Symposium CM1 : New Frontiers in Aberration Corrected Transmission Electron Microscopy

Within the past fifteen years, electron microscopy has been revolutionized by the advent of the aberration corrector. Aberration correction dramatically improves spatial resolution into the sub-ångstrom regime, unlocking previously inaccessible information about material structure. While these recent advances have proven essential to the atomic scale characterization of materials, real-space measurements have remained largely qualitative. In particular, accuracy and precision for scanning transmission electron microscopy (STEM) was significantly hampered by the presence of sample drift and scan distortion. Until recently, this limitation has obscured the capabilities to characterize minute changes to the atomic structure that can ultimately define material properties.

In this talk, I will introduce revolving scanning transmission electron microscopy (RevSTEM). The method uses a series of fast-acquisition STEM images, but with the scan coordinates rotated between successive frames. I will provide a theoretical basis for the approach and demonstrate that the technique is capable of achieving sub-picometer accuracy and enables real-space crystallographic measurements in STEM. Multiple case studies will be presented to demonstrate the power of this new technique to characterize materials. For example, I will show how picometer precise measurements enable the capability to directly observe static atomic displacements within a complex oxide solid solution, highlighting differences in local bonding. Furthermore, using ferroelectric HfO2 thin films, I will also show how RevSTEM images can be used to accurately determine crystallographic parameters in real-space, and to determine the structural origins for spontaneous polarization in these materials. These results open a new world of atomic scale exploration that was previously just beyond our reach.

Single atoms and small atomic clusters offer a range of novel, tunable properties for a number of applications such as selective catalysts [1]. Achieving precise control of the desired properties of these systems first requires an understanding of the interaction between the cluster and its support. Advances in aberration-corrected scanning transmission electron microscopy (STEM) mean that atomic resolution imaging and characterisation is now achievable for many materials. Observing individual atoms and small clusters remains difficult, however, due to low signal-to-noise ratio and beam-induced motions causing blurring during image acquisition. One route around these problems is to acquire rapid image sequences in an effort to reduce the electron dose and also to capture any dynamic behaviour of the atoms. Making use of the spatial and temporal correlations between frames, and using a novel processing method based on singular value thresholding [2], we have developed robust approaches to recover individual atomic positions, including STEM acquisition rates of 10 frames per second or better [3].

We have applied the approach to the study of catalytically important copper, as well as other atoms, on few-layer graphene oxide (GO), where the presence of functional groups on GO may aid the control of deposited clusters by acting as preferential pinning sites. Analysis of an annular dark-field STEM image sequence reveals a range of behaviours, with some strongly-pinned atoms and other more mobile atoms undertaking random walks on the surface. Further investigation of the atom trajectories extracted from the sequence uncovers preferred jump lengths and directions in agreement with discrete sites on a graphene-like lattice. Combined with ab-initio DFT and kinetic Monte Carlo calculations, this analysis provides a new insight into the formation and behaviour of small atom clusters under an electron beam, and the interactions between few-atom catalysts and high surface area supports.

References
[1] Tyo, EC, Vajda, S. (2015). Nat. Nanotechnol. 10, 577-588.
[2] Candes EJ, Sing-Long CA, Trzasko JD. (2013). IEEE Trans. Signal Process. 61, 4643-4657.
[3] Furnival T, Leary R, Midgley PA. (2015). Manuscript in preparation.

Electron beam damage in the S/TEM is a pervasive issue that functions as a physical limit to material studies. Here we discuss damage mechanisms in CeO₂ and several other oxides. Measurements made by electron energy loss spectroscopy and HAADF STEM reveal a unique behavior; the presence of detectable damage depends on exceeding a dose rate threshold (e/nm²s) rather than a critical dose (e/nm²). The active damage mechanism in these materials is the sputtering of oxygen which creates vacancies. In this case, the microscope environment acts as an oxygen reservoir supporting the oxidation and recovery of these materials. Therefore only when the damage rate - a function of the electron dose rate - is sufficiently high will damage begin to accumulate. Otherwise the recovery rate - a function of intrinsic material properties and microscope environment - is sufficient to heal the damage and a detectable amount of damage will not accumulate even when the material is exposed to large cumulative doses. This phenomenon was initially identified in CeO₂ but other materials have been studied and a summary of those results will be presented. Additionally, methods to prevent the onset of damage under high dose rates will be discussed.

Nanodiamonds are of interest in scientific fields ranging from drug delivery and optoelectronics to astrophysics and planetary materials. Common to all of these research areas is a need to understand the atomic-scale properties of the nanodiamonds, including surface termination, and impurity and defect contents, and to relate these to the nanoparticle formation and processing conditions. Low voltage aberration-corrected scanning transmission electron microscopy (AC-STEM) is uniquely-suited to the characterization of individual nanodiamonds with single-atom sensitivity imaging and spectroscopy. Whereas the phase contrast in aberration-corrected high resolution TEM is well-suited to imaging the diamond lattice planes and twin defects, AC-STEM provides superior sensitivity for imaging and spectroscopy of C monolayers and heteroatoms. We recently showed that AC-STEM with energy dispersive x-ray spectroscopy (EDXS) can be used to identify individual impurity atoms associated with nanodiamonds isolated from meteorites, including Si. Silicon impurities are of particular importance, because the SiV^- defect in diamond exhibits strong photoluminescence (PL) at 738 nm that could potentially be exploited for optoelectronic applications, and for detection of diamonds in astrophysical settings. One complication in studying this defect in nanodiamond powders or suspensions is that residual monolayers of amorphous carbon are difficult to impossible to completely remove from the nanodiamonds, and go undetected by most measurements other than dark-field AC-STEM imaging. If Si atoms at nanodiamond surfaces get overcoated in amorphous carbon, they may contribute to the PL signal. Mobility of the Si impurities may also contribute to instabilities in the PL signal.

We are investigating both meteoritic and synthetic nanodiamonds with a Nion UltraSTEM 200, operated at 60 kV. Electron energy loss spectroscopy in both low-loss and core-loss regions provides confirmation of the local sp³:sp² bond distributions, and EDXS allows compositional analysis of individual nanodiamonds down to single impurity atoms. Our measurements of meteoritic nanodiamonds indicate that the Si impurities are distributed across both diamond and amorphous carbon regions. These atoms diffuse rapidly under the electron beam, which suggest surface, rather than bulk, incorporation. Further analysis of the distribution of other impurities, including N and O, and correlation with optical properties are planned.

Imaging biological nanostructures at high resolution using electron beams is a grand challenge due to a wide variety of hurdles. Biomolecules are non-crystalline and as such are not amenable to image processing techniques that exploit lattice periodicity. The soft biomaterials generally have a low atomic number, mainly consisting of carbon, nitrogen, phosphorus and oxygen, and this makes direct imaging difficult due to thickness effects from a supporting substrate. They are usually hydrated and can change their conformation when they are desiccated, and they are vulnerable to damage from electron radiation. To achieve high resolution, they must also be purified so they are free of organic and hydrocarbon contaminants that absorb on the sample surface may obscure their features.
Recent advances, however, are making these biomolecules much more amenable to direct imaging, with techniques ranging from cryo or liquid-cell electron microscopy to preserve hydrated structure, direct electron cameras for low-dose imaging, multiparticle tomographic image processing to calculate average structure, and ultra-thin substrates such as 2-D materials for low-background imaging.
In this work, we employ the 2-D material approach, preparing clean low-background graphene substrates for direct imaging of self-assembled DNA origami. Triangular DNA Origami, with 80 nm sides and an interior triangular hole (Rothemund Nature 440, 297-302 (16 March 2006)), were prepared using a standard self-assembly process. The samples were gel purified, and water was exchanged for the TAE Mg buffer. Samples were deposited on a variety of graphene membranes, including those formed using conventional PMMA-transfer as well as graphene membranes formed using newly developed polymer-free transfer techniques [Whitener et al, under review (2015)]. All samples were imaged using a Cs-corrected Nion UltraSTEM operated at 60 kV. We demonstrate that with gel-purified origami structures on single-sheet h-graphene, we obtain high signal-to-noise MAADF images of the DNA nanostructure. These have a higher signal-to-noise as compared to uranium-stained conventional DNA structures on ultrathin Si or holey-carbon, with a marked improvement in image resolution. The effects of graphene preparation, cleanliness, and tuning of the hydrophobicity of the graphene surface will also be presented. This sample preparation technique paves the way forward in direct imaging of biomolecules at sub-nm resolution in a STEM.

Aberration correction of electron optics has led to routine atomic resolution imaging based on both elastic and inelastic scattering. Z-contrast imaging and electron energy loss spectroscopy (EELS) based on core loss have become powerful tools for characterizing two-dimensional materials such as graphene [1,2]. It is often assumed that interactions involved in low-loss spectroscopy are too delocalized to form atomic resolution images. We have recently demonstrated atomic resolution of STEM images of graphene based on valance electron energy loss spectroscopy (VEELS) taken using a Nion UltraSTEM operating at an accelerating voltage of 60 kV [3]. We have however confirmed these images are a result of inelastic scattering due to VEELS excitations using a combination of density functional theory (DFT) and dynamical scattering theory. Dynamical scattering theory is essential to understand the underlying localization of the observed images, while DFT give us insight into transitions that contribute to the image features.
In this presentation we will extend this work to the study of defects and impurities in two-dimensional materials. We show that impurity atoms lead to spatially resolved intensity variations as a function of energy loss. Using the combined tools of DFT and dynamical scattering we can explain the source of both excess intensity and a reduction in intensity observed for different energy losses. We show that image intensities, due to new states introduced by the presence of the impurity atoms, can be used to differentiate between different atomic configurations.
With the development of improved monochromators for aberration corrected electron microscopes the detailed structure of low-loss spectra can now be observed more clearly than ever before. This work potentially opens a new field of atomic resolution microscopy, allowing the mechanisms underlying image formation to be clearly understood, and hence the properties of two-dimensional materials to be probed in more detail than previously possible.

[1] Krivanek et al., Nature, 462, 7228 (2010).
[2] Zhou et al., Microscopy and Microanalysis, 18, 1342 (2012).
[3] Kapetanakis et al., Phys. Rev. B, 92, 125147 (2015).
[4] This research was supported by DOE Grant No. DE-FG02-09ER46554, the Center for Nanophase Materials Sciences (CNMS), which is sponsored at ORNL by the Scientific User Facilities Division, Office of Basic Energy Sciences, U.S. DOE and by the Office of Basic Energy Sciences, Materials Sciences and Engineering Division, U.S.DOE. Numerical calculations were performed at the National Energy Research Scientific Computing Center (NERSC), which is supported by the Office of Science of the U.S. DOE under Contract No. DE-AC05-00OR22725

Aberration-corrected scanning transmission electron microscopy (STEM) has become a powerful technique for advanced materials characterization of complex nanostructures. Recent progress in the development of quantitative methods now allows us to extract structural, chemical, electronic, and magnetic information from experimental images in 2D as well as in 3D. Most studies using quantitative STEM concern nanostructures which are relatively stable under the incoming electron beam and therefore the atomic structure under investigation can be assumed to remain unchanged under illumination with high electron doses. However, radiation damage becomes increasingly relevant not only in biological studies but also in the study of nanostructures. In a post aberration-correction era, an important challenge is therefore to push the development of quantitative methods toward its fundamental limits. In order to reach this goal, the allowable electron dose needs to be used in the most optimal way. For that purpose, new strategies to optimize the microscope and detector settings and to analyze the experimental images will be demonstrated.

Existing quantification methods treat STEM images as numerical datasets from which unknown structure parameters are estimated by comparison with image simulations or by using parameter estimation-based methods. In order to retrieve the underlying 3D atomic structure of unknown nanoparticles, the use of so-called scattering cross-sections to count the number of atoms along the viewing direction has become a successful technique. The cross-section approach quantifies the total scattered intensity for each atomic column and has been shown to be robust to a wide range of imaging parameters including defocus, source size effects, and aberrations such as astigmatism. Recently, a new electron channelling-based prediction model has been developed that allows one to accurately predict cross-sections of mixed columns opening up possibilities to unscramble hetero-nanostructures in 3D from a limited number of projection images. The high sensitivity of scattering cross-sections in combination with a thorough statistical analysis in principle enables us to count atoms with single-atom sensitivity. However, the need for sufficient statistics in order to obtain reliable atom counts limits its usefulness to study beam-sensitive nanoparticles. To circumvent this problem, a hybrid method combining the benefits of the statistics-based method with the use of image simulations has been developed and applied to study challenging materials. Finally, the effect of electron dose, scan noise, and the choice of detector settings on the atom-counting performance will be demonstrated. It will be shown how to balance atom-counting reliability and structural damage as a function of electron dose. Such a computation allows one to predict the minimum electron budget needed in order to attain maximum precision.

Todd H. Brintlinger¹, Paul A. DeSario², Jeremy J. Pietron², Rhonda M. Stroud¹, and Debra R. Rolison<sup>2

1.</sup> Materials Science and Technology Division, U.S. Naval Research Laboratory, Washington, DC, 20375 USA
^2. Chemistry Division, U.S. Naval Research Laboratory, Washington, DC, USA 20375

With high surface area and efficient photochemistry, titania aerogels are potential photocatalysts for applications such as solar energy conversion¹ and environmental remediation². However, their large bandgaps do not efficiently match the main solar spectrum, and development of titania-based systems active in the visible has been an important recent thrust within materials research. One approach is to incorporate metal nanoparticles with local surface plasmon resonances that interact with incoming photons such that lower energies are absorbed. The effects of the metal nanoparticle size, shape, and distribution on the photocatatyic activity have been well documented for a variety of semiconductors¹. The behavior of the metal nanoparticles once incorporated onto supports is critical, since even small, initially well-dispersed metal nanoparticles can oxidize, aggregate or disperse over time or upon extended illumination and exposure to ambient conditions.

We previously demonstrated that composite precious-metal-loaded aerogels³, in which metal is added as the titania begins to gel, show improved photocatalytic activity over aerogels in which the metals are added after aerogel formation. Here, we extend our investigations of these systems using non-precious metals incorporated within titania aerogels. Following similar synthesis and photocatalytic characterization from our previous work, we find well-dispersed non-precious-metal nanoparticles within a titania matrix. However, the comparatively lower atomic number of non-precious-metals (Z
References:
[1] S. Linic, P. Christopher and D.B. Ingram, Nat. Mater 10, p. 911 (2011).
[2] T. L. Thompson and J. T. Yates, Jr., Chem. Rev. 106, p. 4428 (2006).
[3] J. J. Pietron, R. M. Stroud and D.R. Rolison, Nano Lett. 2, p. 545 (2002).

Silicon drift detectors (SDDs) have opened up a new era in energy dispersive x-ray (EDX) microanalysis at the nanometre and even sub-nanometre scales. The improved detector design allows for larger devices and therefore increased solid angles for x-ray collection. This provides a significant improvement in x-ray count rates such that atomic resolution maps and x-ray tomography are both now possible. In addition, aberration correction in the scanning transmission electron microscope (STEM) allows the use of larger probe-forming apertures to produce increased beam currents in much smaller probes. The combined result provides huge improvements in x-ray counts from small volumes leading to the potential for improved quantitative analysis at the nanometre scale and below.

In the same way that the scattering cross section, σ, can be calculated from ADF image intensity and for ionisation edges in EELS, it is possible to calculate an EDX partial cross section using an approach that demonstrates similarities with the ζ-factor method. Our quantification method is applied to PtCo alloy nanoparticles that have been acid-leached to provide platinum enrichment (or rather cobalt depletion) at the particle surface. The leaching process produces very little change in cobalt composition when investigating the whole particle via coarser analysis techniques such as traditional EDX.

With our newly-developed cross section approach, we show the intended cobalt depletion at the surface is in fact absent at the vertices of the more facetted nanoparticles. Additionally, we can now quantify the levels of cobalt depletion in the first few atomic layers of the particle. We show the leaching produced a localised surface depletion that can only be determined by this high resolution EDX quantification.

Another major advantage of this new technique is the potential for combining and comparing this cross sections with those produced in ADF STEM quantification and ionization edges in EELS, leading towards even further understanding of materials at atomic resolution.

The drive towards smaller and more efficient electronic devices has resulted in an increased need to understand the physical behaviour of nanostructures such as semiconductor nanowire (NW) crystals. The shape and size of the NW crystals offer new prospects, which do not only rely on electron confinement effects but also offer the possibility to manipulate the intrinsic properties by elastic straining of the NWs up to the limit of plastic deformation. In general, material structures with nanoscale dimensions hold big promise with respect to strain engineering because it is observed that smaller structures are more flexible. Here we address the quantum mechanical behaviour in the mechanical elastic limits of III-V NWs using in situ aberration corrected transmission electron microscopy (TEM).

Controlled and systematic experiments combining the different imaging and spectroscopic methods of TEM with the unique methods of in situ TEM have been used to correlate local atomic structure and electrical transport properties of individual GaAs and InAs NWs with different size and morphology. The influences of strain and electron beam excitation have been studied. Such experiments need measurements with highest spatial resolution and yield information that cannot be obtained by any other methods. In situ TEM studies using combined TEM- scanning tunnelling microscope holders allow us to apply high strain and simultaneous measure the change in resistance while imaging the NWs in the TEM.

The bending experiments showed that the NWs are highly flexible. The IV-curves were measured and the results were in accordance with ex situ measurements for the NWs. In situ TEM study show that the resistance of the GaAs NW increases gradually when the NW is bent while the I-V curves remain linear at different strain levels. As the strain is released, the NWs recover their original shape and the I-V curves coincide with the ones obtained without any force applied on the NW in the initial state. The mechanism of this phenomenon can be related to the modification of the band structure of the GaAs NWs due to a change in lattice structure resulting from the mechanical strain. Electron energy loss spectroscopy was used to study the effect of strain on the electronic structure with emphasis on the low energy loss interval of 0 to 50 eV. Electron beam induced current measurements were also performed to study the effect of strain on the diffusion length of the charge carriers.

Aerographite is an ultra light porous material of carbon showing excellent elasticity to large deformations with several tens strain [1]. Its unique characteristics are derived from the three-dimensional interconnected structure of hollow carbon fibers. Conventional aerographite has tetrapod-type network, since the morphology is transferred from tetrapod-type ZnO used as the template in the fabrication process. Recently, urchin-like ZnO particles with micrometer order diameters, so-called ZnO nanorod-microsphere has been fabricated [2]. These particles consist of radially-arranged ZnO nanorods with 50nm diameter. Using this instead of tetrapod-type one, we have developed a new type of aerographite, spiky-shell microparticles. The spherical shell consisting of radially arranged hollow nanorods with 100nm diameters and 5-10nm thicknesses. The spiky-shell morphology can ensure the mechanical strengths of such thin shell particles. In this study, mechanical properties of the particles were evaluated by single-particle-level compressive tests by in-situ electron microscopy.
In a transmission electron microscope, a Si substrate supporting aerographite particles by Van der Waals force and a cantilevered probe for scanning probe microscopy were individually fixed to two stages of nano-manipulator system, one of which is movable driven by piezoelectricity. By operating the manipulator, a single particle on Si substrate contacted to the tip of the cantilevered probe. Structural changes in the particle were observed when stressed in the compressive direction. Relationship between stress and strain was also examined by similar experiments carried out in a scanning electron microscope. An aerographite particle was placed between two parallel cantilevers, and compressed by operating one cantilever with monitoring deflections of cantilevers and changes in shape of the particle simultaneously. Compressive force was estimated from the deflection of cantilevers.
As the result, an aerographite particle fabricated in this study showed excellent elastic behavior under large strain (73% in maximum). Stress-strain curves and in-situ observation suggested the two-step deformation process in the elastic compression; local deformation at contact region and compression of the whole particle. Since individual nanorods did not deform, elasticity of the particles may be derived from the flexible connections between nanorods at their bottom portions. Probability inducing crack increased above 30% strain, but the aerographite particles inducing cracks almost recovered its spherical shapes after unloading. Accumulation of residual strain during the repetitive compression was also evaluated. It indicated that improvement of the crystallinity of graphitic layers by annealing treatment well contributed to improve the performance as the ultraflexible microparticles.

[1] M. Mecklenburg et al. Adv. Mater. 24, (2012) 3486-3490.
[2] Y. Hirota et al. Chem. Lett. 43, (2014) pp. 360-362.

Nanodomain structures in ferroelectrics are attractive due to their applications in ultra-small electric, optical, actuator and memory devices [1-5]. Recently, atomic-resolution in-situ Transmission Electron Microscopy (TEM) has revealed numerous novel ferroelectric nanodomain structures such as the flux-closure array, the self-similar nested bundles and strongly charged domain walls, all of which exhibit extraordinary properties [1,6,7]. In this work we utilized an in-situ mechanical system in TEM to study the nanodomain structures in the free-standing BaTiO₃ nanostructures under both stress and electrical bias loading. Using a high speed direct electron detector on the aberration-corrected TEM, we were able to perform scanning nanobeam diffraction experiments during loading that revealed the domain evolution leading to superelastic deformation in the nanostructures. Here we will present our in situ observations coupled with the nanobeam strain mapping and high resolution investigation of the nanodomains formed during superelastic deformation BaTiO₃ nanostructures.

ACKNOWLEDGEMENTS
This work has been supported by the Natural Science Foundation of Jiangsu Province,China (Grant No.BK20151382) , and the Molecular Foundry, which is supported by the Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.

REFERENCES
1. G. Catalan, J. Seidel, R. Ramesh, J. F. Scott, Rev. Modern Phys. 84, 119 (2012)
2. O. Auciello, J. F. Scott and R. Ramesh, Physics Today, 51, 22 (1998).
3. M. Dawber, K. M. Rabe, J. F. Scott, Rev. Mod. Phys., 77, 1083 (2005).
4. J. F. Scott, Science 315, 954 (2007).
5. H. Lu, C. W. Bark, C. B. Eom, G. Catalan, A. Gruverman, et.al. Science, 336, 59 (2012).
6. Y. L. Tang,Y. L. Zhu,X. L. Ma,S. J. Pennycook, et. al. Science 348, 547 (2015 ).
7. C. T. Nelson, P. Gao, X. Q. Pan, et. al. Science 334, 968 (2011).

The real-time observation of deformation of a nano junction under mechanical stress provides us fresh experimental knowledge toward understanding the mechanisms and tribology at the nanoscale [1]. A monolithic silicon MEMS (microelectromechanical system) was placed and operated in a TEM; this allowed us to measure the deformation and the size of 1-100 nm during tensile and shear testing. Also normal and shear forces were determined from the difference in the displacement of the supporting beam with and without the nano junction.
We prepared tips of bare Si and those coated with Ag, and Au films. Those tips were suspended by flexible beams and driven by electrostatic actuators. The longitudinal actuator brought the tips into contact and formed a junction. Tensile testing was performed by reducing the actuation voltage. An Au junction did not elongate much but a Si junction showed super plastic deformation as long as 2000 %. Molecular dynamic simulation based on a single-crystal/amorphous two-phase model could reproduce the behavior of the Si junction [2].
In the shear deformation testing of Si and Ag junctions, the lateral actuator applied shear force to the junction, whose size was typically 2-5 nm in diameter and 3-10 nm in length. The silicon junction elongated smoothly for 15 nm before breakage. The maximum shear force was 78 nN. Simulation based on molecular dynamics reproduced the behavior accurately [3]. On the other hand, a silver junction elongated for only 5 nm and was broken at 5 nN. Elongation was step-wise like a stick-slip motion. The size of steps (0.3 nm and 0.6 nm) corresponds well to the theoretical sliding distance of 0.29 nm calculated from the lattice spacing along the sliding plane. The mechanical work necessary to break the junction agreed well with the surface energy of newly created surfaces after breakage [4]. Whole shear testing processes of another Ag junction was in-situ observed by TEM from formation, and deformation to breakage with the measurement of normal and shear forces. The adhesion force was the major component in normal force; this results in an apparent negative friction coefficient. Although the contact area changed during the deformation process, the normal force and the shear force did not change much. Therefore, the usual assumption that the normal and shear forces are proportional to the real area of contact does not hold in the single nano junction.

References:
[1] T. Sato, L. Jalabert, H. Fujita, Microelectronic Engineering, vol.112, pp. 269–272, 2013
[2] T. Ishida, F. Cleri, K. Kakushima, M. Mit, T. Sato, M. Miyata, N. Itamura, J. Endo, H. Toshiyoshi, N. Sasaki, D. Collard and H. Fujita, Nanotechnology, 22 355704 (2011).
[3] Tadashi Ishida, Takaaki Sato, Masatsugu Oguma, Takahumi Ishikawa, Noriaki Itamura, Keisuke Goda, Naruo Sasaki and Hiroyuki Fujita, Nano Letters, vol.15, pp. 1476-1480 (2014)
[4] Takaaki Sato, Tadashi Ishida, Laurent Jalabert, Hiroyuki Fujita, Nanotechnology, Vol.23, p.505701, (2012).

Freestanding, amorphous TiAl (45 at.% Ti) thin films (150 nm thick) were subjected to in-situ TEM tensile straining using MEMS based testing stages. Micro strain along the longitudinal and transverse direction was calculated at different stages of loading by measuring geometric changes in the first ring of the selected area diffraction pattern. Simultaneously, the macroscopic stress-strain (σ-ε) response of the film was measured using in-built force and displacement gauges in the MEMS stage.
The micro strain along the longitudinal direction (e₁₁) followed a trend similar to that of the macroscopic strain (ε), but was consistently smaller. Thus, the Young’s modulus calculated using e₁₁ was significantly larger compared to the modulus obtained from the macroscopic σ-ε measurements. To investigate whether this deviation was a result of anelastic strain, which is not captured by the micro strain measurements, we carried out ex-situ deformation experiments at different strain rates. We found that the Young’s modulus was 7% higher when the film was loaded at a strain rate of 10^-2 sec^-1 compared 10^-6 sec^-1, revealing notable anelasticity in these metallic glass films.
In addition, we measured the full width at half maximum (FWHM) of peaks along all directions of the first diffraction ring during straining. The measurements showed that the averaged FWHM, which provides a measure of the variation in the nearest neighbor distances, decreased with increasing stress and did not recover its initial value upon unloading. This suggests that permanent atomic scale structural rearrangements are induced in the metallic glass films due to straining.
This in situ TEM technique gives us the unique capability to measure micro strains from extremely small regions (<1 μm in diameter) of thin film specimens, which is not possible using x-ray or neutron diffraction techniques which are typically used to probe bulk metallic glass specimens.

An understanding of how nanoscale defects nucleate, move within grains, and propagate through multiple grains at high strain rates (up to 10^4/s) in metals is needed in order to explain and predict dynamic behavior and spall strength. To enable in situ observation of dislocations and twins in a TEM at high strain rates, several challenges have been addressed. The time resolution of conventional TEM is usually limited to around 30 frames per second. To achieve the needed time resolution for high strain rates, we use the Dynamic TEM at the Lawrence Livermore National Laboratory which is able to record pictures every 70ns in movie mode. TEM stages for in situ mechanical testing are generally limited to quasi-static strain rates. We have designed a TEM holder capable of deforming samples at strain rates ranging from quasi-static to 10^4/s that utilizes two piezoelectric actuators working in bending to load samples. This system is calibrated and instrumented with strain gages to provide a time-resolved record of the net force and strain applied to the sample. Because of the piezo system’s limitations in terms of force, very small samples are required to achieve the desired strain rate. We have developed a new sample preparation procedure that combines mechanical polishing, femtosecond laser machining, and precision ion milling to form, from bulk samples, 300-µm-wide rectangular specimens with a 25-µm-wide gauge region where the electron transparent area is obtained by ion milling. We will present the latest results of high-strain-rate in situ mechanical tests conducted on copper and magnesium alloys specimens.

This talk will highlight recent advances with in situ Transmission Electron Microscopy (TEM) nanomechanical testing techniques that provide insight into small-scale plasticity and the evolution of defect structures in lightweight alloys and oxide nanostructures. In addition to measuring the strength of small-volumes, measuring the evolution of strain during plastic deformation is of great importance for correlating the defect structure with material properties. Here we demonstrate that strain mapping can be carried out during in-situ deformation in a TEM with the precision of a few nanometers without stopping the experiment. Our method of local strain mapping consists of recording large multidimensional data sets of nanodiffraction patterns during the test. This dataset can then be reconstructed to form a time-dependent local strain-map with sufficient resolution to measure the transient strains occurring around individual moving dislocations.

The development of nano-devices has aroused intensive investigation in the interfacial interaction (adhesion and friction) of two-body-contact at nano-scale. Here, via the joint results of in-situ high resolution transmission electron microscopy (HRTEM), we observe the nucleation and subsequent annihilation of “open stacking fault tetrahedron” (open-SFT), with only three of its four planes covered by staking faults, by deforming the single nano-contact between gold crystals. The direct visualization of dislocation behavior offers novel insights in the nano-scale tribological study.

Previous studies have shown that metal films with similar thickness and grain size but dissimilar texture show significant differences in their mechanical behavior. For instance, ultrafine-grained (UFG) Al films with no preferred texture show lower flow stress and more pronounced nonlinear behavior during unloading compared to films with a bicrystalline microstructure.
To understand the mechanisms of such texture-induced differences in mechanical behavior we performed quasi-static in situ TEM straining of non-textured and bicrystalline UFG Al films with automated crystal orientation mapping (ACOM). The ACOM results show that significant grain rotations, up to ~6, occur in the non-textured films even at small strains (~0.5%) during loading, whereas the grains in the bicrystalline film exhibited significantly smaller rotations. Furthermore, reverse rotation of the grains occurred upon unloading in the non-textured film, which provides a possible reason for the observed nonlinearity in the stress-strain response. Bright field TEM imaging showed that grain contrast changes (indicative of grain rotations) were time dependent, which suggests that diffusive processes could be active in addition to dislocation slip.
The results show that the combination of ACOM and in situ TEM straining can provide a more detailed picture of the complex deformation processes occurring in UFG and nanocrystalline metals.

In situ transmission electron microscopy (TEM) can provide valuable insights into the deformation behavior of nanostructured materials. However it is critical to understand and quantify the effects of the electron beam (e-beam) exposure on the deformation response of such materials. In this study, we investigated the effects of the e-beam on the stress-strain response of nanocrystalline and ultrafine-grained aluminum and gold thin films during in-situ tensile straining. The e-beam accelerating voltage, area and intensity were systematically varied to study the nature and extent of beam-induced artifacts at different beam conditions.

We found that e-beam exposure caused increased dislocation activation and marked stress relaxation in the Al and Au films that spanned a range of thicknesses (80-400 nm) and grain sizes (50-220 nm). The e-beam also caused an unusual necking along the width of the sample, with the extent of necking increasing with the area of the specimen exposed to the beam. Notably, these effects were observed at accelerating voltages well below the radiation damage threshold of these materials. And contrary to expectation, the beam-induced artifacts were more pronounced at lower accelerating voltages.

The experiments, performed on two metals with highly dissimilar atomic weights and properties, suggest that the e-beam can cause significant changes in the deformation behavior of a range of nanostructured materials during in situ TEM straining.

In situ indentation in a high-resolution transmission electron microscope (HRTEM) is used to observe dislocation nucleation and glide in single crystalline TiN thin films. The structural images acquired under load are used to measure lattice strains and the corresponding local stresses are inferred from first-principles computed non-linear elastic stress-strain response. This experimental approach is shown to estimate local resolved shear stresses corresponding to partial or full dislocation nucleation and motion of full dislocations in high strength materials, and validate the first-principles calculated Peierls stresses.

Heterogeneous catalysts often undergo dramatic changes in their structure as the mediate a chemical reaction. Multiple experimental approaches have been developed to understand these changes, but each has its particular limitations. Electron microscopy can provide analytical characterization with exquisite spatial resolution, but generally requires that the sample be imaged both ex situ and ex post facto. Photon probes have superior depth penetration and thus can be used to characterize samples in operando (i.e when they are actively working). But they generally lack spatial resolution and thus give only ensemble average information.

We have taken advantage of the recent developments in closed-cell microscopy methods to develop an approach that allows us to successfully combine electron, x-ray and optical probes to characterize supported nanoparticle catalysts in operando. By measuring the reaction products at each stage of the reaction, we can directly correlate the information that can be obtained from each approach, and thus gain a deep insight into the structural dynamics of the system.

In this work, we will show how a combination of x-ray absorption near edge (XANES) and scanning transmission electron microscopy (STEM) can be used to characterize the changes that occur in a model NiPt bimetallic catalyst during oxidation and reduction. Bimetallics are of broad interest in heterogeneous catalysis as the provide the opportunity to selectively tune reactivity and selectivity. However, the characterization of their structure by averaged probes such as x-ray absorption spectroscopy is comprised by the heterogeneity that such systems may proscribe.

The presentation will focus on the development and application of experimental methods used to describe the morphological changes that occur in this model bimetallic system. These will include high temperature atmospheric pressure electron microscopy, the direct measurement of reaction products using gas chromatography–mass spectrometry and the ability of a newly developed electron microscope for operando microscopy (based on the FEI Talos platform) to characterize bimetallic nanoparticles through energy dispersive x-ray spectroscopy.

Isolated single metal atoms dispersed on high-surface-area supports have recently demonstrated remarkable activity and selectivity for a plethora of catalytic reactions [1-4]. The interaction of the individual metal atoms with the support surface modifies the surface electronic structure of the metal-support ensembles and thus tunes the binding strength of the reactant molecules. Such an approach to engineering the surface electronic structure of high-surface-area support materials can be effectively utilized for developing new and better catalysts with broad applications in chemical transformations, energy and environment. Aberration-corrected scanning transmission electron microscopy (AC-STEM) techniques have proved to be critical in developing single atom catalysts (SACs) [1-4]. With subangstrom electron probe sizes, enhanced probe current, and efficient annular dark-field detectors the AC-STEM becomes a powerful tool for routinely determining the spatial dispersion of metal single atoms, their relative locations with respect to the positions of the surface atoms of crystalline supports, and, to a certain degree, their relative strength of binding to the surfaces of the support materials. It is hypothesized that metal single atoms located at the cation vacancy sites of the support material should be relatively stable even under electron beam irradiation while small metal clusters may be extremely unstable, a consequence of weak anchoring. The challenges in fully understanding the nature of supported metal single atoms include the electron transfer processes and the vertical location of the individual metal atoms with respect to the surface atoms of the support materials. Imaging of single atoms of metal (e.g., Au, Pt, Ir, Pd, etc.) on various types of supports (e.g., Fe₂O₃/Fe₃O₄, ZnO, NiO, Co₃O₄, graphene, etc.) and the catalytic performances of the corresponding SACs will be discussed [5].

References
[1] B. Qiao et al., Nat. Chem., 2011, 3, pp 634–641.
[2] J. Lin et al., J. Am. Chem. Soc., 2013, 135, pp 15314–15317.
[3] H. Wei et al., Nat. Commun., 2014, 5, Article # 5634.
[4] B. Qiao et al., ACS Catal., 2015, 5, pp 6249–6254.
[5] This work was supported by the start-up fund of the College of Liberal Arts and Sciences of Arizona State University and the National Science Foundation under CHE-1465057. The author acknowledges the use of facilities in the John M. Cowley Center for High Resolution Electron Microscopy at Arizona State University.

Linking catalyst structure with activity is a primary goal of much catalysis research. Observation of the catalyst structure at the atomic scale using environmental TEM (ETEM) while catalysis is taking place is a powerful technique for linking activity with structure. To do this well, it is essential to know the activity of the catalyst while it is being observed. The simultaneous observation of the catalytic activity and atomic structure of a catalyst during catalysis in the microscope is known as operando TEM.
Operando TEM can be accomplished by monitoring the gas composition using the complimentary techniques of mass spectrometry and electron energy loss spectroscopy. For CO oxidation over Ru, this capability has been previously demonstrated, but details of the atomic structure of the nanoparticle surfaces could not be obtained without the use of aberration corrected microscopes. Now, with an image-corrected Titan ETEM, it is possible to directly observe the surface structures present on Ru catalyst nanoparticles.
Surface layers were observed on some particles exposed to 2 Torr of a stoichiometric ratio of CO and O₂ at 200°C. Similar surface layers were observed in mixtures of H₂ and O₂ and have been attributed to oxidation of the Ru nanoparticle surfaces. While the surface layers are clearly visible in the image corrected TEM, the exact structure of the surface layers could not be directly interpreted without image simulation [1].
Though Ru for CO oxidation is a well-studied system, there is still debate in the literature regarding the most active state of the catalyst surface, though the general consensus is that some oxidation of the surface occurs. Several oxidized Ru surface structures have been proposed for the (0001) surface of Ru. In the present work, image simulations have been performed using models of these proposed structures. We will present results comparing the simulated images of these and other possible structures to operando images of Ru nanoparticle surfaces obtained during high catalytic activity, to determine the atomic structure of the active surface observed on Ru catalyst nanoparticles in the aberration corrected ETEM.
[1] N.P. Walker, B.K. Miller, & P.A. Crozier, Materials Research Society. (these proceedings)

Iron oxide catalysts were widely used for industrial water-gas-shift reaction, ethylbenzene dehydrogenation, etc. Their catalytic performances strongly depend on their shape and surface structure which may dynamically evolve during the catalyst activation or the catalytic reaction processes. Aberration-corrected environmental transmission electron microscopy (AC-ETEM) and related atomic scale imaging and analytical technqiues enable the investigation of dynamic structural changes of nanostructured catalysts with atomic scale resolution under close to working conditions. In order to understand the evolution of the surface atomic structures of the Fe₂O₃ nanostructured catalysts during the reductive activation process we specifically synthesized α-Fe₂O₃ nanoplates with predominantly {001} facets. The AC-ETEM data clearly revealed the nucleation and growth processes of α-Fe₂O₃ to γ-Fe₂O₃ and finally to Fe₃O₄. During the H₂ reduction process (2 Torr. H₂ at 300°C) the shape of the original α-Fe₂O₃ nanoplates was maintained during the phase transformations. Furthermore, initial atomic resolution images unambigously confirmed that the phase transformations initiate at the corners and atomic steps. Although we have not been able to directly observe the hydrogen atoms these experimetnal data suggests that H₂ preferentially adsorbed and dissociated at the coner and step sites of the α-Fe₂O₃ nanoplates. Such understanding of the detailed atomic scale structures of nanostrcutred catalysts during the activation and catalytic reaction processes signifciantly enhances our understanding of the nature of working catalysts.

We have developed _s correctors for both TEM imaging and STEM probe is retained, and in some aspects enhanced. The ESTEM system is fully compatible with improved high vacuum STEM and TEM, especially for hot stage studies (5) and with window specimen holders for higher gas pressures (6) or liquids (7). This is a variant of the established and widely adopted design of ECELL system [3] with regular microscope aperture disks mounted inside lens polepieces to separate the column into a series of differential pumping zones at gas pressures in the range from mbar to Pa at the specimen to 10^-10 mbar in the gun.
Initial experiments with the system have explored the presence and dynamics of single atoms, clusters and nascent nanoparticles on the support between larger features. The reliable detection of single atoms of Pt on a 4nm model system and on more practical carbon supports extends the scope for effective studies to include single atom tracking. Work has begun to revise existing theories of nanoparticle stability under reaction conditions with the new information reliably accessed here for the first time, extending the single atom knowledge from STM studies [8] into a wider context.
References
1. E D Boyes, M R Ward, L Lari and P L Gai, Ann Phys (Berlin), 6 (2013) 423
2. P L Gai, L Lari, M R Ward and E D Boyes, Chem. Phys. Letts. 592 (2014) 335
3. E D Boyes and P L Gai, Ultramicrosc. 67 (1997) 219
4. P L Gai et al MRS Bulletin, 32 (2007) 1044
5. J Sagar et al, Appl. Phys. Letts. 105 (2014) 032401
6. J Creemer et al, Ultramicrosc. 108 (2008) 993
7. R Kroger, Nature Materials, 14 (2015) 369
8. P Jiang, X Bao and M Salmeron, Acc. Chem. Res. 48 (2015) 1524
The Engineering and Physical Sciences Research Council (UK) is thanked for supporting the program with the strategic research grant EP/J018058/1.

In situ environmental transmission electron microscopy (E-TEM) has recently advanced with technological developments such as aberration correction of lenses, fast detection cameras, and miniaturized specimen containers with various functions. These advances have enabled observation of a variety of phenomena in materials and devices at higher spatial and temporal resolution especially in gases. It is now possible to investigate the essential static and dynamic characteristics of materials and devices by quantitative in situ E-TEM at the atomic scale [1].

We briefly summarize our recent in situ E-TEM studies of the catalysts of Au/CeO₂, Au/TiO₂, Pt/CeO₂ and others using Cs-corrected Titan ETEM G2. It is well-known that gold, the most stable metallic element, shows remarkable catalytic activity for CO oxidation even below room temperature. Gold nanoparticulate catalysts, prepared using the deposition precipitation method exhibited high catalytic activity at room temperature. Systematic acquisition along with both numerical and statistical analyses of the E-TEM imaging data led to the intrinsic catalyst structure in the reaction environments. The quantitative analyses [1] further indicated that the activation sites of oxygen molecules at room temperature are most likely to be at the perimeter interface between gold nanoparticles and metal oxide supports. During the reaction, the perimeter interface is not structurally rigid. Glimpse of gas molecules that interact with the surface of a gold nanoparticle is now possible with Cs corrected E-TEM. We will present in situ Cs-corrected E-TEM analyses of other gold catalysts.

We have observed the oxidation and reduction processes of the surface of Pt nanoparticles by Cs-corrected E-TEM. Atomic layers of Pt oxides, including α-PtO₂ and Pt oxides of other forms, started forming on the preferential facets of Pt nanoparticles at the early stage, entire oxidization on the whole surface of Pt nanoparticles then followed. The oxides were reduced promptly to Pt by adding a small amount of CO or H₂O vapor to the dominant O₂ gas. Electron irradiation during E-TEM observation activates the gases non-thermally, therefore promoting or suppressing the processes at room temperature [2]. We will present the effect of moisture on other catalysts.

It is now realized that a crucial era of in situ E-TEM has started. For quantitative in situ TEM of time-dependent phenomena, for instance dynamic atomic motions in a chemical reaction, quantitative evaluation and removal of the electron-irradiation-induced phenomena that may appear in the background in in situ E-TEM data is required. We think that the robust E-TEM apparatus combined with quantitative methodologies is definitely a necessary condition for the serious applications of in-situ E-TEM in materials and devices.

References.

[1] S. Takeda, Y. Kuwauchi, H. Yoshida, Ultramicroscopy, 151 (2015) 178.

[2] H. Yoshida, H. Omote, S. Takeda, Nanoscale, 6 (2014) 13113.

Photocatalytic water splitting has the potential of producing sustainable clean energy by converting and storing solar energy into H₂ fuel. Photocatalytic materials generally consist of semiconductors with bandgaps larger than 1.8 eV and metals that can be catalytically active to perform water oxidation/reduction reactions. It is now recognized that atomic level in-situ observations of these catalysts are critical for understanding structure-reactivity relationships and deactivation processes such as photocorrosion. This requires that the system be observed not only in presence of reactant and product species but also during in-situ light illumination. Here concerns and applications associated with building a “photo-reactor” inside an aberration-corrected ETEM are discussed. An optical fiber based in situ illumination system has been developed previously for an FEI Tecnai F20 ETEM with light intensity up to 10 suns [1]. Using this system, we have observed detailed structural evolution on TiO₂ photocatalysts during exposure to in situ light and gas environments [2]. Recently we have installed an optical fiber into an FEI Titan environmental transmission electron microscope with an image corrector providing sub-Angstrom resolution. The optical fiber is guided by a fiber holder using the objective aperture port. We will discuss the design and performance on applications related to solar water splitting. NaTaO₃ which is a highly active photocatalyst with a large bandgap semiconductor will be tested as a model material to evaluate the system. It is believed that with the capability of in-situ illumination and superior resolution, deeper understanding of reaction and deactivation mechanisms of various photocatalyst systems will be gained.

[1] B.K. Miller, P.A. Crozier, Microsc. Microanal. 19 (2013) 461–469
[2] L. Zhang, B.K. Miller, P.A. Crozier, Nano Lett. 13 (2) (2013) 679–684
[3] The support from US Department of Energy (DE-SC0004954) and the use of Titan microscope at John M. Cowley Center for High Resolution Microscopy at Arizona State University is gratefully acknowledged.

To move forward with creating novel nano-electronic devices there is a need to understand behavior of electrons in a wide range of materials whith dimensions of the nanometer scale. This is a common task for the industry dealing with electronic or magnetic memories, light-emitting, photovoltaic or multiferroic devices. The important role for development and production of such devices is the characterization of the local magnetic fields. Such characterization is also of high importance for biomedical applications like hyperthermia treatment or local drug delivery, paleomagnetism, environmental magnetism or biomagnetism.
The best tool for the local magnetic fields characterization (in terms of resolution and sensitivity) is the Transmission Electron Microscopy (TEM). There are several methods within TEM which have been developed and successfully used for visualization and quantification of the nanoscale magnetic fields: Lorenz microscopy, Electron Holography and Differential Phase Contrast (DPC). All these methods have their intrinsic constraints, like limited field of view or low resolution, and all of them depend on the presence of equipment additional to a conventional TEM (Lorentz lens, biprism, segmented detector).
The DPC method is the most useful for multiscale imaging, i.e. fast switching between studding of objects of about 10µm down to a few nm in size. However the conventional DPC requires specially designed position sensitive detector(s) and costly hardware solutions.
Here we report a simple generalization of the DPC imaging method to emidiatly extend the capabilities of the majority of existing TEM systems (without modifications) towards multiscale characterization of local magnetic properties of nanomaterials. Our method implies a usage of a unitary (non-segmented) virtual bright field detector in combination with a modified differential phase contrast approach.
The suggested method demonstrates high sensitivity to the local magnetic fields, provides a very large field of view, a few nanometers spatial resolution and in-focus condition. It also, in principle, allows direct quantification of nanomaterials magnetic fields.
The usability of our method both at micro- and nano- scale is tested on the investigation of 2 materials: a) cylindrical Co/Ni nanowires with a high aspect (length to diameter) ratio; b) ordered arrays of Co/Ni nanowires – promising candidates for 3D magnetic memory devices.

Recent developments in in situ liquid cell Transmission Electron Microscopy (TEM) techniques enable direct investigation of nano-systems in relevant environments. By encapsulating the liquid between two electron transparent silicon nitride membranes the sample can be introduced into the TEM column without compromising vacuum. This allows for dynamic nanoscale phenomena to be directly observed in situ at high spatial and temporal resolution under controlled electron dose conditions allowing for imaging and chemical analysis.

In this work we use magnetite (Fe₃O₄) nanoparticles as a platform and alter surface chemistry to systematically study their behavior and the effects of the electron beam in situ. Iron oxides are ubiquitous in natural systems, from biomineralization to the terrestrial carbon cycle, and serve as a platform for a range of engineered applications, for example, Magnetic Particle Imaging (MPI). Interactions between iron oxides, solvents, minerals, small molecules, and tissues are fundamental to these diverse systems. Magnetite nanoparticles are synthesized by thermal decomposition of Fe³⁺oleate, and are single crystalline, monodisperse, and phase-pure to ensure uniform physio-chemical and magnetic properties. As-synthesized particles are terminated with oleic acid and soluble in organic solvents, and are transferred to aqueous phase by coating with an ambiphillic co-polymer then functionalized with positive and negative charged species.

Here we demonstrate application of the in situ liquid (S)TEM cell to study behavior of magnetite nanoparticles with different surface chemistries in organic and aqueous solutions. We show that magnetite in the presence of organic solvent (1-octadecene) is stable up to relatively large electron beam doses (>50 e-/Ųs). However, the same magnetite nanoparticles undergo a dissolution process in the presence of water even under low dose conditions of <1 e-/Ųs. The dissolution process can be tuned, as it is proportional to the total e-beam dose delivered to the sample during the experiment, and dependent on surface chemistry of the functionalized magnetite nanoparticles. We show that nanoparticles with charged functional groups interact with reactive species in solution, accumulating ions at the particle surface, slowing dissolution, and enhancing particle interaction and agglomeration. In this presentation we discuss how to mitigate and utilize the reductive effects of the electron beam, both in the case of magnetite and more broadly for other iron oxide/hydroxide phases, beam sensitive oxide materials, and in general for in situ TEM experiments.

This work was supported by NSF-1334012, NIH 1R01EB013689-01/NIBIB, 1R41EB013520-01, 1R42EB013520-01, and the CII LDRD at PNNL. PNNL is a multi-program national laboratory operated by Battelle for the U.S. DOE under Contract DE-AC05-76RL01830. A portion of the research was performed at the EMSL user facility sponsored by DOE-BER and located at PNNL.

Electron energy loss spectroscopy (EELS) in a new generation of aberration-corrected electron microscopes provides direct images of the local physical and electronic structure inside a material at the atomic scale. The sensitivity and resolution can extend to imaging single dopant atoms or vacancies in their native environments Comparable advances in detector technology are now poised to enable a similar resolution in the measurement of structure and fields in materials. Here we describe a high speed, high dynamic range imaging hybrid pixel array detector (EMPAD - electron microscope pixel array detector) for use in electron microscope applications, especially as a universal detector for scanning transmission electron microscopy. The in-pixel circuitry provides a 1,000,000:1 dynamic range within a single frame, allowing the direct electron beam to be imaged while still maintaining single electron sensitivity. A 1.1 kHz framing rate enables rapid data collection while scanning. The scattering is recorded on an absolute scale, so that information such as local sample thickness can be directly determined. By capturing the entire unsaturated diffraction pattern in scanning mode, the detector can simultaneously capture bright field, dark field, and phase contrast information, as well as being able to analyze the full scattering distribution, allowing true center-of-mass imaging for electric and magnetic field measurements, and opening the way for new multichannel imaging modes.

Transmission electron microscopes primarily employ indirect cameras (IDC) for electron detection in imaging, diffraction, and EELS modes. Such cameras convert high energy incident electrons to photons which, through a fiber optic network or lens, are coupled to a camera, typically a CCD. This indirect detection method inherently limits the camera’s point spread function (PSF) and detection quantum efficiency (DQE). Over the last decade, radiation tolerant CMOS monolithic active pixel sensors, which directly image high energy incident electrons, have been developed which offer improved PSF and DQE in comparison to conventional IDCs. Such direct detection cameras, here abbreviated DDCs, have been successfully utilized in cryo-TEM and in situ TEM for both imaging and diffraction, however, the performance of a DDC has not been evaluated for electron energy-loss spectroscopy (EELS). Here we present a comparison of a DDC (Gatan K2 Summit) and an IDC (Gatan US1000FTXP) for the application of EELS. When the DDC is operated in electron counting mode, we find that the DDC offers increased signal to noise ratio (SNR) compared to the IDC for low electron doses. This increase in SNR is attributed to the DDC’s near zero read-out noise when operated in counting mode. We also find that given a fixed beam energy spread and spectrometer dispersion, the DDC provides enhanced energy resolution due to its narrow PSF. When operated at the low dispersion of 1 eV / channel, the DDC is able to resolve Ti L_2,3 edge splitting in a SrTiO₃ sample while the IDC is not able to resolve the splitting. These improvements offered by the DDC have a wide range of applications including efficient and low-dose spectrum imaging, trace element detection and analysis, and time-resolved EELS.

The advent of commercially-available direct detection cameras (DDCs) for transmission electron microscopy (TEM) offers the opportunity to reduce noise in images and diffraction patterns to levels that approach the Poisson noise of the electron beam. For sufficiently low dose rates, their design can enable significant improvements in detective quantum efficiency (DQE) and modulation transfer function (MTF) when compared to conventional charge-coupled device (CCD) cameras. Existing literature on DDCs is focused predominantly on structural biological applications, where they provide clear advantages under low dose conditions, e.g., typically < 10 e⁻Å⁻² . For example, DDCs have been utilised for resolving high-resolution structural information in biological materials in cryo-electron microscopy. Whereas the characteristics of DDCs at dose rates and spatial resolutions that are applicable to biological materials are already well established, in many other areas of TEM the dose rate can exceed 1000 e⁻Å⁻², while the spatial resolution can vary from nanometers to better than 1 Å. In these contexts, the benefit of DDCs is less clear.
Here, we examine this question in the context of high-resolution phase contrast imaging and off-axis electron holography and demonstrate that the improved MTF and DQE of a DDC result in clear benefits over conventional CCD cameras. For electron holography, we find a significant improvement in the holographic interference fringe visibility and a reduction in statistical error in the phase of the reconstructed electron wavefunction. For high-resolution phase contrast TEM imaging of a BiFeO₃ thin film, we show that images recorded using a DDC reveal individual atomic columns with a good signal-to-noise ratio (SNR) at low dose rates. Further improvement in SNR could be obtained by correlation and averaging over a series of images. As a result of the low camera noise, the correlation of individual images is robust even at low dose rates. Similar findings apply to the multiple acquisition of electron holograms, where averaging leads to an improvement in SNR that is close to the ideal root-N behavior (N being the number of images).
Our results show that DDCs are highly beneficial for the acquisition of high-resolution TEM images and electron holograms at low dose rates, thereby minimising potential specimen damage while maintaining an adequate SNR for analysis

Historically, STEM imaging has been dominated by annular dark field and bright field detection, all of which simply integrate the scattering in large parts of the back focal plane of the specimen. Some work has used split detectors to produce differential phase contrast, which has seen a resurgence of interest in recent years. However, the advent of fast direct electron counting detectors has now allowed the imaging of significant portions of the diffraction pattern for every point on the scan. This promises to allow the full richness of the information present in the back focal plane, and which was traditionally used in conventional TEM, to now be used in the formation of STEM images. There are a huge range of possible applications for this, but in this talk, recent progress at Glasgow and Oxford on STEM imaging will be outlined. This will include physical integration of suitable detectors into microscopes, and hardware and software developments for optimum readout. A number of investigations using different parts of the diffraction pattern and different processing methodologies will be covered, including both the details of the experiment and the data processing, as well as the outcomes for the materials under study. These will include atomic resolution phase contrast imaging using a ptychographic reconstruction method, fluctuation electron microscopy of medium range order in glasses, imaging of biological materials, and studying perovskite octahedral tilting in three dimensions.

Transmission electron microscopy studies of polymeric materials are greatly compromised by the extreme radiation sensitivity that these materials display. In particular, this has inhibited the applicability of high-resolution phase contrast imaging (HR-TEM) approaches for exploring the structure of (semi-)crystalline polymeric materials. This is because HR-TEM generally requires substantial electron doses to obtain the requisite signal to noise in the image. However, with the recent advent of direct electron detector technologies [1], the time is ripe to revisit the question as to whether or not HR-TEM methods can be applied to these important materials systems.
In this work, we demonstrate how the combination of aberration-corrected phase contrast electron microscopy, direct electron detection and low dose methods can be used to provide molecular scale (atomic) resolution images of the structure of an important liquid crystalline polymer, poly-(3-hexylthiopene) (P3HT). In particular, we take advantage of the high detection quantum efficiency of a Gatan K2-IS Direct Electron Detector [2], operating in the “counting mode”. This is a low dose mode wherein every primary electron that that impinges on the detector is located, and both the read-out noise and stochastic distribution of energy deposited in traditional scintillator-based systems are eliminated as sources of noise. This allows an effective doubling of the detectable information beyond the traditional Nyquist frequency limit [3]. Utilization of the counting mode allows the acquisition of a series of very low dose images (of order 1-10 electrons/Å^2) to be acquired sequentially, and cross-correlated to obtain an image with maximum signal to noise. Importantly, the serial acquisition allows identification of the maximum dose that the material can withstand, as image degradation is directly visible. In this presentation, we will focus on issues associated with appropriate image acquisition parameters (accelerating voltage, dose, dose rate) for HR-TEM imaging of P3HT, and associated di-block copolymers made of P3HT & polydimethylsiloxane (PDMS). Insights that can be obtained from the ability to image polymeric chains with these structures, as well as “tie chains” that have been proposed to provide charge transport pathways between nanoaggregated structures will be presented.

A dedicated Hitachi HF3300C microscope, “I2TEM”, installed in CEMES (Toulouse, France) has been specially designed to carry on electron interferometry and in-situ TEM experiments. I2TEM is a 300 kV cold FEG microscope fitted with a multibrism set-up, two stages capability, a GIF quantum ER, a 4k X 4k camera and a Cs-corrector “B-COR” from CEOS.
The first stage location within the objective pole piece allows performing classical HREM experiments while the second location is in a field free region above the objective lens and below the third condenser lens and allows carrying TEM imaging or electron holography in Lorentz mode. Contrary to non-dedicated microscope, I2TEM allows, in Lorentz mode, using apertures in the focal plane of the objective lens (i.e. used as “Lorentz lens”) to select diffracted beams essential to carry Dark Field Electron Holography (DFEH) experiments.
In addition, the B-COR can be adjusted to correct for the objective lens aberrations when excited in HREM or in Lorentz modes at various voltages (60kV, 80kV, 200kV and 300kV). This unique multipolar optical system (also called “Aplanator”) has been specially designed to correct not only for the Cs, the axial coma (B2), the three-fold astigmatism (A2), but also to compensate for the radial and azimuthal off-axial coma. These off-axis corrections are achieved thanks to two additional pair of short hexapoles located in between two image planes inside the corrector. These planes are located between three long hexapoles used to compensate the axial aberrations (Cs, B2, A2, …) like in the standard C-COR. As conventional Cs-correctors allow for correcting most of the important first and second order aberrations confined close to the optic axis in HREM images of few ten of nanometers wide, the Aplanator compensates for aberrations in much larger field of view images (the number of equally resolved point regarding the standard π/4 limit is indeed considerably higher). It is therefore of huge interest for large field of view HREM images recorded with a 4k X 4k camera and for low magnification images or holograms obtained in Lorentz mode.
We will present recent results showing the capacity of the B-COR to correct for the axial and off-axial aberrations of the 11 mm pole piece gap of the I2TEM objective lens and achieve 80pm spatial resolution in HREM mode. Results will also be presented showing the capacity of the Aplanator to correct for the objective lens aberration when used in Lorentz mode where 0.5 nm spatial resolution has been achieved.

We demonstrate the formation of partial dislocations in monolayer graphene at elevated temperatures of 500-700°C with sub-Angstrom resolution aberration corrected transmission electron microscopy [1]. The Shockley partial dislocations act to redistribute strain, providing an energetically more favorable structure to the typically observed 5-7 type perfect dislocation. The partial dislocations were able to diffuse through low energy migration pathways, providing some insights into the atomistic dynamics of graphene under annealing.

This data was acquired using an in-situ high temperature holder (DENS Solutions), which provided responsive and accurate sample heating through a Pt wire heating element, and allowed for the stable acquistion of images over long exposure times while still maintaining single atom resolution. Imaging was performed using Oxford's JEOL 2200 MCO equipped with a monochromator filter, reducing the beam energy spread to 217meV at 80kV [2], giving sufficient spatial resolution to permit the delineation of discrete single atoms of carbon in the graphene monolayer [3-7].

The formation of partial dislocations in atomically two-dimensional materials has not previously been either observed or predicted. We observe their formation at temperature's of 500-700°C through nucleation at existing defect sites, and were able to capture a number of glide and climb migration events. The temperature dependence of partial dislocation formation has potential implications in graphene's high temperature plasticity and mechanical properties. Furthermore, our images reveal a number of striking similarities between the partial dislocation structure and dynamics found in graphene to those modelled for three-dimensional hexagonal cubic materials, such as silicon. These include a kink-assisted migration mechanism, allowing for the low-energy migration of partials through consecutive bond rotations. Thus our results are potentially of interest to researchers operating in either the field of nanomaterials or silicon materials.

[1] Robertson, A. W. et al. Nano Letters, 2015, 15, p5950-5955.
[2] Mukai, M. et al. Ultramicroscopy, 2014, 140C, p37-43.
[3] Robertson, A. W. et al. Nano Letters, 2014, 14, p3972-3980.
[4] Chen, Q. et al. ACS Nano, 2015, 9, p8599-8608.
[5] He, K. et al. Nature Communications, 2014, 5, 3040.
[6] Robertson, A. W. et al. ACS Nano, 2013, 7, p4495-4502.
[7] Warner, J. H. et al. Science, 2012, 337, p209-212.

Metallic interconnecting wires in integrated-circuits devices require good electrical conductivity and reasonable resistance against electromigration for device performance and reliability considerations. Nanotwinned Cu metallization that possesses many spectacular properties has been considered to be a good interconnect material. In general, two twinned crystals are separated by a Σ3 coherent twin boundary (TB). For a Cu nanowire (NW) prepared by pulsed electrodeposition, most of TBs are across the NW and in the direction perpendicular to the NW growth direction. In this study, we report the observation of a special twinning structure in Cu NWs by Cs-corrected transmission electron microscopy (TEM). The TB structure and lattice orientation were identified by high-resolution TEM images. Unlike typical triple junctions formed by three grain boundaries, we found that the intercepted TB structure consists of three coherent TBs and one relaxed TB. The stability of TB junction structure will be discussed in the study.

Aberration correction has revolutionized the capabilities of scanning transmission electron microscopes (STEMs) on several fronts. Some of these capabilities are now being used to their full potential, while the likely game-changing impact of others is yet to be felt.
Efficient elemental and chemical mapping, with atomic resolution and single atom sensitivity, are two areas that are maturing. They have been made possible by large beam currents in atom-sized electron probes – several hundred pA and beyond in 1-1.5 Å diameter probes in aberration-corrected cold field emission STEMs - and by the improved capabilities of electron energy loss spectroscopy (EELS) and energy-dispersive X-ray spectroscopy (EDXS) instrumentation [1,2]. Because of its higher collection efficiency, EELS gives a smaller minimum detectable mass (MDM) than EDXS for the same electron dose. EDXS typically gives, because of its reduced spectrum background, the smaller minimum detectable mass fraction (MDMF). When used with lower energy beams that avoid knock-on radiation damage, the two techniques can detect single impurity atoms in many different materials. Comparing experimental energy-loss near-edge structures (ELNES) with theoretical predictions can also determine the density of unoccupied states of the probed atom (or atomic column), and hence the atomic valence (chemical mapping) and the atom’s near-neighbor environment. EDXS enjoys the advantage that it can detect single atoms even in thicker samples [3], in which the EELS spectrum background makes single atom identification harder.
Ultra-high energy resolution (<10 meV) EELS is an area that continues to yield fundamental surprises. First demonstrations that monochromated STEM can detect vibrational spectra from different materials were made less than two years ago [4,5]. A new frontier is now being opened: damage-free EELS, achieved by positioning a nm-sized probe a few tens of nm away from the sample, and collecting vibrational spectra in an “aloof geometry” [6,7].
Another promising future area is combining single atom imaging with single atom manipulation, using either the atom-sized electron beam itself or in combination with different chemical atmospheres to remove atoms, and atomic deposition or ionic bombardment to add new ones.
This talk will summarize progress in the key areas, with emphasis on the most recent results.
[1] D.A. Muller et al., Science 319 (2008) 1073-1076.
[2] T.C. Lovejoy et al., Microsc. Microanal. 21 (Suppl. 3, 2015) 339-340.
[3] R.M. Stroud et al., submitted to Nature Materials (2015).
[4] O.L. Krivanek et al., Nature 514 (2014) 209-212.
[5] T. Miyata et al., Microscopy 63 (2014) 377–382.
[6] R.F. Egerton, Ultramicroscopy 159 (2015) 95–100.
[7] P. Rez et al., submitted to Nature Communications (2015).

High spatial resolution detection of hydrogen-oxygen bonds is important for many materials systems including catalysts, proton conductors, minerals, clays, aerosols as well as life science systems. In many systems, the accumulation of hydrogen species at surfaces, interfaces and grain boundaries plays a critical role in functionalities related to chemical activity and charge transport. Detecting the presence of oxygen-hydrogen bonds in materials with traditional electron energy-loss spectroscopy with high spatial resolution is extremely challenging due to both radiation damage and the problems associated with unambiguous identification of an O-H bond in the low-loss spectrum. With the enhanced energy resolution of recently developed monochromators, it is now possible to probe vibrational as well as electronic excitations at the nanometer level with electron energy-loss spectroscopy (EELS) [1]. The delocalized nature of the low-loss spectrum also makes it possible to use the aloof beam spectral acquisition mode to probe bonding while dramatically reducing electron beam damage. Here we show that this new form of vibrational spectroscopy can be used to detect the presence of hydrogen and oxygen bonds in materials. To investigate the feasibility of OH detection, a series of hydroxide and hydrates have been investigated [2]. The OH stretch mode varied from 430 - 452 meV depending on the local molecular environment around the OH species. The unambiguous vibration fingerprint of the hydrogen-oxygen bond can be correlated with highly localized structural and morphological information obtained from nanoparticles and nanostructures.

References
[1] O.L. Krivanek et al. Nature 514, 209-212 (2014).
[2] P.A. Crozier et al, Microscopy and Microanalysis Proceeding, 2015.
[3] The support from US DOE (DE-SC0004954) and ASU’s John M. Cowley Center for High Resolution Electron Microscopy is gratefully acknowledged.

In the last two years the incorporation of newly-developed monochromators and spectrometers in aberration-corrected electron microscopes has allowed: (i) the improvement of energy resolution of atom-sized electron beams down to about 9 meV and (ii) the optimization of signal collection associated with inelastic scattering at this energy resolution. These quantitative advances inevitably produce qualitative changes in experimental results that might change the way we think about the behavior of atomic scale structures. At Rutgers University, we have obtained optical phonon spectra from several nanomaterials, down to 40 meV in energy. Scattering cross sections increase inversely with the energy loss, so intensities are very strong, reaching 1% of the no-loss intensity for the lowest energies we have measured. In this talk we will present our vibrational EELS results associated with two different types of nanostructured oxides. We have observed macroscopic optical modes, including surface phonons – Fuchs-Kliewer modes – and we are characterizing their behavior as a function of probe position and distance from surfaces in the aloof scattering geometry. The nanostructure geometry supports a variety of vibrational modes. Aloof scattering is very long ranged when electronic screening of the coulomb potential is weak, and thus it often produces a signal that is very similar to IR at large impact parameters. Also, when electrons directly traverse bulk material they can impose large non-dipole polarization fields, driving excitations which are normally IR-silent. Thus as has often been the case, new measurement capabilities are resulting in qualitatively new and exciting ideas about the nanoscale behavior of materials.

We acknowledge financial support of the U.S. Department of Energy, Office of Science, Basic Energy Sciences under Award # DE-SC0005132

lightweight, flexibility, and low-cost manufacturing processes, the power conversion efficiency of these devices is still much too low for them to be utilized in real-world engineering applications. It is commonly agreed that the key to improving these devices is to understand the donor/acceptor (D/A) interface within the photoactive blend layer of an OPV device, but while much work has been done in understanding the morphology of this interface, there remains a lack of knowledge concerning the interface’s electronic structure. The determination and understanding of the electronic structure of this interface is vital as it could allow for the development of better performing OPVs.
Due to its ability to measure the optoelectronic properties of a material with both high spectral and spatial resolutions, electron energy-loss spectroscopy (EELS) in the scanning transmission electron microscope (STEM) is an ideal technique for determining the electronic structure of the D/A. However, when exposed to a monochromated electron beam, the organic materials comprising the photoactive blend layer are susceptible to beam damage, which may lead to broken bonds, changes in the local chemistry, etc., making it impossible to know if the EELS measurements made are representative of the material. Therefore, it is first necessary to determine how to collect EELS data without these organic samples incurring beam damage. This has been accomplished by optimizing the STEM-EELS acquisition conditions in an attempt to limit the beam dose whilst also comparing the results of the EELS measurements to similar experiments conducted using variable angle spectroscopic ellipsometry.
Monochromated (<100meV) STEM-EELS has been conducted on four organic materials commonly used in OPVs, including C₆₀, copper phthalocyanine (CuPc), [6,6] phenyl C₆₁ butyric acid methyl ester (PCBM), and poly(3-hexylthiophene) (P3HT), demonstrating that it is possible to acquire EELS data for these beam sensitive organic materials utilizing a monochromated electron beam and achieving high energy resolution. Furthermore, EELS measurements have been made at the D/A interfaces for two bilayer OPV architectures (CuPc/C₆₀ and P3HT/PCBM) by utilizing rapid spectrum imaging acquisition techniques, signifying that it is also possible to collect EELS data for these beam sensitive materials with high spatial resolution. Using the signals measured for the organic standards coupled with the optimized EELS acquisition methods, peaks in the EELS spectra of these bilayers have been reliably attributed to signals from the donor, acceptor, or a blend of the two materials. Further examination of the collected data will yield information about the electronic structure of the D/A interface so that the relationship between the charge generation at the interface and the efficiency of the OPV device may be understood.

More than 15 years have passed since the first successful development of a Cs-correction system for a modern high-resolution TEM [1]. This type of corrector now is standard for high-resolution TEM and STEM. One important aspect of the first decade of aberration correction has been the development of new high-resolution TEMs and STEMs dedicated to aberration correction due to an increased overall stability.
Just after the introduction of commercially available Cs-corrected TEM/STEM the discussion started if one can further reduce the resolution limiting aberrations in order to achieve a new landmark of resolution in the range of 50 pm. The chromatic aberration Cc was assumed to be the major resolution limiting parameter and, hence, had to be cancelled. The benefits of Cc-correction are highest at low energies where the relative energy width DE/E is largest (E < 100 keV).
With the first Cc/Cs-corrected TEM the defined goals with respect to resolution could be achieved [2] although an increase of an unexplainable image spread in TEM mode was noticed. After thorough investigation over more than three years the origin of this effect finally could be explained theoretically and experimentally [3]. The image spread is caused by thermal magnetic noise produced by freely moving electrons which are present in any conductive material. The resulting noise (Johnson noise) depends on the temperature and the type of material and currently seems to be unavoidable.
A new project concentrating on high resolution at low energies (20 keV ≤ E ≤ 80 keV) was initiated by U. Kaiser, Univ. Ulm, Germany [4], for which a new Cc/Cs-corrector has been designed. The SALVE corrector is optimized for low energies, a favorable C₅ of around +5 mm, a large field of view and minimum effects from Johnson noise. As the base instrument a FEI Titan (S)TEM has been selected and installed at our premises in June 2015. Currently, the corrector is installed and the alignments are carried out. When working with this system we got the impression that the goal of this project to achieve an usable acceptance angle of 50 mrad at all high tensions is well feasible. Johnson noise seems to be the primary challenge for Cc-corrected systems due to the necessarily large beam diameters within the corrector. An optimization, however, is possible as it has been done for the Cc/Cs-corrector installed at the PICO instrument at ER-C, Jülich/Germany [5]. Recently, we could “upgrade” the PICO corrector by reducing the beam diameter, optimizing the multipole elements, and increasing their excitation. The modification now is installed at ER-C and an improvement of the attainable resolution and stability has been demonstrated.
References
[1] M. Haider et al, Nature 392 (1998) p 768.
[2] C. Kisielowski et al., Micr. and Microanal. 14 (2008) p 469.
[3] S. Uhlemann et.al., Phys. Rev. Lett. 111 (2013) 046101.
[4] URL: http://www.salve-project.de/home.html.
[5] J. Barthel and A. Thust, Ultramicroscopy 111 (2010) p 27.

For a conventional TEM, the Scherzer defocus is typically used to maximize phase contrast and resolution. For a TEM with a spherical aberration (C_s) image corrector, correcting C_s towards zero improves resolution. However, this also dramatically reduces phase contrast due to the uncorrected chromatic aberration (C_c) damping envelope. To balance resolution and phase contrast, Jia et al introduced the negative C_s imaging (NCSI) method in which a negative medium C_s value and a corresponding Lichte defocus are chosen to gain strong phase contrast with sufficient resolution and facilitate direct structural mapping. With C_s close to zero, we have found that C_c correction can be used to play an important role to improve resolution for both phase-contrast and amplitude-contrast HREM as a small C_c value can improve resolution without compromising phase contrast. In addition, C_c correction has an even more substantial effect on the amplitude contrast transfer function where C_c correction offers the possibility to exploit amplitude contrast imaging (ACI) in HREM. Under ACI conditions, atomic resolution channeling contrast can be realized, allowing us to obtain directly interpretable high-resolution electron microscopic images with discrimination between light and heavy atomic columns. Using this imaging approach, we have successfully visualized the atomic structure in a BaTiO₃/CaTiO₃ superlattice with high spatial accuracy and discrimination between Ba and Ca columns, providing direct visualization of the Ca and Ba associated oxygen octahedral tilt that controls ferroelectric behavior in these superlattice structures. Combined with the first-principles calculations, we found that a metastable “interface phase” of CaTiO₃ with large ferroelectric polarization is stabilized by the mechanical and electrical boundary conditions of the BaTiO₃/CaTiO₃ superlattice. Under this new mechanism, a large number of perovskites with the CaTiO₃ type structure will become good candidates for novel highly-polar multiferroic materials.

*This work was performed at the Center for Nanoscale Materials, a U.S. Department of Energy Office of Science User Facility under Contract No. DE-AC02-06CH11357.

The advent of chromatic aberration-correction in the transmission electron microscope (TEM) offers new prospects for high-resolution imaging at low voltages and energy-selective imaging. Recent improvements in the setup of the achroplanatic CEOS CCOR corrector allow for sub-ångstrom resolution at an acceleration voltage of 50 kV and enhanced optical stability over the timespan of several minutes provides for reliable energy-filtered transmission electron microscopy (EFTEM) spectrum image data at atomic resolution. Experimental examples of low-voltage high-resolution and energy-filtered images of complex oxides, thin layered materials and nanoparticles obtained with Jülich’s chromatic aberration corrected microscope “PICO” will be presented to demonstrate the unique optical properties of the CCOR.
Atomic-scale transmission electron microscopy pushed towards low electron energy by virtue of chromatic aberration correction opens a new horizon for direct imaging of atomic details of nanostructures at reduced radiation damage. Here it is used to analyse the structure and defects in atomic sheets of 2D materials and catalytic hybrid nanostructures.
EFTEM showing atomic detail becomes practicable because of the negligible chromatic focus spread after chromatic aberration correction. The achroplanatic CEOS CCOR corrector allows to record elemental maps on a large field of view with large energy windows, which is essential for the dose-efficient acquisition of atomic resolution images formed by the weak inelastic core-loss scattering. Aspects of atomic resolution EFTEM will be discussed. The quantification of EFTEM maps towards atomic resolution chemical composition maps is in general complicated by the preservation of elastic contrast emerging from elastic scattering. Thin specimen and careful choice of contrast transfer settings yield directly useful qualitative elemental maps on the atomic scale.

In the most recent generation of transmission electron microscopes, chromatic aberration correction promises to provide improved spatial resolution and interpretability when compared with the use of spherical aberration correction alone, as the improved temporal damping envelope of the objective lens, especially at lower accelerating voltages. The reduced dependence of image resolution on energy spread in a chromatic aberration corrected microscope offers benefits for conventional bright-field and dark-field imaging as a result of the decreased influence of inelastic scattering on spatial resolution, even when using zero-loss energy filtering. Less refocusing is also necessary when moving between regions of different specimen thickness, while for energy-filtered TEM chromatic aberration correction allows large energy windows and large objective aperture sizes to be used without compromising the spatial resolution of energy-loss images.

Here, we assess the benefit of combined chromatic and spherical aberration correction of the Lorentz lens of the Titan PICO TEM in Forschungszentrum Juelich for magnetic-field-free imaging with the conventional microscope objective lens switched off. We use Fourier transforms of spherical and chromatic aberration corrected lattice images taken in Lorentz (magnetic-field-free) conditions to demonstrate a spatial resolution of better than 0.5 nm. We present an experiemntal and theoretical study of the benefit of chromatic and spherical aberration correction for studies of magnetic microstructure in materials, both at domain walls and at surfaces and interfaces in materials. We also discuss the factors that presently limit spatial resolution in Lorentz mode in the Titan PICO microscope and propose approaches that can be used to resolve these limitations.

Solid oxide fuel cells (SOFCs) can be operated using carbonaceous fuels by adding an internal reforming layer to the anode. Developing intermediate temperature devices (500°C - 750°C) would solve many issues that currently limit SOFC applications [1]. Ni/Gd co-doped ceria anode materials have shown promise for operating at intermediate temperatures and maintaining good catalytic performance [2]. Ni serves as the fuel reforming catalyst but may also result in carbon deposition onto the Ni metal surface [3]. Carbon deposition becomes a major problem as carbon can coat the surface and render the Ni inactive. Carbon can also dissolve into the bulk Ni metal and cause stress which may lead to fractures and failure [3]. Ceria has been shown to oxidize carbon on its surface through release of oxygen atoms from the lattice [2]. Therefore, Ni/Gd co-doped ceria may inhibit carbon deposition while maintaining good reforming activity.
To investigate this process, we have performed reforming tests over a model Ni/Gd doped ceria catalyst used to reform methane. Ex situ and in situ environmental transmission electron microscopy (ETEM) was employed using an FEI Titan aberration corrected transmission electron microscope (TEM) to investigate the atomic level processes that result in carbon formation. Imaging and energy dispersive x-ray spectroscopy (EDS) were performed on a 15wt% Gd, 8wt% Ni co-doped ceria catalyst to investigate the morphology and structure of the material, as well as the location of carbon deposition on the catalyst during the partial oxidation of methane reaction. In ex situ studies, graphite formation was seen on Ni particles that were separated from the particle aggregates. In areas where Ni particles were in intimate contact with Gd-doped ceria (GDC), small to no carbon formation was seen. This suggests that the proximity of the Ni particles to the GDC regions influences the carbon formation. In order to further study carbon deposition, in-situ ETEM experiments will be performed under reactive gas conditions to directly observe the structure of the particles and the spatial distribution of carbon deposition; results will be presented [4].

[1] Brett, D. J. L. et al, Chemical Society reviews 37 (2008), p. 1568–78.
[2] Zhou, Y. et al, Physical Chemistry Letters 9 (2010), p. 1447.
[3] Atkinson, A. et al, Nature Materials 3 (2004), p. 17–27.
[4] We gratefully acknowledge support of NSF grant DMR-1308085 and ASU’s John M. Cowley Center for High Resolution Electron Microscopy.

Colloidal metal nanoparticles are considered promising for catalysis, sensing, and spectroscopy applications [1]. To tailor the functionality of these materials, it is essential to understand factors that influence their growth behavior and morphological evolution. It has been demonstrated that shapes influence surface properties [2], but a better understanding of noble metal nanoparticle growth mechanisms is needed for rational shape-controlled synthesis [3].
In situ liquid cell microscopy has recently been utilized to directly image the nucleation and growth of nanocrystals in order to elucidate growth mechanisms and reaction kinetics [4,5]. In our work, the radiolytic effects of the electron beam are used to reduce the metal precursor resulting in nanocrystal nucleation and growth [6]. By using in situ liquid cell STEM, we observe dendritic gold nanostructured materials from HAuCl4 as a controlled experiment. We investigate how two different capping agents, cetyltrimethylammonium bromide (CTAB) and sodium citrate, affect Au nanocrystal growth morphology. The experiments were performed in a static fluid stage (Protochip Poseidon500) with an FEI Titan S/TEM operating at 300kV. Electron dose and concentration of surfactants were varied in order to observe their effects on the in situ nanocrystal growth and shape evolution. We show how spherical gold nanoparticles grew slowly when subjected to a relatively high concentration of CTAB (50 mM ) at a low dose rate. In contrast, at low concentration of CTAB gold particles with a larger size were formed in spherical, polyhedral, and cubic-like shapes. In addition, the growth rates and nucleation rates of the nanocrystals were dependent on the initial concentrations of CTAB and electron dose rates. In future work, we will deepen our analysis of the interactions between CTAB, sodium citrate and HAuCl4 in order to shape-control nanocrystal synthesis and correlate the specific effect of the capping agents on the evolving facet dynamics during nanocrystal growth.
References:
1.Tao, A. R., Habas, S., & Yang, P. (2008). Small, 4(3), 310–325.
2.Grzelczak, M.,Perez-Juste, J.,Mulvaney, P.,Liz-Márzan, L. M. (2008).Chem. Soc. Rev. 37, 1783−1791.
3.Xia, Y., Xia, X., & Peng, H.-C. (2015). Journal of the American Chemical Society.137, 7947-7966.
4.Kraus, T., & de Jonge, N. (2013). Langmuir, 29(26), 8427–8432.
5.Liao, H.-G., Niu, K., & Zheng, H. (2013). Chemical Communications, 49(100), 11720–7.
6.Park, J. H., Schneider, N. M., Grogan, J. M., Reuter, M. C., Bau, H. H., Kodambaka, S., & Ross, F. M. (2015). Nano Letters. 15, 5314-5320.

The spatial and energy resolution are the two of the most crucial parameters in electron energy-loss spectroscopy in scanning transmission electron microscopy (STEM-EELS). With ultrahigh energy resolution the width and the tail of the zero loss peak is significantly reduced which proves to be essential for analyzing the peak shifts of surface plasmons of metallic nanostructures, especially for metals with surface plasmon energies < 1 eV. The new generation aberration-corrected monochromated STEM-EELS has demonstrated an energy resolution of ~ 9 meV and thus made phonon spectroscopy possible inside the electron microscope [1]. With such instrument, we investigated the various degrees of coupling between metal nanoparticles and semiconducting nanostructures. The size dependence of surface plasmonic resonance energies of metallic nanoparticles (e.g., Ag, Au and Cu) have been investigated. The coupling of these metal nanostructures with large band-gap semiconductors (e.g., ZnO) and insulators (e.g., MgO) can be studied by monochromated STEM-EELS. By choosing different types and sizes of metal nanoparticles one can tune the energies of their surface plasmons to probe their interaction strength with the chosen semiconductors. For example, weak and strong plasmon-exciton coupling between Ag nanoparticles and ZnO nanowires have been investigated [2] and the observed resonance phenomenon was interpreted by coupled oscillator model. The size dependent resonance energies of localized surface plasmons of Ag, Au and Cu with particles sizes from ~20 nm to the quantum region of ~1-2 nm are being investigated. These results provide a fundamental understanding of the energy transfer mechanisms in metal/semiconductor nanocomposites systems [3].

1. Krivanek, O. L.; Lovejoy, T. C.; Dellby, N.; Aoki, T.; Carpenter, R. W.; Rez, P.; Soignard, E.; Zhu, J.; Batson, P. E.; Lagos, M. J.; Egerton, R. F.; Crozier, P. A. Nature 2014, 514, 209-212.
2. Wei, J.; Jiang, N.; Xu, J.; Bai, X.; Liu, J. Nano Lett. 2015, 15, 5926-5931.
3. This work was supported by the start-up fund of the College of Liberal Arts and Sciences of Arizona State University. The authors gratefully acknowledge the use of facilities within the LeRoy Eyring Center for Solid State Science at Arizona State University.

Superconductor (SC) ferromagnet (FM) Nb/Co multilayers have been produced by magnetron-sputtering with a 100 nm thickness of Nb and 5, 10, 20 nm of Co. The superconducting properties have been investigated by electric transport measurements. It was found that the thicker Co layers decrease the superconducting transition temperature (T_c) less than the thinner ones. In order to understand this unexpected behavior, the microstructure of the layers has been investigated by means of Atomic Force Microscopy (AFM), conventional and aberration corrected Transmission Electron Microscopy (TEM) and by STEM energy-dispersive X-ray spectroscopy (EDS). It was found that the decisive parameter which determines the effect of magnetic layers on T_c of the superconducting layers is not only the roughness (R), but the ratio of the roughness to thickness (d_Co) of the magnetic Co layer, δ=R/d_Co. For δ>1 the magnetic stray field of the magnetic layers is the main reason for the T_c reduction.

The development of the high-resolution transmission electron microscope makes the study of materials characteristics from qualitative to quantitative. The quantitative study is to obtain quantitative characteristics of constituted atoms on the nanometer or smaller scale related to the materials properties, such as atomic displacements, element distributions, and local component concentrations. At the early stage of the quantitative high-resolution transmission electron microscopy (HRTEM), besides the insufficient microscope resolution, it was also seriously affected by imaging distortions due to the contrast transfer function (CTF) modulation and dynamical scattering effect, resulted in it generally on the nanometer scale. The extraction of quantitative information on the atomic scale still requires considerable additional efforts.

The application of spherical aberration (C_s) corrected transmission electron microscopes makes the quantitative HRTEM truly enter into the atomic scale. Compare to the scanning transmission electron microscopy, the HRTEM has a shorter image acquisition time and a higher signal-to-noise image ratio. The former helps to reduce outside disturbances at the time of taking images and thus avoid the formed image distortions. The latter contributes to observe some weak, but extremely important changes of the image contrast, such as in the defective specimen regions. Furthermore, the negative C_s imaging technique, which has the even stronger contrast, has been used frequently in the quantitative HRTEM. The technique is using the optimum focus image under the negative C_s condition in combination with image calculations.

Although the optimum focus image can eliminate image distortions caused by the CTF modulation and has the resolution on the sub-angstrom scale nowadays, the dynamical scattering effect, which has seriously influence on the image contrast, is still unavoidable. In the quantitative HRTEM, the analysis of the image contrast is critical because it can be used to determine atom positions (the peaks of the image contrast), elemental distributions, local component concentrations (the contrast is highly sensitive to the atomic number accumulated along atomic columns), etc. It has already known that, due to the dynamical scattering effect, the structure image contrast changes differently with the specimen thickness for differently weighted atoms. Unfortunately, this rule is qualitative and can be only used for distinguishing the atomic species. Thus, it is still not clear the impact of dynamical scattering effect on the image contrast for the quantitative analysis of C_s-corrected images. Here, to resolve this problem, the multislice image simulation was used for the cubic SiC crystal. The optimum images simulated under the positive and negative C_s conditions and the structure images obtained after processing of simulated images with non-optimum focus were examined.

Nano beam diffraction in the TEM is a very versatile tool to characterize the local atomic structure and orientation of crystalline materials as well as the atomic packing in amorphous solids, which can be further combined with EDX analysis for compositional information. The possibilities this approach provides for high-end materials characterization will be discussed using the following examples:
For crystalline materials automated crystal orientation mapping (ACOM) has been well established over the last couple years for ex-situ [1] and in-situ investigations [2]. We have been using this approach to characterize the lithiation state in partially charged, beam sensitive Li_xFePO₄ (LFP). The ACOM analysis will be compared with an EFTEM analysis at the Fe-L edge, the Li-K/Fe-M edge and in the plasmon range showing excellent agreement. The results will be discussed in terms of dose, reliability and achievable information, e.g. crystallographic orientation coherence or mismatch at the interface between lithiated and delithiated LFP.
Pair distribution analysis [3] is a simple but powerful approach to access the information of the short-range ordering in amorphous materials such as metallic glasses. We have extended the 1D pair distribution function to 3D data cube by scanning nano beam diffraction to map structural variations in glass nanocomposites using multiple linear least square fitting. Using Sc₇₅Fe₂₅ and Sc₁₀Fe₉₀ nanoglasses [4] as examples, we could not only show the differences between the two glasses in terms of dominant Sc-Sc vs. Fe-Fe bonding, but also identify local differences confirming the nanoglass structure with different Sc-Sc/Fe-Fe and Fe-Sc contributions at the core and at the interface regions in agreement with compositional variations measured by STEM-EDX/EFTEM.

[1] E.F. Rauch, E. F., L Dupuy, Metallurgy and Materials (2005) 87–99
[2] A. Kobler et al., Ultramicroscopy (2013) 128, 68-81.
[3] D.J.H. Cockayne, Annual Review of Materials Research (2007) 37(1), 159–187.
[4] J.X. Fang, et al., Nano Letters (2012) 12(1), 458–463.
[5] Support through the EU project Hi-C and the Karlsruhe Nano Micro Facility is gratefully acknowledged.

Study of defects formation and behavior under different stimulus in two dimensional materials is very important to understand its effects on materials’ mechanical, physical and chemical properties. In this talk, dynamics and stability of extended defects in few – layered hexagonal boron nitride are investigated at different temperatures from 450 to 900 °C using an aberration – corrected transmission electron microscope (TEAM 0.5). The evolution in the size, shape and edge configuration (zigzag versus armchair) of triangular defects over time and at different temperatures is discussed.

Recent advances in STEM EELS monochromation now allow for routine ultra-high energy resolution of 15 meV or better in the low-loss region [1]. This capability opens up the opportunity of detecting subtle signal changes in the bandgap region as well as inter-band and surface states. Here we apply this technique on different oxide nanoparticle systems including wide bandgap semiconductors and UV absorption photocatalysts. Local bandgap changes have been determined by scanning the focused electron beam across an interface. Surface states were probed using aloof beam techniques. The measurements give insights on the local electronic structures which could significantly affects the properties of the catalysts since charge transfer and catalytic reactions take place on the surface/interfaces. This study focuses primarily on TiO₂ and Ta₂O₅ based photocatalyst systems functionalized with Ni/NiO surface co-catalysts due to their interesting catalytic behaviors [2,3]. Well-defined MgO nanocubes were also investigated because of their strong interaction with water. Line scans across NiO/Ta₂O₅ interface has shown local bandgap shifts when the beam is positioned at different distance away from the interface. Spectral simulation of inter-band states were performed using both dielectric models and density of states calculation to correlate the experimental results to the theoretical electronic structures of the materials.

[1] O.L. Krivanek et al, Nature 514 (2014), 209.
[2] L. Zhang et al, J. Phys. Chem. C 119.13 (2015), 7207.
[3] Q. Liu et al, Appl. Catal. B: Environ.172 (2015), 58.
[4] The support from US Department of Energy (DE-SC0004954) and the use of NION microscope at John M. Cowley Center for High Resolution Microscopy at Arizona State University is gratefully acknowledged.

Resistive switching phenomena in binary and complex metal oxides have attracted great interest for applications in next-generation non-volatile memory devices. Recently, nanoparticles have been adopted as switching materials, both for use in smaller devices and as model systems for understanding resistive switching mechanisms in nanoscale materials [1,2].
Here, we investigate the origin of resistive switching behavior in individual nanoparticles that have an electrical bias applied to them in situ in the transmission electron microscope (TEM). We use current-voltage (I-V) measurements in the TEM to study resistive switching mechanisms in real time while simultaneously applying electron energy-loss spectroscopy (EELS) and off-axis electron holography to the same particles. The latter technique allows projected electrostatic potential distributions in materials to be recorded with nm spatial resolution [3], providing information about the role of heterointerfaces and conducting nanofilaments during switching processes.
We study TiO₂ nanoparticles that were synthesized using a sol-gel method and then vacuum annealed. We observe reproducible forming-free bipolar switching behavior in individual nanoparticles that are contacted electrically in the TEM using a moveable electrical probe. I-V curves obtained from single particles are correlated with local EELS and electrostatic potential measurements, providing direct information about nanoscale microstructural, compositional and electrical changes occurring single nanoparticles during resistive switching.

[1] E. Goren et al. Appl. Phys. Lett., 105, 143506 (2014)
[2] Q. Hu et al. Semicond. Sci. Tecnol., 30, 015017 (2015)
[3] M. R. McCartney et al. Annu. Rev. Mater. Res., 37, 729 (2007)

The integration of dissimilar materials is highly desirable for many different types of applications but often challenging to achieve in practice. The unrivalled imaging (and spectroscopic) capabilities of the aberration-corrected electron microscope enable enhanced insights to be gained into the local structure and chemical bonding at heterostructured interfaces. To illustrate some of these ongoing issues, this talk will describe recent investigations of oxide/oxide, oxide/semiconductor, and semiconductor/semiconductor interfaces.

Ni-based superalloys are an integral part of jet turbine engines, especially because of their high temperature strength and stability. Given the elevated temperatures and longer hold times at which current engines are operated, a fundamental study of creep deformation mechanisms is crucial to ensuring ongoing development and safety of these parts. Under such operating conditions, higher diffusion rates can lead to stacking fault formation and segregation, which can drastically change the creep properties of these alloys.
The superalloy in this study has minor variations in composition compared to that of the commercially available ME3 superalloy. Compression creep tests showed improved properties over the ME3 alloy, requiring further study of the different creep mechanisms. It was shown that the hexagonal D0₂₄ η phase forms along isolated superlattice extrinsic stacking faults (SESFs) with preferential segregation of solute atoms to the fault. The formation of this phase requires a reordering process that is aided by the elevated temperature and thus diffusion rates.
Within the D0₂₄ crystal structure, it has been proven through first principles calculations that site-specific segregation is energetically preferred. To characterize this segregation along the SESFs, atomic resolution energy dispersive X-ray spectroscopy (EDX) was performed using an aberration corrected scanning transmission electron microscope (STEM). Rigorous quantification of the experimental data conclusively shows that the η phase exhibits site-specific ordering, as well as preferential segregation of certain elements from the surrounding γ′ matrix. The increased creep resistance in this alloy can be attributed to SESF formation and the subsequent reordering process, inhibiting twin formation.
As a further confirmation of the segregation within the fault, high angle annular dark field (HAADF) STEM image simulations were performed on structures that had been relaxed through the use of first principles calculations. Statistically representative structures were generated using the quantified EDX results for both segregated and randomized structures. Simulated images of the η phase in γ′ are shown to be in very close agreement with those taken of the experimental alloy.

Cassiterite (SnO₂) is a wide-band n-type semiconductor having many applications in the field of optoelectronics, e.g. for transparent conductive films, gas-sensors, solar cells, etc. For some applications, such as the production of hierarchic branched structures with high surface to volume ratio, twinning is preferred over single-crystals. Twinning on (101) planes of cassiterite is common in natural crystals and it is also frequently observed in synthetic materials produced under specific conditions (reducing atmosphere), ¹ or with the addition of certain dopants (oxides of Fe_, Sb, Nb, Ta).² In order to exploit twinning, the origin and mechanism of twin formation must be understood. Generally it is accepted that twins form either by deformation (high pressure), transformation (polytypic transitions) or growth. The formation of growth twins has been shown to be chemically induced.³ They are triggered by the presence of small amounts of specific solute elements during crystal growth. These elements are incorporated into the twin boundary and cause exaggerated growth of the whole crystal along the twin plane. Twins can also form by topotaxial replacement reactions, where a precursor phase recrystallizes into one or more products in oriented manner,⁴ or by self assembly through oriented attachment of crystals.⁵ Differences between the five twin types reflect in their coherency, local structure and chemistry of the interface. Deformation twins only exhibit local displacements within a fixed sublattice, whereas transformation twins show disrupted sublattice stacking across the twin plane. In addition to disrupted stacking, chemistry induced twins contain solute elements at structurally defined positions along the twin boundary plane. In addition to the above features, topotaxial twins may contain oriented inclusions of the precursor phase, whereas for twins formed by oriented attachment, deviations from ideal crystallographic relation are observed. To distinguish between the twin types, accurate structural and chemical analyses are necessary. Here, corrected STEMs appear to be the most appropriate tool. In our study of (101) twins in cassiterite (SnO₂) we used natural and synthetic samples. The twin boundaries were analyzed on a JEM 2010F TEM/STEM and a JEM-ARM200 CF with aberration-corrected STEM (both JEOL, Japan). We have shown that natural twins contain coherent inclusions of magnetite and hercnyite, with ferrous iron, indicating reducing conditions during their growth. In synthetic Nb-doped SnO₂, the formation of twins may be related to charge compensation due to the Nb⁵⁺ incorporation into the cassiterite lattice by reduction of Sn⁴⁺, which in effect triggers the twinning.¹
Ref: ¹Zheng JG et al. (1996) JApplPhys 10:7688. ²Kawamura F et al. (1999) JAmCeramSoc 82:774. ³Daneu N and Rečnik A (2012) Acta PetrolMineralAbstractSer 7:32. ⁴ Rečnik et al. (2015) ContribMineralPetrol 169:19. ⁵Penn RL and Banfield JF (1998) AmMin 83:1077.

Au Yolk-Shell or rattle-like structures have increasingly attracted scientific attention due to their potential application on catalysis, drug delivery and energy storage. In this work a combined seed mediated method with galvanic replacement is used to obtain Au yolk-shell structures from initial core shell Au@Ag templates. On the first step, Au truncated octahedral seeds of 23 nm in size are synthesize in the presence of CTAC/CTAC. Core-shell Au@Ag nanocubes are obtained by the addition of Ag ions at 50°C. The core-shell nanocubes are highly sensitive to radiation damage, after 20 minutes of interaction with the electron beam; superficial reconstruction of high indexed facets is observed. The subsequential addition of gold to a solution containing the Au@Ag nanocubes template triggers galvanic replacement between the Ag⁰ on the shell and the metallic [AuCl₄]^-1 species. Truncated cubes and octahedral yolk-shell structures are formed by the fast oxidation of the Ag shell and the diffusion of the Au ions onto the {100} faces. Nanobeam diffraction was employed to analyze the crystalline structure of the faceted nanoparticles. Atomic resolution imaging of Au yolk-shell structures was done using Z-contrast HAADF-STEM. Electron Tomography of a single truncated cube was acquired from -62° to +62°; 3D reconstruction visualization was obtained by the stack of images after a Simultaneous Iterative algorithm method.

Two-dimensional (2D) transition-metal dichalcogenide (TMD) materials, e.g. MoS₂ and WS₂, have attracted a great attention for applications in flexible electronics, optoelectronics and catalysis on account of their unique electronic, optical and chemical properties. Their remarkable properties can be even further tuned via substitutional alloying. There are several reports on growth of monolayer alloys with tunable band-gaps using different techniques. However, there is no detailed study on impacts of alloying on the atomic structure of monolayer crystals. Here, we have grown single-layers of W_xMo_1-xS₂ alloys through annealing MoS₂/WO_x powders in the presence of sulfur vapor at 800 °C and atmospheric pressure. Owing to single-layer nature of the synthesized alloys and using advanced aberration-corrected transmission electron microscopy (AC-TEM) we can directly image their atomic structure and measure their local optical properties with high spatial resolution. In this study, we use aberration-corrected scanning transmission electron microscopy (AC-STEM) to visualize the atomic structure of single-layer W_xMo_1-xS₂ triangles and show how the chemical composition varies across the monolayer domains. We also investigate the local optical properties as a function of dopant concentration via low-loss electron energy loss spectroscopy (EELS) using a monochromated electron source. Understanding how dopants are distributed in the host lattices at the atomic-level and how they affect the optical properties of monolayers can provide fundamental information, by which next-generation tunable electronics and optoelectronics can be designed.

The properties of functional materials are controlled by the microstructure and local chemistry. With the implementation of aberration correctors, electron microscopy entered a completely new era, enabling studies of sensitive materials at atomic resolution using lower acceleration voltages.

Several different types of sensitive functional materials have been investigated; (i) natural layered hybrid functional materials, (ii) bio-inspired functional materials and (iii) man-made doped functional materials.
(i) Many natural materials are highly complex composite or hybrid materials, in which an organic matrix and inorganic crystalline components are closely linked together forming exceptional architectures. The microstructure and chemical composition of rodents incisors have been characterized using a combination of imaging and analytical TEM techniques. Highly enriched surface layers covering the teeth were discovered, consisting of multiple iron containing varieties [1]. Electronic structure investigations showed that iron is in predominantly 3+ valence state.
(ii) Unique morphologies of natural materials combined with excellent physical and mechanical properties [2] have inspired man-made modern materials design. Although the calcium-phosphate system is widespread, several fundamental aspects are not fully understood. In our study, different organic molecules were employed to stabilize unique calcium-phosphate neuron-like structures [3] consisting of a dense core and thin filaments stretching radially out from the center in a circular form. High-resolution TEM and electron diffraction data suggest that these structures are amorphous. Energy-loss near-edge fine structures of the Ca-L_2,3 and O-K edges acquired from different neuron-like structures were compared to the spectra from standard Ca-phosphate compounds [4].
(iii) The incorporation of the rare-earth dopant Dy into the prototypical 3D topological insulator Bi₂Te₃ has been studied with the intent to achieve higher ferromagnetic ordering temperatures and higher magnetic moments [5]. TEM lamellae of high crystallinity films were prepared by ultramicrotomy. The characteristic crystal structure formed by the stacked quintuple layers separated by van der Waals gaps was resolved. Energy-dispersive X-ray lines-scans were acquired traversing the van der Waals gap between the adjacent quintuple layers. The Dy-L intensity profile revealed the substitutional incorporation of Dy atoms on Bi sites and the absence of Dy in the van der Waals gap.

References:
[1] V Srot et al.: Microsc Microanal 21-S3 (2015), 1143
[2] UGK Wegst and MF Ashby: Philos Mag 84 (2004), 2167.
[3] M Espanol et al.: J Mater Chem B 2 (2014), 2020.
[4] V Srot et al.: Microsc Microanal 21-S3 (2015), 1539.
[5] SE Harrison et al.: Scientific Reports (2015), accepted.
[6] The research leading to these results has received funding from the European Union Seventh Framework Programme [FP7/2007-2013] under grant agreement no. 312483 (ESTEEM2).

Nanoscale ferroelectrics are expected to exhibit various exotic domain configurations, such as the full flux-closure pattern. These flux-closure domains should be switchable and may give rise to an unusually high density of bits as well as undergo vortex-polarization phase transformation. They are also predicted to be potentially useful as mechanical sensors and transducers. Similar domains are well known in ferromagnetic materials, and their topological properties and dynamics are under intense investigation. However, in ferroelectric materials, particularly in tetragonal ferroelectrics, the coupling of polarization to spontaneous strain would be so pronounced that formation of a closure-quadrant with its resultant severe disclination strains could be impossible. We have observed not only the atomic morphology of the flux-closure quadrant but also a periodic array of flux-closures in ferroelectric PbTiO₃ films, mediated by tensile strain on a GdScO₃ substrate. Using aberration-corrected scanning transmission electron microscopy, we directly visualized an alternating array of clockwise and counter-clockwise flux-closures, whose periodicity depends on the PbTiO₃ film thickness. The results provide a new similarity between ferroelectric and ferromagnet, and extend the potential of employing epitaxial strain for modulating ferroelectric domain patterns. Designs based on controllable ferroelectric closure-quadrants could be fabricated for investigating their dynamics and flexoelectric responses, and in turn assist future development of nanoscale ferroelectric devices such as high-density memories and high-performance energy-harvesting devices.

Structure and composition analysis are crucial for understanding any material and for the respective development of new applications e.g. in materials, life science, medicine and biomimetics. Fast chemical analysis from the mm to the atomic scale can be carried out by energy dispersive X-ray spectroscopy (EDS) in the electron microscope (SEM/FIB/STEM). The wide spatial resolution window is a crucial condition for chemically characterizing not only nanoparticles but also their surroundings and distribution in a larger environment. Examples of analysis approaches for any type of nanoparticles, including or applicable to bio-generated metal nanoparticles, will be shown in this contribution and used to explain the available technology, which has substantially evolved in recent years.
Combinations of standard EDS with STEM or SEM were used to study particular chemical elements connected to the intake of iron or to sulfur storage by specialized bacteria. Single EDS detectors in combination with aberration corrected transmission electron microscopes allow the routine characterization of core-shell and more complicated magnetic nanoparticles. Even single atoms of Sulfur and Silicon in a carbonatious substrate can be identified [1].
Multiple detectors and annular detector arrangements ensure large collection angles and minimize shadowing and absorption effects in the spectra [2]. In transmission electron microscopy a high collection angle enables speed allowing the investigation of beam sensitive samples or 3D analysis. In SEM high topography bulk samples can be investigated fast and sometimes in a close to natural state. Examples are Ca oxalate crystals in the agave leaf and nano-clay particles, embedded in a porous polymer matrix. TEM-samples can be investigated in an SEM as well. Using the so-called T-SEM approach nanoparticle types and their distribution over large areas can be statistically evaluated [3, 4].
[1] T.C. Lovejoy et al., Appl. Phys. Lett. 100 (2012) 154101.
[2] R. Terborg et al., Microsc. Microanal.16 (Suppl. 2) (2010) 1302-1303.
[3] S. Rades et al. RSC Adv., 2014, 4(91), (2014) 49577-49587.
[4] D.-V. Hodoroaba et al., IOP Conf. Ser. MSE, EMAS (2015), accepted, in press.

Intermetallic compounds have been the subject of considerable interest as structural and functional materials in a wide range of applications. They are used as engineered materials themselves or as second phase components for high performance. It is important that these materials be fully characterized to permit effective alloy development to meet the requisite balance of properties for a given application. One important parameter is the degree to which these compounds are ordered. This is particularly important in a new series of materials known as high entropy alloys (HEA), or compositionally complex alloys (CCA), which contain typically four to six alloying elements, at or near to equi-atomic concentrations. Many of these CCAs contain an intimate mixture of a disordered bcc phase and an ordered phase with the B2 crystal structure. These ordered phases contain reasonable concentrations of many of the alloying elements in the alloy, and so it is of interest to know how the elements are partitioned to the two sublattices in the ordered structure and to what degree anti-site defects are tolerated, i.e., what is the degree of order in these compounds. An approach to the determination of this latter parameter has been afforded by the combination of aberration-corrected S(TEM) instruments, where electron probe sizes less than interatomic spacings may be achieved and coupled with x-ray energy dispersive spectrometers making use of SDD detectors and large collection angles. This approach has been adopted in the present research, making use of a FEI Themis™ with Super-X™. Experimentally, the compositions of the individual sublattices are determined from spatially-resolved XEDS measurements and subsequent data analysis, and these compositions are plotted onto an Ordering Tie-Line diagram, from which the degree of order is deduced.
Effort was focused on the determination of the degree of order of the B2 phase present in the CCA CoCrCuFeNiAl_1.5. Initially, however, a “standard” sample was employed, namely the compound NiAl, where the sublattice occupancy is well known. This talk will present the results of applying the approach described above to the determination of the degree of order of the B2 compound NiAl, and also the ordered B2 phase in the CCA. Account is taken of the expected electron delocalization and also spectral contributions from phonon scattered electrons, which was done using the procedures described by Allen and co-workers. More serious difficulties were encountered when using Gaussian functions to effect extraction of integrated intensities from maxima, located at the positions of the atomic columns, in the signal profiles derived from the XEDS data. The custom of selecting the “best fit” for the Gaussian functions did not yield a unique solution. Possible solutions to this problem will be discussed, and an assessment of the degree to which these permit a quantitative determination of the degree of order will be presented.

Spherical aberration (Cs) correctors have changed our way of doing electron microscopy. In this presentation, we will show different semiconductor multilayers studied by Scanning Transmission Electron Microscopy (STEM) using a probe Cs-corrected transmission electron microscope (TEM). Images have been obtained on a FEI-TITAN Ultimate TEM and analysed using the Template Matching Analysis (TeMA) and the quantitative STEM method.
The usage of aberration corrector provides sharper image details which can be tracked by image recognition software with much higher accuracy.
From the intensity, the chemical content and the thickness of the sample can be estimated.
From the distance between spots the strain in the layers can be measured.
The results obtained on SiGe/Si and ZnTe/CdSe multilayers will be presented.
Recording series of Z-contrast images allows improving the signal to noise ratio and minimizing greatly scanning errors. In addition, these series allowed us to observe the diffusion of individual Ge atoms near the SiGe layers. During their diffusions, these Ge atoms can occupy interstitial positions. Experimental results have been compared to ab-initio calculations.

Electron energy loss spectroscopy is an invaluable technique to study the detailed structure and the chemical state of materials at unprecedented spatial resolution. In today’s modern aberration-corrected electron microscopes, it is possible to tackle problems requiring the highest energy resolution, down to 10-30meV[1], and highest spatial resolution, down to the Ångström level, so that atomic resolved spectroscopy with high spectroscopic sensitivity and resolution can be obtained.

In this presentation, various examples of applications of aberration-corrected electron energy loss spectroscopy will be given. First of all, the detection of low-loss features in plasmonic nanostructures and nanoantennas, down to the mid-infrared part of the electron energy loss spectrum, will be shown by imaging resonances down to 0.17eV [2,3] and hybridization effects with strong field enhancements between nanostructures. We will then highlight detailed structural and analytical work in a number of alloy nanoparticles systems ranging from fundamentals of phase stability and surface segregation to catalysts used for fuel cells. We will present test cases where single monolayer segregation is observed in nanoalloys [4] and graphene [5] used for fuel cell catalyst applications, using a combination of high-angle annular dark-field STEM imaging, EELS elemental mapping and simulations. We will show an example of in-situ electrochemistry studies displaying the evolution of individual nanocatalysts during cyclic voltammetry [6] and identical location high-resolution imaging whereby it is possible to clearly show the dissolution of Pt from catalyst nanoparticles. This powerful technique can also be used to study of the structure and substitutional effects from single atom dopants in phosphors [7], metallic alloys [8], high-temperature superconductors [9] and semiconducting nanowires used for LED applications [10].

[1] E.P. Bellido et al., Microscopy and Microanalysis 20, 767-778 (2014)
[2] D. Rossouw et al., Nano Letters 11, 1499-1504 (2011)
[3] D. Rossouw, G.A. Botton, Phys. Rev. Letters 110, 066801 (2013); S. J. Barrow et al. Nano Letters 14, 3799-3808 (2014); Y. Liang, D. Rossouw et al., Journal of the American Chemical Society 135, 9616-9619 (2013)
[4] S. Prabhudev et al., ACS Nano 7, 6103-6110 (2013), and ChemCatChem (in press) DOI: 10.1002/cctc.201500380.
[5] S. Stambula et al., Journal of Physical Chemistry C 118, 3890-3900 (2014)
[6] G.-Z. Zhu et al., Journal of Physical Chemistry C 118, 22111-22119 (2014)
[7] G.-Z. Zhu et al., Phys. Chem. Chem. Phys. 15, 11420-11426 (2013)
[8] M. Bugnet, A. Kula, M. Niewczas, G. A. Botton, Acta Materialia 79, 66-73 (2014)
[9] N. Gauquelin et al., Nature Communications 5, 4275 (2014)
[10] S.Y. Woo et al., Nanotechnology 26, 344002 (2015) and Nano Letters 15, 6413-6418 (2015)

Since GaN rose to technological prominence twenty-five years ago, convergent beam electron diffraction (CBED) has been used to determine the polarity of this non-centrosymmetric material. In the past decade aberration corrected scanning transmission electron microscopes have made it possible to determine the polarity of these materials by imaging the N columns directly using the annular bright field (ABF) technique. Variation exists in the literature concerning polarity determined by these two methods, even when the nanowires (NWs) are grown by the same technique under similar conditions on nominally identical substrates. In an effort to understand these differences we have measured the polarity of a number of GaN nanowires and films both directly, with ABF, and indirectly, with CBED. Both techniques have been supplemented with image simulations. Samples for the study were grown by plasma assisted molecular beam epitaxy on HVPE template GaN and Si(111) substrates. The growths on Si(111) were initiated with AlN layers followed by a variety of AlN and GaN buffer layers. CBED and ABF measurements were made along either the [1-100] or [11-20] zone axes. Surprisingly the two methods do not always yield the same polarity on the same sample. ABF imaging was performed on NWs grown under six different sets of conditions. Four of these samples showed nitrogen polarity; the other two showed mixed polarity. For the mixed polarity NWs the inversion domain boundaries are parallel to the growth direction. The predominance of N polarity in the NWs is consistent with the recent suggestion that this polarity is required for catalyst-free NW nucleation.[1] In some instances, however, these measurements are inconsistent with CBED measurements on the same sample set, which suggested gallium polarity. There are several factors which may contribute to errors in polarity determined from CBED images. Incomplete knowledge of specimen thickness may cause incorrect polarity determination, since CBED patterns in GaN are sensitive to specimen thickness. The lower spatial resolution of CBED relative to ABF imaging may not allow differentiation of polarity changes such as those observed in wires with mixed polarity in this study. In addition, stacking faults or structure polytype defects may complicate CBED pattern interpretation. Finally, the very small dimensions of nanowires may limit the accuracy of the diffraction due to the reduced volume of material contributing to the pattern. The influence of these various factors and results on thin film samples will be discussed.

Most types of point defects, including high atomic-number substitutional impurities, interstitial impurities, and self-interstitials, have been directly imaged, with single-defect sensitivity, by either transmission electron microscopy (TEM) or scanning TEM (STEM). However, imaging point defects that decrease, rather than increase local density, such as vacancies, remains a significant challenge.
Here we report an approach for imaging single La vacancies in LaMnO₃ by combining frozen phonon multislice simulations with high precision high-angle annular dark-field (HAADF) STEM experiments. Simulations show that the vacancy causes a La column intensity reduction and small shifts in the position of the neighboring atomic columns, both of which are large enough to be detectable in the high precision HAADF STEM experiments, and both of which depend on the vacancy depth along the column due to probe channeling. Candidate images of single La vacancies with both column intensity reduction and neighboring atomic columns position shifts statistically consistent with simulations were found in experimental images of LaMnO₃ thin films grown on DyScO₃.
A Bayesian statistical model comparing experimental image features (intensity and column shifts) to simulations was applied to each single La vacancy candidate to calculate the probability of a vacancy existing in the atomic column and, if it exists, the probably that it occupies a center depth along the column. Preliminary results suggest that we can localize single vacancies to within 2 unit cells.

In this contribution the application of aberration-corrected electron microscopy for the characterization of novel semiconductor heterostructures and complex oxide materials will be discussed. Amongst the materials to be addressed in the talk are bismuth vanadate, which is used as a photocatalytic material for energy applications, epitaxially grown Ge/Si heterostructures and a novel perovskite-type multiferroic material. Aberration-corrected (scanning) transmission electron microscopy in combination with electron energy-loss spectroscopy is employed to locally characterize electronic and structural defects. For BiVO₄, it is shown that a reduced surface shell leads to the formation of an n⁺-n homojunction which has a detrimental effect on its photocatalytic properties. The importance of the advanced resolution of aberration-corrected microscopes for understanding defects in materials is demonstrated in the characterization of a unique dislocation structure in Ge microcrystals for which conventional electron microscopy imaging leads to a wrong result. Moreover, the enhanced resolution enables precise measurement of atomic column positions and thus for measuring the symmetry of unit cells. This is important for the assessment of the strain state in epitaxially grown multilayered oxide materials. Combining the information about the strained unit cell with electron energy-loss spectroscopy makes it possible to relate the strain state with the occupation of oxygen sites. Besides these materials science applications, current and future limits of aberration-corrected electron microscopy will be discussed with emphasize on the electron-atom interaction and its consequences for high resolution imaging at reduced microscope acceleration voltages.

Phase separation of In_xGa_1-xN alloys into Ga-rich and In-rich regions was first predicted by Ho and Stringfellow and later observed by a number of research groups. InGaN samples grown at high temperatures as typically applied in metal organic chemical vapour phase deposition are particularly prone to phase separation. As the indium concentration controls the optical emission properties of InGaN it is important to quantify the degree of phase separation in a thin InGaN film. However, due to the radiation sensitivity of InGaN to beam damage by fast electrons, as observed by O’Neill et al. and Smeeton et al., high electron fluxes as typically used in extended high-resolution imaging in transmission electron microscopy (TEM) or core-loss electron energy-loss spectroscopy (EELS) may lead to erroneous results.
Low-loss EELS can yield spectra of the plasmon loss regions at much lower electron fluxes and so potentially prevent or at least reduce beam damage in the TEM. Unfortunately, due to their delayed edge onset, the low energetic transitions of gallium (Ga 3d transitions yield M_4,5 peaks at 23.8 and 28.5eV) and indium (In 4d transitions yield N_4,5 peaks at 20.0 and 25.9eV) overlap with the plasmon peak which shift from 19.35eV for GaN to 15.5eV for InN.
Here we demonstrate a method to quantify phase separation in InGaN thin films by studying the low-loss region in EELS by fitting both plasmon and core losses over the whole energy range of 13-30eV. Phase separation is shown to lead to a broadening of the plasmon peak and the overlapping core losses, resulting in an unreliable determination of the indium concentration from analyzing the plasmon peak position alone if phase separation is present. In this study we have applied Lorentz fitting for the main plasmon peak. By considering the chemical shift of the InGaN plasmon peeak with chemical composition, as well as corresponding shifts of Ga 3d and In 4d transitions, artificial spectra for each indium concentration have been constructed. Finally, multiple linear least-squares regression can be performed to fit experimental spectra as weighted superpositions of spectra corresponding to pure GaN, pure InN and an In_xGa_1-xN alloy. For x=0.3 and x=0.59, the relative contributions of the binary compounds are negligibly small and indicate random alloys. For x_nom.=0.62 we found strong indication for phase separation, manifested by the need to include significant contributions of GaN and InN to explain the broad peaks in some of the spectra where energy-dispersive X-ray spectroscopy suggested =0.68.
While the initial experiments reported here have been undertaken in a conventional Schottky field-emission transmission electron microscope, aberration corrected instruments with their improved over-all stability will allow us to perform the same studies with improved spatial resolution in STEM spectral imaging mode, as small electron beams can then be combined with longer acquisition times without too high drift.

Quantitative high-angle annular dark-field (HAADF) imaging in scanning transmission electron microscopy (STEM) places experimental images on an absolute intensity scale normalized to the incident beam intensity for direct comparison with image simulations. It has been used to count the number of atoms in the atomic columns and to determine the depth position of dopant atoms in thin samples. A major challenge is to improve the signal to noise ratio and thus the interpretability of quantitative HAADF images for defect analysis. For example, when determining the depth positions of dopants in a sample, certain configurations have an uncertainty range that spans multiple positions, even when the precision is less than a unit cell, even for ideal cases such as heavy dopant atoms that have fairly high visibility [1].
Here we introduce a new method [2], variable-angle HAADF-STEM (VA-HAADF), which uses the angular information of the signal within the atomic-number-sensitive HAADF mode that addresses these challenges. For example, for imaging dopant atoms, the angular dependence of the scattering depends on the dopant atom depth position. Thus, by combining the depth information from multiple detector settings using VA-HAADF, the uncertainty in the dopant depth position measurements is significantly reduced. Using image simulations, we show that the combined information from just two detectors already reduces the uncertainty in the dopant depth position measurement and can uniquely identify atomic configurations that are indistinguishable with a single detector setting. We experimentally demonstrate the VA-HAADF using Gd-doped SrTiO₃ films where the Sr columns contain between zero and two Gd dopant atoms and that are imaged in HAADF mode using two different collection angles. We discuss the opportunities for future developments and applications.
[1] J. Hwang, J. Y. Zhang, A. J. D'Alfonso, L. J. Allen, and S. Stemmer, Phys. Rev. Lett. 111, 266101 (2013).
[2] J. Y. Zhang, J. Hwang, B. J. Isaac, and S. Stemmer, Sci. Rep. 5, 12419 (2015).

Atomic-resolution core-level electron energy-loss spectroscopy (EELS) has become an extremely powerful tool for probing nanomaterials. Performed in a scanning transmission electron microscope (STEM), the technique uses an atomic-sized electron beam, which is rastered across the sample while the EELS signals are collected. The technique enables mapping of the chemical composition and electronic bonding of materials at the atomic scale. However, it is often applied at a qualitative level only. Moreover, the chemical maps acquired using EELS are prone to artifacts which can lead to erroneous interpretations, particularly when heavy-scattering elements are present. In this work, we demonstrate that atomic-resolution EELS chemical mapping can be made quantitative. We also demonstrate a simple procedure that corrects the contrast in the experimental maps so that they can be directly interpreted without artifacts.

Artifacts in the EELS maps arise primarily from elastic scattering of the electron probe within the sample. Such elastic scattering can dominate the contrast, especially in the cases of heavy atomic columns, lower energy losses, or smaller EELS collections angles. As a result, artifacts such as volcano-like structures and contrast reversals have been reported. We have developed an approach to remedy this problem, which uses the simultaneously acquired elastic signal to correct the artifacts. This approach has been demonstrated using pure Si and (BaTiO₃)₈/(SrTiO₃)₄ heterostructures. The corrected maps exhibit essentially-pure chemical contrast that it is directly interpretable.

Quantitative atomic-resolution chemical mapping refers to the ability to perform element-specific column-by-column atom counting. This technique is considerably more challenging than quantitative annular dark-field (ADF) imaging (which is now a relatively common quantitative atomic-resolution imaging technique) due to the greater complexity of the underlying inelastic core-loss scattering. Here we show that absolute quantification is possible under certain criteria using a DyScO₃ single crystal specimen. Through the acquisition of carefully calibrated EELS maps and ADF images on an absolute intensity scale, together with double-channeling simulations of the core-loss signals, we show that both the contrast and the absolute intensities in the experimental Dy-M₅ EELS maps have excellent agreement with the simulations. On the other hand, at lower energy losses, we find only semi-quantitative agreement in the Sc-L_2,3 and Dy-N_4,5 maps, signalling a break down of the adopted physical model and suggesting possible improvements.

Overall, with the advances in instrumentation and electron scattering simulation methods, we demonstrate that atomic-resolution EELS chemical maps can be made quantitative and directly interpretable. This capacity has significant implications for the materials communities concerned with phenomena occurring at the atomic or nanometer length scales.

In combination with the latest hardware such as aberration-corrected scanning transmission electron microscopes (STEMs) and high-speed detectors, the advances in the recent software developments allow us to acquire large-scale datasets such as multidimensional image series and spectrum images (SIs) via electron energy-loss spectrometry (EELS) and X-ray energy dispersive spectrometry (XEDS). Although the trend to acquire large-scale datasets is desired, it is more challenging to deal with (or even to analyse) the large-scale datasets. For examples, extraction of unknown features and estimation of dominant trends which are just initial steps of data analysis toward quantification are challenging enough. If the datasets are relatively noisy, which is very common for atomic-resolution EELS/XEDS SIs, data analysis would be much harder tasks. Multivariate statistical analysis (MSA) is one of efficient approaches to analyse the large-scale datasets in terms of feature identification and extraction. Since a use of MSA is relatively straightforward, various MSA approaches have been applied to SIs as data-mining and noise-reduction tools. Despite that the MSA approach is very efficient and useful, it may create unexpected artifacts especially in higher noise conditions. Since these artifacts might mislead results, it is essential to explore the sensitivity and the validation of the MSA approach. In this study, the MSA sensitivity is evaluated by using SIs simulated to match atomic-resolution analysis conditions. Then, procedures to improve the MSA sensitivity will be proposed.

In order to evaluate the MSA sensitivity, X-ray spectra and EELS spectra were simulated using NIST Desktop Spectrum Analyzer package and Gatan EELS Advisor, respectively. In the simulation, the acquisition time and the beam current were set as typical values for atomic resolution analysis in aberration-corrected STEMs. Once spectra were simulated, Poisson noise was applied and relatively small-scale SIs were synthesized by adding a different spectrum with slightly modified composition to a single pixel. By applying MSA, the detection sensitivity of the single pixel variation MSA was evaluated. The evaluation results from simulated Sis indicate that the PCA reconstruction process may modify absolute signals when the signal variation is in the random noise level or below.

There are two approaches to improve the MSA sensitivity: (1) reduction of random noise and (2) enhancement of true variations. The former requires modifications in experimental conditions (higher currents and longer acquisitions). Conversely, the latter can be achieved by PCA analysis to divided small segments within a SI. Now this approach has developed as local PCA. The division can be made (a) spatially and (b) spectrally. Especially the spectrally local PCA is useful for EELS SIs since the background intensity varies significantly. Advantages of the local PCA approach will be addressed.