cm2sr [1]
This brightness value is almost the same as that calculated for the 1 MV FE-TEM. The calculated maximum brightness is 3.3 ×1010 A/cm2sr for the 1.2 MV FE-TEM. We expect it to reach 5×1010 A/cm2sr by increasing the angular current density of the cleaner FE-tip under ultra high vacuum condition (3.0×10-10 Pa) [2].

References
[1] T. Kawasaki et al.  J. Electron Microsc. 49 (2000) 711-718.
[2] K. Kasuya et al. submitted to  J. Vac. Sci. Technol. B.

 


This research was supported by the Japan Society for the Promotion of Science through the FIRST Program initiated by the Council for Science and Technology Policy.

Fig. 1: Schematic view of the illumination system and different ray paths A, B, and C.

Fig. 2: Calculated brightness and probe current as a function of the excitation of the magnetic lens in terms of IN(V1)-1/2 , where I is the lens current, N is the number of turns of the coil (1700), and V1 is the FE extraction voltage    (5 kV).

Type of presentation: Poster

IT-1-P-1693 Measurement of current density distribution in shaped e-beam writers

Horáček M.1, Bok J.1, Kolařík V.1, Urbánek M.1, Matějka M.1, Krátký S.1
1Institute of Scientific Instruments AS CR, v.v.i., Brno, Czech Republic
mih@isibrno.cz

The ZrO W(100) Schottky cathode is used in our e-beam writing system working with a rectangular-shaped electron beam. The homogeneous angular current density distribution is crucial for quality of exposures of the shaped beam lithography systems. Two basic types of the angular emission distribution can be observed in dependence on the microscopic final end form shape of the emitter tip, with bright centre and more common dark centre [1]. The stable operation of the cathode thus stable end form shape requires a delicate balance of parameters inside the gun which however can slightly change during cathode life time. This implies the necessity of analysing and periodical monitoring the current density distribution in e-beam. Four methods enabling this measurement are presented.
First we implemented a method based on the modified knife-edge approach [2], when a part of the scanned element of the beam is blanked out and the current within the remaining "open" part is measured. The 2D information of the current distribution is obtained by stepwise opening of selected segments. The measurement error analysis was made and necessary measurement averaging in each segment were used in order to reduce the random error of the current [3]. The size of the scanned element was 6 × 6 µm2, a maximum usable segment for one shot in our lithography system (Fig. 1).
The current distribution obtained by the knife-edge method was compared with a method using a luminescent screen. The YAG:Ce single-crystal screen was irradiated by the e-beam stamp of the 6 × 6 µm2 and the areal light emission was recorded by a magnifying optical system with a CCD camera. The emitted light intensity is directly proportional to the e-beam current, thus the current density distribution can be compared with other measurements methods. However, the absolute measurement is hardly possible (Fig. 2).
Next the same e-beam stamp of the 6 × 6 µm2 was scanned over Faraday cup opening. The advantage of this method is uniform distribution of the measurement error instead of the modified knife-edge method. The absolute value of the current density is affected by the demagnification of the electron optics during measurement (Fig. 3).
Another method is based on evaluation of developed electron resist exposed by the 6 × 6 µm2 separate shaped e-beam stamp using atomic force microscope. The depth of the developed resist depends on the spread of the energy in the electron resist. The real current density distribution was obtained by the deconvolution of the developed resist with electron scattering model (Fig. 4).

References

[1] K-Liu et al., J. Vac. Sci. Technol. B 28, C6C26 (2010).
[2] M. Sakakibara et al., Jpn. J. Appl. Phys. 46, 6616 (2007).
[3] J. Bok et al., J. Vac. Sci. Technol. B 31, 31603-1 (2013).


The authors acknowledge the support from MEYS CR (LO1212) together with EC (ALISI No. CZ.1.05/2.1.00/01.0017), the TACR project No. TE01020118 and institutional support RVO:68081731.

Fig. 1: Modified knife-edge method.

Fig. 2: Luminescent screen method.

Fig. 3: Faraday cup method.

Fig. 4: Electron resist exposure method.

Type of presentation: Poster

IT-1-P-1958 Quantitative measurement of the OAM spectra of electron vortex beams.

Clark L.1, Béché A.1, Guzzinati G.1, Verbeeck J.1
1EMAT, University of Antwerp, Antwerp, Belgium
laura.clark@uantwerp.be

Electron vortex beams have been subject to a great level of interest since their first demonstration only a few years ago [1]. Much of the interest in the field stems from their potential to measure magnetic transitions within a sample, at a previously unreachable scale. While much progress has been made, in producing electron vortices of high purity, high intensity and atomic scale, research into the required counterpart towards full experimental application, of orbital angular momentum (OAM) measurement, has not yet matured to its full potential [2-4].

In the last 12 months, the first methods to measure the OAM make-up of an electron vortex beam have been demonstrated [5-7]. However, the methods presented thus far, are limited to only those cases where the input beam is in a single vortex state, and do not allow measurement of the relative weightings of vortex states in a beam . Indeed, a generic electron wave can be seen as a superposition of multiple vortex modes and the weight of each of these modes can in principle be measured.

We introduce here an experimental technique able to measure the relative weightings of 5 or more OAM modes within an input beam, through the use of a multi-pinhole interferometer (MPI). This is a technique which has recently been used to measure the strength and location of optical vortices, but which is easily adaptable to practical implementation in a TEM, placing an MPI aperture in the SA plane, below the sample.

Experimental results are shown, having measured the OAM spectrum of pure l={-1,0,+1,+2} centred vortex beams, enabling the first quantitative discussion of their experimental purity. We further demonstrate the so-called mode broadening effect, by measuring the changes in OAM composition as a vortex beam is shifted away from the central axis of measurement.

This application of an MPI within a TEM has enabled measurement of an approximate OAM spectrum in the SA plane. We give experimental evidence alongside theoretical models, enabling rapid discrimination of different orders of vortex beams even if the electron beam consists of a superposition of different OAM modes. This capability serves as a promising tool to measure OAM exchanges in the interaction of electrons with a sample.

[1] Bliokh, KY, et al. PRL 99.19 (2007): 190404
[2] Verbeeck, J., et al. Nature 467.7313 (2010): 301-304
[3] Clark, L., et al. PRL 111.6 (2013): 064801
[4] Béché, A, et al, Nature Physics 10.1 (2014): 26-29
[5] Guzzinati, Giulio, et al. arXiv: 1401.7211 (2014)
[6] Saitoh, K, et al. PRL 111.7 (2013): 074801
[7] Shiloh, Roy, et al. arXiv: 1402.3133 (2014)


We acknowledge funding from the European Union under the FP7 program: ERC Starting Grant No. 278510-VORTEX and Integrated Infrastructure Initiative Reference No. 312483-ESTEEM2.

Fig. 1: Plot of idealised electron vortex beam – brightness represents intensity, and hue represents phase

Fig. 2: A five-pinhole multi-pinhole interferometer, enabling measurement of OAM modes in the set l={-2:+2}

Fig. 3: Experimental diffraction pattern from an l=+1 vortex centred on the MPI

Fig. 4: Autocorrelation function produced from the experimental diffraction pattern.

Type of presentation: Poster

IT-1-P-1961 Development of Phase Contrast Scanning Transmission Electron Microscopy

Iijima H.1, Minoda M.2, Tamai T.2, Kondo Y.1, Hosokawa F.1
1EM Business Unit, JEOL Ltd., 3-1-2 Musashino, Akishima, Tokyo 196-8558, Japan, 2Department of Applied Physics, Tokyo University of Agriculture and Technology, 2-24-16 Nakacho, Koganei, Tokyo 184-8588, Japan
hiiijima@jeol.co.jp

Phase contrast transmission electron microscopy (P-TEM) is a powerful tool to enhance the image contrast of transparent materials such as ice-embedded biological specimens and polymer materials. In P-TEM, a phase plate is placed at the back-focal plane (BFP) of the objective lens (OL). It gives a phase shift for scattered electron waves, resulting in a change of phase contrast transfer function (PCTF) from sine to cosine type. Eventually, phase variation of specimens is converted into intensity variation. Among various types of phase plates, a carbon film phase plate with a small central hole is the most practical1. However, there is a serious issue that high-density electron beam (cross-over) on the phase plate causes the charging and/or the alteration of the phase plate, resulting in decreasing the life time of the phase plate.

To overcome this issue, we are developing phase contrast scanning transmission electron microscopy (P-STEM). Figure 1 shows the schematics of P-TEM and P-STEM. According to the reciprocity theorem, the same contrast appears in the P-TEM and the P-STEM if a phase plate is placed at a front-focal plane (FFP) of an OL in P-STEM. In P-STEM, a cross-over is not formed on the phase plate, so that improvement of the phase plate life time is expected. In our experiments, we used a field emission electron microscope (JEM-2100F) equipped with a Schottky electron source, to obtain a coherent small probe on a specimen. Phase plate is placed on a condenser lens aperture plane conjugate to the FFP of the OL.

On the other hand, it is well known that the small detection angle is needed to obtain good phase contrast in STEM imaging. Figure 2 compares a conventional bright-field STEM and a P-STEM images of amorphous carbon film with different detection angle shown in Fig. 1. And Fourier transforms of the conventional bright-field STEM image and the P-STEM image with β = 4 mrad show the sine shape. By contrast, that of the P-STEM image at β = 0.3 mrad shows the cosine shape, which proves that the P-STEM can be achieved with small detection angle.

[1] R. Danev and K. Nagayama, J. Phys. Sci. Jpn. 70 (2001) 696.


This development was supported by the program for "Development of Systems and Technologies for Advanced Measurement and Analysis" under JST.

Fig. 1: Schematic of P-TEM (left) and P-STEM (right). The phase plate is placed at the BFP of the objective lens in P-TEM and the FFP of the objective lens in P-STEM.

Fig. 2: Conventional bright-field STEM and P-STEM images of amorphous carbon film. All images are taken close to focus. (a) Conventional bright-field STEM image. (b) P-STEM image with β = 4 mrad. (c) P-STEM image with β = 0.3 mrad. (d)-(f) Fourier transforms for images shown above. Scale bars; 10 nm in (a)-(c), 4 nm-1 in (d)-(f).

Type of presentation: Poster

IT-1-P-2090 Contrast enhancement of phase objects by using Phase Contrast Scanning Transmission Electron Microscopy

Minoda H.1, Tamai T.1, Iijima H.2, Hosokawa F.2, Kondo Y.2
1Department of Applied Physics, Tokyo University of Agriculture and Technology, 2EM Business Unit, JEOL Ltd
hminoda@cc.tuat.ac.jp

It is well known that an interaction between electron waves and molecules composed of light elements such as biological molecules is very weak. Therefore, it is very difficult to obtain their high contrast image in transmission electron microscopy (TEM). Contrast enhancement of the phase objects by using a phase plate was proposed at the middle of the 20th century [1], but it was realized at the beginning of 21st century [2]. In the pioneering work by Nagayama, a carbon thin film with a hole in its center is used as a phase plate (PP) and it was placed at a back focal plane (BFP) of the objective lens (OL). A role of the PP is giving a phase shift to scattered wave by means of the mean inner potential of the PP material. Electron waves having a phase shift interfere with electron waves without phase shift. Accordingly, phase image would be able to be visualized.

Applying the principle of reciprocity to scanning transmission electron microscopy (STEM), imaging optics of the STEM is equivalent to that of a conventional TEM. Therefore, a phase contrast scanning transmission electron microscopy (P-STEM) can be used to enhance phase contrast of the phase objects. In the present study, a PP can be set on the condenser lens aperture (CLA) plane that is optically equivalent to a front focal plane (FFP) of an OL. The P-STEM image which enhances image contrast could be obtained by getting an appropriate optical condition. Figure 1 show an example of the comparison of (a) the conventional STEM bright field image and (b) the P-STEM image. Ferritin molecules were used as a specimen. This comparison clearly shows contrast enhancement in P-STEM. In this paper, the results obtained by sung phase contrast microscopy to the STEM mode are introduced.

[1] F. Zernike, Physica 9 (1942) 686.

[2] R. Danev and K. Nagayama, J. Phys. Sci. Jpn. 70 (2001), 696.


This development was supported by SENTAN, JST.

Fig. 1: A comparison of (a) C-STEM and (b) P-STEM images of ferritin molecules.  The contrast enhancement in P-STEM is evident.

Type of presentation: Poster

IT-1-P-2325 AC- Voltage Operated Schottky Electron Source

Yada K.1, Saito Y.1
1Daiwa Techno Systems Co.,Ltd
yada@daiwatechno.co.jp

Introduction: Zr-O/W100 Schottky electron source has been widely used in electron beam instruments because of its favorable properties such as 1)vacuum technological tolerance at its operational condition, 2) rather long life time and high brightness. It is still required, however, that a vacuum must be better than 10 -8 Pa and stability of high voltage must be better than 10 -5 when DC high tension and electro-magnetic lens systems are used in the instrument. We tried to find promising materials. Among them, we selected BaZrO3 and SrZrO3 and tested their thermal field emission properties with both DC and AC high tension powers.

Results: Field emission tips of 110- and 100-oriented W wire were made by electrolytic method and powder of BrZrO3 or SrZrO3 was pasted near the apex as usual. Thermal field emission patterns obtained by DC and AC voltage are very similar and crystal facets are indexed very easily. Fig.1 and Fig2 show emission patterns of BaZrO3/W(110) and BaZr3(W100) obtained by AC and DC operation ,where optimal working temperature is 800 degree C. Similar results are obtained with SrZrO3(W)100 cathode but its optimal temperature is little higher than the case of BrZrO3.

As advantages of AC operation of present schottky electron source, followings are concluded:

1)Schottky shield is not necessary because of low working temperature of the emission materials.

2)Emission beams can be focused, deflected and stigmated by using electrostatic lens,deflector and stigmtor, respectively.

3)Work function of newly adopted materials here is so small that working temperature is fairly low (800-850 degreeC). Consequently ,energy spread of the beam will be narrow.

4) As commercial AC electric supply can be used without any rectifire or stabilizer, factor cost will be fairly reduced.

5)We think that the present Schottky electron source is the best selection for a generation of strong and small X-Ray source of projection X-ray microscope.


Fig. 1: Fig.1

Fig. 2: Fig.2

Type of presentation: Poster

IT-1-P-2198 The effect of Detector Thickness on Direct Detector Performance

Clough R. N.1, Moldovan G.2, Kim J. S.1, Kirkland A. I.1
1Department of Materials, University of Oxford, UK, 2Oxford Instruments NanoAnalysis, High Wycombe, UK
robert.clough@materials.ox.ac.uk

Direct detection refers to a detection system where signal is generated in the sensor chip directly by the imaging electrons; indirect systems generate photons in a scintillator from the imaging electron and it is these photon which are coupled to the sensor chip that generate signal. One of the key advantages of a direct detection system is the possibility of producing thin detectors; these are desirable as a thin detector has improved detection performance in terms of Modulation Transfer Function (MTF) and Detective Quantum Efficiency (DQE) [1]. This improvement arises from the fact that many electrons will pass all the way through the sensor and escape the detector system generating signal along the way, before large lateral scattering has occurred.

We have taken a prototype CMOS based direct detector featuring full frame resolution of 1024 by 1024 pixels, with a pixel size of 20µm and readout of 30fps [2]. Two different versions of the detection chip were produced. The first is a 20µm thick p- active layer on a p+ substrate mechanically thinned to 50µm. The second was made from silicon on insulator (SOI) wafer with a 20µm device layer with the handle wafer removed using a chemical etch. For each of these detectors the MTF and DQE were measured using standard techniques [3] at 80 and 200kV. Here we shall present the characterisation data along with images of gold particles on an amorphous substrate to show how thinner detectors lead to improved detector performance, allowing images taken at lower magnification to have improved resolution.

[1] G. McMullan, et al, Experimental observation of the improvement in MTF from backthinning a CMOS direct electron detector, Ultramicroscopy, 109 (2009).
[2] A.J. Wilkinson, et al, Direct Detection of Electron Backscatter Diffraction Patterns, Phys. Rev. Lett. 111 (2013).
[3] R. R. Meyer, et al, Experimental characterisation of CCD cameras for HREM at 300kV, Ultramicroscopy, 85 (2000).


We would like to acknowledge Dr T. Anaxagoras and Prof. N. Allinson from the University of Sheffield for provision of CMOS wafers, and C Wilburn of Micron Semiconductor Ltd. for chip packaging.

Fig. 1: MTF of a 20µm thick detector at 80 and 200kV.

Fig. 2: Au on amorphous Carbon at 80kV and 120,000x magnification taken with a 20µm thick sensor chip.

Fig. 3: Au on amorphous Ge at 200kV and 120,000x magnification taken with a 20µm thick sensor chip.

Type of presentation: Poster

IT-1-P-2263 Maximising Phase Contrast in Aberration-corrected STEM using Pixelated Detectors

Yang H.1, Pennycook T. J.1,2, Nellist P. D.1,2
1University of Oxford, Department of Materials. Parks Rd, Oxford, OX1 3PH, UK, 2EPSRC SuperSTEM Facility, Daresbury Laboratory, WA4 4AD, UK
hao.yang@materials.ox.ac.uk

For imaging weak phase biological specimens, phase contrast imaging using elastically scattered electrons provides the most information for a given amount of radiation damage as compared to electron inelastic scattering as well as X-ray and neutron scattering [1]. In scanning transmission electron microscopy (STEM), most phase information from weak scattering objects lies inside the bright field disc of the convergent beam electron diffraction pattern, which can be reconstructed using the method described by Rodenburg et al [2]. In this work we show that, compared to alternative modes including annular bright field (ABF) and differential phase contrast (DPC), phase contrast using a pixelated detector generates higher contrast in reconstructing the phase and therefore enjoys a higher dose efficiency in imaging weak phase objects.

With zero aberrations, any centrally symmetric detector will give no contrast for a weak phase object, as the two sides of disc overlapping regions in the convergence beam electron diffraction pattern are pi out of phase under weak phase approximation, and cancel each other when integrated using a central symmetrical detector geometry. Therefore, asymmetric detector geometries like DPC are expected to have higher phase contrast than ABF. In DPC, the quadrant detector can be divided into more segments with different collection angles, and the contrast transfer function is found to depend on the collection angles used, therefore the detector geometry of DPC can be further optimized to collect the maximum phase information per detected electrons. A pixelated detector provides even greater flexibility over where the information in the bright-field disc is retrieved from for each spatial frequency in the image.

Simulations have been done using an arbitrary weak phase specimen whose maximum atomic potential equals to that of a carbon atom, and has a Gaussian shape with a full width half maximum (FWHM) of 1nm. The artificially high width of the object is designed to test the lower spatial frequency transfer. The reconstructed phase with a dose as low as 50 electrons/Å2 and Nyquist resolution of 4.6Å still shows an interpretable feature (Figure 1). This dose is close to the critical dose of 5-50 electrons/Å2 for imaging biological specimen. In contrast to using a pixelated detector, neither ADF, ABF (Figure 2) nor DPC (Figure 3) show any recognizable structure feature under the same dose of 50 electrons/Å2. The formation of image contrast in ABF relies the presence of aberrations for a weak phase object, and here we are assuming an aberration-corrected microscope with zero residual aberrations.

[1] Henderson, R. Quarterly Reviews of Biophysics 28, 171-193 (1995).

[2] Rodenburg, J. M. et al. Ultramicroscopy 48, 304-314 (1993).


The authors would like to acknowledge financial support from the EPSRC (grant number EP/K032518/1) and the EU Seventh Framework Programme: ESTEEM2.

Fig. 1: Figure 1: Phase retrieval using a pixelated detector. (a) Schematic of a high speed pixelated detector. (b) A weak phase object with a maximum phase change of 0.15 radian. Reconstructed phase (c) assuming no noise, (d) with shot noise and a dose of 50 electrons/Å2. The scale bar is 5nm.

Fig. 2: Figure 2: Simulated (a,b) ADF and (c,d) ABF images of the weak phase object. The intensity is normalized to the number of incident electrons. (a)(c) assume no noise in image, and (b)(d) consider shot noise with the electron dose being 50 electrons/Å2. The scale bar is 5nm.

Fig. 3: Figure 3: Differential Phase Contrast (DPC) imaging using a 4-quadrants detector in (a). The simulated STEM DPC images using both (b,c) A-C quadrants, and (d,e) B-D quadrants, where (b)(d) assume noise free, and (c) (e) considers shot noise with the electron dose being 50 electrons/Å2. The scale bar is 5nm.

Type of presentation: Poster

IT-1-P-2346 Effects of dielectric substrate on localized surface plasmon in a silver nano-particle

Fujiyoshi Y.1, Nemoto T.1, Kurata H.1
1Institute for Chemical Research, Kyoto University, Kyoto, Japan
fujiyoshi@eels.kuicr.kyoto-u.ac.jp

Recently localized surface plasmons (LSPs) which are collective oscillation of conduction electrons of metallic nano-particles (NPs) attract researchers in nano-optics because of strong optical confinement and electric field enhancement, leading to many applications including biochemical sensors and surface-enhanced Raman spectroscopy (SERS) etc. Since the dielectric environment around the NP affects the property of LSPs, it is important to elucidate the effects of dielectric materials supporting NPs on LSPs.
In the present work, we examined special distributions of LSP excited on a silver NP supported by MgO substrate using electron energy loss spectroscopy (EELS) combined with scanning transmission electron microscopy (STEM). Spectral imaging (SI) data were acquired along the direction parallel to the MgO surface supporting a silver NP, which enabled us to observe the intensity distribution of LSP excitation as a function of the distance from the silver NP/MgO interface. The experiment was performed by an aberration corrected STEM (JEM-9980TKP1) equipped with a cold-FEG.
Figure 1 and 2 show a HAADF image of silver NP on MgO substrate and its LSP map extracted from SI data, respectively. From the HAADF image the NP can be regarded as a sphere. When a spherical metal particle is isolated in vacuum, the excitation probability of LSP should distribute isotropically around the particle. However, the LSP map in Fig. 2 shows anisotropic distribution, that is, the intensity at the top surface of silver NP is strong compared to that at other positions, which means that the effect of dielectric substrate is remarkable. In order to interpret such anisotropic distribution, we simulated the electromagnetic field induced in the silver NP on MgO substrate using finite-difference time-domain (FDTD) method.
Figure 3 shows the spatial distribution of field calculated by assuming the incident plane waves polarized perpendicular (a) and parallel (b) to the substrate surface. When the polarization of incident wave is perpendicular to the substrate, the field strength in the NP on MgO is enhanced compared to that in the isolated NP as shown in Fig. 3(a), which corresponds to the observed strong excitation at position A in Fig. 2. In case of the parallel polarization the field strength in the NP on MgO is weakened (Fig. 3(b)), corresponding to the observed intensity at position B in Fig. 2. Therefore, the anisotropic distribution of the LSP excitation in silver NP on MgO surface can be attributed to the direction of electric polarization induced in the NP depending on the electron positions.


Fig. 1: HAADF image of a silver NP supported on MgO surface.

Fig. 2: LSP map extracted from the energy range from 3.2 to 3.6 eV in the SI data.

Fig. 3: Spatial distribution of electromagnetic field calculated by FDTD simulations. Incident plane waves were assumed to be polarized parallel (a) and perpendicular (b) to the substrate surface. Solid and broken lines correspond to the intensity profiles for an isolated silver NP and the silver NP supported on MgO surface, respectively.

Type of presentation: Poster

IT-1-P-2410 Thermal Emission Properties of GdB6 Cathode

Saito Y.1, Yamagishi K.1, Yada K.1
1Daiwa Techno Systems Co.,Ltd.
saito@daiwatechno.co.jp

Introduction: Hexa boride of lanthanum (LaB6) has been widely used in electron beam
instruments because of its higher brightness than that of tungsten hairpin cathode. But it might
be that there are better materials than LaB6. Among many borides of lanthanide, hexa-borides
of Ce and Gd are promising from the existing data[1] based on Richardson-Dushman equation as
shown in Table 1. So we tested electron emission properties of GdB6.


Table 1
              A      φ(eV)       AT2         I(A/cm2)
LaB6      29      2.66     93960000   3.316363
GdB6    0.84     2.06     2721600     4.607963
GdB6     9.3      2.55     30132000   2.16234
GdB6      10      2.58     32400000   1.915988


R-D Eq. I=AT2EXP(-φ/kT), A:R-D constant, k:Boltsman cont, T:temperature(1800k), I:electron
density


Results: Fig.1 shows photograph of Ta wire covered with GdB6 powder where the central part is
slightly protruded. Fig.2 is beam pattern of GdB6 cathode at working temperature when the
cathode is installed in a scanning electron microscope. Fig.3(a) shows SEM image of ZnO
particles obtained with tungsten hairpin cathode and Fig.3(b) shows that obtained with GdB6
cathode. It is seen that image quality of (b) is superior to that of (a). It is also clear that emission
performance of present GdB6 powder cathode is nearly equal or little better than that of LaB6
single crystal. Sintered GdB6 cathode is now under examination to compare with the single
crystal LaB6 cathode.


Reference:[1]Japan-Soviet Communication “Emission Characteristics of Materials” pp.96-81 by
V.S.Fomenko, Published by Naukova Dunka, Kiev 1970


Fig. 1: Fig.1

Fig. 2: Fig.2

Fig. 3: Fig.3(a)

Fig. 4: Fig.3(b)

Type of presentation: Poster

IT-1-P-2469 Electron differential phase microscopy with an A-B effect phase plate

Tanji T.1, Ikeda U.2, Niimi H.2, Usukura J.1
1EcoTopia Science Institute, Nagoya University, Nagoya, Japan 1, 2Graduate School of Nagoya University, Nagoya, Japan 2
tanji@esi.nagoya-u.ac.jp

    Observations of week phase objects, such as thin films of light elements, thin polymer films, biological sections etc., are available by electron phase microscopy[1]. Many of phase plates utilized are thin film types. Some electrostatic types have been developed, but they are not so general, because the fabrication of the filter with fine structures is very difficult. The mainstream of todays phase plate is the thin film type. This type of the phase plate, however, has some disadvantages, i.e. control of the film thickness, charging up, contamination and so on. We adopted the phase plate with a magnetic thin filament which generates the vector potential around itself by an Aharonov-Bohm (A-B) effect. The filament type phase plate with the A-B effect was proposed and constructed firstly by Nagayama. This type of the phase plate generates the differential phase contrast in the image, and has a longer life time than the thin film type. Any clear differential effect, however, has scarcely reported so far.
    We will report that the effect of a phase plate consisting of a Wollaston platinum filament of 1 µm in diameter covered with ferromagnetic material, Nd-Fe-B of 5 nm thick, deposited by Pules Laser Deposition. The filament with a clean surface selected by SEM is mounted on a single hole Cu grid. The phase difference in the both side spaces of the filament measured by electron holography shows 1.5 rad as shown in Fig.1. Being set on the aperture holder, the phase plate is inserted in the back focal plane of the objective. Figure 2 shows images of a colon bacillus stained with Pb. Fine structures can be observed clearer in the image using the phase plate than in the image taken ordinarily. The direction of the differentiation is shown by the arrowhead.

Refernce

[1]K. Nagayama, Another 60 years in electron microscopy: development of phase-plate electron microscopy and biological applications, Journal of Electron Microscopy, 60(2011) S43-S62.


Fig. 1: (a) Electron phase map reconstructed by electron holography. (b) A line profile along the arrow head in (a) which is averaged along the long side of the rectangle. The phase difference is about 1.5 rad. between both sides of the filament.

Fig. 2: mages of a colon bacillus stained with Pb taken at under-focus condition without the filament(a) , and in-focus with the filament(b).

Type of presentation: Poster

IT-1-P-2476 Effect of a phase plate on TEM imaging

Edgcombe C. J.1
1TFM Group, Dept of Physics, University of Cambridge
cje1@cam.ac.uk

The type of phase plate that has been most widely reported (eg [1]) consists of a plain disc of material such as carbon, of controlled thickness, with a central hole to pass the direct beam. Images made with this type of plate show bright outlines or halos around certain features [2, 3].  Analysis of geometrical imaging has shown how these halos occur.  It is necessary to consider the response to all spatial frequencies that are present in a typical object. In principle this can be done straightforwardly by Fourier transforming the object phase to find its spatial frequency distribution at the back focal plane (BFP), multiplying by the response of the phase plate and further transforming to find the image distribution.

The response has been found [4] for a weak phase object consisting of a circular disc of radius b, centred on the microscope axis. The phase plate is assumed to advance the phase of components with angular frequencies greater than a value q0 , defined as
q0 = 2 (pi) r2 ⁄ λf
where r2 is the radius of the central hole in the plate for the direct beam, λ is the electron wavelength and f is the focal length of the lens. The resulting image intensity is shown in figure 1 for a phase advance of (pi)/2 and a range of values of B = q0b. The object is imaged with little overshoot when B is less than about 1. Reported results [5] agree with this transition value for B.

The step changes at radius b are always imaged fully but as B increases, the low-frequency components are progressively lost from the image and for B > 1, the mean intensity across a step falls to the background value.  The full range of the step is maintained, so the intensity changes from +(half the range) at radii just less than b, to –(half the range) just outside the step. Thus a bright halo or outline is produced just outside the boundary r = b, for objects with B > ~1.  The darker central patch for B = 8 agrees with observation [3]. The maximum object diameter that corresponds to Bmax , the maximum B for accurate imaging, is
2bmax = 2Bmax ⁄ q0 = λ Bmax f ⁄ (pi) s2
where s2 is the radius of the hole needed to pass the direct beam.  To increase the size of object that can be imaged accurately, it will be necessary to reduce s2 or increase the focal length of the objective lens.

References
Danev R and Nagayama K 2001 Ultramicroscopy 88 243-52
Fukuda Y, Fukazawa Y, Danev R, Shigemoto R and Nagayama K 2009 J Struct Biol 168 476-84
Danev R and Nagayama K 2011 Ultramicroscopy 111 1305-15
Edgcombe C J 2014 Ultramicroscopy 136 154-9, http://dx.doi.org/10.1016/j.ultramic.2013.09.004
Hall R J, Nogales E and Glaeser R M 2011 J Struct Biol 174 468-75


Fig. 1: Figure 1. Image intensities produced by a (pi)/2 phase plate with fixed q0 for uniform disk objects with a range of diameters 2b.  Responses are shown for values of B = q0b (increasing from top left) of 0.2, 0.5, 1, 2, 5 and 8.

Type of presentation: Poster

IT-1-P-2510 Toward electron polarizators

Grillo V.1,2, Karimi E.3, Balboni R.4, Gazzadi G. C.1, Frabboni S.1,5, Mafakheri E.1,5, Tang W. X.6,7, Boyd R. W.3,8
1CNR-Istituto Nanoscienze, Centro S3, Via G Campi 213/a, I-41125 Modena, Italy, 2CNR-IMEM, Parco delle Scienze 37a, I-43100 Parma, Italy. , 3Department of Physics, University of Ottawa, 150 Louis Pasteur, Ottawa, Ontario K1N 6N5, Canada, 4CNR-IMM Bologna, Via P. Gobetti 101, 40129 Bologna, Italy, 5Dipartimento FIM, Universitá di Modena e Reggio Emilia, Via G. Campi 213/a, 41125 Modena, Italy, 6College of Materials Science and Engineering, Chongqing, 400044, China, 7School of Physics, Monash University, Clayton, VIC, 3800, Australia, 8Institute of Optics, University of Rochester, Rochester, New York 14627, USA
vincenzo.grillo@cnr.it

We describe the experimental and theoretical improvements toward the realization of an efficient electron spin polarizator. The initial proposed polarizator [1] was based on the spin-orbit conversion of a vortex beam [2] to a beam with a defined polarization. The conversion occurred within a compensated quadrupolar Wien Filter (WF).

The theoretical improvements are supported by simulations of the beam-field interaction through a new multislice for propagation including spin [3]. The experimental steps are based on the introduction of phase holograms to produce e-beams close to ideal Bessel beams [4]. To improve the flexibility and feasibility of the polarizer we have considered different possible alternative design: e.g. when the pitch fork hologram is positioned below the WF it is possible to obtain simultaneously the 2 polarized beams and switch between them [3]. Alternative design permit also to remove the electric fields. We have also studied the higher order corrections of the WF by magnetic multipoles of higher order and calculated the possible effects of the fringing fields: the efficiency in the selection of the polarized states increases with the order of the vortex and consequently of the multipoles in the WF.

Fig 1 is an example of simulation of the wavefunction after a WF for a beam at 15 KeV (e.g. for SPLEEM and low voltage TEM applications ) for 2 initial spin state. The brightness is proportional to the wave intensity, the phases encoded in the color. Due to the spin orbit coupling different spin are transformed, inside the WF, in different phase factors and orbital momentum. Only the center of the state |ℓ=0,↑> has stationary phase and therefore contributes to the intensity at the center of a pupil in far field diffraction.

For this simulation we corrected the asymmetric aberrations by multipolar elements but still obtained a strong phase oscillation beyond a radius dependent of the size of the field that must be further corrected to obtain maximal efficiency.

Fig 2 a,b is an example of phase hologram described in its thickness map and overall pattern. This pattern reaches an efficiency of 40%. In fig 2c an example of Bessel beam with ℓ=2 is shown. These beams, in the diffraction plane (see fig d), transform to narrow rings. This strongly reduce the demand of lateral stability of the fields and the problems of phase oscillations described in fig. 1

[1] E. Karimi et al. Phys Rev. Lett 108, 044801 (2012)
[2] J. Verbeeck et al Nature 467, 301 (2010).
[3] E. Karimi et al Ultramicroscopy 138, 22 (2014)
[4] V. Grillo et al. Phys. Rev. X 4, 011013 (2014)


Fig. 1: Wavefunction after a Wien filter for a beam at 15KeV. The initial beam had ℓ=1 and 2 spin states were considered. The final spin state are also separately plotted. The external phase oscillation are due to residual aberrations.

Fig. 2: Example of phase hologram: the thickness profile in Si3N4 a) and the full pattern b)are shown. Example of Bessel beam with ℓ=2 in the Fresnel c) and Fraunhofer d) regime

Type of presentation: Poster

IT-1-P-2578 Design of a monochromator for aberration-corrected low-voltage (S)TEM

Mukai M.1, Omoto K.1, Sasaki T.1, Kohno Y.1, Morishita S.1, Kimura A.1, Ikeda A.1, Somehara K.1, Sawada H.1, Kimoto K.2, Suenaga K.3
1JEOL Ltd., Akishima, Tokyo, Japan, 2National Institute for Material Science (NIMS), Tsukuba, Ibaraki, Japan, 3National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki, Japan
mmukai@jeol.co.jp

Low-voltage analytical electron microscope equipped with delta-type aberration correctors for image- and probe-forming lens system [1] was developed under a project “Triple-C phase-1” to study the atomic structures of carbon materials sensitive to the damage by irradiation of electrons. It enabled us to reveal the characters of graphenes by EELS [2] and to visualize and specify an encapsulated single metal atom in a fullerene [3]. However this microscope was equipped with a cold field emission gun to obtain high brightness therefore its energy resolution remains at approximately 0.3 eV.
For the next challenges, we have started to develop a new type of low-voltage aberration-corrected analytical electron microscope equipped with a monochromator working at 15-60 kV under a project “Triple-C phase-2”, whose targeted energy resolution is better than 25 meV. Fig. 1(a) and 1(b) show an appearance of the microscope and a configuration of components inside the cover.
The developed monochromator employs a double Wien-filter system, arranged between the extraction anode of Schottky source and the accelerator, which is similar configuration to previous design [4]. The electron trajectories from the electron source to the plane of the exit crossover of the monochromator are calculated as shown in Fig. 2. Electron trajectories are set to be symmetric to the plane of energy-selection slit so that the energy-dispersion formed by the first Wien-filter at a slit plane is cancelled by the second Wien-filter at an exit plane as a consequence of the double Wien-filter system. Thus, after the monochromator, the electron probe is achromatic and the energy spread is controllable by choosing the width of the slit, independently on the probe size. In addition, the setting of the monochromator and the electron trajectories inside the monochromator are independent of the change of the accelerating voltage since the accelerator of the electron gun is located after the monochromator and the potential along the optical axis inside the monochromator is kept constant.
We intend to evaluate the performances of the developed low-voltage monochromated electron optical system and the enhancement of spatial resolution arising from a small chromatic aberration in TEM at low accelerating voltage with large scattering cross-section and small specimen damage by reducing a primary electron energy.

References
[1] H. Sawada, et al.: J. Electron. Microsc. 58 (2009) 341.
[2] K. Suenaga and M. Koshino, Nature 468 (2010) 1088.
[3] K. Suenaga, et al.: Nature chemistry 1 (2009) 415.
[4] M. Mukai, et al.: Ultramicroscopy (2014) accepted.


This work is supported by Japan Science and Technology agency, Research Acceleration Program.

Fig. 1: Computer graphics of a low-voltage aberration-corrected analytical electron microscope equipped with the developed monochromator, (a) An appearance with the cover and (b) a column of the microscope inside the cover.

Fig. 2: (a) Calculated trajectories along optical axis from source to exit of monochromator, (b) beam shapes at slit plane with an energy-dispersed 1st focus and (c) beam shapes at exit plane with an achromatic 2nd focus. The red lines and the green lines show the trajectories having different energies of 1 eV inside of the monochromator.

Type of presentation: Poster

IT-1-P-2604 Holographic generation of Electron quasi-Bessel beams

Frabboni S.1,2, Grillo V.2,3, Karimi E.4, Balboni R.5, Gazzadi G. C.2, Mafakheri E.1,2, Boyd R. W.4,6
1Dipartimento FIM, Università di Modena e Reggio Emilia, Via G. Campi 213/A, 41125 Modena, Italy , 2CNR-Istituto Nanoscienze, Centro S3, Via G Campi 213/a, I-41125 Modena, Italy, 3CNR-IMEM, Parco delle Scienze 37a, I-43100 Parma, Italy. , 4Department of Physics, University of Ottawa, 150 Louis Pasteur, Ottawa, Ontario K1N 6N5, Canada, 5CNR-IMM Bologna, Via P. Gobetti 101, 40129 Bologna, Italy, 6Institute of Optics, University of Rochester, Rochester, New York 14627, USA
stefano.frabboni@unimore.it

Recently the attention of electron microscopy community has been attracted by the generation of electron beams by means of holographic element that allows to shape the electron wavefront through a modulation of the phase or amplitude transmittance. This new degree of freedom has already demonstrated huge potentialities in application with electron vortex beams [1]. In this contribution we discuss the case of the quasi-Bessel beams obtained as a coherent superposition of conical plane waves along a closed ring of finite angular aperture [2].
Fig 1a shows the simulated transverse distribution of the electron Bessel beam at the first order of diffraction propagating, in the Fresnel region, from the hologram shown in b). In Fig 1c is reported the scanning electron microscope image of the nanofabricated phase hologram with a zoom-in image of the central region shown in the upper inset. The hologram is obtained from of a FIB-milled silicon nitride membrane, which is almost transparent to the 200keV electron beam [3]. Different depths modify the local projected potential; thus, electrons see different effective paths at grooves. In Fig 1d the distribution of the diffracted electrons in the Fraunhofer region of propagation, is reported. In the first order of diffraction, the Bessel beam forms a ring in the far-field. Due to the limited number of grooves of the hologram, the ring, typical of the Bessel beam, is convoluted with the Airy function of the hologram aperture, thus forming a quasi-Bessel beam. In Fig 1e is shown the measured transverse intensity distribution of the quasi-Bessel beam of the zeroth order generated by the hologram shown in Fig 1c, in Fresnel regime. In Fig 1f the experimental radial intensity distribution of the Bessel beam, blue solid curve, is compared with simulations by varying the convergence of the beam incident on the hologram plane, thus showing the effect of the partial coherence on the Fresnel ring contrast.
Bessel beams have many interesting properties, namely resistance to diffraction and the smallest spot diameter compared to other ordinary type of beams that could be exploited in STEM tomography. In Fig 2 is reported the diffraction free range of the quasi Bessel beam shown in Fig 1c.


[1] J. Verbeeck, H. Tian, and P. Schattschneider, Nature 467, 301 (2010).
[2] V. Grillo, E. Karimi et al. Phys. Rev. X 4, 011013 (2014)
[3]V. Grillo, G.C. Gazzadi, E. Karimi et al. Appl.Phys. Lett. 104, 043109 (2014)


M.E. acknowledges  the support  of SPINNER 2013.

Fig. 1: Computer generated hologram and electron Bessel beams of the zeroth order.

Fig. 2: Propagation of Bessel beams of the zeroth order in the Fresnel regime.

Type of presentation: Poster

IT-1-P-2685 Low Voltage Mini TEM

Coufalová E.1, Mynář M.1, Štěpán P.1, Drštička M.1, Sintorn I. M.2, 3
1DELONG INSTRUMENTS a.s., Brno, Czech Republic, 2Centre for Image Analysis, Uppsala University, Sweden, 3Vironova AB, Stockholm, Sweden
michal.drsticka@dicomps.com

On the basis of experience with the low voltage transmission electron microscopy at 5 kV, which is intended for the study of samples with low contrast (organic matters), we tried to design a TEM optimized in many aspects:
1) Maintaining relatively low voltage to keep up high contrast.
2) The use of such energy, which would open the possibility to increase the resolution of the system to the area of atomic (molecular) resolution using the monochromatization of the primary beam and Cs correction in future.
3) Practical standpoints – reasonable dimensions, resistance to external influences.
4) Energy sufficient for the transmissivity of electrons through samples of "standard" thickness.
It turned out to be suitable to base such electron-optical system on the use of magnetostatic (the objective lens) and electrostatic (projection system) elements. For the above reasons, we have chosen a range of energy of 10-25 keV. This choice enables to maintain the concept of combination of electron-optical and light-optical magnification, which leads to a significant reduction of the dimensions of the unit and solving simultaneously the problem of TEM image digitalization. It emerged that the working energy of 25 keV is the highest possible energy, at which there is no degradation of the applicable high light-optical magnification due to scattering in the single crystal fluorescent screen.
Using light lenses with large numerical aperture (up to 0.95), we achieve a high collection efficiency of the light from the screen. Also, the level of the light signal is high enough at 25keV energy. We have verified that the electron-optical system can be operated in several modes:
1) TEM at 25 keV
2) STEM at 15, 10 keV
3) DIFF at 25keV
The first experimental results confirm the assumptions obtained by electron-optical simulations, in particular the expected resolution in various modes.
It is further confirmed that the contrast inevitably decreases at the energy of 25 keV compared to the lower energies, however, it is still significantly higher than in the energy area of above 50 keV. Even thin sections for which there is no significant increase of chromatic aberration provide sufficient contrast in the image at this working energy. This brings the opportunity to study both stained and unstained samples at low radiation damage.
This version has been optimized for identification of viruses – samples prepared with negative staining and fixation. It allows mobility of the device, and is equipped with user friendly control system with a simple concept that provides remote control resources to allow to be controlled by upper level image analysis software for automatic virus recognition (Kylberg and Sintorn EURASIP J. on Image and Video Processing 2013, 2013:17).


The work has been supported by Eurostars Programme of EUREKA and European Community.

Fig. 1: The body on MiniTEM on a standard desk

Fig. 2: Section of the column

Fig. 3: ATCC and rota viruses stained with 2%Uac in TEM mode at 25 keV

Fig. 4: ATCC and rota viruses stained with 2%Uac in STEM mode at 10 keV

Type of presentation: Poster

IT-1-P-2688 Energy analyzer for point electron sources

Kolařík V.1, Coufalová E.1, Mynář M.1, Drštička M.1
1DELONG INSTRUMENTS a.s., Brno, Czech Republic
michal.drsticka@dicomps.com

We have built an energy analyzer for characterization of parameters of various types of point emitters, electron guns, and illumination blocks of electron columns. It can be also used for characterization of electron monochromators, and for studying the influence of electron – electron interaction on the beam energy spread.
The concept of the analyzer is very simple and physically straight, based on dispersion characteristics of magnetic prism: It is configured for measuring energetic spread of emitters with the virtual source size between 1 nm and 50 nm independently of the electron source distance, it means any design or type of electron gun can be measured.
The theoretic resolution of the analyzer is:
• < 15 mV for the virtual source size of 50 nm,
• < 3 mV for the virtual source size of 15 nm.
The image of virtual source is focused only in the dispersion direction (see Fig. 2). The dispersion of the magnetic prism in this plane is about 3 µm/V at the output edge of the prism. The optical set guarantees the resolution of electron spectrometer on the level of 10 mV or better, the use of slit aperture provides the capability of statistical evaluation of 2048 spectra (pixel columns).
Although the dispersion itself is relatively small (units of µm/V), the analysis is possible at the level of units of mV, because the source image size in the spectral plane is in units of nm. The dispersion plane can be enlarged electron-optically so that it is projected onto a screen with the size accessible for imaging by high-quality light optics (the dispersion and source image are magnified in the same proportion).
The significant input parameters that determine the resulting energy resolution are the virtual source size and used aperture angle. We illustrate on the chart that the effect of the virtual source size for cold field emission and Schottky cathode is in a significant range of aperture size under the resolution of a light objective lens with NA as high as 0.95 in this arrangement.
The high energy resolution of the electron-optical part can be used for very effective monochromatization of en electron beam.
Reference: V. Kolařík, M. Maňkoš, L. H. Veneklasen, Close packed prism arrays for electron microscopy, Optik 87, No.1(1991)


The work was supported by ”Electron Microscopy“ Competence Centre of Technology Agency of the Czech Republic

Fig. 1: Section along the optical axes

Fig. 2: Optical scheme

Fig. 3: The influence of the aperture on the energy resolution for CFE and Schottky emitter at dispersion of 2.8 µm/V in relation to the optical limit (at NA = 0.95, M=40×, pixel size = 7.4 µm)

Fig. 4: Examples: calibration of measurement, profiles - CFE energy spread – 0.34 eV, Schottky emitter energy spread – 0.58 eV

Type of presentation: Poster

IT-1-P-2710 Spiral phase plates for electron vortices

Béché A.1, Winkler R.2, Planck H.3, Hofer F.3, Verbeeck J.1
1EMAT, University of Antwerp, Groenenborgerlaan 171, 2020 Antwerp, Belgium, 2Center for Electron Microscopy, Steyrergasse 17, 8010 Graz, Austria, 3Institute for Electron Microscopy and Nanoanalysis, Graz University of Technology, Steyrergasse 17, A-8010 Graz, Austria
armand.beche@uantwerpen.be

Vortex beams have been recently developed in electron optics and generate a lot of interest due to their potential ability of retrieving magnetic information down to the atomic scale [1, 2]. Several techniques are now available to produce such beams like the holographic mask [2] or the more recent magnetic needle [3]. In this work we propose to extend the idea of Uchida & Tonomura [1] by creating a spiral phase plate with smoothly increasing thickness.
The phase plate should be composed of a light material to prevent too much absorption from the plate itself and be ideally thicker than 100 nm at its highest point to allow a smooth increase of thickness. Focused electron beam induced deposition (FEBID) is an ideal tool to realize such structures as it can deposit functional materials with high spatial resolution. In the present case, ultrathin silicon nitride (SiN) was successfully used as substrate to fabricate SiO2 spiral phase plates as shown in Fig. 1. In order to prevent unwanted scattering from the central hole in the spiral, it was filled with a small amount of platinum via FEBID.
The phase plate was then introduced into the Qu-Ant-TEM, an FEI Titan3 transmission electron microscope, operated in Lorentz mode, to achieve a large field of view with extended spatial coherence conditions. Carefully illuminating the phase plate with a uniform electron beam and looking in the far field, typical features of vortex beams were recorded. Fig. 2 displays a through focus series of the resulting beam which reveals the presence of a doughnut like intensity pattern with the destructive interference centre of the vortex beam.
In order to quantify the orbital angular momentum (OAM) carried by the outgoing beam, electron holography was performed at the edge of the phase plate. By measuring the phase shift between the thickest and thinnest area, the total OAM was estimated to be 0.6 (Fig. 3).
Further tuning of this setup provides another method for creating atomic sized electron vortex beams with the advantage of providing a single vortex beam that is easy to obtain in a standard TEM.

[1] Uchida M. & Tonomura A., Nature Letters (2010), 464, p737-739.
[2] Verbeeck J. et al., Nature Letters (2010), 467, p.301-304.
[3] Béché A. et al., Nature Physics (2014), 10, p. 26-29.


This work was financially supported by the European Research Council under the 7th Framework Program (FP7), ERC grant 246791 COUNTATOMS, ERC Starting Grant 278510 VORTEX and Integrated Infrastructure Initiative No. 312483 ESTEEM2.

Fig. 1: (a) TEM view of the spiral phase plate with the central hole filled with platinum. (b) Atomic force microscope image revealing the thickness profile of the phase plate.

Fig. 2: Far field through focus series of an electron beam evenly illuminating the phase plate. The destructive interference area in the middle of the fully condensed beam (black hole) is typical of a vortex beam.

Fig. 3: (a) Phase image of the edge of the phase plate, between the thickest and thinnest area of the spiral, acquired by holography. (b) Phase profile taken along the dotted area displayed in (a) revealing a total phase shift of 4 rad corresponding to a total OAM of ~0.6.

Type of presentation: Poster

IT-1-P-2835 Projector lens and CCD camera distortions in a Hitachi HF-3300 TEM

Denneulin T.1, Gatel C.1, Houdellier F.1, Hÿtch M. J.1
1CEMES CNRS, 29 rue Jeanne Marvig, 31055 Toulouse, France.
hytch@cemes.fr

In a transmission electron microscope (TEM) the projector lenses are known to introduce large-scale distortions. The magnification and the rotation in the image can vary up to 5% and 2° across the field of view [1]. Therefore the accurate mapping of any physical field (strain, magnetic or electric field) using high resolution TEM or holography requires the calibration of those distortions. The method used here does not add noise to the phase image and alleviates the need for a reference hologram.
We have investigated the projector and the CCD camera distortions on a recently installed aberration-corrected HF-3300 Hitachi TEM (I2TEM-Toulouse). The distortions were measured using off-axis electron holograms acquired in the vacuum. A double biprisms setup was used to remove the Fresnel fringes [2]. The voltages of the biprisms were set so that the interference pattern fills entirely the 4k Gatan CCD camera. Two holograms with a different orientation of the biprisms were acquired in order to reconstruct the 2D strain field using geometrical phase analysis (GPA) [3]. Before GPA calculation, the reconstructed phase images were fitted using a 4th order polynomial to remove the noise.
The influence of the magnification and the values of P1 and P2 was investigated. It was found that the distortions are mainly dependent on the value of P2. Fig. 1 shows the strain field obtained for 4 different values of P2. Increasing P2 is equivalent to “zoom” into the distortion pattern. The variations across the image are then lower for high values of P2. At a nominal magnification of ×1.5M (P2 is equal to 5.3 A) the mean dilatation Δxz varies from 0 to 3% and the rigid body rotation ωxz varies from 0 to 1° from the center to the corner of the image.
According to the theory [1], Δxz and ωxz should be circular shaped. However it can be noted that the rotation image is slightly triangular shaped. After analysing the ronchigram of the camera provided by Gatan [4] we found that this is due to the low frequency distortions of the camera (see Fig. 2 after correction of the camera distortions). We then created an artificial ronchigram for correcting both the projector and the camera distortions. The procedure will be detailed during the presentation. Fig. 3(a) is an example of dark-field hologram acquired on a SiGe layer grown by epitaxy on a Si substrate. Without correction (Fig. 3(b)) the reconstructed phase image exhibits some variations in the substrate and the phase ramps in the layer are slightly distorted. Those artifacts are removed after correction (Fig. 3(c)).

[1] F Hüe et al, J. Electron. Microsc. 54(3) (2005), 181–190
[2] K Harada et al, Appl. Phys. Lett. 84(17) (2004), 3229–3231
[3] MJ Hÿtch et al, Ultramicrosopy 74 (1998), 131–146
[4] P Mooney, private communication


This work received financial support from the European Union under the Seventh Framework Programme under a contract for an Integrated Infrastructure Initiative Reference 312483-ESTEEM2 and the European Metrology Research Programme (EMRP) Project IND54 Nanostrain. The EMRP is jointly funded by the EMRP participating countries within EURAMET and the European Union.

Fig. 1: Distortions of the projector lenses as a function of the displayed value of P2. The strain field was calculated by geometrical phase analysis after fitting the phase images reconstructed from the holograms. From left to right are shown the horizontal εxx, the vertical εzz, the shear εxz strain, the mean dilatation Δxz and the rotation ωxz.

Fig. 2: Distortions (for P2 = 6.0 A) obtained after correcting the phase images with the CCD camera ronchigram.

Fig. 3: (a) (004) dark-field electron hologram of a SiGe layer epitaxially grown on a Si substrate. (b) Phase image reconstructed from the hologram without correction. (c) Phase image reconstructed with correction of the projector and camera distortions.

Type of presentation: Poster

IT-1-P-2975 Design and Characterization of a Single-Atom Electron Column

Lin C. Y.1,2, Chang W. T.1, Hsu W. H.1,3, Lai W. C.1,2, Chen Y. S.1, Hwang I. S.1
1Institute of Physics, Academia Sinica, Taipei, Taiwan, 2Department of Physics, National Taiwan University, Taipei, Taiwan , 3Department of Materials Science and Engineering, National Tsing Hua University, Hsinchu, Taiwan
cylin@phys.sinica.edu.tw

        It has been shown that noble-metal covered W(111) single atom tips (SATs) can be reliably prepared [1,2]. We have demonstrated full spatial coherence of electron beams emitted from the SATs [3]. Thus, single atom electron sources are suitable for phase retrieval imaging methods, such as holography and coherent diffractive imaging. We have proposed a SAT-based low-keV electron microscope that allows different imaging modes, as shown in Fig. 1. For this purpose, we plan to build an electron column with the capability to accelerate electron beams to 1~5 keV and a focused beam spot smaller than 100 nm. The column is composed of two parts: an electron gun and a condenser lens.

        The electron gun consists of a SAT, an extractor/suppressor, and an acceleration electrode. The tip is mounted on a holder that can be translated, tilted, and rotated in nanometer scale by piezo-positioners. Therefore, the tip-lens alignment can be done in vacuum without alignment coils. We have recorded the opening angles of the electron beams. As shown in the inset of Fig. 2, the emitter can be moved to different positions with the piezo-positioners and the corresponding beam profiles are recorded. Fig. 2 shows the half opening angles of the beams at an electron energy of 2.5 keV measured at different extraction voltage and different separations. Clearly the beam opening angle varies with the tip position. When the tip is positioned at about -2.5 mm, the half opening angle can be smaller than 1 mrad. We also find that the suppressor design that is often used in normal field emitters is not effective in reducing the beam divergence for the SAT emitter.

        The condenser lens consists of a limiting aperture, an einzel lens, and an octupole stigmator. We used Simion 8.1 software to simulate the lens parameters and determine the aperture diameter. In our simulations at the electron energy of 2.5 keV and the working distance of 2 mm, a spot size of 140 nm is obtained when the limiting aperture of 100 μm is used; a spot size of 20 nm is obtained when the limiting aperture of 20 μm is used. Fig. 3(a) shows the whole assembly of our instrument. As shown in Fig. 3(b), we have obtained a diffraction pattern on a small region of a suspended CVD graphene, which show two domains with different orientations. We are also designing a microcolumn based on the MEMS technique. Our ultimate goal is to determine the atomic structures of few-layer two-dimensional structures such as graphene and one-dimensional structures such as carbon nanotubes and bio-molecules.

References

[1] H. S. Kuo et al, Nano Lett. 4(12) (2004), p. 2379.

[2] H. S. Kuo et al, J. J. Appl. Phys. 45 (2006), p. 8972.

[3] C. C. Chang et al, Nanotech. 20 (2009), p. 115401.


This work is supported by National Science Council of ROC and Academia Sinica.

Fig. 1: Schematic of a multi-mode low-keV electron microscope

Fig. 2: Beam divergence, measured with the half opening angle at an electron energy of 2.5 keV, versus the extractor/suppressor voltage at different emitter positions. The inset is the schematic for characterization of the beam profile at different emitter positions.

Fig. 3: (a) Illustration and photo of a low-keV electron microscope (b) The diffraction pattern of a CVD graphene sample

Type of presentation: Poster

IT-1-P-2984 Wire corrector for aberration corrected electron optics

Nishi R.1, Ito H.2, Hoque S.2
1Osaka University, Osaka, Japan, 2Hitachi High-Technologies Corporation, Ibaraki, Japan
rnishi@uhvem.osaka-u.ac.jp

The wire corrector on the analogy of multipole correctors was proposed by H. Ito [1]. Two-parallel line current (Fig. 1) makes the magnetic field similar to that of a quadrupole as shown in Fig. 2. When using two-parallel line current, the filed cancels at the rotation symmetric axis and the two-parallel line current can generate quadrupole magnetic field because each magnetic field has opposite rotation direction of magnetic flux.
The wire corrector is only arranged by parallel line currents without using any magnetic materials, so it can be easily and simply fabricated and arranged in comparison to a conventional multipole. Adverse effect of hysteresis of magnetic material does not exist and homogeneity of magnet property is not needed. Magnetic field can be controlled by superimposition of parallel line currents. In actual layout, the wire corrector is configured to a coil shape in addition to the parallel currents with infinite length, but the effect of a coil shape can be reduced by consideration of its shape. Applying constant current to a main coil, fine adjustment of magnetic field can be performed by applying current to a sub coil. The wire corrector is valuable to the aberration corrected electron optics with high precision alignment and reproducibility.
When using the wire corrector of N=2, the magnetic field is similar to quadrupole field but the magnetic field is expanded in a series which also contains octapole field as a higher order term, as shown in Eq.(1) inset of Fig.1. Due to the wire corrector has octapole component, the wire corrector has possibility of simultaneous correction of spherical aberration in addition to chromatic aberration. Symmetric curved ray optical system constituted by combining both components of a deflector and the wire corrector of N=2, is expected that chromatic and spherical aberration is potentially corrected in such configuration.
The combination of the round lenses and the wire correctors of N=3 decreases the spherical aberration [2]. This shows the wire correctors of N=3 worked as a hexapole. The wire corrector has a potential of consisting an easy-to-use aberration corrector.

[1] Hiroyuki Ito et al, USP 7,872,240 B2 (date of patent: Jan. 18, 2011).
[2] H. Rose, Optik, 85 (1990) 19.


A part of this work of calculation was done by Dr. Eric Munro and Dr. John Rouse in Munro's Electron Beam Software Ltd.

Fig. 1: The wire corrector consisting of two parallel line currents (N=2).

Fig. 2: Magnetic flux in the wire corrector (N=2).

Type of presentation: Poster

IT-1-P-3022 Measuring the Orbital Angular Momentum of Electron Vortex Beams by Forked Grating

Saitoh K.1, Hasegawa Y.2, Hirakawa K.2, Tanaka N.1, Uchida M.3
1EcoTopia Science Institute, Nagoya University, 2Department of Crystalline Materials Science, Nagoya University, 3Advanced Science Research Laboratory, Saitama Institute of Technology
saitoh@esi.nagoya-u.ac.jp

After the first report of the production of an electron vortex beam, an electron traveling in free space with orbital angular momentum (OAM) [1], electron vortex beams have been attracting a great attention owing to the unique physical property and application to a new microscopy in materials science [2]. In the present paper, we show the how the electron vortex beams are diffracted by forked gratings and how the OAMs of the electron vortex beams are transferred to each of the diffracted waves (Fig.1(a)). [3].

Figures 1(b) and 1(c) show a schematic diagram of the experimental setup of the present study. The binary masks of the spiral zone plates [Fig. 1(d)] and the forked gratings [Fig. 1(e)], fabricated from 200 nm thick PtPd films using a focused-ion-beam instrument (Hitachi FB-2100). The spiral zone plates and forked gratings were inserted into the condenser lens aperture position and selected-area aperture position, respectively, of a transmission electron microscope (JEOL JEM-2100F), which was operated at an acceleration voltage of 200 keV.

Figures 2(a) and 2(b) show electron vortex beams with OAMs of 10h and -10h, respectively, produced by the spiral zone plate. Each of the electron beams show a ring composed of 10 peaks at the center [4]. Figure 2(c) shows an electron diffraction pattern for an incident electron vortex beam with m = 10h. The diffraction pattern shows a series of diffracted rings, as indicated by the arrows. The central ring, composed of 10 peaks, is the transmitted beam with m = 10h. The 1st- and -1st-order diffracted electron beams show similar ringlike features, but have 11 and 9 peaks, respectively. This indicates that the electron OAMs of the 1st- and -1st-order diffracted beams are 11h and 9h, respectively. Figure 2(d) shows an electron diffraction pattern for an incident electron vortex beam with m = -10h. The pattern shows a series of diffracted rings as in Fig. 2(c), but is horizontally inverted from that shown in Fig. 2(c). The transmitted (0th-), 1st-, and -1st-order diffracted rings show 10, 9, and 11 peaks, respectively, indicating that the electron OAM of the 1st- and -1st-order beams are -9h and -11h, respectively. Our results indicate that the forked grating with a Burgers vector of b = 1 transfers not only linear momentum but also OAM, where the electron OAM transfer of the nth-order diffracted electron beam is nh. This diffraction property could be used as an electron OAM analyzer, as the nth-order diffracted beam shows a normal peak.

References

[1] M. Uchida and A. Tonomura, Nature 464, 737 (2010).

[2] J. Verbeeck, H. Tian, and P. Schattschneider, Nature 467, 301 (2010).

[3] K. Saitoh et al, Phys. Rev. Lett. 111, 074801 (2013).

[4] K. Saitoh et al., J. Electron Microsc. 61, 171 (2012).


The present work was partly supported by the Grant-in-Aid for Scientific Research (A) (No. 23241036), the Ministry of Education, Culture, Sports, Science and Technology, Japan, and the Mitsubishi Foundation.

Fig. 1: (a) Schematic drawing of the present experiment. (b),(c) Ray-path diagrams of the experimental setups for incident electron vortex beams of m = 10h (b) and m = 10h (c). (d) A spiral zone plate with a 20 μm diameter introduced to the condenser lens system (e) A forked grating with a 30 μm diameter introduced to the image-forming lens system.

Fig. 2: Incident electron vortex beams with m = 10h (a) and m = -10h (b), and diffraction patterns of the vortex beams with m = 10h (c) and m = -10h (d) generated by the forked grating shown in Fig. 1(e). 

Type of presentation: Poster

IT-1-P-3222 Contrast enhancement in TEM imaging by use of a central beam stop

Zandbergen H.1, Xu Q.1
1Kavli Institute of Nanoscience, Delft University of Technology, Delft, The Netherlands
h.w.zandbergen@tudelft.nl

In TEM in life science, beam damage is the most important limitation. This is also the case for materials science samples like graphene, polymers and hybrid materials. On the imaging side an important boost is expected from the introduction of a phase plate. Phase plates have been researched over several decades and no easy to use system has emerged yet, indicating that is not easy task. Given the importance of efficient imaging, it is clearly necessary to explore other routes. We have explored [1] the possibility of dark-field imaging for contrast enhancement in which we have tried to block the central beam [2] and leave as many of the diffraction beams un-blocked.
Central beam block apertures (the abbreviation DF-000 is used in this abstract) in the shape of Mercedes star (see Figure 1) were made with a FIB. In our experiments we have observed no sign of charging, possibly due to the DF-000 shape. In central disk should preferably be smaller than the frequency, g, one wants to observe, which is of course much smaller for biological samples than for most materials samples. Our DF-000 removed frequencies corresponding to d-spacings of 8.7 Å and larger. In the presentation we will report how far we can decrease the size of central disk without charging problems and with still good blocking of the central beam.
For the drilling of holes in exfoliated graphene without contamination build-up we heated the graphene to 600°C. TEM experiments were done at 300keV and post-specimen aberration correction at 600°C. Figure 1c and 1d shows high-resolution and DF-000 images of multilayer graphene (4-5 layers). A hole was made in this sample using an e-beam. This hole can be seen very well in the DF-000 image and only faintly in the BF image. In both cases one can see that the graphene lattice continues up to the edge of the hole. The gradual decrease in thickness is clearly visible in the DF-000 image and not at all in the BF image. Thus we can obtain in the DF-000 image high-resolution information with a similar resolution limit as the BF image.
Figure 2 shows several images of graphene with three holes with varying size imaged at various focus values, showing that the bright field images in the range from -1500 to + 1500 nm show hardly any contrast and none at zero focus. On the contrary, the contrast in the DF-000 taken at 0 focus shows the largest contrast and in particular the smallest delocalization. In this case selecting ~ zero focus is easy by minimizing the blurring in the image.

1. Zhang C, Xu, Q, Peters PJ, and Zandbergen, H, Ultramicroscopy 134, 200 (2013)
2. Cowley, J., Acta Crystallographica Section A 1973, 29, 529-536


Fig. 1: (a) SEM image of the DF-000 objective aperture used to stop the central beam. (b) shows a typical DF-000 aperture in diffraction space. (c) and (d) show HREM images taken without an aperture and with the DF-000 image taken from the exactly the same area.

Fig. 2: Images from the same area of single layer graphene with three holes showing the effect of focus in BF and in DF-000 modes. No DF-000 at defocus values of -1000 and -3000 nm are given because these are too blurred. The two small holes are only barely visible in the BF image taken at -3000 nm and not at all in the other three BF images.

Type of presentation: Poster

IT-1-P-6004 Determinationof geometrical form factor of emitter from Schottky plot

Emura Y.1, Murata H.1, Rokuta E.1, Shimoyama H.1, Yasuda H.2, Haraguchi T.2
1Faculty of Science and Technology, Meijo University, 2PARAM Corporation
hkmurata@meijo-u.ac.jp

In this paper we report preliminary experimental results on a LaB6 Schottky emission electron gun, which also includes our new findings that the electric field strength on the emitter surface can be estimated experimentally from the Schottky plot whose slope depends not on the work function but only on the reciprocal of the emitter temperature. According to the theoretical considerations on the Schottky emission, if the values of log10 j (j: emission current density) are plotted as a function of √F (F: field strength on the emitter surface), then the graph becomes a straight line with the slope of 1.913/T (T: emitter temperature), which is known as “the theoretical Schottky plot”. In experiment, on the other hand, the beam current I is measured as a function of the extraction voltage Va. Thus, the slope of “the experimental Schottky plot” is different from that of “the theoretical Schottky plot”. From I = j × ΔS (ΔS: emission area on emitter surface), the vertical axis of “the experimental Schottky plot” is expressed as log10 j + log10 ΔS, which means the graph is moved parallel to the vertical direction without changing the slope of the graph. We mark a new scale on the horizontal axis of “the experimental Schottky plot” in order that the slope may be equal to 1.913/T. Then, the new horizontal axis should be graduated in √F. This procedure makes it possible to relate the field strength F directly to the extraction voltage Va as F = β Va, where β is the geometrical form factor of the emitter.

The Schottky emission experiment has been done in the ultra-high vacuum chamber, using the experimental circuit shown in Fig. 1. A flat top LaB6 emitter is embedded into a rhenium conical sheath, and is heated by a tungsten hairpin filament, as shown in Fig. 1.

The beam current IF was measured as a function of the extraction voltage Va for a constant emitter temperature T = 1600 K by the Faraday cup placed behind the fluorescent screen. In Fig. 2, the values of log10 IF are plotted as a function of √Va. It can be seen that the plot is almost a straight line, which indicates that the emission is under Schottky emission mode. Figure 2 (a) and (b) also show emission patterns observed on the fluorescent screen at Va = 2 kV and 5 kV, respectively.

Figure 3 shows the above procedures, where the new horizontal axes scaled in √F are placed in addition to the original horizontal axes scaled in √Va. From the relation between F and Va, we have found that β = 99.6 [1/cm] for T = 1600 K. We have also performed the field calculation for the experimental system shown in Fig. 1. According to the calculation, the geometrical form factor β has been found to be β = 95.0 [1/cm], which is in good agreement with that estimated from “the experimental Schottky plot”.


Fig. 1: Experimental circuit for measuring the beam current by Faraday cup.

Fig. 2: Experimental Schottky plot for emitter temperature T = 1600 K. Emission patterns observed on the fluorescentscreen at Va = 2 kV (a) and 5 kV (b).

Fig. 3: Schottky plot for T = 1600 K for determination of geometrical formfactor β. A newhorizontal axis scaled in the square root of F is added so that the slope ofthe Schottky plot may be equal to 1.913/T.

Type of presentation: Poster

IT-1-P-6020 Design and realisation of variable C shaped structured illumination

Mousley M.1, Thirunavukkarasu G.1, Babiker M.1, Yuan J.1
1Department of Physics, University of York, Heslington, York, YO10 5DD, United Kingdom
mgm514@york.ac.uk

Structured illumination is a new development in electron microscopy, with the advantage such as longer column channeling distances in crystals by donut-shaped illumination of atomic scale vortex electron beams [1]. In this paper, we introduce a controlled way to realize C shaped structured illumination. Analytical equations determining the parameters of the C shaped illumination pattern have been derived using phase gradient analysis, allowing independent control of the C-opening angle and radius of the C shape. Experimentally, we have used computer generated hologram (CGH) method to generate C shaped structured illumination in a 200 keV transmission electron microscope. Both amplitude and phase CGH masks have been used and comparisons with simulations show a strong match between the theoretical results and the experimentally recorded electron microscope images. C-shaped illumination has promises in potential applications such as electron beam lithography for production of metamaterials which utilise split ring resonance structures [2]. Physical dimensions of the artificial electromagnetic resonance structures as small as nanometres should now be possible. Furthermore the orientation of the C shape illumination can be readily identified, allowing the easy identification of the Faraday rotational effects of the vortex beams [3].

[1] H. Xin and H Zheng (2012) Microscopy and Microanalysis, Vol. 18, p711-9

[2] D. R. Smith, W. J. Padilla, D. C. Vier, S. C. Nemat-Nasser and S. Schultz (2000) Phys. Rev. Lett. 84, p4184-7.

[3] C. Greenshields, R. Stamps, S. Franke-Arnold (2012) New J Phys. 14, 103040


We wish to thank the UK Engineering and Physical Science Research Council (EPSRC) for financial support to this research by a grant (EP/J022098) and M. Ward of Leeds Electron Microscopy and Spectroscopy Centre, University of Leeds for the help with focused ion beam experiment.

IT-2. High resolution TEM and STEM

Type of presentation: Invited

IT-2-IN-2458 Advanced scanning transmission electron microscopy with segmented annular all field detector

Shibata N.1,2
1The University of Tokyo, 2JST-PRESTO
shibata@sigma.t.u-tokyo.ac.jp

In this talk, I will review our recent and on-going findings from our exploration of new atomic-resolution imaging modes using an area detector which is capable of atomic-resolution STEM imaging [1]. One possibility is atomic-resolution differential phase contrast (DPC) imaging [2]. It has been reported that, to a good approximation, DPC STEM images represent the gradient of the object potential (= fields) taken in the direction of the diagonally opposed detector segments, provided the object scatters weakly [3-5]. Here, we show atomic-resolution DPC STEM images of SrTiO3 observed from the [001] direction [2]. Fig. 1(a) shows the orientation relationship between the SrTiO3 crystal and the detector segments used in this study. The probe-forming aperture angle was 23 mrad and the polar angle range of the detector segments was 15.3 to 30.6 mrad. Fig. 1(b) shows the experimental difference image and its intensity profile projected over the vertical direction in the image. The simultaneous ADF STEM image and its intensity profile are used for reference since the peaks in ADF image are a well-established indicator of the true atomic positions. Fig. 1(c) shows the results of corresponding image simulation. It is clear that the DPC STEM profile has a node (zero crossing) at the atom location. The profile is antisymmetric about this point, reflecting the reversal of the electric field direction across the atom along the direction of diagonal detector segments. Combined with detailed image simulations, atomic-resolution DPC STEM is found to provide information on the local electric field distribution in the vicinity of the atomic columns. Some application results of DPC STEM imaging for ferroelectics and their interfaces will be presented.
Another possibility is annular bright-field (ABF) imaging and its derivatives. Fig. 2(a) shows a schematic of the ABF detector geometry. We form “enhanced” (e)ABF images [6] by simply taking the difference between ABF images and the corresponding BF images using the area detector. As shown in Fig. 2(b), we find that light element imaging can be selectively enhanced by this process. We anticipate that the area detector will offer still further possibilities for new atomic-resolution STEM imaging modes useful for material characterization.

References
[1] N. Shibata et al., J. Electron Microscopy 59, 473 (2010).
[2] N. Shibata et al., Nature Phys., 8, 611 (2012).
[3] N.H. Dekkers and H. de Lang, Optik, 41, 452 (1974).
[4] H. Rose, Ultramicroscopy, 2, 251 (1977).
[5] W.C. Stewart, J. Opt. Soc. Am., 66, 813 (1976).
[6] S.D. Findlay et al., Ultramicroscopy, 136, 31 (2014).


I deeply thank S.D. Findlay and Y. Ikuhara for their collaboration in materials characterization and Y. Kohno, H. Sawada and Y. Kondo for their collaboration in the detector development. This work was supported by the PRESTO, JST. A part of this work was conducted in Research Hub for Advanced Nano Characterization, The University of Tokyo. 

Fig. 1: (a) Schematic illustration showing the relationship between the crystallographic orientation of SrTiO3 and the two detector segments. (b) The DPC STEM image formed by subtracting the signal in detector segment Y from that in detector segment X and its image intensity profile [2]. 

Fig. 2: (a)Schematic illustration of BF and ABF detector geometry. (b)ABF and eABF images of LaTiO3 observed from [001] direction [6].

Type of presentation: Invited

IT-2-IN-2576 Insights into Materials Properties with Quantitative STEM and EELS

Botton G. A.1, Bellido E. P.1, Bugnet M.1, Dudeck K. J.1, Gauquelin N.1, Liu H.1, Prabhudev S.1, Rossouw D.1, Scullion A.1, Stambula S.1, Woo S. Y.1, Zhu G. Z.1
1Department of Materials Science and Engineering, McMaster University, Hamilton, ON, Canada
gbotton@mcmaster.ca

The development of aberration correctors in scanning transmission electron microscopy (STEM) has dramatically improved the analytical “toolkit” of materials scientists. In particular, when combined with electron energy loss spectroscopy (EELS), STEM makes it possible to detect compositional and spectroscopic changes at the atomic level that can be used to understand the structure, and ultimately the performance of materials. Here we present some examples of quantitative STEM and EELS as applied to the study of graphene-based materials, complex nanoparticles used in electrocatalysts, and the defects generated in implanted Si and plasmonic structures.

An FEI Titan microscope was used for this work. With this system, we imaged Pt atoms on multilayer graphene nanosheets (GNS) and demonstrate that single Pt atoms are stabilized during atomic layer deposition on N-doped GNS. Quantitative analyses of images show that the single atoms are located at GNS edge steps and that the doping strongly suppresses the growth of Pt clusters (Figure 1a, b) [1]. Similarly, quantitative images have been used to detect atomic displacements on PtFe intermetallic core-shell nanoparticles that exhibit very high specific activity compared to pure Pt [2,3]. Not only is elemental mapping at the atomic scale possible, but the high beam current and fast spectrometers also allow the acquisition of maps with large sampling of the nanostructure. This is illustrated in the study of PtRu nanocatalysts used in fuel cells where Ru core-Pt shell structures are very clearly mapped (Figure 1c).

Beyond the “simple” deduction of the distribution of elements in nanostructures from maps, quantification is essential to understand the detailed structure of defects and correlate compositional measurement with the optical response of materials. The detailed quantification of the atomic position of a defect, in this case a so-called {311} defect [4] generated by the implantation of ions in Si [4,5] shows that an excellent agreement is obtained between the experimental atomic positions and molecular dynamics simulations (Figure 3) [4] with an accuracy of better than 0.05nm for more than 100 atomic columns. Similarly, quantitative analysis of SiGe alloys has allowed us to deduce compositional fluctuations and interdiffusion in proximity of interfaces [6].

[1] S. Stambula et al., J. Phys. Chem. C, on-line (2014), DOI: 10.1021/jp408979h
[2] S. Prabhudev et al., ACS Nano 7, 6103-6110, (2013)
[3] M.C.Y. Chan et al, Nanoscale 4 (22), 7273-7279, (2012)
[4] K.J. Dudeck et al., Physical Review Letters, 110, 166102 (2013)
[5] K.J. Dudeck et al., Semiconductor Science and Technology, 28, 125012, (2013)
[6] G. Radtke et al., Physical Review B 87, 205309, (2013)


The authors are grateful to NSERC for supporting this research. The microscopy was carried out at the Canadian Centre for Electron Microscopy, a National facility supported by NSERC and McMaster. We are grateful to Paolo Longo (Gatan Inc.) for the help in setting up the Quantum 966 spectrometer.

Fig. 1: HAADF STEM image of single Pt atoms and clusters stabilized on N-doped GNS. Raw signals (a), edges and atoms detected (b). Green arrows point to the few Pt atoms (pink) stabilized on GNS terraces (edges labeled in yellow) [1]. (c) elemental maps of PtRu core-shell nanoparticles

Fig. 2: HAADF STEM image of a {311} defect in Si (a) and the deduced atomic positions (crosses) in (b) with the expected atomic positions deduced by molecular dynamics calculations [4].

Type of presentation: Invited

IT-2-IN-2893 Extending the capabilities of high-resolution STEM:measuring depth dependent strain using optical sectioning and aberration-free phase contrast imaging of low-Z materials

Nellist P. D.1,2, Yang H.1, Lozano J. G.1, Hirsch P. B.1, Pennycook T. J.1,2
1Department of Materials, University of Oxford, UK, 2SuperSTEM Laboratory, Daresbury, UK
peter.nellist@materials.ox.ac.uk

The development of aberration correction in scanning transmission electron microscopy (STEM) has had a major impact on spatial resolution and analytical capability. Unsurprisingly, alongside these developments come further complications but also opportunities. The increased numerical aperture allowed by aberration correction leads to a reduced depth of focus (DOF), which in a modern instrument may be just a few nanometres, and typically less than the sample thickness. The increased numerical aperture of the probe converging optics also leads to a larger bright-field (BF) disc in the detector plane, and as a result much of the scattering by the sample remains in the BF disc. In this presentation we will two explore STEM imaging modes that make use of each of these effects to provide aberration-corrected STEM with new capabilities.


The reduced DOF means that in principle a three-dimensional (3D) data-set can be recorded as a focal series of images. In practice, a confocal configuration is generally required. At atomic resolution, however, nanometre-scale depth resolution is also available in the conventional STEM configuration [1]. For dislocations in GaN viewed end-on we show the detection of depth-dependent Eshelby twist displacements associated with screw dislocations. We also show that ADF STEM optical sectioning can be used to measure the screw displacements parallel to the dislocation line for dislocations lying in the plane of the TEM sample, and we use this effect to measure the dissociation reaction of mixed dislocations in GaN. Despite the channelling of the probe, the depth sensitivity persists, and Fig. 1 shows how a simple weighted potential model is a reasonable approximation to a full channelling simulation.


Use of a pixelated detector to record the entire BF disc in the detector plane as a function of probe position results in a 4D data set. A phase contrast image can be retrieved from this data set using a processing method proposed by Rodenburg et al [2]. Interference between the BF disc and a diffracted disc leads to intensity in the overlap region that oscillates with respect to probe position. Figure 2 shows the magnitude and phase of that oscillation for a bilayer graphene sample. From such data a full phase contrast image can be retrieved and we compare the sensitivity of this imaging mode with alternative techniques such as annular bright-field and differential phase contrast. The data is also an excellent instrument diagnostic, and effects such as aperture charging, residual aberrations and the effect of chromatic aberrations can also be observed.

[1] P.D. Nellist and P. Wang, Annual Review of Materials Research 42 (2012) 125-143.
[2] J.M. Rodenburg, B.C. McCallum and P.D. Nellist, Ultramicroscopy, 48 (1993) 303-314.


This research has received funding from the EPSRC and the EU 7th Framework Programme under Grant Agreement 312483 - ESTEEM2 (Integrated Infrastructure Initiative–I3) and was partly performed at the EPSRC National Facility for Aberration-Corrected STEM.

Fig. 1: Aberration-corrected ADF STEM simulated images along the a lattice direction for a 10 nm thick sample of GaN containing a screw dislocation lying parallel to [0001] in the mid-plane of the foil. For each defocus , the left panel shows a full frozen phonon calculation using the QSTEM code and the right panel a simple weighted potential approach.

Fig. 2: The (a) amplitude and (b) phase of the interference observed in the BF disc for one particular spatial frequency with respect to probe position in the 4D data set. The data was recorded from bilayer graphene at 60 kV using a Nion UltraSTEM 200 with a convergence angle of 30 mrad.

Type of presentation: Oral

IT-2-O-1555 Imaging of light elements by annular dark-field Cs-corrected STEM

Lotnyk A.1, Poppitz D.1, Gerlach J. W.1, Rauschenbach B.1
1Leibniz Institute of Surface Modification (IOM), Leipzig, Germany
andriy.lotnyk@iom-leipzig.de

Nowadays, many crystalline lattices can be imaged directly at atomic resolution in Cs-corrected STEM. Recently, it was shown that light and heavy elements in crystalline lattices can be detected with an ABF method1 or with a double-detector STEM method.2 However, imaging of atomic columns of light elements by ADF method remains challenging. Particularly, the observation of light element columns at the interface between two different materials is still a difficult issue. In this work, we were able to detect directly and simultaneously the N and C atomic columns at the GaN-SiC interface and within the GaN and SiC materials. Additionally, the O atomic columns in a SrTiO3 single crystal were also observed by our method. We have studied the influence of imaging conditions on the appearance of N and C atomic columns in the GaN and SiC materials. The obtained results are discussed and are supported by image simulations.

       The GaN thin film for this study was grown on 6H-SiC(0001) substrate by ion-beam assisted molecular beam epitaxy. STEM experiments were performed on a probe Cs-corrected Titan3 G2 60-300 microscope operated at 300 kV. A probe forming aperture of 20 mrad was used. Cross-sectional samples for STEM work were prepared by FIB technique. To improve the surface quality of the TEM specimens and to reduce the samples thicknesses, a focused low-energy argon ion milling (NanoMill system) was applied.3 Ion energies from 900 eV down to 200 eV were applied to remove implanted Ga ions and amorphous regions caused by the FIB. Image simulations were performed with the xHREM/STEM software package.

       Figures 1 and 2 show the results of our work.4 We found that by adjusting the settings of  HAADF detector and defocus value in STEM, the light element columns at the GaN-SiC interface and within the w-GaN, 6H-SiC and SrTiO3 lattices can be imaged using only a single HAADF detector. We concluded that image simulations for interpretation of atomic-resolution STEM images are only necessary when the probe forming aperture angle overlaps the inner angle of an annular STEM detector or when a complex defect structure is observed in a studied TEM sample. Our method works well using either ADF or HAADF detector, because their angular ranges and defocus values can be easily adjusted on any Cs-corrected STEM. Thus, on TEM systems equipped with only one HAADF detector, the technique can be used without any doubt and upgrades to an ABF detector.

1. S.D. Findlay, N. Shibata, H. Sawada et al., Appl. Phys. Lett. 95, 191913 (2009).
2. Y. Kotaka, Appl. Phys. Lett. 101, 133107 (2012).
3. D. Poppitz, A. Lotnyk, J.W. Gerlach, B. Rauschenbach Acta Mater. 65, 98 (2014).
4. A. Lotnyk, D. Poppitz, J.W. Gerlach, B. Rauschenbach Appl. Phys. Lett. 104, 071908 (2014).


The financial support of the European Union and the Free State of Saxony (LenA project; Project No. 100074065) is greatly acknowledged.

Cs-corrected: aberration-corrected; STEM: scanning transmission electron microscopy; ABF: annular bright-field; ADF: annular dark-filed; HAADF: high-angle ADF; FIB: focused ion beam; w-GaN: wurtzite-type GaN; i: detector inner angle; o: detector outer angle.

Fig. 1: (a) Atomic-resolution STEM image of the GaN-SiC interface taken with a HAADF detector (i20.4-o124.6 mrad) and schematic representation of w-GaN and 6H-SiC lattices along the [2-1-10] zone axis. (b) and (c) Simulated images of w-GaN and 6H-SiC, respectively, at 5 nm underfocus. The TEM sample thickness is measured to be about 16 nm.

Fig. 2: High-resolution STEM images of SrTiO3 acquired with (a) ADF (i19.1-o106.5 mrad) and (b) ABF (i10.1-o19.1 mrad) detectors. The insets in (a) and (b) show the SrTiO3 structure viewed along the [001] zone axis. The TEM sample thickness is measured to be about 60 nm.

Type of presentation: Oral

IT-2-O-1645 Atomically Resolved 3D Shape Determination of an MgO Crystal from a Single HRTEM Image

Jia C. L.1,2,3, Mi S. B.1,4, Barthel J.3,5, Wang D.1, Dunin-Borkowski R. E.2,3, Urban K. W.2,3, Thust A.2,3
1International Center of Dielectric Research, The School of Electronic and Information Engineering, Xi'an Jiaotong University, Xi'an 710049, China, 2Peter Grünberg Institute, Forschungszentrum Jülich GmbH, 52425 Jülich, Germany, 3Ernst Ruska-Centre for Microscopy and Spectroscopy with Electrons, Forschungszentrum Jülich GmbH, 52425 Jülich, Germany, 4Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China, 5Central Facility for Electron Microscopy, RWTH Aachen University, Ahornstr. 55, 52074 Aachen, Germany.
a.thust@fz-juelich.de

High-resolution transmission electron microscopy (HRTEM) allows one to investigate the structure of matter on an atomic level [1]. However, most atomic structure characterizations obtained by HRTEM were so far restricted to the determination of atomic column positions in the image plane perpendicular to the incident electron beam. Due to the fact that the depth resolution of the TEM technique along the beam direction is inferior to its lateral resolution, full 3D structure determinations on an atomic level remain highly challenging. The 3D structure retrieval problem can be solved with tomographic methods, where a multitude of images is acquired from different observation directions. Such multi-image approaches are very demanding at atomic resolution due to instrumental instabilities [2] and due to a possible radiation damage of the object. Alternatively, single-image approaches, where only one exposure is taken along a crystallographic zone axis, have been successfully used to count the number of atoms in crystalline columns running parallel to the beam direction. However, a full 3D determination of the crystal shape would additionally require a highly accurate determination of all column positions along the beam direction, which has not been achieved so far with the single-image approach.

We demonstrate that the full 3D shape of a thin MgO crystal can be determined in a nearly unique way from a single HRTEM image (Fig. 1). Our 3D determination of the crystal shape is based on refining an atomic structure model (Fig. 2) in such a way that a HRTEM image simulated on the basis of this model fits best to the experimental image. In contrast to the usual simplifying assumption of flat lower and upper object surfaces in conjunction with a single global defocus value [3], our structure refinement is executed now locally column-by-column, allowing also for atomically corrugated object surfaces. The comparison between simulation and experiment is made on the basis of absolute image intensity values [4]. A crucial part of our procedure is an extended statistical confidence test which yields detailed quantitative statements on the uniqueness and the reliability of the retrieved 3D crystal shape.

References:

[1] K.W. Urban, Science 321 (2008) 506.
[2] J. Barthel and A. Thust, Ultramicroscopy 134 (2013) 6.
[3] C.-L. Jia et al, Microsc. Microanal. 19 (2013) 310.
[4] A. Thust, Phys. Rev. Lett. 102 (2009) 220801.


Fig. 1: High-resolution image of the edge of an MgO crystal taken along the [001] zone axis with a CS-corrected FEI Titan 80-300 electron microscope at 300 kV accelerating voltage. The 3D shape reconstruction was performed at the area indicated by the dashed box.

Fig. 2: 3D structure model retrieved from the boxed area in Fig. 1. Red balls indicate Mg atoms, yellow balls O atoms, purple balls indicate formally half-occupied Mg positions, and green balls formally half-occupied O positions.

Type of presentation: Oral

IT-2-O-1857 Towards a quantitative exit wave function: the influence of phonon scattering

Liberti E.1, Kim J. S.1, Kirkland A. I.1
1Department of Materials, University of Oxford, Parks Road, Oxford, OX1 3PH, UK
emanuela.liberti@materials.ox.ac.uk

To provide the utmost quantitative information about the atomic structure of the specimen is the ultimate challenge sought by modern high-resolution transmission electron microscopy. At the specimen exit surface, quantitative structural information is embedded in the object complex wave function, which can be recovered, with atomic resolution, from a focal (or tilt) series of aberration corrected HRTEM images [1]. Nonetheless, the quantitative information that is obtained from the exit wave is often in disagreement with imaging simulations. This disagreement is in effect a contrast mismatch, or Stobbs factor, which accounts for a reduction of the experimental image contrast by a factor of three with respect to the calculations [2]. The scattering of phonons following the electron beam-specimen interaction is amongst the possible causes of the Stobbs factor [3].
In this contribution, we discuss the role of phonon scattering in the quantification of the exit wave function of a single layer of graphene. For this idealized object, the contribution of the thermal phonon scattering to the total elastic scattering can be directly investigated by quantifying the exit wave function at different temperatures. For the imaging simulations, the influence of thermal motion upon modeling of the elastic scattering is studied quantitatively, using both the absorptive potential and frozen phonon approaches, addressing the role of the Debye-Waller factor in predicting the thermal displacement of graphene atoms. Experimentally, the exit wave function is recovered in the linear imaging approximation, in both heating and cooling conditions, as well as at room temperature.
To conclude, we present, and discuss, the comparison between the quantitative exit wave functions, obtained in both calculated and experimental approaches.

[1] A.I. Kirkland, S. J. Haigh, Jeol news, 44 (2009) 6 – 11.
[2] M.J. Hÿtch, W.M. Stobbs, Ultramicroscopy 53 (1994) 191 – 203.
[3] A. Howie, Ultramicroscopy 98 (2004) 73 – 79.


The research leading to these results has received funding from the European Union Seventh Framework Programme under Grant Agreement 312483-ESTEEM2 (Integrated Infrastructure Initiative–I3).

Type of presentation: Oral

IT-2-O-1921 Assessment of lower-voltage TEM performance using 3D Fourier transform of through-focus images

Kimoto K.1, Ishizuka K.2
1National Institute for Materials Science, Tsukuba, Japan, 2HREM Research Inc., Higashimatsuyama, Japan
kimoto.koji@nims.go.jp

The performance of an aberration-corrected TEM is determined by the information limit that is often demonstrated using Young's fringe method. However Young's fringe method could show unexpected high frequency information due to the non-linear terms as pointed out by several researchers [1,2]. The three-dimensional (3D) Fourier transform (FT) of through-focus TEM images allows us to discriminate between the linear and the non-linear imaging terms [3,4]. The linear imaging terms are observed on twin Ewald spheres in the 3D FT using an amorphous specimen. Here, we use the 3D FT of through-focus TEM images for the assessment of two low-voltage TEM systems.

Two spherical-aberration-corrected microscopes were assessed and compared. One was a Titan3 (FEI) equipped with a monochromator and a spherical aberration corrector for image forming (CEOS, CETCOR) operated at an acceleration voltage of 80 kV. The energy spread of the electron source was 0.1 eV under monochromated condition. The other microscope, the TripleC microscope, was equipped with a cold field-emission gun (CFEG) and the spherical aberration corrector developed for the TripleC project. This microscope was operated at 60 and 30 kV [5], and the energy spread was 0.3-0.4eV.

Figure 1 schematically shows various 3D data processed in this study [6]. Acquired through-focus TEM images are stacked as a function of the defocus z (Fig. 1a). The 3D Fourier transform Iuvw (Fig. 1c) of through-focus images shows two paraboloids called Ewald spheres, attached at the origin. The information limit can be estimated as an observable range of the Ewald spheres.

The signal of Ewald spheres depends on various factors, such as atomic scattering factors, a specimen structure, thickness, and the modulation transfer function of an imaging device; therefore, the quantitative evaluation of diverse TEM systems is not straightforward. Here we apply the tilted incidence in the 3D Fourier transform method (Fig. 2) to normalize those factors. We evaluate the spatial frequency at which information transfer decreases to 1/e2 (Fig. 3). It was found that the energy spread of the electron source is the major limiting factor even in a monochromated TEM [7].

[1] M. Haider et al., Microsc. Microanal. 16 (2010) 393. [2] J. Barthel, et al., Phys. Rev. Lett. 101 (2008) 200801. [3] Y. Taniguchi, et al., J. Electron Microsc. 40 (1991) 5. [4] M. Op. de Beeck et al., Ultramicrosc. 64 (1996) 167. [5] H. Sawada et al., Ultramicrosc. 110 (2010) 958. [6] K. Kimoto et al., Ultramicrosc. 121 (2012) 31. [7] K. Kimoto et al., Ultramicrosc. 134 (2013) 86.


We thank Drs. Nagai, Freitag, Sawada, Sasaki, Ohwada, Sato and Suenaga for invaluable discussions. This work is supported by Nanotechnology Platform of MEXT and Research Acceleration Program of JSPS.

Fig. 1: Schematics of (a) through-focus TEM images Ixyz, (b) stack of 2D FTs Iuvz, and (c) 3D FT of the through-focus images Iuvw. Since Iuvz and Iuvw are complex, their moduli are shown in gray scale. The cross section Ivz is similar to the Thon diagram. Two Ewald spheres attached at the origin are observed in the 3D Fourier space Iuvw.

Fig. 2: Cross sections of 3D FTs under on-axial and tilted incidence conditions. (a) Titan3 (80kV) and (b) TripleC (60kV).

Fig. 3: Information limit of (a) monochromated Titan3 (80kV), TripleC at 60kV (b) and 30kV (c).

Type of presentation: Oral

IT-2-O-2157 An alternative normalization method for quantitative STEM

Martinez G. T.1, Jones L.2, Béché A.1, Verbeeck J.1, Nellist P. D.2, Van Aert S.1
1Electron Microscopy for Materials Science (EMAT), University of Antwerp, Antwerp, Belgium, 2Department of Materials, University of Oxford, Oxford, United Kingdom
gerardo.martinez@uantwerpen.be

Techniques such as annular bright-field (ABF) or high-angle annular dark-field (HAADF) scanning transmission electron microscopy (STEM) have become widely used in quantitative studies because of the possibility to directly compare experimental and simulated images when the experimental data is expressed in units of ‘fraction of incident probe’ [1]. This is achieved by subtracting by the amplifier’s ‘black-level’ normalizing the experimental image by the mean sensitivity of the annular detector. Since the detector response is spatially inhomogeneous [2], a ‘detector sensitivity’ profile needs to be included in image simulations in order to account for these irregularities. Unfortunately, the quantification procedure now becomes both experiment and instrument specific, with new simulations needing to be carried out for the specific response of each instrument’s detector. This not only impedes the comparison between different instruments but can also be computationally very time consuming.

In this work, we propose an alternative method for normalizing experimental data in order to compare these with simulations that consider a homogenous detector response. To achieve this, we determine the electron flux distribution reaching the detector by means of a camera length series, which is then used to determine the corresponding weighting of the detector response. Figure 1a) shows the detector scan and b) its corresponding active area. The electron flux reaching the active area of the detector is shown in Figure 1 c), which was determined using a camera length series (Figure 2). Next, after normalizing this flux profile to unity, it is multiplied pixel-wise with the experimental detector map, Figure 1d), in which the detector response inhomogeneity is clearly observed. By integrating Figure 1d), we obtain an overall ‘flux-weighted detector sensitivity’ value, which can be used for the experimental data normalization. To validate the proposed methodology, we simulated a [100] oriented Pt crystal using the StemSim software under the frozen lattice approach [3]. The simulations considered homogeneous and inhomogeneous detector sensitivities for 60 – 190 mrad detector acceptance angles. Figure 3 shows that the total intensity for a simulation considering inhomogeneous detector sensitivity followed by electron flux weighting (analogous to experimental conditions) is in perfect agreement with simulations performed with homogeneous detector sensitivity (the ideal case).

[1] J. M. Lebeau and S. Stemmer, Ultramicroscopy 108 (2008), p.1653-8
[2] K. MacArthur, L. Jones, and P. Nellist, Journal of Physics: Conference Series (2013)
[3] A. Rosenauer and M. Schowalter, Springer Proceedings in Physics, vol. 120 (2007), p. 169–172.


Funding from the FWO Flanders, the EU FP7 (312483 - ESTEEM2), and the UK Engineering and Physical Sciences Research Council (EP/K032518/1) is acknowledged.

Fig. 1: Proposed flux-weighted normalization steps: a) experimental detector map, b) detector active area, c) determined flux pattern using camera length series, and d) flux-weighted sensitivity resulting from product of plots a) and c).

Fig. 2: Measured electron flux distribution from simulated camera length series. Using this plot, Figure 1c) is computed for the detector active area.

Fig. 3: Total scattered intensity for homogeneous (blue) and inhomogeneous (black) detector sensitivity. Red circles correspond to the total scattered intensity of inhomogeneous detector sensitivity after electron flux weighted normalization.

Type of presentation: Oral

IT-2-O-2221 Channeling Effects on the Accuracy of HAADF STEM Quantification of Bimetallic Catalyst Nanoparticles

MacArthur K. E.1, Lozano-Perez S.1, Ozkaya D.2, Nellist P. D.1
1Department of Materials Science, University of Oxford, Oxford, UK , 2Johnson-Matthey Technical Centre, Reading, UK
katherine.macarthur@materials.ox.ac.uk

Quantification of high angle annular dark-field scanning transmission electron microscope (HAADF STEM) images uses atomic resolution images as data sets to extract sample composition and thickness information from. A new quantification method based on calculating the scattering cross-section (CS) of each atomic column has been shown to be more robust to microscope image parameters.1 Using an automated code,2 this analysis involves converting images to an absolute scale through detector normalisation3, integrating over each atomic column within an image and multiplying by pixel area.
Whilst mathematically robust to microscope imaging parameters there are many other factors which affect the accuracy of quantification results. Channelling occurs when the columns of atoms in a crystal are aligned parallel to the incident electron beam and they act like miniature lenses providing an extra focusing effect on the probe. The subsequent atoms in the atomic column see a more focused probe than the first atom; resulting in them supplying increased scattering out to the detector. Along the length of the column, oscillations in intensity are seen, much as though the electrons are propagating in a waveguide. The whole column may therefore have a different scattering CS than the sum of the individual CSs of its constituent atoms. The ordering of atom types within an atomic column also affects the overall CS. Comparably another process known as de-channelling provides cross-talk between neighbouring columns of atoms. Cross-talk occurs when part of the probe is scattered and becomes channelled by a neighbouring column of atoms and then scattered out to the detector, thereby contributing information to the signal from neighbouring columns.
Atomic resolution requires viewing a crystal down a low order zone axis; any sample mis-tilt away results in a reduction in the channelling contribution and therefore a loss in CS, Figure 1. By 4̊ of tilt the effects of channeling are almost completely lost, whilst some atomic resolution remains. Top-bottom effects in the bimetallic columns are also diminished by 4̊ mis-tilt. At small tilts, however, there is a plateau region where the CS is independent of tilt, the size of which is dependent on probe convergence angle size, Figure 1. We believe the robustness to tilt when imaging on axis is more beneficial than the potential composition information gain from tilting far off a zone axis. This is particularly the case for nanoparticles which tilt under the beam. Combining with spectroscopy techniques will be necessary for gaining compositional information.

1 H E et al, Ultramicroscopy 133 (2013), p109-19
2 The Absolute Integrator is free for academic use from www.lewyjones.com/software/
3 JM LeBeau et al, Nano Letters 10 (2010), p4405-8


The research leading to these results has received funding from the European Union Seventh Framework Programme under Grant Agreement 312483 - ESTEEM2 (Integrated Infrastructure Initiative–I3), and from the EPSRC (grant number EP/K032518/1).

Fig. 1: The CS of a column of 7 Pt atoms in a crystal plotted against tilt away from a <110> zone axis. (a) Pt, blue, with the top or bottom atom replaced with Co, red and green, show a plateau region before dropping with tilt. With no channelling the CS would be 7x1 Pt atom, dotted red. (b) Tilt plot with different probe convergence angles.

Type of presentation: Oral

IT-2-O-2283 Measuring surface atom structures in Pt and Au nanocatalysts with high precision STEM imaging

Yankovich A. B.1, Berkels B.2,3, Dahmen W.2,4, Binev P.2, Sanchez S. I.5, Bradley S. A.5, Voyles P. M.1
1University of Wisconsin – Madison, USA, 2University of South Carolina, USA, 3Universität Bonn, Germany, 4RWTH Aachen, Germany, 5UOP LLC a Honeywell Company, USA
ayankovich@wisc.edu

TEM and STEM aberration correctors make sub-Ang resolution imaging routine. Once atoms are resolved, the question is how precisely can their positions be measured? TEM and STEM regularly achieve precision smaller than the resolution, but STEM encounters practical limits, such as image distortions from instabilities, before reaching the signal to noise ratio (SNR) fundamental precision limit. Combining multiple frames improves SNR and precision. Rigid registration is a common approach, but it does not correct for all types of instabilities. We have developed a non-rigid (NR) registration scheme for STEM images that accounts for all types of image distortions caused by instabilities during acquisition[1, 2].

Fig. 1 shows the results of the NR registration and averaging of a series of 512 HAADF STEM images of GaN. We show that sub-pm precision is achieved by fitting each Ga column to a 2D Gaussian, calculating the interatomic separations as shown in the histograms in Fig. 1(c) and (d), and using the standard deviation as the precision. The sub-pm precision in the x and y directions (0.74 and 0.85 pm) is reproducible and is 5-7 better than rigid registration.

A multislice simulated HAADF STEM image of a Si [110] dislocation core, shown in Fig. 2(a) was used to create an image series that includes distortions representative of real experiments, including thermal drift, floor vibrations, acoustic noise, electromagnetic fields, and electronic instabilities. Fig. 2(b) shows the NR registered and averaged image of the distorted series, demonstrating that inhomogeneous strain is preserved by NR registration.

NR registering and averaging STEM images allows for pm-scale measurements of surface atom bond length variation in Pt nanoparticles, which are prototypical noble metal catalysts. NP’s surface structure is crucial to their chemical activity but measuring it is extremely challenging. Fig. 3(a) shows that a Pt nanocatalyst exhibits pm-scale contraction of atoms at a (1-11)/(-1-11) corner and expansion of a (1-11) facet, with very little lateral displacement. Standardless atom counting on the same NR registered STEM image shows that the Pt NP is between 1 and 8 atoms thick with <1 atom uncertainty, as shown in Fig. 3(b). High precision in both positions and thickness are enabled by the extremely high SNR after NR registration. In general, STEM imaging with pm-precision will aid in understanding atomic displacement fields important in catalysis, defects, interfaces, and ferroic materials.

[1] Berkels et al, Ultramic. 138, 46 (2014).

[2] Yankovich et al, “Picometer-Precision Analysis of STEM Images of Pt Nanocatalysts” under review (2014).


We acknowledge funding from the Department of Energy, Basic Energy Sciences (DE-FG02-08ER46547), NSF (DMS 1222390), USC’s Special Priority Program SPP 1324, and the Excellence Initiative of the German federal and state governments, and the UW Materials Research Science and Engineering Center (DMR-1121288).

Fig. 1: (a) The first of 512 HAADF STEM images of GaN [11-20]. (b) Average of 512 frames after NR registration. The red dots are the positions of the columns identified by fitting. (c) and (d) Histograms of the X and Y separation measurements from (b).

Fig. 2: (a) Simulated HAADF STEM image of a Si [110] dislocation core model displaying an inhomogeneous strain field. (b) NR registered and averaged HAADF STEM image after the simulated image in (a) was made into an image series with typical distortions representative of experimental series.

Fig. 3: (a) Average of 56 HAADF STEM images of a Pt [011] nanoparticle after NR registration. The red arrows show magnified displacement vectors of the surface atoms and the displacement magnitudes are labeled in white. (b) The number of atoms in each column determined by comparing the experimental absolute intensities to simulations.

Type of presentation: Oral

IT-2-O-2611 The detection of single dopant atoms by high-resolution off-axis electron holography

Cooper D.1, Mayall B.1, McLeod R.2, Haigh S.3, Dunin-Borkowski R. E.4
1CEA-LETI, Minatec, 17 rue des Martyrs, 38054 Grenoble, Cedex 9, France, 2CEA-INAC, Minatec, 17 rue des Martyrs, 38054 Grenoble, Cedex 9, France, 3School of Materials The University of Manchester, Manchester M13 9PL, U.K., 4Ernst Ruska-Centre for Microscopy and Spectroscopy with Electrons, D-52425 Jülich, Germany
david.cooper@cea.fr

In 2003, a paper was published that discussed the need for a technique to detect single dopant atoms. It was stated that “single dopant atom detection is critical to device design, but it can also unravel complex and unexpected phenomena which may also open up new areas of materials exploration”. In 2014, it is still highly challenging to measure the locations, chemical identities and electrostatic potentials of single dopant atoms [1].

The technique of off-axis electron holography in the transmission electron microscope (TEM) involves the use of an electron biprism to interfere an electron wave that has passed through a thin specimen with a reference wave, in order to form an interference pattern that can be used to determine the phase shift of the electrons. Although electron holography has been used for many years to measure dopant potentials in semiconductors, soon devices will become so small that measurements of the electrostatic potentials of individual dopant atoms may be required.

Based on simulations, the expected step in phase shift across a single ionized P atom in Si is ~2π/1000 radians. This level of sensitivity can be reached easily in electron holographic measurements at low spatial resolution if long acquisition times are used. However, it is more of a challenge at atomic resolution. Here, we demonstrate progress towards the detection of single dopant atoms using electron holography. Figure 1(a) shows an electron hologram acquired at 80 kV using an aberration-corrected FEI Titan Ultimate TEM equipped with a high brightness gun, a monochromator and a single biprism. A careful choice of microscope lens settings allows holograms to be acquired with excellent interference fringe contrast and fine fringe spacing. Figure 1(b) shows an intensity profile extracted from the hologram, while Fig. 1(c) shows the experimentally measured phase resolution plotted as a function of interference fringe spacing, demonstrating that the conditions required to detect single dopant atoms are within reach if large numbers of phase images are added together.

Figure 2(a) shows part of an off-axis electron hologram of a thin MoS2 crystal recorded using an interference fringe spacing of 40 pm. The corresponding reconstructed phase image in Fig. 2(b) has a spatial resolution of 0.12 nm, while the line profile in Fig. 2(c) demonstrates that individual atomic columns with a spacing of 0.12 nm can be resolved. We are presently working towards the acquisition of signals from single dopant atoms in graphene and silicon and comparing our results with scanning TEM images. Great care is required to optimize specimen preparation and to minimize radiation damage, electron beam induced charging and contamination.

[1] Castell et al. Nature Materials 2, 129-131 (2003)


DC and RDB thank the ERC for the starting grant “Holoview” and the advanced grant “IMAGINE” respectively.

Fig. 1: (a) An off axis electron hologram acquired with a fringe spacing of 50 pm (b) profile of the fringe intensity showing a contrast of 33 %. (c) Experimentally measured phase resolution as a function of fringe spacing (spatial resolution is 2-3 times the fringe spacing).

Fig. 2: (a) Detail of an off-axis electron hologram of a MoS2 crystal with a fringe spacing of 40 pm (b) reconstructed map of the electrostatic potential profile and (c) profile showing that the 1.2 A spaced atoms have been resolved.

Type of presentation: Oral

IT-2-O-2414 A Method to Analyse the Chemical Composition in (InGa)(NAs) based on Evaluation of HAADF Intensity in STEM

Grieb T.1, Müller K.1, Mahr C.1, Cadel E.2, Beyer A.3, Talbot E.2, Schowalter M.1, Volz K.3, Rosenauer A.1
1Institute of Solid State Physics, University of Bremen, Otto-Hahn-Allee 1, 28359 Bremen, Germany, 2Groupe de Physique des Matériaux (GPM) UMR 6634, Normandie Université, Université et INSA de Rouen–CNRS, Av. de l’Université, BP 12, 76801 Saint Etienne du Rouvray, France, 3Materials Science Center and Faculty of Physics, Philipps University Marburg, Hans-Meerwein-Straße, 35032 Marburg, Germany
rosenauer@ifp.uni-bremen.de

InxGa1-xNyAs1-y is of technological interest for laser diodes in telecommunication and solar cells as both, In and N, lower the semiconductors band gap to emit or absorb in the infra-red spectral range. It was shown for ternary materials that an unknown chemical concentration (eg. of In in InGaN [1]) can be determined by high-angle annular dark field (HAADF) scanning transmission electron microscopy (STEM). For this purpose experimental HAADF intensities are compared with simulated ones. The experimental intensities are normalized to the total beam intensity which allows for determining the thickness in regions with known chemical composition. In this contribution this method is extended to evaluate the quaternary system InxGa1-xNyAs1-y. As a specific HAADF intensity cannot be allocated to a pair of concentrations (x,y) in a unique way, further information is needed. To this end, the local strain state is additionally determined from the high-resolution HAADF-STEM image.

The HAADF intensities were simulated with a frozen-lattice multislice approach implemented in the STEMsim software [2], considering thermal diffuse scattering (TDS). It was shown that for (In)GaNAs, besides TDS, Huang-scattering at static-atomic displacements (SADs) has to be taken into account [3]. SADs are distortions of the atomic lattice due to different covalent radii of In and Ga as well as As and N. The SADs were computed by relaxing the supercells using Keating's valence force field parametrization [4] in the LAMMPS code [5]. Fig. 1. shows the ratio of the simulated HAADF intensity of InGaNAs and GaAs versus specimen thickness for different In and N concentrations. For thicknesses above approx. 50 nm the intensity ratio increases not only with In but also with N concentration, although N has a smaller atomic number than As. This effect reveals the strong influence of additional scattering at SADs. An MOVPE grown InGaNAs/GaAs quantum-well sample is characterized by the outlined method. The mean concentrations of 32 % In and 2 % N (see concentration profiles in Fig. 2) are in good agreement with the results from XRD (marked by arrows). In addition, atom-probe tomography was applied to this sample, and the corresponding In profile is also shown in Fig. 2. Both, profile shape and mean concentration are in good agreement with the HAADF-STEM results.

[1] Rosenauer et al., Ultramicroscopy 111 (2011) 1316.

[2] A. Rosenauer and M. Schowalter, Springer Proc. Phys. 120 (2007), 169.

[3] Grillo et al., Phys. Rev. B 77 (2008), 054103.

[4] P. N. Keating, Phys. Rev. 145 (1966), 637.

[5] S. Plimpton, J. Comput. Phys. 117 (1995), 1.


We thank the DFG under contracts SCHO 1196/3-1, RO 2057/8-1 and GRK1782.

Fig. 1: Ratio of the simulated HAADF intensity for InGaNAs and GaAs (material contrast) as a function of specimen thickness for different indium concentrations (color) and nitrogen concentrations (line style). The HAADF intensity increases for specimen thicknesses above 50 nm with In and N concentration due to Z-contrast and scattering at SADs.

Fig. 2: Determination of the chemical composition of an InGaNAs layer embedded in GaAs. Concentration profiles from averaging concentration maps (HAADF analysis: indium and nitrogen) and from atom probe tomography (only indium). Concentrations derived from HRXRD are marked by arrows.

Type of presentation: Oral

IT-2-O-2538 Optimizing Phase Contrast Imaging in Aberration Corrected TEM

Kahl F.1, Hartel P.1, Linck M.1, Müller H.1, Haider M.1
1Corrected Electron Optical Systems GmbH, Heidelberg, Germany
frank.kahl@gmx.net

Since the realization of the first aberration corrected TEM [1], the number of corrected TEMs is still rapidly growing. Two key benefits are boosting the tremendous success of spherical aberration correction: vanishing delocalization and improved point resolution limit. The latter is achieved by using CS=C3 as additional optimization parameter to increase the aperture radius where the phase shift distribution (phase plate) of the elastically scattered electrons stays close to the optimum of +π/2 or −π/2 for dark or bright atom contrast, respectively. Various efforts have been made to optimize the phase plate, e.g. [2, 3], employing different measures for the distance of real and ideal phase plate as criterion for optimization. However, the different criteria lead to similar results.

Many users still use a π/4-limit for each aberration coefficient separately to assess the corrected state. However, advanced criteria such as minimizing the integrated mean quadratic deviation [3] or minimizing the largest deviation from the ideal phase over the aperture are much more sensible. While designing the SALVE II corrector [5] we used the latter criterion for optimizing the phase plate to assess imaging quality.

Fig. 1 shows the phase shifts generated by all axial aberration coefficients up to fifth order. Compensation schemes can be applied for aberrations of same multiplicity but different orders. In Fig. 1 potential partners are arranged within one column. The highest-order coefficient and the aperture size determine the optimum values for the lower-order coefficients (e.g. multiplicity 2: S5 given by corrector design; S3A1 optimized during alignment). The procedure for optimizing coefficients of non-zero multiplicity is similar to optimize C3 and defocus C1 for a given C5, except that the deviation from zero instead of +π/2 or −π/2 is minimized.

The performance of the SALVE II corrector for 40 kV is shown in Fig. 2. The phase plate (a) corresponds to the output of the CEOS software after aberration correction. The π/4-circle is misleadingly small as it is largely determined by C3. In image (b) only the sum of all non-round contributions is shown. Optimizing C1 for given C3 and C5 yields passband (c). Only with a full compensation scheme (d, e) for all fourth- and fifth-order aberrations using all adjustable lower order aberrations, a passband of up to 50 mrad can be achieved.

References:

[1] M. Haider et al, Nature 392 (1998), 768-769.

[2] O. Scherzer, Ber. Bunsen-Gesellschaft phys. Chemie 74 (1970), 1154-1167.

[3] M. Lentzen, Microsc. Microanal. 14 (2008), 16-26.

[4] M. Born, E. Wolf, Principles of Optics, 6th edition (Cambridge university press, Cambridge), p. 468.

[5] SALVE II project, <http://www.salve-project.de>.


none

Fig. 1: Table of axial aberrations visualized as phase plates. At the aperture edge ±6π is adopted; the phase is wrapped to [-π (black); π (white)[. Aberrations of same multiplicity but different order can partly compensate each other.

Fig. 2: Left: Phase plates for measured aberration coefficients at the SALVE II microscope operated at 40 kV. The passband in (c) demonstrates the reduced contrast transfer due to non-round residual aberrations. Right: With a suitable compensation scheme a passband up to 50 mrad can be achieved. The aperture radius of all phase plates is 75 mrad.

Type of presentation: Oral

IT-2-O-2644 Low-voltage and energy-filtered chromatic aberration-corrected high-resolution TEM on the PICO instrument

Houben L.1,3, Luysberg M.1,3, Barthel J.2,3, Mayer J.2,3, Dunin-Borkowski R. E.1,3
1Peter Grünberg Institut 5, Forschungszentrum Jülich, Germany, 2Gemeinschaftslabor für Elektronenmikroskopie, RWTH Aachen, Germany, 3Ernst Ruska-Centre for Microscopy and Spectroscopy with Electrons, Forschungszentrum Jülich, Germany
l.houben@fz-juelich.de

The advent of chromatic aberration correction in the transmission electron microscope (TEM) offers new prospects for high-resolution imaging at low accelerating voltages and for energy-filtered TEM (EFTEM). Examples of low voltage and energy-filtered images of complex oxides, thin layered materials, and nanoparticles will be presented, demonstrating the unique optical properties of an achroplanatic CEOS CCOR Cc/Cs corrector on Jülich’s chromatic aberration-corrected “PICO” microscope.
Atomic-resolution imaging at low voltages, currently down to 50 kV, allows high-resolution studies of radiation-damage-sensitive nanomaterials, such as CdSe/CdS nanostructures obtained from a cation exchange reaction, graphene and carbon nanotubes. It is also beneficial for the study of organic ligands, ligand-stabilised materials and materials that are functionalized with organo-metallic compounds.
The ability to acquire dose-efficient atomic-scale EFTEM elemental maps using the achroplanatic CCOR corrector on this microscope with a large field of view and large energy windows results from the fact that the chromatic focus spread is negligible after chromatic aberration correction. Figure 1 shows an example of an atomic-resolution elemental map of Ca obtained from a thin TEM foil of a calcium-titanate/strontium-titanate multilayer. Figure 2 shows a structural and compositional modulation in a (CeS)1.2CrS2 misfit-compound nanotube, which comprises alternating hexagonal CrS2 and rock-salt CeS sheets that have a repeat period of 11.2 Ångstrom.
The quantification of EFTEM elemental maps to provide atomic-resolution information about the local chemical composition of a specimen is complicated by the preservation of elastic contrast due to elastic scattering, which gives rise to thickness and defocus dependent contrast with fine details at all energy losses. Optical stability over minutes of collection time and careful image alignment and background subtraction are also required to obtain meaningful and reliable atomic-scale EFTEM elemental maps.


The authors thank J. Schubert (Forschungszentrum Jülich) and L. Penchakarla and R.Tenne (Weizmann Institute of Science) and M. Bar Sadan (Ben Gurion University) for kindly providing the materials used in this study

Fig. 1: High-resolution EFTEM images of a CaTiO3/SrTiO3 [001] multilayer sample taken at 300 kV. Ca L23 (a) pre-edge image, (b) post-edge image and (c) background-subtracted map. (d, e, f) Noise-reduced images obtained by averaging over 5x5 periods in the CaTiO3 layer, revealing Ca on the A sites of the pseudo-cubic perovskite lattice.

Fig. 2: Atomic-resolution micrographs and spectroscopic images of a CeCrS3 nanotube. (a) Schematic view of the CeCrS3 misfit lattice and a tubular structure. (b) HRTEM image and (c) magnified region of (b) with crystallographic projections superimposed. (d) EFTEM maps showing alternating Ce and Cr signals and (e) magnified region of (d).

Type of presentation: Oral

IT-2-O-2658 Atomic resolution secondary electron imaging and simulation of the SrTiO3 (001) c(6x2) surface reconstruction

Ciston J.1, Brown H. G.2, D'Alfonso A. J.2, Koirala P.3, Lin Y.3, Ophus C. L.1, Inada H.4, Zhu Y.5, Allen L. J.2, Marks L. D.3
1National Center for Electron Microscopy, Lawrence Berkeley National Lab., Berkeley, USA, 2School of Physics, University of Melbourne, Parkville, Victoria, Australia, 3Department of Materials Science and Engineering, Northwestern University, Evanston, USA, 4Hitachi High Technologies Corporation, Ibaraki, Japan, 5Condensed Matter Physics and Materials Science, Brookhaven National Laboratory, Upton, USA
jciston@lbl.gov

Aberration-corrected scanning transmission electron microscopes (STEM) enable simultaneous collection of atomically resolved signals relating to coherent scattering (bright field and annular bright field imaging), structural information based on thermal scattering to large angles known as high-angle annular dark-field (HAADF) imaging, bonding information using electron energy-loss spectroscopy (EELS), and element identification using both EELS and energy dispersive x-ray spectroscopy. Atomic resolution imaging based on secondary electron (SE) signals was demonstrated in 2009 [1], but the technique is only slowly growing in use. While these SE signals are highly surface sensitive due to the narrow escape depth of electrons with energies <50eV [2], atomic resolution SE imaging of surface structures that differ from a simple bulk crystal termination has not been previously demonstrated.


We have recently imaged the c(6×2) reconstruction on the (100) surface of single crystal SrTiO3 (Fig 1) through simultaneous atomic resolution SE and HAADF STEM with complementary HREM imaging. The ability to simultaneously record surface sensitive SE and bulk dominated HAADF signals at atomic resolution makes the problem of surface structure registration to the bulk lattice highly tractable, which is a distinct advantage over other scanning probe methods. By inspection it is clear that the registration of the previously reported structure, primarily refined from surface x-ray diffraction and scanning tunneling microscopy (STM) experiments [3], is incorrect. Interpretation of the experimental SE measurements from first principles is now possible using a recently developed quantum mechanical model to simulate SE images. This approach takes into account the probability and angular distribution of electrons that are ejected from atoms in the specimen when ionization of both core and semi-core electrons occurs [4]. Our preliminary simulations of a newly proposed structure of the SrTiO3-<100>-c(6×2) reconstruction are in good agreement with the bulk-subtracted experimental SE data (Fig 2), and consistent with previously reported data from STM, Auger spectroscopy, and x-ray diffraction measurements. The structure solved by SE imaging is also stable in density functional theory simulations, and is on the thermodynamic convex hull of known reconstructions on SrTiO3 <100>.


[1] Y Zhu et al., Nat. Mater. 8 (2009) p. 808
[2] A Howie, J. Microsc. 180 (1995) p.192
[3] CH Lanier, et al., Phys. Rev. B 76 (2007) 045421
[4] HG Brown et al., Phys. Rev. B 87 (2013) 054102


A portion of this work was performed at NCEM, supported by the Office of Science, Basic Energy Sciences of the U.S. Department of Energy under Contract No.: DE-AC02-05CH11231.

Fig. 1: (a) Weak beam dark field image of SrTiO3 001 c(6x2) single crystal with g=(200) showing atomically flat terraces (b) Transmission electron diffraction pattern of the c(6x2) reconstruction with c2mm symmetry acquired off-zone to minimize bulk dynamical diffraction

Fig. 2: (a) Bulk subtracted experimental secondary electron image of a 6×2 unit cell surface reconstruction on a <100> SrTiO3 substrate averaged over ~600 unit cells with c2mm symmetry enforced. The 200-keV probe had a convergence semi-angle of 25 mrad. (b) Prelimenary simulation of the result in (a) using the (projected) surface reconstruction indicated.

Type of presentation: Oral

IT-2-O-2674 Fast imaging with inelastically scattered electrons by off-axis chromatic confocal electron microscopy

Zheng C. L.1, Zhu Y.1, Lazar S.2, Etheridge J.1
1Monash University, Victoria, Australia, 2FEI Electron Optics, Eindhoven, The Netherlands
changlin.zheng@monash.edu

Imaging with inelastically scattered electrons is an important method for studying the composition and electronic properties of materials down to the atomic scale [1]. In this work, we describe an approach for fast mapping of inelastically scattered electrons using a scanning transmission electron microscope in a confocal mode, without using a spectrometer. We develop an off-axis scanning confocal electron microscope configuration using a double spherical-aberration corrected STEM/TEM. The electron probe is focused onto the sample at a significant angle to the optic axis of the imaging lens (Fig 1) and the probe-corrector retuned to form an atomic-scale electron probe in the specimen plane. Under the effect of the chromatic aberration of the imaging lens system, electrons with a chosen energy loss, the confocal energy, Ec, can be focused to a confocal point on the detector plane, while electrons of all other energies, including the zero loss electrons, will be chromatically defocused at that plane. In addition, the tilting of the incident beam laterally shifts the object exit wave in the back focal plane of the imaging lens, introducing an energy-related lateral displacement of the defocused probe. The inelastically scattered electrons are then chromatically dispersed both parallel and perpendicular to the optic axis, effectively separating electrons with different energies. In particular, electrons with the confocal energy can be detected selectively using an integrating detector. Using a synchronized set of scan-descan coils, these confocal electrons can remain focused on the detector as the electron probe is scanned across the specimen (Fig 1).

We illustrate the method with nanoscale core-loss chemical mapping of silver (M4,5) in an aluminium-silver alloy and atomic scale imaging of the low intensity core-loss La (M4,5@ 840eV) signal in LaB6 (Fig 2). The scan rates are up to 2 orders of magnitude faster than conventional STEM spectrum imaging methods recorded by CCD, enabling a corresponding reduction in radiation dose and improvement in the field of view [2]. Moreover, this off-axis chromatic confocal configuration offers the potential for fast nanoscale three-dimensional chemical mapping when coupled with the improved depth and lateral resolution of the incoherent confocal mode [3].

[1] R. F. Egerton, Electron energy-loss spectroscopy in the electron microscope (Plenum Press, New York, 1996), 2nd edn.

[2] C. Zheng, Y. Zhu, S. Lazar, J. Etheridge, Physical review letters, accepted (2014).

[3] T. Wilson and C. Sheppard, Theory and practice of scanning optical microscopy (Academic Press, London ; Orlando, 1984)


Funding is acknowledged from the Australian Research Council Grants DP110104734 and LE0454166.

Fig. 1: (a) Optical diagram of off-axis scanning confocal electron microscopy. (b) Chromatic defocused probe image of amorphous carbon. The refocused chromatic confocal energy is centered at energy loss of 300 eV. Chromatic confocal point is indicated by the arrow.

Fig. 2: Atomic resolution off-axis SCEM map of lanthanum M4,5 (~840 eV) core loss electrons in LaB6. The pixel dwell time is 1.5 ms with image size 256 x 256.

Type of presentation: Oral

IT-2-O-2881 Putting a New Spin on Scanning Transmission Electron Microscopy

LeBeau J. M.1, Sang X.1, Grimley E. D.1
1North Carolina State University, Raleigh, North Carolina, USA
jmlebeau@ncsu.edu

The full text of the abstract is not available. Please contact the presenting author.


Type of presentation: Oral

IT-2-O-2885 Three-dimensional location of a single dopant with atomic precision by aberration-corrected ADF STEM

Ishikawa R.1,2, Lupini A. R.2, Findlay S. D.3, Taniguchi T.4, Pennycook S. J.5
1University of Tokyo, Japan, 2Oak Ridge National Laboratory, USA, 3Monash University, Australia, 4National Institute for Materials Science, Japan, 5The University of Tennessee, USA
ishikawa@sigma.t.u-tokyo.ac.jp

Impurity doping is the key technology for enhancing physical and chemical properties in semiconductors. These functional dopants usually take the form of isolated single atoms, and the materials properties have strong sensitivities to the doping concentration, spatial distribution and three-dimensional location of the dopants. The recent development of aberration-corrected electron microscopy has allowed the determination of the two-dimensional spatial distribution of single dopants with atomic spatial resolution. However, this resolution has been achieved only in the lateral directions, and the last dimension, depth, has not yet achieved atomic resolution.

Here we use quantitative annular-dark field scanning transmission electron microscopy (ADF STEM)[1] to directly visualize isolated single Ce dopants accommodated in bulk w-AlN single crystals[2], exhibiting strong visible-light photo-luminescence. Through combining with frozen phonon image simulations, we determine the three-dimensional location of the Ce dopant with single atomic-layer precision in depth[3].

On the basis of the mean signal value comparison between the experiment and the simulations, we estimate the number of atoms per column, and the atomic-resolution thickness map is shown in Figure 1a. During sequential acquisition, we observed a single Ce dopant jump from X to Y through the interstitial site (Figure 1b-d). For the two columns of X and Y, we performed image simulations of all the possible dopant configurations in depth. In the thicker specimen, it may be difficult to uniquely determine the depth location of a single dopant owing to strong dynamical intensity oscillation. To overcome this issue, we implemented multi-component analysis such as mean signal value, maximum peak intensity and profile fitting. As shown in Fig. 1e, the experimental profile at atom X is well matched with that of the simulation of dopant location to a 9 unit-cell depth. And similarly, atom Y is located to a depth of 8 unit-cells. To develop more general method, we also analyze the same data set with Bayesian statistical model, which does not require a priori knowledge of the number of atoms. And we obtained the same depth locations of Ce dopant. By tracking a single dopant, we could begin to determine the three-dimensional atom diffusion path within bulk materials.

References

[1] R. Ishikawa, A.R. Lupini, S.D. Findlay and S.J. Pennycook, Microsc. Microanal. 20, 99 (2014).

[2] R. Ishikawa, et al., Sci. Rep. 4 3778 (2014).

[3] R. Ishikawa, A.R. Lupini, S.D. Findlay, T. Taniguchi and S.J. Pennycook, Nano Lett., (2014) in press.


R.I. acknowledges support from JSPS Postdoctoral Fellowship. A.R.L. acknowledges support by the U.S. DOE. S.D.F. acknowledges support under the Discovery Projects funding scheme of the Australian Research Council (Project No. DP110101570). T.T. acknowledges support by a Grant-in-Aid for Scientific Research on Innovative Areas "Nano Informatics" (Grant No. 25106006) from JSPS.

Fig. 1: Figure 1. Sequentially acquired Z-contrast images of w-AlN viewed along the [11-20] direction, (a) atomic-resolution thickness map, (b) averaged over frames of 1-19, (c) frame 20, (d) frames of 21-40. (e) Z-contrast profiles obtained from atom X (exp.) and the simulations of Ce locations to 8, 9, 10-unit cells.

Type of presentation: Oral

IT-2-O-3220 3D strain in HAADF – STEM images

Guerrero-Lebrero M. P.1, Bárcena-González G.1, Guerrero E.1, Liu Y.3, Kepaptsoglou D. M.4, Ramasse Q.4, Li L.3, Lazarov V. K.2, Galindo P. L.1
1Department of Computer Science and Engineering, Universidad de Cádiz, 11510 Puerto Real, Spain, 2Department of Physics, University of York, Heslington, York, United Kingdom, 3Department of Physics, University of Wisconsin, Milwaukee, WI 53211, USA, 4SuperSTEM Laboratory, STFC Daresbury Campus, Warrington, WA4 4AD, United Kingdom
maria.guerrero@uca.es

Strain mapping can be used to analyze materials at the atomic-column level, measuring local displacements and strain, and so revealing lattice translations, dislocations and/or rotations. Several methodologies have been developed to determine 2D strain field mapping from HRTEM images, either in real space (peak finding) [1, 2] or in Fourier space (geometrical phase analysis, GPA) [3]. Since 3-dimensional strain is independent of the image plane it might be ideal to gain insight into the behavior, shape and deformations of nanomaterials.

First, a HAADF focal series of 93 images (between 0nm and 14nm at steps of 0.15nm) of epitaxial Bi2Se3 (0001) thin films grown by atomic layer molecular beam epitaxy was taken with a Nion UltrastemTM 100 transmission electron microscope at 100 kV. Then, strain mapping of 23 images (from 19 to 41) was calculated using the Peak Pairs Analysis (PPA, [2]) plug-in for DigitalMicrograph available from HREM Research Inc.

Figure 1 shows one image of this focal series at a depth of 7nm, figure 2 shows the corresponding 2D strain map of this slice. It can observe that the screw dislocation position (red circle in figure 2) does not correspond to the apparent intensity change in HAADF STEM image (red circle in figure 1). The dark area in Figure 1 can be interpreted as a triangular spiral characterized by atomically smooth terraces, the way in which this material grows [4].

Figure 3 shows the 3D strain reconstruction where the screw dislocation movement, shape and tilt can be observed. Eshelby-Stroh twist [5] in the screw dislocation can be recognized, the upper part of the dislocation rotates in clockwise direction and the lower part turns in a counter – clockwise direction. The dislocation tilt has been estimated to be 6.3º with regard to the optical axis.

[1]Kret, S., Ruterana P., Rosenauer A., Gerthsen D. Extracting quantitative information from high resolution electron microscopy. Phys. Status Solidi (b) 227(1):247-295 (2001)

[2]Galindo, P. L, Sławomir, K., Sanchez, A.M., Laval, Y., Yañez A., Pizarro, J., Guerrero, E., Ben, T., Molina, S.I. The Peak Pairs algorithm for strain mapping from HRTEM images. Ultramicroscopy 107:1186-1193 (2007)

[3]Hÿtch, M. J., Snoeck, E., Kilaas, R. Quantitative measurement of displacement and strain fields from HREMicrographs. Ultramicroscopy 74:131–146 (1998)

[4]Liu.Y, Li, Y. Y., Rajput, S., Gilks, D., Lari, L., Galindo, P.L., Weinert, M., Lazarov V. K., Li, L.. Tuning Dirac states by strain in the topological insulator Bi2Se3. Nature Physics. (2014)

[5] Eshelby, J.D., Stroh, A.N. CXL. Dislocations in thin plates, Philosophical Magazines Series 7 42, 1401 (1951)


Fig. 1: HAADF Bi2Se3 slice at 7nm thickness. Apparent screw dislocation position is marked by the red circle.

Fig. 2: exx strain map that corresponds to the image in the Figure 1 calculated using PPA software [2].

Fig. 3: 3D strain reconstruction of a Bi2Se3 screw dislocation. Positive and negative strains are shown in red and blue respectively and the dashed red line represents the optical axis. The reconstruction makes it possible to collect a great deal of information about the dislocation motion.

Type of presentation: Oral

IT-2-O-3292 Position resolved single electron response of the HAADF-STEM detector and improved method for intensity normalisation

Schowalter M.1, Krause F. F.1, Grieb T.1, Mehrtens T.1, Müller K.1, Rosenauer A.1
1Institut für Festkörperphysik, Universität Bremen, Bremen, Germany
schowalter@ifp.uni-bremen.de

Quantification of HAADF-STEM images as demonstrated in [1,2] is based on normalising the image intensity with respect to the incident electron beam and comparison with image simulations. For that the electron beam is scanned over the detector and the intensities Iout outside and Idet on the detector yield the normalized intensity Inorm=(I-Iout)/(Idet-Iout).

Recently, it has been shown that accidental electrons can hit the HAADF detector, although the electron beam is scanned in a specimen free area [3]. From such a “vacuum image” the number of counts caused by a single electron can be inferred and intensity can be scaled in units of electrons per pixels which enables an alternative way for STEM image quantification [3] and error estimation based on electron statistics.

In this contribution we show that accidentally impinging electrons cause artifacts in the normalization of image intensity using the detector scan technique (DST) [1,2]. We introduce an improved DST which is able to avoid such errors. In addition, we demonstrate a method for measuring single electron signals as a function of detector position.

The red line in Fig. 1 depicts a linescan through an HAADF image of an a-C wedge evaluated using the conventional DST. The normalised intensity exhibits a significant shift towards negative intensities in vacuum. This can be attributed to accidentally impinging electrons [3], whose dose is different for a detector scan (image mode) and a vacuum scan (diffraction mode). To account for this difference we suggest to replace Iout in the numerator by the intensity Ivac obtained from a vacuum scan. The result of this is shown by the blue line in Fig. 1, where the intensity in the vacuum region vanishes. Fig. 2 shows histograms of vacuum images for different dwell times, revealing a large peak at 9900 due to the background level of the detector as well as further peaks corresponding to one or more electrons per scan position. Different dwell times yield fundamentally different curves so that Ivac depends on dwell time. Therefore, vacuum scan and image scan must be performed with the same dwell time.

We also measured the spatially resolved response of the detector to a single electron by drastically decreasing the beam current and taking a series of 256 detector scans with 2048 by 2048 pixels. The position of the single-electron peak was measured in bins of 16 by 16 pixels and the position of the zero-electron peak was subtracted. Fig. 3 nicely depicts the position sensitive single-electron response.

[1] J. M. LeBeau and S. Stemmer, Ultramicroscopy, 108, 1653 (2008).
[2] A. Rosenauer et al., Ultramicroscopy, 109, 1171 (2009).
[3] R. Ishikawa, et al., Microscopy and Microanalysis, 20, 99 (2014).


Fig. 1: Normalized intensity along a linescan. Normalization was done using Iout and Idet as derived from a detector scan (red) as well as using the background level from the vacuum image (blue).

Fig. 2: Log. of the freq. of intensities in vacuum images for different dwell times. Each curve shows peaks corresponding to one or more electrons impinging on a certain pixel. The intensity is normalised with respect to the dwell time, so that the distance between e.g. one-electron peak and zero-electron peak is inversely proportional to the dwell time.

Fig. 3: Two-dimensional map showing the one-electron response as a function of position on the detector.

Type of presentation: Oral

IT-2-O-3488 Linking Thickness, Channelling and Secondary X-ray signals in Atomic Resolution Scanning Transmission Electron Microscopy

Weyland M.1, Findlay S. D.2, D'Alfonso A. J.3, Allen L. J.3
1Monash Centre for Electron Microscopy, Monash University, Melbourne 3800, Australia, 2School of Physics, Monash University, Melbourne 3800, Australia, 3School of Physics, The University of Melbourne, Melbourne 3010, Australia
matthew.weyland@monash.edu

Quantification of EDX signals at atomic resolution can be treated by separation into two components; the scattering of electrons prior to ionisation and the subtleties of X-ray generation, emission and collection. Significant progress has recently been made concerning the first of these factors, with Forbes1 showing that consideration of both elastic and thermal scattering is required to explain anomalous contrast variations. However, true quantification requires a similarly detailed approach to the X-ray side of the system, with proportionality between signal and composition dependent on a multitude of factors including scattering cross-section, X-ray Fluorescence yield, Adsorption and detector geometry. Kotula has demonstrated a reference based approach2, scaling signals to averages from areas of known chemistry and Kothleitner recently showed the use of a ‘non-channelling’ (off-axis) approach to scale signals for quantification3. Both of these approaches offer a potential solution, but one of the main limitations is a lack of experimental data linking thickness, channelling and collected signal for a known specimen and well characterised instrument.

Results will be presented of a systematic study between thickness and EDX signal for known crystal structures and compositions. These will be matched with image simulation taking into account elastic and thermal scattering. The results presentenced will be carried out using a dual aberration corrected FEI Titan3, with well-defined probe illumination conditions, fitted with a standard 30mm2 ultra-thin window Si(Li) detector (0.13 sr) and a new 60mm2 windowless SSD detector (0.3 sr). Thickness will measured by position averaged convergent beam electron diffraction (PACBED), with EDX spectra acquired scanning over the same specimen area. Results will be presented from several specimens including Strontium titanate, GaAs/InGaAs radial nanowire heterostructures and Al-Cu alloys. By recording data from multiple areas with different thicknesses, trends between thickness, X-ray signal and channelling condition and its implications for quantitative high resolution EDX will be explored.

1. B. D. Forbes, A. J. D’Alfonso, R. E. A. Williams, R. Srinivasan, H. L. Fraser, D. W. McComb, B. Freitag, D. O. Klenov and L. J. Allen, PRB 86 024108, 2012
2. P. G. Kotula, D. O. Klenov and H. S. von Harrach, M&M 18(4), 2012
3. G. Kothleitner, M. J. Neish, N. R. Lugg, S. D. Findlay, W. Grogger, F. Hofer and L. J. Allen, PRL 112(8) 085501, 2014


The Australian research council is acknowledged for financial support through grants DP130102538 and LE0454166 (FEI Titan3).

Type of presentation: Poster

IT-2-P-1522 Modification of an existing laboratory room to house a Cs corrected microscope.

Papworth A. J.1, Nellist P. D.1
1The Department of Materials, The University of Oxford, Oxford OX1 3PH UK.
adam.papworth@materials.ox.ac.uk

The first Cs corrected microscopes became generally available at the beginning of the 21st century. Cs corrected microscopes require very tight environmental conditions, which often means that when purchasing a Cs corrected microscope you also have to build a new building as well. An example of purpose built high resolution microscope laboratory is SuperSTEM in the UK. However, it is possible to meet the environmental conditions by converting existing rooms, removing the need for a new building, and therefore making ownership of a Cs corrected microscope more affordable.

The required environmental conditions fall into four groups; Electromagnetic force (EMF), Temperature, Acoustic and Vibration, where the biggest cause of instabilities can come from outside interferences, such as trains, power cables and general road traffic. In most cases these outside interferences can be mitigated, for example moving power cables; however trains and traffic cannot be relocated. This paper outlines measures that can be made to minimise the environment factors by careful design and choice of equipment.

The environmental targets set by the design team were as follows; EMF AC and DC <0.5mG, mechanical displacement (vertical and horizontal) <0.3µm, acoustic noise for all frequencies with a flat field response microphone <60db, room temperature and air movements targets were set as; temperature 20oC ±0.2 hr-1 fluctuation, air flow within the room was to be vertical with a minimum air flow of 100mm sec-1. These targets were considered as reasonable to obtain while also meeting the requirements of the Cs corrected microscopes, which were under consideration at that time of planning.

The final design of the rooms, equipment, anti-vibrational block and services gave the following results. The EMF measurement gave an AC X and Y of 0.05mG and Z 0.2mG with no significant DC component. The acoustics were compromised by noise coming from the floor above with a maximum of 55db at 120Hz. Room temperature was measured at 20oC ±0.08 over a five hour period with a 2.5kW load. The room remain within specification, even when the door was left open for two hours. The isolation block showed no external vibration being measured from the roads or surrounding buildings above the normal background. Vibration measurements were also taken during the night as wel


Type of presentation: Poster

IT-2-P-1525 Cs CORRECTED ATOMIC RESOLUTION TEM IMAGES OF THE HUMAN TOOTH ENAMEL CRYSTALS

REYES-GASGA J.1,3, TIZNADO-OROZCO G. E.2, BRÈS E. F.1
1Unité des Matériaux et Transformation (UMET). Université de Lille 1, Sciences et Technologies. Bâtiment C6. 59650 Villeneuve d’Ascq. Lille, France., 2Unidad Académica de Odontología. Universidad Autónoma de Nayarit. Edificio E7, Ciudad de la Cultura “Amado Nervo”, C.P. 63190 Tepic, Nayarit, Mexico., 3Permanet Address: Instituto de Física, UNAM. Circuito de la Investigación s/n. Cd. Universitaria, 04510 Coyoacán, Mexico D. F., México
jreyes@fisica.unam.mx

Aberration-corrected HR-STEM and HAADF images of the human-tooth-enamel crystallites are presented. These spatial and energy resolutions images have allowed to get information on the physical meaning of the central dark line (CDL) defect which leads to the anisotropic dissolution of the crystals.
Human tooth enamel is composed in 95% of hydroxyapatite crystals (HAP, Ca10(PO4)6(OH)2) which are elongated-plate-like of 30 to 60 nm wide and 100 to 200 nm long, approximately [1]. They are organized in microns-sized structures named “rods” or “prisms” that go from the enamel–dentin junction to the enamel surface (figure 1). The chemical analysis in the micron range of enamel by different analytical techniques, mainly EDS spectroscopy, has indicated the existence of trace elements. Thus, carbonated hydroxyapatite (c-HAP) with Na, Mg, Cl, as trace elements, has been stabled for these crystals [2].
When observed with the Transmission Electron Microscope (TEM), the enamel crystallites show a structural defect of 1 to 1.5 nm width in their central region approximately, the Central Dark Line (CDL) (figure 2), whose structure and role in the enamel structure itself is unknown yet [3, 4].
Several studies have shown that the CDL favors their anisotropic dissolution [4, 5]. During the carious process, for example, the enamel crystals are destroyed in a systematic fashion: first a series of hexagonal holes aligned along the [11-20] are observed, then the holes develop anistropically along the [0001] direction and cross the whole crystals [4, 5].
Enamel crystals are electron beam sensitive, other important parameter against the HRTEM observation (figure 3). Therefore, the use of low electron doses is critical during the study of the CDL. Therefore aberration corrected HR-STEM is the appropriated equipment for carrying out the chemical and structural analyses of the enamel crystallites.
Human tooth enamel samples were obtained from permanent non-carious human molar teeth, extracted for orthodontic or periodontal reasons. Samples were prepared in the FIB-FEI QUANTA 200 3D equipment using the two beams system.
JRG thanks to DGAPA-UNAM (contract IN106713), CONACYT and PASPA-DGAPA-UNAM for sabbatical support.

References
1. R.Z Le Geros, Calcium Phosphates in Oral Biology and Medicine ed H M Myers (San Francisco, CA: Karger). 1991.
2. G.E. Tiznado-Orozco et al., J. Phys. D: Appl. Phys. 42 (2009).
3. J. Reyes-Gasga et al., J. Mater. Sci. Mater. Med. 19, 877-882 (2008).
4. E. Brès et al., Journal de Physique, 51, C1-97-102, (1990).
5. E. Brès et al., Ultramicroscopy, 12, 367-372 (1984).


This research has received funding from the European Union Seventh Framework Programme under Grant Agreement 312483-ESTEEM2 (Integrated Infrastructure Initiative–I3).

Fig. 1: TEM image of human tooth enamel crystals inside the micron-sized structure named “prism”.

Fig. 2: Magnification of one of the human tooth enamel crystals shown in figure 1. The row indicates the presence of the “central dark line”.

Fig. 3: HRTEM image of a human tooth enamel crystals aligned along the [0001] direction. The arrow indicates the electron beam damage.

Type of presentation: Poster

IT-2-P-1600 Phase Contrast Transfer Function for Differential Phase Contrast in High Resolution Local Electric Field Measurements

Majert S.1, Kohl H.1
1Physikalisches Institut und Interdisziplinäres Centrum für Elektronenmikroskopie und Mikroanalyse (ICEM), Westfälische Wilhelms-Universität Münster, Wilhelm-Klemm-Straße 10, 48149 Münster, Germany
s_maje03@uni-muenster.de

Differential Phase Contrast (DPC) is a contrast mechanism that can be utilized in the Scanning Transmission Electron Microscope (STEM). Since the advent of DPC, the technique has been used to image magnetic fields within a specimen [1]. To this end, a ring detector is divided into four quadrants and the direct electron beam is placed within the ring, only overlapping a small part of the detector. In a classical interpretation, the direct beam is slightly tilted by the magnetic fields in the specimen, so that subtraction of different detector segement signals yields DPC. Recently, this DPC geometry was also employed to investigate local electric fields with high resolution [2,3].

To determine whether this interpretation of DPC is still valid in high resolution, the wave nature of the electrons has to be taken into account. This can be done by calculating the Phase Contrast Transfer Function (PCTF) [4] to examine the contrast mechanism. For DPC, the PCTF should be proportional to the spatial frequency k=2π/λ whereas a PCTF constant as a function of the spatial frequency k would indicate conventional phase contrast.

Assuming an ideal lens, which is a good approximation for an aberration corrected STEM, the PCTF for a weak phase object can be calculated using elementary geometry. A cut through the two dimensional PCTF, evaluated for the parameters of a local electric field measurement, is shown in fig.1. It is striking that the area in which the PCTF is proportional to k is rather small (up to ca. 0.2 1/Å as seen in fig. 2), indicating that, for high spatial frequencies, DPC would not occur. While this is unproblematic at low resolutions (where the configuration described above leads to an improved signal to noise ratio [5]), it suggests that under these conditions the classical model is not valid for high spatial frequencies and the detector setup is therefore not suited for high resolution DPC applications.

The calculated PCTF shows that, for the given parameters, DPC is limited to spatial frequencies of about 0.2 1/Å. We are currently looking for possibilities to increase the resolution by optimizing the detector geometry.

[1] J. N. Chapman et al., Ultramicroscopy 3 (1978) 203
[2] M. Lohr et al., Ultramicroscopy 117 (2012) 7
[3] N. Shibata et al., Nature Physics 8 (2012) 611
[4] H. Rose, Ultramicroscopy 2 (1977) 251
[5] J. N. Chapman et al., IEEE trans. on magn. 26 (1990) 1506


Fig. 1: PCTF L(k) for an ideal microscope with an acceleration voltage of 300 kV, an aperture angle of 21.6 mrad, an inner detector angle of 21.0 mrad and an outer detector angle of 40.7 mrad, corresponding to the configuration in high resolution local electric fields measurements.

Fig. 2: Low spatial frequency region of the PCTF in fig.1, showing that DPC only occurs for spatial frequencies k smaller than ca. 0.2 1/Å.

Type of presentation: Poster

IT-2-P-1633 Lifetime of the aberration-corrected optical state in HRTEM

Barthel J.1, Thust A.2
1Central Facility for Electron Microscopy, RWTH Aachen, Germany, 2Peter Grünberg Institute, Forschungszentrum Jülich GmbH, Germany
ju.barthel@fz-juelich.de

The technique of high-resolution transmission electron microscopy (HRTEM) experienced an unprecedented progress through the introduction of hardware aberration correctors, and by the improvement of the achievable resolution to the sub-Ångström level. As a consequence, the required precision level to measure and to adjust the optical properties of transmission electron microscopes has become increasingly demanding. A second consequence of this development, which has received little attention so far, is that aberration correction at a given resolution requires additionally a well-defined amount of optical stability. We investigate the qualification of a variety of high-resolution electron microscopes to maintain an aberration-corrected optical state in terms of a lifetime.

A comprehensive statistical framework is introduced for the estimation of the optical lifetime [1]. The temporal evolution of the twofold astigmatism is extracted from a series of images recorded over several minutes (Fig. 1). The twofold astigmatism serves as representative indicator for the optical stability since it is one of the most volatile image aberrations, has a strong influence on the image contrast, and can be measured rapidly in a simple experiment [2]. A model-based evaluation method was developed, which allows us to distinguish between two major components of astigmatism fluctuations, a random walk and a constant drift. A very useful output of the model-based evaluation is a probability curve (Fig. 2), which informs the operator about the chance to still work in an aberration-corrected state after a given timespan.

Optical stability evaluations for different high-resolution microscopes reveal surprisingly short lifetimes on the order of a few seconds up to a few minutes. The observed short lifetimes denote a critical limitation of the timespans between aberration measurement, aberration correction and the actual imaging. Therefore further investigations and technical developments are necessary in order to stabilize electron microscopes with respect to their sub-Ångström qualification. Since the topic of optical stability turns out to be of similar importance as the topic of resolution itself, we recommend to include a routine assessment of the optical stability in acceptance tests for high-resolution microscopes operating in the discussed resolution regime. For this purpose, the lifetime evaluation procedures developed in this work have been implemented in a user friendly and freely downloadable software [3].

References:

[1] J. Barthel, A. Thust, Ultramicroscopy 134 (2013), p. 6.
[2] J. Barthel, A. Thust, Ultramicroscopy 111 (2010), p. 27.
[3] J. Barthel, http://www.er-c.org/barthel/pantarhei/, (Feb 2014).


J.B. gratefully acknowledges funding within the core facilities initiative of the German Science Foundation (DFG) under the grant number MA 1280/40-1.

Fig. 1: Evolution of the twofold astigmatism extracted from images of amorphous carbon. Already after one minute the astigmatism fluctuations violate the π/4 limit for 300 keV electrons and for a microscope resolution of 0.8 Å.

Fig. 2: Decay of the probability to still work in an aberration-corrected state, evaluated from the astigmatism fluctuations shown in Fig. 1.