Symposium TC4 : Advances in Spatial, Energy and Time Resolution in Electron Microscopy

Annular dark-field (ADF) scanning transmission electron microscopy (STEM) is a powerful technique for materials characterization of complex nanostructures. Recent progress in the development of quantitative methods allows us to extract reliable structural and chemical information from experimental images in 2D as well as in 3D. In quantitative STEM, images are treated as datasets from which structure parameters are determined by comparison with image simulations or by using parameter estimation-based methods [1]. So-called scattering cross-sections, corresponding to the total scattered intensity for each atomic column, are very sensitive for the number of atoms contained in an atomic column and have been shown to be robust to a broad range of imaging parameters [2, 3]. This explains its success to count the number of atoms with single atom sensitivity, which is of great importance to help retrieving the 3D atomic structure [4]. Recent examples demonstrate this potential when combining atom-counting with (i) discrete tomography, (ii) depth sectioning or (iii) an energy minimization approach requiring only a single image as an input [5, 6].

Reducing the number of images by avoiding tilt tomography is of great help when studying beam-sensitive nanostructures. However, the tolerable electron dose is often still orders of magnitude lower than what is typically used for atomic resolution imaging. To understand the effect of electron shot noise, scan noise, and radiation damage, a statistical framework is proposed that allows us to balance atom-counting reliability and structural damage as a function of electron dose. Furthermore, a hybrid method for atom-counting, in which the advantages of a statistics-based and image simulations-based method are efficiently combined in one framework, is suggested. This method is particularly useful for low-dose image acquisitions and is therefore of importance to quantify atomic structures in their native state with the highest possible precision.

Finally, new developments to extend atom-counting from homogeneous to heterogeneous materials are presented. For heterogeneous materials, small changes in atom ordering in the column have an effect on the cross-sections, significantly complicating the analysis. To circumvent the need for time-consuming image simulations to compute scattering cross-sections of mixed columns, an atomic lensing model is proposed based on the principles of the channelling theory. The benefits of this model to unravel the 3D composition at the atomic scale will be demonstrated [7].

[1] S. Van Aert et al. IUCrJ 3 (2016) 71.
[2] S. Van Aert et al. Ultramicroscopy 109 (2009) 1236.
[3] H. E et al., Ultramicroscopy 133 (2013) 109.
[4] S. Van Aert et al., Physical Review B 87 (2013) 064107.
[5] S. Van Aert et al, Nature 470 (2011) 374.
[6] L. Jones et al., Nano Letters 14 (2014) 6336.
[7] K.H.W. van den Bos, accepted for publication in Physical Review Letters (2016).

We have developed a high-precision Z-contrast STEM imaging technique based on non-rigid registration of a series of short exposure images capable of locating the position of single atomic columns in a high-resolution STEM image to better than 1 pm [1]. We have used high-precision STEM to image the surface distortions of a Pt nanocatalyst on a SiO₂ support and to image single La vacancies within the bulk of a LaMnO₃ thin film. The Pt nanocatalyst exhibits a large contraction of ~40 pm at the corner between two {111} facets, and an outward bulge varying from 4 to 12 pm along the adjacent flat {111} surface. La vacancies in LaMnO₃ are identified by both the decrease in scattered intensity they create in the Z-contrast STEM image and the shifts in the positions of the surrounding cation atomic columns. The surrounding cation shifts are only detectable using high-precision STEM and are crucial to distinguishing a vacancy inside the thin STEM sample which is likely to be preserved from the original film, from a vacancy on the STEM sample surface, which is likely to be introduced by sample thinning. By comparing experimental data on atomic column intensities and positions to multislice simulations from first-principle derived vacancy structures using a Bayesian statistical model, we can localize the vacancies inside the sample depth, usually to an uncertainty of one atomic layer. These experiments detect a concentration of randomly distributed La vacancies consistent with the thin film growth conditions for the sample.

[1] A. B. Yankovich, B. Berkels, W. Dahmen, P. Binev, S. I. Sanchez, S. A. Bradley, A. Li, I. Szlufarska, P. M. Voyles, Nature Communications 5, 4155 (2014).

If oxides alone show intriguing physical properties, such as colossal magnetoresistance, high T_c superconductivity or multiferroicity, when combined and confined they show the emergence of unexpected electronic, magnetic, phononic or topological phenomena. These emergent states reflect the competition between different underlying instabilities, broken symmetries, interfacial doping or interface induced structural modifications. Advanced microscopy and spectroscopy techniques enable fundamental understanding of such heterostructures, allowing probing interfaces in real space with atomic resolution. This constitutes a critical task in oxides as the system properties depend on slight structural distortions in the oxide lattice at the interfaces or around point defects.
This talk will briefly review some state-of-the-art applications to oxide heterostructures using annular bright-field (ABF) combined with high angle annular dark field (HAADF) imaging modes in the aberration-corrected scanning transmission electron microscope (STEM), which are perfectly capable to visualize both cation and oxygen atomic positions in oxide heterostructures. Examples will include the identification of the specific distortions of the A, B, and O sub-lattices of the ABO₃ perovskite structure, which allowed measuring in real space the tilts and the amplitude of the polar distortions of 3, 5 and 7 unit cell thick LaAlO₃ layers grown on SrTiO₃. The structural and chemical characterization of an unforeseen point-defect complex including both Cu and O vacancies that leads to the formation of ferromagnetic clusters in YBa₂Cu₃O_7-x high Tc superconductors will also be discussed.
Primary collaborators: G. Herranz, M. Stengel, R. Mishra, R. Guzman, E. Bartolome, M. Valvidares, M. Scigaj, F. Sánchez, J. Fontcuberta, M. Roldan, T. Puig, X. Obradors, M. Varela

Despite being of wide commercial use in devices,^[1] the orders of magnitude increase in resistance that can be seen in some semiconducting BaTiO₃-based ceramics, on heating through the Curie Temperature (T_C), is far from well understood. Current understanding of the behavior hinges on the role of grain boundary resistance that can be modified by polarization discontinuities that develop in the ferroelectric state.^[2,3] However, direct nanoscale resistance mapping to verify this model has rarely been attempted and the potential approach to engineer polarization states at the grain boundaries that could lead to optimised positive temperature coefficient (PTC) behavior is strongly underdeveloped.

Here we present direct visualization and nanoscale mapping in a commercially optimised BaTiO₃-PbTiO₃-CaTiO₃ PTC ceramic using Kelvin probe force microscopy, which shows that, even in the low resistance ferroelectric state, the potential drop at grain boundaries is significantly greater than in grain interiors. State-of-the-art aberration-corrected scanning transmission electron microscopy and electron energy loss spectroscopy reveal new evidence of Pb-rich grain boundaries symptomatic of a higher net polarization normal to the grain boundaries compared to the purer grain interiors. Due to the clear chemical and intrinsic change in symmetry identified at the grain boundaries, a quantitative analysis of the crystal field splitting (CFS) was carried out; the remarkable trend seen is suggestive of octahedral distortion transitioning across single grain boundaries. These results validate the critical link between optimised PTC performance and higher local polarization at grain boundaries and suggest a novel route towards engineering devices where an interface layer of higher spontaneous polarization could lead to enhanced PTC functionality.

[1] B. M. Kulwicki, Journal of Physics and Chemistry of Solids. 1984, 45(10), 1015-1031.
[2] W. Heywang, Journal of the American Ceramic Society. 1964, 47(10), 484.
[3] G. H. Jonker, Solid-State Electronics. 1964, 7(12), 895-903.

Lithium ion rechargeable batteries have been used as a power source for electronic devices and hybrid vehicles. The present batteries contain liquid electrolytes, which includes some safety issues such as leakage. For the development of a full solid-state Li-ion rechargeable battery, it is required to explore solid-state electrolytes with faster Li ion conductivity. Inorganic solid electrolytes with excellent thermal and electrochemical stability are one of the good candidates for solid-state electrolytes. Among various solid-state electrolytes, Li_3xLa_(2/3)-xTiO₃ (LLTO) single crystal shows excellent conductivity of 1.1×10^-3 Scm^-1 (x ~ 0.11), which is close to that of liquid electrolytes. Polycrystalline materials are feasible for more practical usage, however Li-ion conductivity of polycrystalline becomes one or two orders of magnitude smaller than that of single crystals. This significant drop of ionic conductivity in polycrystalline should be related to the grain boundaries (G.B.s) and therefore it is necessarily to understand how individual grain boundaries affect local ionic conductivity. In this study, we have investigated LLTO grain boundary structures by using scanning electron microscopy (SEM) equipped with electron backscatter diffraction (EBSD) detector and atomic-resolution scanning transmission electron microscopy (STEM).

The grain orientations of polycrystalline LLTO (x ~ 0.11) were mapped by using SEM-EBSD, where Σ5 G.B. was found to be the most frequently observed G.B. among other coincidence site lattice G.B.s. To understand the mechanism behind decreasing ionic conductivity, the local atomic structures of the Σ5 G.B. were characterized by annular dark-field (ADF) STEM imaging. We identified that the Σ5 G.B. has the orientation of symmetric tilt Σ5 [100]/(013), and the boundary plane has a glide symmetry with Ti-column termination. Furthermore, the Σ5 G.B. has some specific structural features in the distributions of interatomic distances and the sites of La atoms. We have carefully measured the interatomic distances between La-La and Ti-Ti columns with several picometer precision. It was found that the bond lengths in the vicinity of the G.B. core region show 6% lattice expansion compared with that in the bulk. The local lattice distortion is confined within a single layer adjacent to the G.B. core. On the boundary plane, distinct La enrichment in the Li-rich columns is also observed, which may prevent Li ion conductivity across the boundary. Detailed discussion will be given in the presentation.

Acknowledgement. A part of this work was supported by the Research and Development Initiative for Scientific Innovation of New Generation Batteries (RISING2) project of the New Energy and Industrial Technology Development Organization (NEDO), Japan, and was also conducted in the Research Hub for Advanced Nano Characterization, The University of Tokyo, under the support of "Nanotechnology Platform" (Project No.12024046) by MEXT, Japan

In this talk we will discuss the structural and physical properties of oxide based ferromagnetic (FM) /ferroelectric (FE) tunnel junctions where FE BaTiO₃ (BTO) tunnel barriers are sandwiched between FM La_0.7Sr_0.3MnO₃ (LSMO) electrodes. LSMO/BTO heterostructures grown by high pressure O₂ sputtering tend to display symmetric interfaces, which help stabilizing head-to-head domain walls within the FE layer. Atomic resolution scanning transmission electron microscopy observations combined with electron energy-loss spectroscopy and theory confirm the presence of such domain walls, associated with the presence of O vacancies. We will show how a confined electron gas arises at the wall, which highly affects the tunneling transport across the junction. Resonant tunneling assisted by the discrete levels of the FE quantum well results in strong quantum oscillations of the tunneling conductance. Harnessing the electronic properties of such domain walls within ferroelectric tunnel barriers may pave the way towards future devices with new functionalities. Acknowledgements: Research at UCM sponsored by Spanish MINECO MAT2014-52405-C02-01, MAT2015-66888-C3-3-R and Fundación BBVA. Research at ICMM sponsored by MINECO MAT2014-52405-C02-2-R. Electron microscopy observations at Oak Ridge National Lab supported by the U.S. DoE, Basic Energy Sciences, Materials Sciences and Engineering Division, and through a user project supported by ORNL’s CNMS User Program, which is also sponsored by DOE-BES.

Nanostructures with periodic phase separation hold great promise for creating 2D and 3D superlattices with extraordinary functionalities. Understanding the mechanisms driving superlattice formation demands an understanding of the underlying local atomic structure. However, nanoscale structural fluctuations intrinsic to these superlattices have insufficient long range perfection to diffract to high order, making them extremely difficult to detect and characterise with conventional diffraction-based structure determination methods. A direct imaging method is necessary to probe the local structure. However, this is also challenging, requiring sensitivity to picometre atomic displacements of light elements.

Using aberration-corrected scanning transmission electron microscopy (STEM), we have developed an optimized atomic-level pseudo-bright-field condition to image and measure the oxygen octahedra in perovskite oxides. We used electron scattering calculations to determine detector collection angles that enable octahedral tilts and associated oxygen displacements to be measured and mapped with picometre precision and robustly over large specimen thicknesses up to 150 nm.

Applying this real-space octahedral-tilt mapping to Li_0.5–3xNd_0.5+xTiO₃, a promising solid electrolyte in Li-ion batteries, revealed an unconventional superlattice with 2D modulated octahedral tilts. A mathematical description of the octahedral-tilt modulation was derived based on the quantitative tilt maps, which explicitly identified the high-order harmonic character of the modulation. Using simultaneous annular-dark-field (ADF) imaging, we also mapped the lattice parameters unit-cell by unit-cell, uncovering highly-localized strain associated with the tilt modulation. Furthermore, we demonstrate the tunability of the tilt modulation by changing Li stoichiometry. Amazingly, we observe a reversible annihilation/reconstruction of the tilt modulation correlated with delithiation/lithiation process, suggesting this structural transformation is associated with Li-ion conduction in this promising Li-ion conductor [1].

The above observations are largely inaccessible from conventional diffraction analysis [2], and lead to an unprecedented mechanically-coupled tilting competition model to explain the superlattice formation [1]. This real-space approach to quantify local octahedral structure and correlate it with strain can be applied to other advanced oxide systems.

This work was supported by the Australian Research Council (ARC) grants DP110104734 and DP150104483 and a Monash University IDR grant. The FEI Titan³ 80-300 S/TEM at Monash Centre for Electron Microscopy was funded by the ARC Grant LE0454166.

[1] Y. Zhu, R. L. Withers, L. Bourgeois, C. Dwyer & J. Etheridge Nature Materials 14 (11), 1142-1149 (2015).
[2] A. M. Abakumov, R. Erni, A. A. Tsirlin, M. D. Rossell, D. Batuk, G. Nénert & G. V. Tendeloo, Chem. Mater. 25, 2670–2683 (2013).

Layered complex oxides offer an unusually rich materials platform for emergent phenomena through many built-in design knobs such as varied topologies, chemical ordering schemes, and geometric tuning of the structure. A multitude of polar phases are predicted to compete in Ruddlesden-Popper(RP), A_n+1B_nO_3n+1, thin films by tuning layer dimension (n) and strain; However, direct atomic-scale evidence for such competing states is currently absent. Using aberration-corrected scanning transmission electron microscopy with sub-Ångstrom resolution in Sr_n+1Ti_nO_3n+1 thin films, we demonstrate the coexistence of antiferroelectric, ferroelectric, and new ordered and low symmetry phases. We also show, the atomic rumpling of the rock salt layer, a critical feature in RP structures that is responsible for competing phases, is directly imaged; Exceptional quantitative agreement between electron microscopy and density functional theory is demonstrated. The study shows that layered topologies can enable multifunctionality through highly competitive phases exhibiting diverse phenomena in a single structure.
By a direct comparison between atomic-scale imaging and density functional theory [1], this work shows that the Ruddlesden-Popper thin films under large tensile strain are a host to a large number of competing phases at room temperature that can possess ferroelectric, antiferroelectric, ferrielectric, paraelectric, and other complex phases. Layered oxides such as the Ruddlesden-Popper structures studied here provide a natural mechanism to decouple adjacent perovskite blocks through rumpling across a rock salt layer that is directly imaged in this work and shows exceptional agreement with DFT predictions. A combination of a highly degenerate energy landscape consisting of many competing phases, combined with decoupled 2-dimensional polar blocks separated by rock salt layers could be a basis for the design of new classes of digital relaxors and electrocaloric materials. While the conventional approach of achieving phase competition in oxides is largely through inducing chemical disorder, this study suggests alternate paths by tuning a layered structure with topology and strain knobs. The quantitative sub-Å imaging and DFT study presented here provides a template for the materials-by-design paradigm by interfacing atomic-level theory and atomic-level experiments in a tight feedback loop

The properties of lithium battery are strongly dependent on the behavior of lithium ions during charge/discharge process. Since this behavior determines the stability, lifetime and reliability, direct visualization of Li site is needed to understand the mechanism of the properties. Recently we have proposed that annular bright field (ABF) scanning transmission electron microscopy (STEM) is very powerful technique to produce images showing both light and heavy element columns simultaneously. In this technique, an annular detector is located within the bright-field (direct-scattered) region, and the columns display absorptive-type contrast. In this study, light elements in several lithium battery related materials such as LiFePO₄, LiCoO₂, La_2/3-xLi_3xTiO₃ (LLTO), La_(1-x)/3Li_xNbO₃ (LLNO) are directly observed by ABF STEM, and the mechanism of lithiation/delithiation is discussed based on the observation results. The properties of thin-film batteries is also influenced by the atomic structures of the embedded interfaces, such as electrode/electrolyte and electrode/current-collector interfaces, as well as the grain and domain boundaries. Detailed analyses of these interface structures, which provides insights into formation mechanisms of the interfaces and the effects of microstructure on electrochemical properties, are essential for understanding the mechanism of lithiation /delithiation and for obtaining the guideline to design the thin film devices. In this study, the epitaxial growth mechanism of a typical cathodic LiMn₂O₄ thin film is investigated by exploring the detailed structural and compositional variations in the vicinity of the film/substrate interfaces.
Acknowledgement. A part of this work was supported by Toyota Motor Co., JSPS and NEDO.

Numerous physical and chemical properties of materials are determined by point defects, which require the analytical tools with sufficient sensitivity and spatial resolution to detect the location, spatial distribution, and atomic configuration of point defects. Scanning transmission electron microscopy (STEM) has been successfully applied to study various kinds of point defects. To date, the types of point defects studied in STEM include heavy dopants in light host atoms, interstitial atoms, or materials with relatively high defect density. However, the observation of certain types of point defects, e.g. vacancies or light elements, remains challenging, mainly because of the low contrast of these defects. Here, we carry out quantitative scanning transmission electron microscopy (QSTEM) using a high angle annular dark field detector (HAADF) where the experimental image intensity is normalized to the incident electron intensity.¹ This approach enables truly quantitative interpretation of image intensities using the direct comparison with the frozen phonon image simulations.
We demonstrate the detection of Sr vacancies in molecular beam epitaxy (MBE) grown SrTiO₃ film on SrTiO₃ substrate using QSTEM and image simulations. The use of multiple detector settings in STEM, known as variable-angle (VA) HAADF-STEM², significantly enhances the contrast and interpretability of different defect configurations and improves the measurement precision for concentrations or locations of dopant atoms. In addition, we employ multiple successive image acquisition with rigid registration to enhance the signal-to-noise ratio of STEM images and reduce image distortions caused by the instabilities in STEM. Lastly, picometer precision measurement of atomic column locations reveals the relaxation of local atomic structure around Sr vacancies, which support the QSTEM results of the vacancy detection. The findings in this work provide new opportunities for point defects study and can be applicable to a wide range of crystal structures in general.


References
1. J. M. LeBeau, S. D. Findlay, L. J. Allen and S. Stemmer, Physical Review Letters 100 (20), 206101 (2008).
2. J. Y. Zhang, J. Hwang, B. J. Isaac and S. Stemmer, Scientific Reports 5, 12419 (2015).

We propose a new scanning transmission electron microscopy (STEM) technique that can realize the three-dimensional (3D) characterization of vacancies, lighter and heavier dopants with high precision. While TEM-based techniques have provided unmatched spatial resolution, depth information is required to determine the 3D positions of individual point defects. In the past, efforts have been made to acquire the depth information of individual atoms by using probe channeling information combined with quantitative high angle annular dark field (HAADF) STEM. In the present work, using multislice STEM imaging and diffraction simulations, we show that selecting a small range in low scattering angles can make the contrast of the defect-containing atomic columns substantially more depth-dependent. The origin of the depth-dependence is the de-channeling of electrons due to the existence of a point defect in the atomic column, which creates extra “ripples” at low scattering angles. We show that, by capturing the de-channeling signal with a narrowly selected ADF angles (e.g. 20-40 mrad), the contrast of a column containing a point defect in the image can be substantially enhanced. The effect of sample thickness, crystal orientation, probe convergence angle, and experimental uncertainty will also be discussed. Our new technique can therefore create new opportunities for highly precise 3D structural characterization of individual point defects in functional materials.

Electron energy-loss spectroscopy is a well-established technique for the characterisation of surface plasmon resonances [1-11]. EELS plasmon measurements are usually performed in a scanning TEM (STEM), providing a unique way of exciting localized surface plasmon modes with (sub-)nanometer spatial precision. With this technique, it was for example demonstrated that new, hybrid modes appear when plasmon resonators are closely spaced [12], and it was even shown that such a hybrid mode can induce controlled electron tunneling between the resonators [13]. Earlier, an approach was presented for measuring the quality factor and the dephasing time of surface plasmon resonances, using monochromated STEM-EELS [14]. In this approach, EELS plasmon spectra are regarded as Fourier-transforms of oscillations in the time-domain. This current paper will further expand on that earlier work and demonstrate that the EELS spectrum can be a rich source of information on the dynamics of surface plasmons. By performing systematic series of measurements, it is even possible to probe information of the very early, transient stages of plasmon oscillations. Acknowledgements:
The National Research Foundation (NRF) is kindly acknowledged for supporting this research under the CRP program (award No. NRF-CRP 8-2011-07). This work results from close collaborations with Joel K.W. Yang (SUTD, Singapore) and Wu Lin & Ding Wen Jun (IHPC, Singapore).
[1] C Powell and JB Swan, Phys. Rev. 115, 869–875 (1959).
[2] ZL Wang, JM Cowley, Ultramicrosopy 21, 347–366 (1987).
[3] F Ouyang, PE Batson and M Isaacson, Phys. Rev. B 46, 15421–15425 (1992).
[4] J Nelayah et al., Nature Phys. 3, 348 – 353 (2007).
[5] M Bosman et al., Nanotechnology 18, 165505 (2007).
[6] B Schaffer et al., Phys. Rev. B 041401 (2009).
[7] B. Ögüt et al. ACS Nano 5 (8) 6701-6706 (2011).
[8] H Duan et al. Nano Lett. 12, 1683–1689 (2012).
[9] D Rossouw and GA Botton, Phys. Rev. Lett. 110, 066801 (2013).
[10] M Bosman et al. Scientific Reports 4 5537 (2014).
[11] A Teulle et al. Nature Materials 14, 87-94 (2015).
[12] PE Batson, Phys. Rev. Lett. 49, 936–940 (1982).
[13] SF Tan, Science 343 (6178), 1496-1499 (2014).
[14] M Bosman et al. Scientific Reports 3 1312 (2013).

Scanning transmission electron microscopy (STEM) boosted by aberration-correction technology has made possible the direct imaging and characterization of atomic and electronic structures at localized volumes of many materials and devices. In STEM imaging, a very finely focused electron probe is scanned across the specimen and the transmitted and/or scattered electrons at each raster position are detected by the post-specimen detector(s) to form images. It is well known that the STEM image contrast is strongly dependent on the detector geometries, and in turn we gain flexibility in determining the contrast characteristics of the STEM images by controlling the detector geometry. For example, by using segmented-type detectors, we can image local electromagnetic fields inside materials through differential phase contrast (DPC) imaging techniques. Recently, we have developed a second-generation high-speed segmented-type detector that is capable of atomic-resolution STEM imaging with sub 0.5Å spatial resolution. Atomic-resolution DPC STEM imaging is possible using the detector, which enables the direct visualization of (projected) atomic electric fields in both crystalline materials and even in isolated single atoms. Having applied DPC STEM to the characterization of materials and devices, the local electromagnetic field maps inside materials can be obtained in real space. We found that DPC STEM imaging is very powerful to directly characterize many interesting internal electromagnetic structures such as pn junctions in semiconductor devices[1], polar oxide interfaces and magnetic Skyrmions[2]. Furthermore, other types of atomic-resolution imaging may be possibe by utlizing the segmented-type detector, which will be discussed in the presentation.
[1] N. Shibata et al., Sci. Rep., 5, 10040 (2015).
[2] T. Matsumoto et al. Sci. Adv. 2, e1501280 (2016).

In the scanning transmission electron microscope (STEM), the commonly-used imaging modes, for example annular dark-field (ADF), make use of detectors that sum the intensity over a region of the detector plane. Recent development in detector technology have resulted in cameras with frame-speeds that can exceed 1 kHz, and can therefore be used to record the detail in the detector plane for each probe position during a STEM scan without inordinate increase in dwell time to record a four-dimensional data set. We present experiments performed using the pnCCD (S)TEM camera, a direct electron pixelated detector from PNDetector, mounted on a JEOL ARM200-CF aberration corrected microscope. The detector has a grid of 264x264 pixels and operates at a speed of 1000 frames-per-second (fps) or higher. ADF images can be recorded simultaneously.
Here we make use of ptychography to form a quantitative image of the phase shift of the electron wave due to the sample. The combination of simultaneous phase imaging and Z-contrast is used to solve the previously unknown structure of a complex nanostructure [1]. We compare the signal to noise ratio of ptychography as a function of electron dose with a range of other phase imaging approaches, including differential phase contrast STEM and conventional HRTEM approaches [2,3]. We show that residual lens aberrations can be detected and corrected post-acquisition, and that 3D information can be reconstructed from a single tilt. We examine the robustness of the ptychography approach to dynamical electron diffraction [4].

[1] T.J. Pennycook et al., Ultramicroscopy, 151 (2015) 160-167.
[2] H. Yang, T.J. Pennycook, P.D. Nellist, Ultramicroscopy, 151 (2015) 232-239.
[3] H. Yang et al, Nature Communications, in press.
[4] The authors acknowledge funding from the EPSRC (grant numbers EP/K032518/1 and EP/K040375/1) and the EU Seventh Framework Programme: ESTEEM2.

Assembly of molecular clusters and nanoparticles in solution is now recognized as an important mechanism of crystal growth in many materials, yet the assembly process and attachment mechanisms are poorly understood. To achieve this understanding we are investigating nucleation and assembly of iron and titanium oxides using in situ and ex-situ TEM, and the forces that drive oriented attachment between nanocrystals and the factors that control them via AFM-based dynamic force spectroscopy (DFS). Our hypothesis is that attachment is due to reduction of surface energy and the driving forces that bring the particles together are a mix dipole-dipole interactions, van der Waals forces, and Coulombic interactions. Therefore they can be controlled via pH, ionic strength (IS), and ionic speciation. In-situ TEM using a custom-designed holder and fluid cell to obtain sub-nanometer resolution shows that, in the iron oxide (ferrihydrite) system, primary particles interact with one another through translational and rotational diffusion until a near-perfect lattice match is obtained either with true crystallographic alignment or across a twin plane. Oriented attachment then occurs through a sudden jump-to-contact, after which the interface expands through ion-by-ion attachment at a curvature-dependent rate. Analysis of the acceleration during attachment indicates it is driven by electrostatic attraction with about one unit of charge on each particle driving the event. Ex-situ TEM analysis shows that the TiO₂ nanowire branching occurs through attachment of anatase nanoparticles to rutile wires on a specific crystallographic plane for which the anatase-to-rutile transformation leads to creation of a twin plane. DFS measurements of the forces between (001) crystal basal planes of mica, and (001) planes of TiO₂ show that the forces have strong relationship to pH, IS, and crystal orientation

Elucidating the fundamental mechanisms and kinetics of nanocrystal growth necessitates the utilization of high spatial resolution imaging techniques that are capable of directly imaging individual nucleation and growth events. In situ liquid cell microscopy, offers the advantage of imaging the dynamics of electron beam induced nanocrystal growth. The concentration and diffusion of radiolytic species, generated by electron beam, is known to affect the nanocrystal growth mechanisms and kinetics in liquid cell experimentation. Using a custom-built scan generator, coupled to an aberration-corrected scanning transmission electron microscope (STEM), we perform controlled and time-resolved electron beam irradiation studies to induce localized deposition of metallic nanocrystals from organometallic precursors. In addition to electron beam induced transformation, we further explore how temperature affects the nanocrystal growth kinetics by employing liquid cell heating devices. Combining time-resolved imaging datasets with quantitative image analysis algorithms, factors controlling chemical transformations were revealed.

It is a current imperative to characterize of interfaces relevant to energy, (bio-) medical and sensing applications under realistic operation conditions. . However, most of the powerful surface sensitive characterization techniques such as X-ray photoelectron spectroscopy (XPS), Auger electron spectroscopy (AES), and electron microscopy usually require high or ultrahigh vacuum conditions for their application. This is a limiting factor since most of the practical interfacial processes take place in liquids and under high pressure gas environments. To overcome this “pressure gap” limitations the membrane based environmental cells have been recently developed ^{1, 2 . 3} Such a membrane has to have high electron transparency and be molecularly impermeable to isolate liquid or gaseous sample from the rest of the UHV chamber. Graphene is ultra-thin molecularly impermeable materials with high mechanical strength and high transparency for electrons in a wide energy range and is therefore the best material platform for fabrication of environmental cells for SEM, XPS and other characterizations.

Along these developments, we report on fabrication of the environmetal cell based on bilayer graphene which is compatible to high and ultra-high vacuum environments. The liquid retention time in such a cell is longer than few hours. With this technique, we charactirised the liquid samples with SEM, EDS, SAM, XPS and PEEM. Based on the SEM and EDS characterization, we calculated the secondary electron yields from liquid water. The electronic structure of liquid water was characterized by SAM, XPS, and PEEM techniques. With this cell, we also demonstrated electrochemical deposition of metal on the surface of graphene working electrode from aqueous electrolyte. This work provides a new method to improve the characterization of liquid materials for energy and environments researches.


1. J. D. Stoll and A. Kolmakov, Nanotechnology, 2012, 23
2. J. Kraus, R. Reichelt, S. Guenther, L. Gregoratti, M. Amati, M. Kiskinova, A. Yulaev, I. Vlassiokiv and A. Kolmakov, Nanoscale, 2014.
3. Velasco-Velez, J. J., Pfeifer, V., Hävecker, M., Weatherup, R. S., Arrigo, R., Chuang, C.-H., Stotz, E., Weinberg, G., Salmeron, M., Schlögl, R. and Knop-Gericke, A. ,. Angew. Chem. Int. Ed., 54: 14554–14558, 2015

The assembly process of nanoparticles from individual atoms, and nanostructures from nanoparticles in solution is fundamental for materials engineering and “bottom-up” fabrication of functional nanodevices.
Using dynamic in situ TEM imaging [1-4] in liquids, I will describe how nanoparticles form in solution and how these nanoparticles interact with each other. First, I will discuss how phase separation of a solution containing Au ions into solute-rich and solute-poor phases leads to formation of Au nanocrystal through a pathway that does not follow classical nucleation theory (CNT). Namely, I will show that there are multiple steps that lead to formation of nuclei from which nanocrystal grow (Figure 1A). These steps are: 1) phase separation of liquid solution into solute-poor and solute-rich phases, from which 2) an amorphous nanoparticles which serve as a precursor for nuclei, emerge. This is followed by 3) crystallization of amorphous nanoparticles into a crystalline nuclei.
Next, I will demonstrate multiple routes for self-assembly of nanoparticles into nanostructures. We will describe the types of chemical and physical interactions and their effect on assembly dynamics. Specifically, I will describe the role of hydration, vdW, hydrogen bonding and capillary forces on guiding the self-assembly of nanoparticles
Our findings highlight the role of physical and chemical forces in material synthesis and self-assembly of nanoparticles.
References:
[1] M. J. Williamson, R. M. Tromp, P. M. Vereecken, R. Hull, F. M. Ross, Nature Materials 2 (2003), p. 532.
[2] H. Zheng, R. Smith, Y. Jun, C. Kisielowski, U. Dahmen, A. P. Alavisatos, Science 324 (2009), p. 1309.
[3] N. de Jonge, D. B. Peckys, G. J. Kremers and D. W. Piston, Proc. Natl. Acad. Sci. U.S.A., 106 (2009), p.2159
[4] U. Mirsaidov, H. Zheng, D. Bhattacharya, Y. Casana, P. Matsudaira, Proc. Natl. Acad. Sci. U.S.A. 109 (2012), p. 7187.
[4] U. Anand, J. Lu, N. D Loh, Z. Aabdin, U. Mirsaidov, Nano Lett. 16 (2016), p. 786–790.
[5] G. Lin, X Zhu, U. Anand, Q. Liu, J. Lu, Z. Aabdin, H. Su, U. Mirsaidov, Nano Lett. 16 (2016), p. 1092.
[6] This work was supported by the Singapore National Research Foundation’s Competitive research program funding (NRF-CRP9-2011-04), and the Ministry of Education of Singapore (MOE2015-T2-1-007)

Surface plasmons and hot electrons are generated in certain metals such as Au, Ag, Al due to the interaction of photons or electrons with the surface electrons. Recently, optical methods have been used to show that hot electrons generated by surface plasmons can trigger various chemical reactions at room temperature. (1-3) Understanding the reactions promoted by localized surface plasmon resonance (LSPR) is important for the design of better and more effective surface plasmon or hot electron enabled catalytic systems for a wide range of energy and environmental applications. However, a number of important questions related to this type of reaction processes remain unclear due to the complexity of the reaction kinetics. Nanoparticles often respond dynamically to the gaseous environment, because gas absorption on the surfaces affects the free energy of the exposed surfaces that can lead to particle surface oscillations. (4, 5) Since the LSPR is known to strongly depend on size, shape and dielectric environment of the particles, any change in either of them will also affect the spatial distribution and the frequency of the LSPR. Therefore, one of the fundamental questions that needs to be answered is how the change of particle morphology (shape), as well as dielectric environment, during photocatalytic reactions may affect the spatial distribution of LSPR on the nanoparticles and if that can be correlated with the locations where the reaction occurs.
In light of the complexity of the reaction process, we combined an ensemble of techniques to characterize LSPR-promoted chemical reactions. In particular, we focus on the LSPR promoted dissociation of hydrogen using Al nanoparticles. An aberration-corrected environmental transmission electron microscope (ETEM) is used to generate movies and time-resolved electron energy-loss spectra (EELS) to monitor the crystallographic and chemical change in the particle as well as LSPR locations and shifts. Under gaseous environments we observe oscillatory morphological change in the Al nanoparticle. EELS imaging, with different energy dispersions, is used to acquire both elemental and LSPR maps of the same particle. These combined spectrum images show correlations between the LSPR profile and chemistry of the nanoparticle. The comparison between the EELS mapping of the particle, with and without a H₂ environment, also shows the effect of the environment on the LSPR generated in the nanoparticles locally. This combined approach to decipher LSPR-promoted reactions provides time-resolved, atomic-scale information on the reaction kinetics and improves the understanding of the dynamics of LSPR promoted relations.

1. Sil D, et al. (2014) ACS nano 8(8):7755-7762.
2. Mukherjee S, et al. (2012) Nano letters 13(1):240-247.
3. Thomann I, et al. (2011) Nano letters 11(8):3440-3446.
4. Vendelbo S, et al. (2014) Nature materials 13(9):884-890.
5. Imbihl R & Ertl G (1995) Chemical Reviews 95(3):697-733.

Growth experiments, carried out in situ within the transmission electron microscope, generate unique information that clarifies mechanisms or quantifies the properties of a material as it grows. Improvements in spatial and temporal resolution are welcome in such experiments. Yet higher resolution generally implies increased dose to the sample, and can also require compromises in terms of the in situ capabilities of the microscope. Here we discuss the tradeoffs involved in carrying out different types of growth experiment at high resolution. In environmental TEM, where growth takes place by exposing the sample to reactive gases, we present examples where a high dose rate interacts with the gas environment, particularly with the residual species present, to change reaction pathways, as compared with the same process imaged under a better controlled vacuum. We show that ultra-high vacuum design mitigates such beam-induced changes. However, optimal vacuum design is challenging to integrate with high-performance electron optics. We therefore discuss other strategies to reduce these effects. In liquid cell TEM, reactions such as electrochemical growth are imaged in a thin, enclosed liquid layer. Radiolysis of water by the electron beam creates reactive species that can alter solution chemistry and reaction kinetics. We present radiolysis calculations that include spatial variations in dose rate, and describe the importance of low dose imaging techniques. We discuss the use of chromatic aberration correction or energy filtering to improve resolution at fixed dose by allowing higher signal to noise in the thicker samples often used in liquid cell experiments. For both liquid cell and environmental growth experiments, modern detectors can improve time resolution, reduce dose through higher sensitivity, and track drift and the progress of beam damage. We anticipate an immense overall impact on in situ experiments and exciting future opportunities for understanding materials growth processes.

The ability to characterize high-energy transient intermediates in molecular reactions has been a hallmark of physical chemistry, as the structural information revealed by time-resolved techniques has provided insight into the underlying mechanisms of chemical reactions. Nanocrystals, on the other hand, participate in numerous important chemically-driven structural transformations (e.g. growth, dissolution, phase transitions), but rigorous methods for monitoring the dynamics of these processes or characterizing short-lived metastable species are lacking. In this work, we capture direct structural evidence of anisotropic nanocrystals transitioning through high-energy intermediate morphologies as they are oxidatively etched. These observations are enabled by transmission electron microscope-based characterization of nanocrystals dispersed in solvent trapped between sheets of single-layer graphene. This so-called graphene liquid cell allows for the structural evolution of nanoparticles to be observed with nanometer-scale spatial resolution in real time on single particles in a manner that preserves the three-dimensional ligand and solvation environment of the native colloid. When a range of different anisotropic gold nanoparticles are exposed to an oxidizing environment triggered by the electron beam, they undergo structural transformations to higher-energy particle shapes as they are etched and ultimately dissolved completely. Pseudo-one-dimensional nanoparticles including rods, dumbbell-shaped rods, and pentagonal bipyramids first experience a constant rate of tip-selective oxidation that generates particles with sharp features. These exposed areas of high curvature are more reactive and act to accelerate etching, further sharpening the tips of the particles, resulting in the transient stabilization of high-energy intermediates through a positive feedback mechanism. Three-dimensional polyhedral nanocrystals including cubes and rhombic dodecahedra similarly experience tip-selective oxidation but transition into more well-defined tetrahexahedron-shaped particles. These intermediate structures possess high-energy features ({310} surface facets), and exist transiently as kinetic products before being completely dissolved. Monte Carlo simulations are used to corroborate the qualitative picture of oxidation reactions transitioning through short-lived metastable species and shed light on the mechanistic details. These data reveal that the observed transformation is governed by a coordination number-based step-recession mechanism, which is the microscopic reverse process of the well-known step-flow mechanism of crystal growth. This work provides important insight into anisotropic nanocrystal reactivity and highlights the depth of information that can be gained through investigations of dynamic nanoscale processes.

The nanostructured electrodes have unique advantages, particularly in rate capability, over their micron-sized counterparts for applications in next-generation lithium ion batteries, owing to shortened ionic transport path. In some of the systems, Li incorporation may occur through non-equilibrium routes, which radically boosts power density. Developing advanced nanostructured electrodes requires better understanding of the thermodynamics and kinetics of Li intercalation in individual nanoparticles during the operation of batteries. However, since the local intercalation processes are transient and vary with the structural change of electrode materials with lithiation, few techniques are available for real time tracking such dynamic processes at relevant time and spatial resolution, particularly for those in nanostructured electrodes. Here, we report our recent results from the development and application of in-situ TEM methods, combined with synchrotron X-ray measurements and ab-initio calculations, in unraveling the complex and dynamic processes of Li intercalation in single nanoparticles of working electrodes. Examples will be provided to the studies of Li intercalation in Li_xFePO₄ and Li₄Ti₅O₁₂ nanoparticles upon lithiation, showing the direct evidences of the coupling between local Li intercalation with phase transformations. The observation may offer a generic explanation of the non-equilibrium intercalation and electrochemical reactions in intercalation-type electrodes.

This last decade, fundamental studies of electrochemical phenomena have been slowed down by a lack of effective in situ experimental setup, which is able to clearly identify structural modifications inside and at the surface of electrode materials. The evolution of microstructures, the appearance of cracks and porosities and the transformation of crystalline phases have to be properly investigated in order to get a better insight into the influence of charge/discharge processes in battery materials and the reaction mechanisms implied in electrochemical storage.
In order to monitor microstructural evolution dynamically during electrochemical cycling, we developed a micro-scale battery set-up implemented within a FIB/SEM instrument [1]. Secondary particles are strongly used by industry for positive electrode fabrication. However, so far just a few works were focused on their morphological evolution during electrochemical cycling. Here, single secondary particles of cathode oxide (NCA and NMC) materials with a size of 5-10 μm are attached to the metal pin of the micromanipulator via a conductive carbon bridge. The micromanipulator allows moving the particle in the chamber and immersing into an ionic liquid electrolyte, which is deposited on a counter electrode, i.e. Li-metal. Electrochemical measurements are carried out using ultra low current instrument. After immersing into electrolyte, the single particle of active materials is cycled in galvanostatic mode with a steady current (1nA), which corresponds to a C-rate of about 1 based on the particle volume and the theoretical capacity [2].
We succeed to carry out in situ experiments inside the FIB/SEM chamber using ionic liquid electrode getting high quality electrochemical measurements in galvanostatic and impedance modes for single cathode particle. We studied structural modifications of individual particle after each in situ charge/discharge cycle by FIB slicing and SEM imaging. Using AMIRA software for reconstruction and segmentation steps, we quantified the formation of cracks as a function of cycle number and extracted 2D skeleton and tortuosity. Evolution of the discharge capacity was correlated with cracks and porosities appearance inside cathode materials. Impedance measurements suggested an increase of Li diffusion inside the particle that is relied on the formation of cracks, which induces an enhancement of discharge capacity. The changes of structural parameters that are induced by cycling were extracted from 3D reconstruction and linked with electrochemical properties [3]. Then, 3D structural data was compared to that obtained by 3D Transmission X-ray microscopy tomography [4], which is made in the APS synchrotron at ANL.

[1] D. Miller al et. Advanced Energy Materials, 3(8), 1098-1103, 2013
[2] A. Demortiere et al. Advanced Energy Materials, 2016 (in submission)
[3] Y.C. Chen-Wiegart et al. ElectroChem. Comm., 28, 127-130, 2014
[4] F. Tariq et al. J. of Power Sources, 248, 1014-1020, 2014

Localized corrosion is a ubiquitous problem among passivating metals, and has historically been challenging to study in the Transmission Electron Microscope (TEM) due to surface chemistry changes during sample preparation. We have developed a thin film corrosion cell and characterized its electrochemical behavior as compared to an analogous 3-D corrosion cell. Using Liquid Cell TEM, we are able to observe corrosion phenomena in real time, with nanoscale resolution, under relevant environmental conditions. Pit localization has often been thought of as a stochastic process for pure metals, and data from this study will add to the understanding of how underlying grain structure may contribute to local breakdown. We have directly observed pit growth in situ as well as apparent local growth of polycrystalline deposits. Evidence from Auger Electron Spectroscopy chemical analysis suggests these are metallic aluminum, which supports a mechanism for pit metastability involving the loss of metal ions from the pit solution. Observed holes have been seen to take on either a fractal-like pattern or discrete circular shapes, however, in areas observed in situ, only circular-type holes have grown. We have observed that the shape that pits evolve into during growth is related to the chloride concentration, and also that electron beam irradiation can additionally influence growth shape. To capture the initiation of localized corrosion events – challenging because of the relatively high, compared to grain dimensions, growth distance between successive frames in the image sequence – we are systematically varying temperature and sample thickness. Overall, we present an application of in situ TEM to measurement of localized corrosion phenomena with the goal of adding to the current understanding of corrosion pit initiation and repassivation.

We acknowledge NSF for funding this research (DMR-1309509), cleanroom and characterization facilities in the center for Materials, Devices, and Integrated Systems at RPI, and Ray Dove, Robert Planty, and Deniz Rende for technical assistance.

Liquid cell electron microscopy (EM) is a developing technique that allows us to apply the powerful capabilities of the EM to image and analyze materials immersed in liquid. We can examine liquid based processes in materials science and physics that are traditionally inaccessible to EM. The liquid/ bias cell consists of silicon nitride window on silicon support called E-chip, which separate the liquid from the microscope vacuum and confine it into a layer that is thin enough for TEM imaging. The importance of liquid cell microscopy in electrochemistry is that liquid cell experiments enable direct imaging of key phenomena during battery operation and relate the structural and compositional changes to the electrochemical behaviors^1,2.

In our study we carried out in-situ liquid TEM studies of LiFe_0.5Mn_0.5PO₄ (LFMP) nanoplatelets synthesized with colloidal procedure, while are very promising for ingrate high rate batteries, however The microscopic mechanism of the phase transition from LiMPO₄ to MPO₄ (M=Fe, Mn) and its dependence on particles size still under debate^3,4,5. The synthesized LFMP NPs were studied using advanced TEM: high angle annular dark field (HAADF)-STEM, EDX and EELS mapping to find crystal structure and atomic distributions. HR-HAADF-STEM showed Li/ (Fe, Mn) anti-site defects. EFTEM and ELLS mapping show the coexistence and homogenous distribution of Fe, Mn, and P elements on single particle. For in-situ experiments LFMP nanomaterials based cathode deposited onto a glassy carbon working electrode in E-chip used to encapsulate conventional electrolyte LP30 (LiFP₆/EC/DMC), the assembly were followed with TEM: HAADF-STEM and spectroscopy (EFTEM-EELS) during cycling.

A key mechanistic aspect in the performance of Li-ion battery electrodes is how Li ions intercalate and deintercalate form electrode during cycling. In this study, we probe in real time the evolution of individual grains of LFMP in the native environment of a battery in a liquid cell TEM. The in-situ electrochemistry reproduces well-established, we performed cyclic voltammetry and galvanostatic cyclic inside the TEM. Particles are seen to delithiate in EFTEM images during charging and discharging (lithium-rich and lithium-poor): we used 5eV EFTEM to obtain spectroscopic mapping of nanoparticles, where the rich and poor lithium particles appear with different contrast, brighter for FMP and darker for LMFP. HRTEM, EDX, and EELS were useful techniques to study the local structural and compositions transformation of this nanomaterial.
In this work we demonstrate the unique ability of a liquid cell in-situ TEM to shed the light on the progress of lithium transport across individual particles during cycling of LMFP NPs based cathode material using a conventional electrolyte.
[1] Frances M. Ross Science 2015
[2] Megan E. Holtz et al. Nano letters 2014. [3] Andrea Paolella et al. Nano letters 2014
[4] Anderson A et al. J. Power Sources 2001. [5] Burch D et al. Electrochem Acta 2008

In recent years, electron microscopy has progressed significantly for the study of heterogeneous catalysts at the atomic-scale. While electron microscopy has mainly been conducted under the high vacuum conditions residing in the electron microscopes, the development of differentially pumped electron microscopes and closed, electron-transparent gas cells enables the introduction of reactive gas environments as well. Hereby, electron microscopy opens up capabilities for monitoring catalysts at atomic-resolution during the exposure to gasses at pressures of up to atmospheric levels and temperatures of up to several hundred centigrade and for correlating the time-resolved observations with concurrent measurements of the catalytic functionality. These advancements are extraordinary beneficial to research in heterogeneous catalysis, because information about surface dynamics and reactivity is made accessible and establishes a foundation for “live” observations of structure-sensitive functional behavior at the atomic level. In this contribution, I will outline such electron microscopy advances for the study of catalysts at atomic resolution and in catalytically meaningful environments and showcase how the dynamic observations can elucidate the role of gas-surface interactions on the working catalysts.

Fundamental molecules such as ethylene and propene are the building blocks of modern chemical industry, and to produce these molecules via energy efficient catalytic conversion is of a central interest in both industry and heterogeneous catalysis. The Te-containing MoVNbTeO_x (Te-M1) among a family of Mo-V-O-based complex oxides is, by far, the best catalyst for ethane to ethylene transformation via oxidative dehydrogenation (ODH). However, an independent and valid reaction mechanism for ethane ODH over M1 has not been established to date. One of the greatest challenges is to access the structural information of the catalytic active sites under the operation condition, without which our understanding of the atomic processes and the elementary pathway of ethane ODH over the active sites of Te-M1 will remain confined to conceptual knowledge.
Among various structural characterization tools, environmental ADF-STEM offers unique directly interpretable imaging of potential catalytic sites at the atomic level. In this work, we attempt to shed some light on the atomic mechanism of ethane ODH by tracking and quantifying the evolution of the atomic configuration of the Te-M1 (001) basal plane at reaction temperature 400 ^oC and in the presence of four systematically chosen atmospheres: 1) microscope high vacuum of 10^-8 mbar, 2) 0.9 mbar inert helium, 3) 0.9 mbar oxygen and 4) 0.9 mbar ethane reaction gas mixture. To achieve high accuracy and sound sampling statistics in quantifying the evolution of the Te-M1 structure, an automated line profile analysis capable of robustly locating over 200 structure units with effective background subtraction and atomic displacement calculation was developed and applied to each ESTEM image. This in-depth analysis revealed for the first time a picture of the working structure of a thin Te-M1 crystal: a robust pentagonal {Nb(Mo)₅} framework supporting a highly thermally and chemically dynamic Te site and it’s neighboring active centers under the ethane ODH operation condition. Complementary DFT modeling suggests that these volatile Te components of the M1 may serve as “activators” promoting electron transfers and producing surface O^- radicals that are known capable of abstracting the first H of an alkane. Considering that neither Te nor Te-oxides alone is catalytically active for alkane ODH, this new activation mechanism could actually reflect a special coordination surroundings provided by the M1 matrix hexagonal channel. This realization of a “mesostructure” instead of individual active site could open new strategies for designing highly active catalysts.
This research is part of the Chemical Imaging Initiative conducted under the Laboratory Directed Research and Development Program at Pacific Northwest National Laboratory (PNNL). PNNL, a multiprogram national laboratory, is operated by Battelle for the Department of Energy under Contract DE-AC05-76RLO1830.

Abstract not available.

Reduction and oxidation (redox) processes at metal-oxide surfaces play an important role in heterogeneous catalysis. In reactive environments the oxide surface can undergo substantial reconstruction and exchange oxygen atoms with the gas phase, thus changing the structural coordination and oxidation state of the surface metal ions that determine its catalytic properties. Observations of surface redox processes made by transmission electron microscopy (TEM) at atomic-resolution and sensitivity and in meaningful gaseous environments will therefore be useful for establishing a detailed understanding of metal-oxide catalysts.[1]

Herein, we describe the use of TEM techniques to gain insights into the properties of vanadium oxide supported on anatase titanium dioxide nanoparticles (VO_x/TiO₂),[2] a materials system with application for the selective catalytic reduction of NO_x emissions to the environment. The catalytic properties of VO_x/TiO₂ can be tuned by e.g. the VO_x layer thickness, and the structure and faceting of the TiO₂ support, but atomic-scale understanding of the surface VO_X phase remains elusive. However, new insights were obtained by combining atomic-resolved imaging and electron energy loss spectroscopy (EELS) of the VO_x /TiO₂ catalyst under alternating reducing and oxidizing atmospheres. The investigation was carried out using electron dose-fractionation strategies for the illumination in order to suppress beam-induced alterations of the sensitive catalyst.[3,4] For high-resolution imaging, this was done by means of focal series, which provided additional benefits for e.g. post-acquisition aberration correction through exit wave reconstruction.

The observations revealed that the outermost atomic layers of the VO_x/TiO₂ catalyst underwent reversible changes in the structure and oxidation state when shifting from oxidizing to reducing environments. Specifically, the VO_x surface transformed between a ordered and disordered state concomitant with a reduction from V(V) to V(III). Thus, the cationic sub-lattice dynamically restructured in response to oxygen exchange at the surface and, surprisingly, the restructuring was found to depend on both the VO_x loading and the supporting facet of the underlying TiO₂.[2] This structure dependency provides a first basis for rationalizing the morphology-dependent catalytic properties reported for VO_x/TiO₂ catalysts, and highlights the importance of extending surface characterizations from single crystal model systems to industrial style materials.

[1] S. Helveg, J. Catal. 328 (2015) 102
[2] M. Ek et al., Submitted 2016
[3] S. Helveg et al., Micron 68 (2014) 176
[4] M. Ek et al., Adv. Struct. Chem. Imag. 2 (2016) 4

One of the grand challenges in science has been to watch atomic motions during structural transitions, i.e. watch atomic motions in real time. Due to the extraordinary requirements for simultaneous spatial and temporal resolution, it was thought to be an impossible quest and has been previously discussed in the context of the purest form of a gedankenexperiment. With the recent development of ultrabright electron sources capable of literally lighting up atomic motions, this experiment has been realized (Siwick et al. Science 2003). Increased source brightness, has enabled the study of photoinduced intermolecular charge transfer process in organic systems (Gao et al Nature 2013), as well as cyclization reactions with bond formation and conserved stereochemistry used in synthetic strategies (Jean-Ruel et al JPC B 2013). One observes the innumerable possible nuclear motions collapse to a few key reaction modes. Even more dramatic reduction in complexity has been observed for the material, Me₄P[Pt(dmit)₂]₂, which exhibits a photo-induced metal to metal electron transfer process. This study represents the first full atom resolved structural dynamics with sub-Å and 100 fs timescale resolution (Ishakawa et al Science 2015). At this resolution, without any detailed analysis, the key large-amplitude modes can be identified by eye. We now are beginning to see the underlying physics for the generalized reaction mechanisms that have been empirically discovered over time. The “magic of chemistry” is this enormous reduction in dimensionality, due to the extremely large anharmonicity in the barrier crossing region, that ultimately makes chemical concepts transferrable. How far can this reductionist view be extended with respect to complexity? In this respect, atomically resolved protein functions provide a definitive test of the collective mode coupling model (Miller Acc. Chem. Research 1994) to bridge chemistry to biology, which will be discussed as the driving force for this work.

Transition metal dichalcogenides (TMDCs) have attracted enormous attention owing, in part, to thickness-dependent band-gap tunability and variation in charge-ordering transition temperatures with number of layers. Accordingly, the transient response of charge carriers to femtosecond photoexcitation in these materials has been vigorously studied, the results of which have provided insights into their fundamental ultrafast photoresponse and electronic relaxation dynamics. In contrast, little attention has been paid to the coinciding structural (i.e., atomic and morphological) response, and essentially no experimental work has been aimed at probing how individual defects affect the observed nanoscale dynamics, especially in relation to the large spot sizes typically employed in pump-probe experiments. Here, I will share some of our recent results on femtosecond electron imaging of defect-modulated structural dynamics in TMDCs. I will begin by describing our experimental setup, which consists of an ultrafast transmission electron microscope capable of reaching sub-picosecond timescales and with which ultrafast electron imaging and diffraction can be performed. I will then describe our results on imaging the lattice response of TMDCs to femtosecond photoexcitation; especially the generation, launch, and dispersion of in-plane and c-axis acoustic phonon modes. Additionally, I will discuss our efforts to understand in detail how energy deposited in the lattice via photoexcitation evolves in space and time via multi-mode excitation, coupling, and decay. Finally, I will conclude by sharing some of our recent observations on photoinduced oscillations of individual metal nanocrystals, with apparent oscillatory frequencies matching those measured with transient absorption through plasmon-frequency modulation.

The real-space observation of acoustic-phonon generation, propagation, and decay in atomically-ordered but morphologically-disordered nanoscale specimens was recently achieved with ultrafast electron microscopy (UEM). Among other insights, this work illustrated that the study such processes on native spatiotemporal scales has the potential to increase understanding of nanoscale thermal-transport phenomena. Moving forward, it will be important to carefully investigate the physical origin of photoinduced phonon generation, as well as the image-forming contrast mechanisms in order to develop quantitative descriptions. Here, we describe the characterization of photo-generated acoustic waves in single-crystal specimens using a host of imaging conditions in bright-field and dark-field modalities. Acoustic phonons in the 20 to 50 GHz range were excited in a germanium specimen with femtosecond laser pulses, the responses of which were systematically studied as a function of specimen orientation, electron-beam conditions, and pump-probe delay in order to determine the contrast mechanisms giving rise to the observed propagating waves. In addition, the nanoscale heterogeneity of the specimen is seen to have an effect on – and in some cases is dominant in determining – the frequency, velocity, wavelength, and spatial localization of the phonons. Here, we also report the observation of nanoscale wave interference, the understanding of which is important to the development of phononic waveguides. The results reported here provide insight into the underlying contrast-forming mechanisms of acoustic-phonon generation and evolution in UEM, which will be important for the development of models relating such experimental results to thermal-energy transport in defect-laden and nanostructured materials.

This talk will be based on Formation of Monatomic Metallic Glasses Through Ultrafast Liquid Quenching, Nature, Vol.512, 177 (2014) by Li Zhong, Jiangwei Wang, Hongwei Sheng, Ze Zhang, and Scott X. Mao. We report an experimental approach to vitrify monatomic metallic liquids by achieving an unprecedented high liquid quenching rate of 10¹⁴ K/s under in-situ transmission electron microscope. Under such a high cooling rate, melts of pure refractory body-centered cubic (bcc) metals, such as liquid tantalum and vanadium, are, for the first time, successfully vitrified to form MGs suitable for property interrogations. With the in-situ TEM observation we investigated the formation condition and the thermal stability of the as-obtained monatomic MGs. The availability of monatomic MGs being the simplest glass formers offers unique possibilities to study the structure and property relationships of glasses. Our technique also exhibits great control over the reversible vitrification-crystallization processes, suggesting its potential in micro-electro-mechanical applications. The ultra-high cooling rate, approaching the highest liquid quenching rate attainable in the experiment, makes it possible to explore the fast kinetics and structural behavior of supercooled metallic liquids within the nano- to pico-second regimes.

Environmental transmission electron microscopes (ETEM) and TEM holders with windowed reaction cells, both now commercially available, enable in situ measurements of the dynamic changes occurring during gas-solid and/or liquid-solid interaction. The combination of atomic-resolution images and high spatial and energy resolution has successfully revealed the nucleation and growth mechanisms for nanoparticles, nanowires, carbon nanotubes and the functioning of catalyst nanoparticles. While TEM-based techniques are ideally suited to distinguish between active and inactive catalyst particles and identify active surfaces for gas adsorption, we still must answer the following questions: (1) are our observations, made from an area a few hundred nanometers in extent, sufficiently representative to determine the mechanism for a particular reaction? (2) Is the reaction initiated by the incident electron beam? (3) Can we determine the sample temperature accurately enough to extract quantitative kinetic information? And (4), can we find efficient ways to make atomic-scale measurements from the thousands of images collected using a high-speed camera. The lack of global information available from TEM measurements is generally compensated for by using other, ensemble measurement techniques such as x-ray or neutron diffraction, x-ray photoelectron spectroscopy, infrared spectroscopy, Raman spectroscopy etc. However, it is almost impossible to create identical experimental conditions in two separate instruments to make measurements that can be directly compared. Therefore, we have designed and built a unique platform that allows us to concurrently measure atomic-scale and micro-scale changes occurring in samples subjected to identical reactive environmental conditions by incorporating a Raman Spectrometer into the ESTEM. We have used this correlative microscopy platform i) to measure the temperature from a 60 µm² area using Raman shifts, ii) to investigate light/matter interactions in plasmonic particles iii) to act as a heating source, iii) to perform concurrent optical and electron spectroscopy such as cathodoluminescence, electron energy-loss spectroscopy (EELS) and Raman. We have developed an automatic image processing scheme to measure atomic positions, within 0.015 nm uncertainty, from high-resolution images, to follow dynamic structural changes using a combination of algorithms publically available and developed at NIST. This method has been proven to capture the crystal structure fluctuations in a catalyst nanoparticle during growth of single-walled carbon nanotube (SWCNT). Details of the design, function, and capabilities of the optical spectrum collection platform and image processing scheme will be illustrated with results obtained during in situ measurements.

With the development of semiconductor technology, the 10 nm feature size is approaching. It is thus quite essential to explore more precise nanofabrication and characterization method to evaluate the shape/structure stability and possible new properties of sub-10-nm material components, especially under external field such as strain, electric, or thermal fields. Here we review our recent progress in atomic resolution nanofabrication and dynamic characterization of individual nanostructures and nanodevices based on the idea of "setting up a nanolab inside a transmission electron microscope". The electron beam can be used as a tool to induce nanofabrication on the atomic scale. Additional probes from a special-designed holder provide the possibility to further manipulate and measure the electric/mechanical properties of the nanostructures in the small specimen chamber of a TEM. Recently, the optical signal also was introduced into the electron microscope to enrich the coverage of investigation inside the “multifunctional nanolab”. All phenomena from the in-situ experiments can be recorded in real time with atomic resolution.

References:
[1] L. Sun, F. Banhart, et. al., Science 312, 1199 (2006)
[2] J. R-Manzo, M. Terrones, et.al., Nature Nanotechnology 2, 307 (2007)
[3] X. Liu, T. Xu, et al., Nature Communications 4, 1776 (2013)
[4] H. Qiu, T. Xu, et al., Nature Communications 4, 2642 (2013)
[5] X. Li, X. Pan, et al., Nature Communications 5, 3688 (2014)
[6] X. Guo, G. Fang, et al., Science 344, 616 (2014)
[7] J. Sun, L. He, et al., Nature Materials 13, 1007 (2014)
[8] W. Zhou, K. Yin, et al., Nature 528, E1 (2015)

Low-voltage aberration-corrected electron microscopy has enabled the the direct imaging of the exact atomic structure in atomically thin materials made of light elements, such as graphene, hexagonal boron nitride, or molybdenum disulfide. Using scanning transmission electron microscopy (STEM), we have studied 2D materials that have been treated for the generation of defects, synthesized in amorphous form, or decorated with molecules. We have also analyzed free-standing heterostructures of single-layer graphene and single-layer hexagonal boron nitride, where in absence of a substrate the interaction between the layers leads to periodic out of plane deformations. Moreover, we observed and quantified the dynamics of point defects under irradiation. Since in-situ experiments for atomic-level manipulation of these surface-only materials inevitably involve open bonds and reactive sites that can trap mobile adsorbates or residual gas from the vacuum, we have recently set up an aberration-corrected low-voltage STEM system where not only the column is built in ultra-high vacuum (UHV) technology, but also integrated with UHV sample preparation and manipulation and sample transfer in vacuum. Initial results from this system will also be shown.

We present a scanning transmission electron diffraction technique capable of mapping the size and distribution of nanoscale crystallites in beam-sensitive soft materials at a resolution on the order of 20 nm. In addition, the method is capable of mapping the local orientation of the crystalline regions and spatially resolving the local degree of crystallinity of the material investigated. Initial studies were performed on a model polymer blend, which is a 50:50 w/w mixture of semicrystalline poly(3-hexylthiophene-2,5-diyl) (P3HT) and amorphous polystyrene (PS). The sample is scanned and a diffraction pattern is acquired for each probe location. An automated algorithm filters, aligns and analyzes the diffraction patterns and an orientation map is composed from the relative orientations of the spots. The technique successfully provides the means to reconstruct dark field images and crystal orientation maps from a single scan, mitigating the effects of beam damage on sensitive specimens. Further studies of polymer thin films have utilized high speed direct electron detectors (Gatan K2) to increase the spatial resolution of the technique through both speed of acquisition and sensitivity of low signal/noise diffraction patterns in soft materials. Our presentation will address the factors contributing to the spatial resolution of the technique, including beam damage, acquisition speed and data analysis.

Electron microscopy of self assembled block polymers is a real space method for determination of the periodic structures and is complimentary to small angle x-ray scattering that uses information in reciprocal space to determine the sample morphology. Transmission electron microscopy (TEM) images are 2D projections of the specimen along the incident beam direction. When the sample is comprised of simple 1D or 2D periodic microdomain structures such as lamellae and cylinders, image interpretation is straightforward. For more complex 3D assemblies, the analysis becomes more involved due to overlap of features in the projections. In the past years, 3D tomographic reconstructions from a tilt series have be used to create 3D models of complex structures such as the gyroid network. However, depending on the particular approach, a missing wedge of information or, if two orthogonal tilt axes are used, a missing cone of information, limits the fidelity of the reconstruction. A symmetry based approach using multiple axes of tilt starting with high symmetry projections can improve the resolution and fidelity of the reconstruction. An alternative approach is to sequentially ion mill a series of thin slices of the sample using FIB while employing SEM imaging between each slice to create a deck of images for the reconstruction. In order for this technique to be successful, the specimen needs to have strong contrast between a low emitter and a high secondary electron emitter, such as occurs for polystyrene-polydimethyl siloxane diblocks or for polystyrene-polyisoprene diblocks for which the unsaturated block has been stained with a heavy Z stain such as OsO₄. Of particular interest are the chiral microdomain morphologies such as hexagonally packed twisted cylinders (H* phase) and the single (I432) or double gyroid (Ia3d) cubic network structures. We report reconstructions of the detailed geometry interface separating the two types of domain.

Biologists have long strived to link macromolecular structure to physiological
function. Of particular interest is the possibility to visualize dynamics of whole cells,
soluble protein complexes and membrane proteins in near-native environments without
freezing and with nanoscale resolution. As liquid cell experiments in transmission
electron microscopy (TEM) have increased in popularity, several bottlenecks have been
discovered that limit biological in-situ TEM experimental capabilities and repeatability.
In this talk I will discuss the design, current progress and applications of 2 new platforms
that enable directed flow, controlled chemical gradients and mixing, low-dose focusing
and imaging, and greater sampling efficiency. I will highlight results from both
soft/biological samples and materials/chemistry systems to demonstrate the versatility of
the platforms for enhancing the usability of in-situ holders while simplifying
interpretation and increasing repeatability of liquid cell results. Finally, future
applications for coupling these platforms with Dynamic TEM and other pump-probe
instrumentation that will push the temporal resolution for in-situ bioimaging into the tens
of microsecond and faster timescales will also be described.

Crystal twinning can take place without any crystal deformation by external shear stress, as extensively studied in various metals and alloys (annealing twins) and also as frequently observed in many oxide minerals during crystal growth (growth twins). Because these twins are two-dimensional planar interfaces, quadruple junctions form when phase boundaries intersect the twinning planes during phase separation. A key aspect that should be taken into account at the quadruple junctions is that four interfacial tensions from the phase boundaries and the twin interfaces should satisfy their force equilibrium. Therefore, a complex correlation between the force balance and the coherency elastic strain is anticipated when the lattice coherency is maintained at the quadruple junctions. By taking olivine-type LiFePO₄, which is used for cathodes in Li-ion batteries, as a model crystalline system, we show that a comprehensively different phase-separation morphology is induced in order to release the high coherency strain confined to quadruple junctions, at which twin boundaries and phase boundaries intersect with each other. High-temperature in situ transmission electron microscopy (S.-Y. Chung et al, Nature Phys. 5, 68 (2009); Nano Lett. 12, 3068 (2012); J. Am. Chem. Soc. 135, 7811 (2013); ACS Nano 9, 327 (2015)) revealed that phase boundaries with a new crystallographic orientation emerged over twinned crystals to provide strain relaxation at quadruple junctions in contrast to the phase separation in twin-free crystals. The high coherency strain and the formation of different phase boundaries can be understood in terms of the force equilibrium between interface tensions at a junction point (S.-Y. Chung et al, Nature Commun. 6, 8252 (2015)). Visualizing the quadruple points at atomic resolution in a crystalline system, our experimental observations emphasize the impact of nanoscale multiple junctions on the morphology evolution during phase separation.

Chalcogenide-based phase change materials (PCMs) are a class of materials that may be rapidly switched between amorphous and crystalline phases by laser or Joule heating. Their distinct optical and electrical properties make them useful for optical and resistivity-based memory devices. For memory applications, crystallization of the amorphous phase is the data-rate-limiting step and must be achieved within nanoseconds. Crystallization kinetics impact switching speed, but separate nucleation and crystal growth rates are difficult to measure experimentally during highly-driven, high-temperature laser- or current-induced crystallization. Nucleation and growth rates during laser heating may be orders of magnitude greater than what may be characterized with conventional microscopy techniques (e.g. thermionic- or field-emission TEM, AFM, optical microscopy).

This work describes the use of multi-frame dynamic transmission electron microscopy (DTEM), a photo-emission TEM technique with nanosecond-scale time resolution to study the crystallization kinetics of Sb-rich PCMs during laser heating. DTEM is distinguished from other photo-emission TEM techniques in that each electron pulse contains enough electrons to form an image or diffraction pattern, so it is suitable for studying irreversible phase transformations. The DTEM at Lawrence Livermore National Laboratory generates multiple electron pulses spaced over several microseconds. An electrostatic deflector directs each pulsed image to a different part of the CCD camera, overcoming the refresh rate of the camera. Up to nine frames may be captured during a laser-induced transformation. Multi-frame DTEM has been used to study crystallization kinetics in the PCM GeTe [1].

GeSb₆Te has only recently been considered for use as a PCM, but it shows great potential for application in solid-state memory devices because of its high crystallization speed [2]. Multi-frame DTEM experiments show the nucleation rate in GeSb₆Te is very low relative to other PCMs, but the growth rates are very high – up to 10.8 m/s for amorphous as-deposited films and far higher for an amorphous film subject to sub-threshold laser annealing before crystallization [3]. The crystallization is reminiscent of explosive crystallization of elemental semiconductors both in the magnitude of the growth rate and in the resulting crystalline microstructures [4]. The in-situ imaging also reveals a connection between a sudden drop in crystal growth rate and the resulting microstructure, which suggest a change in growth mechanism with temperature or temperature gradient.

References
[1] M. K. Santala et al., Appl Phys Lett 102 (2013) 174105.
[2] H. Y. Cheng et al., J Appl Phys, 115 (2014) 5.
[3] M. M.Winseck et al., Dalton Trans 45 (2016) 9988.
[4] M. K. Santala et al., Appl Phys Lett 102 (2015) 174105.

Measuring the dynamic motion of ultrafast vibrations in nanostructures is imperative to understanding and tuning mechanical, electrical, and transport properties of new nanomaterials. Ultrafast electron microscopy (UEM) and dynamic transmission electron microscopy (DTEM) have enabled real-space and reciprocal-space observations of nanostructure motion on nanosecond to femptosecond time scales. However, these techniques involve complex laser illumination schemes coupled with complicated electron optical systems. Here we demonstrate a new electron beam detection method for measuring 1D vibrations of acoustically excited nanostructures. We developed a new electron beam-based technique that utilizes a stationary, focused electron beam to detect 1D nanostructure motion via vibration-induced changes in the sample mass-thickness. The electron beam detection method was implemented in a scanning electron microscope (SEM) and proof of principle experiments were performed by detecting 1D vibrations of an atomic force microscopy (AFM) cantilever. We positioned the last 100 – 200 nm of a piezoelectric driven AFM tip vibrating in free space perpendicular to the propagation direction of a stationary ~3 nm diameter electron probe. The wedge shape of the vibrating AFM tip caused an oscillatory change in the sample thickness encountered by the stationary electron beam, which produced an oscillation in the amount of electron scattering detected by an annular scintillator detector positioned below the AFM tip. By only detecting the signal resulting from electron scattering at a single point, we relax the restraints of conventional SEM for imaging fast processes, i.e. the slow serial imaging process. We show that the detection sensitivity increases with decreasing absolute sample thickness due to the inverse exponential sample thickness scaling for electron scattering into an annulus. The technique is currently capable of detecting sub-nanometer amplitude vibrations at frequencies up to ~1 MHz. While we only detect vibrations on microsecond time scales here, the detection speed is only limited by the bandwidth of the scintillator detector; nanosecond detection will be possible with faster scintillating detectors and amplification electronics. We expect this technique will find abundant applications in measuring angstrom-scale 1D ultrafast vibrations in nanostructures induced by other external stimuli, such as heating and electrical biasing.

Electrical wind force is an important element in electrical switching behaviors of phase-change materials. It has been reported that the electrical wind force influences the motions of dislocations, which determines the degree of order-disorder states existing in phase-change materials. In this presentation, we discuss electrical wind force-driven behaviors occurring in phase-change materials. In the first part of the presentation, we report atomic mass-transport behaviors as DC voltage biases are applied in line-shape Ge2Sb2Te5 (GST) devices. As the electrical current density reached 3-4 MA/cm2 by DC voltage bias, a directional mass transport was identified by forming asymmetric surface morphology on the line-shape GST devices, such that hillock structures are created in the middle of the line whereas void structures are formed at the entrance of (+) electrode side of the line-shape device. The mass transport of GST occurs as a solid-state behavior by electrical wind force that leads physical scattering between electrons and atoms of GST compound, followed by momentum transfer. In the second part, we extend the understanding of the roles of electrical wind force to electrical switching behaviors of GST. We studied the effects of electric voltage pulses on crystalline-to-amorphous phase transition of GST by in situ transmission electron microscopy (TEM). Electrical voltage pulse plays a critical role by creating dislocations through heat shock process: Rising edge of the pulse produces vacancies by heating, whereas during rapid cooling, atomic vacancies are condensed into dislocation loops. As the dislocations feel the electrical wind force, they become mobile and glide in the direction of hole-carrier motion. Continuous increase in the density of dislocations moving unidirectionally leads to dislocation jamming, which eventually induces the crystalline-to-amorphous phase transition. We interpret it through one-dimensional traffic model in which the increase of dislocation density exceeding a certain threshold point induces a catastrophic jamming of dislocations. Our understanding about dislocation induced amorphization suggests that the transition from crystalline to amorphous in phase-change materials may not require a melting process.

Atomically thin 2D materials such as graphene and the transition metal dichalcogenides (TMDs) show a host of promising opto-electronic properties, and could very well pave the future for next generation of nano-devices. Tuning the properties of interest by controlling the distribution and density of defects, is an exciting opportunity to adapt the functionality of the 2D material, much like what is being in the semiconductor industry. In this work, we explore the effects of focused helium ion beam irradiation on the structural, optical and electrical properties of 2D materials, via high resolution scanning transmission electron microscopy, (STEM) Helium Ion Microscopy, (HIM) Scanning Probe Microscopy, (SPM) and Raman mapping. By imaging the material before the He ion irradiation, we establish the baseline atomic configuration of the surface. We then expose the same area of the sample in the HIM, to introduce defects locally with high precision, thereby locally tuning the resistivity and transport properties of the material. By utilizing local crystallography analysis through a High Performance Computing (HPC) system delivered by Bellerophon Environment for Analysis of Materials, we can quantitatively extract the density and the type of the defects induced by the helium ion beam. This information, in conjunction with Density Functional Theory calculations, SPM measurements, as well as Raman mapping; allows us to begin to unravel the effect of individual defects on the macroscopic properties of 2D materials.

Acknowledgements
Research supported by: Oak Ridge National Laboratory’s Center for Nanophase Materials Sciences (CNMS), which is sponsored by the Scientific User Facilities Division, Office of Basic Energy Sciences, U.S. Department of Energy. This study used the resources of the Oak Ridge Leadership Computing Facility at the Oak Ridge National Laboratory, and analysis under support of applied mathematics program at the DOE.

Developments in scanning transmission electron microscopy (STEM) have opened up new possibilities for time-resolved imaging at the atomic scale. Recent examples include a study of the diffusion of dopant atoms in semiconductors [1] and, using environmental STEM, in situ studies of catalytic reactions [2]. Rapid imaging of single atom dynamics brings with it a new set of challenges. High frame rates and long total acquisition times mean novel methods are needed for handling and processing “big data” sets. Further, the need for short exposure times can lead to severe problems with noise.

We are developing novel methods to address the challenges and opportunities posed by rapid time-resolved imaging. By exploiting the spatio-temporal correlations between frames, it is possible to considerably improve the signal-to-noise ratio using singular value thresholding [3]. Crucially, by employing robust procedures to automatically estimate the noise and motion characteristics, it is possible to optimize the process with little user input. The identity and positions of individual atoms in the denoised data can then be determined using a newly-developed intensity-based classification algorithm. The positions in each frame are then linked to form a set of atomic trajectories. Building on the theme of automation, the classifier can be trained using simulated STEM images to process long image sequences, where manual identification and tracking would be prohibitive.

We have applied these methods to investigate the diffusive behaviour of single atoms under the electron beam on both crystalline and disordered substrates. By incorporating prior information gained from experiment and from density functional theory calculations, the classifier is able to robustly identify and track atoms from frame-to-frame. The algorithm then updates its knowledge of the system as new frames are analysed, opening up the possibility for “real-time” tracking of atoms as data is acquired. Development of these methods highlights the potential for combining time-resolved STEM imaging with theory, and together forms a powerful approach for understanding the dynamic behaviour of materials at the atomic scale.

References
[1] Ishikawa R, et al. (2014). Phys. Rev. Lett. 113, 155501.
[2] Gai P, et al. (2014). Chem. Phys. Lett. 592, 355-359.
[3] Furnival T, Leary RK, Midgley PA. (2016). Ultramicroscopy. doi:10.1016/j.ultramic.2016.05.005

Time-resolved stroboscopic imaging in ultrafast electron microscopy (UEM) is based on the repeated stimulation of reversible dynamics using an ultrafast laser system. Because of this, photothermal heat accumulation and dissipation limit the operational frequencies accessible during stroboscopic experimentation due to the potential generation of responses that are long-lived relative to the duration between pulses (e.g., crystallization, melting, etc.). Higher operational frequencies are often desired for UEM experimentation due to the increase in photoelectron current, which ultimately dictates image acquisition time. Understanding the specimen-specific response to laser excitation can provide insight into the spatiotemporal temperature profile which plays a significant role in the design of UEM experiments and the accessible specimen parameter space. Here, we describe the UEM operational limitations of a representative thin film specimen supported on a silicon nitride (SiN) membrane TEM grid and irradiated in situ with a femtosecond laser system. Various TEM techniques were used to characterize the laser-induced crystal nucleation and growth of a SiN/Ta/TbCo/Ta thin-film specimen, an optomagnetic structure that has been shown to exhibit ultrafast all-optical magnetic switching. A geometry-specific laser-heating model that accurately depicts photothermal accumulation and transport was developed to quantify the temperature rise in the representative thin-film specimen, as well as provide a basis for future UEM experiment design and expand on the accessible specimen parameter space. Furthermore, a procedure for determining the in situ laser power at the specimen, relevant for all UEM and in situ laser-excitation work, was developed. Moving forward, work will focus on developing and incorporating photothermal-accumulation models into the study of other specimen geometries (e.g., wedges, nanoparticles, nanorods, etc.) in order to generalize the approach and aid in the design of optimized UEM specimens and experiments such that spatiotemporal resolutions can be increased.

Shu Fen Tan, Utkarsh Anand and Utkur M. Mirsaidov<sup>1,2,3,4

1</sup>Centre for Bioimaging Sciences and Department of Biological Sciences, National University of Singapore, 14 Science Drive 4, Singapore 117543
²Department of Physics, National University of Singapore, 2 Science Drive 3, Singapore, 117551
³NUSNNI-Nanocore, National University of Singapore, 5A Engineering Drive 1, Singapore, 117411
⁴Centre for Advanced 2D Materials and Graphene Research Centre, National University of Singapore, 6 Science Drive 2, Singapore 117546

*corresponding authors, E-mail: [email protected]

The assemblies of nanoparticles exhibit distinctive properties than its isolated forms. Nanoparticles with tunable gap sizes are good candidates for plasmonics [1-2] and catalysis [3] studies. These structures are challenging to fabricate by top-down nanofabrication techniques such as electron-beam lithography (EBL), because of resolution limitations and up-scaling issue. On the other hand, self-assembly serves as a robust ‘bottom-up’ method to align nanoparticles into desired arrays. The assembly process is often characterized by the ‘quench-and-look’ approach or indirect dynamic spectroscopic methods [4]. As such, the intermediate steps and the how the nanoparticle polarization, surface molecules participate in controlling the final assembled structures still remains largely unresolved.

Here, we demonstrate the use of in-situ liquid cell electron microscopy to study the self-assembly of anisotropic nanoparticles; Au nanorods with linker molecules; cysteamine at high temporal and spatial resolution. By simply tuning the concentration of the linker molecules, we observe two different assembly modes of nanorods: end-to-end and side-by-side assemblies as reported elsewhere [5]. We further investigate how these interactions can be used to align Au nanorods into such assemblies to attain the lowest total energy state by monitoring the nanorods dynamics as a function of time. Understanding the physical and chemical interactions that govern the self-assembly could potentially lay the foundation for rational design of desired assembled nanostructures for application in catalysis, opto-electronics and drug delivery.

We acknowledge the support from Ministry of Education of Singapore (award No. MOE2015-T2-1-007).

References
1. S. F. Tan et al., Quantum plasmon resonances controlled by molecular tunnel junctions. Science 343: 1496-1499, 2014.
2. K. J. Savage et al., Revealing the quantum regime in tunnelling plasmonics. Nature 491: 574-577, 2012.
3. G. Prieto et al., Towards stable catalysts by controlling collective properties of supported metal nanoparticles. Nature Mater. 12: 34-39, 2013.
4. M. Grzelczak et al., Shape control in gold nanoparticle synthesis. Chem. Soc. Rev. 37: 1783-1791, 2008.
5. Liu et al., End-to-end and side-by-side assemblies of gold nanorods induced by dithiol poly(ethylene glycol). Appl. Phys. Lett. 104: 253105, 2014.

The strong optical response of the prototypical transition metal dichalcogenide, molybdenum disulfide (MoS₂), makes it of particular interest for nano-scale photodetectors and flexible electronics. To inform the design of such devices, the complete photo-induced structural response – from picosecond (ps) phonon excitation to microsecond mechanical motion – would ideally be characterized on the appropriate spatiotemporal scales. To this end, we discuss here our development and use of femtosecond (fs) and nanosecond (ns) bright-field electron imaging to directly follow the continuous photoexcited structural evolution of a single MoS₂ flake. The photo-induced structural response was studied by mapping the trajectories of low-index diffraction-contrast features (e.g., bend contours and moiré fringes) in real-space, bright-field, ultrafast electron microscopy (UEM) image sequences with time steps of 1 ps, 10 ps, and 10 ns. In the initial few hundred picoseconds, we track the displacement of a single diffraction contrast feature as a function of time and find that the fast Fourier transform of this oscillation yields a single frequency of 50 GHz, in accord with the excitation of an acoustic-phonon mode. As reflected in the UEM diffraction-contrast response, these high-frequency, acoustic-phonon modes quickly scatter and interfere with one another before damping over the first 200 ps and relaxing into lower frequency (~1 GHz) oscillations. Via nanosecond UEM imaging, we find that these 1-GHz oscillations are a precursor to still lower-frequency, whole-flake mechanical oscillations. Within this regime, three distinct vibrational modes, with frequencies of a few to several MHz and with (1/e) lifetimes of tens of microseconds, are observed. This work demonstrates the ability of UEM to study complete ultrafast processes over multiple timescales and provides insight into the potential to incorporate the photoinduced structural response of few-layer MoS₂ into nano-scale, ultrafast applications.

The coating layer of galvannealed steels, which are widely used in applications in automotive and building industries, consists of a lamellar series of intermetallic compounds in the Fe-Zn system; Γ (Fe₃Zn₁₀), Γ₁ (Fe₅Zn₂₁), δ_1k (FeZn₇), δ_1p (Fe₁₃Zn₁₂₆) and ζ (FeZn₁₃). The crystal structures of all the intermetallic compounds are complex and consist of Fe-centered Zn₁₂ icosahedra, which share their faces or vertices with one another. Because deformation properties of these intermetallic compounds determine the press formability response of the galvannealed steels, we have been investigating deformation behavior of each intermetallic phase through compression tests of polycrystalline/single-crystal micropillar specimens. Recently, we have recently revealed that the Γ, δ_1p and ζ phases exhibit compression plasticity in which slip systems are selected so as not to destroy the Fe-centered Zn₁₂ icosahedra. This implies that the Fe-centered Zn₁₂ icosahedra might behave as if they were large-sized atoms during the plastic deformation. In the present study, we have observed the atomic arrangement of the cores of dislocations and stacking faults bounded by partial dislocations in some of the deformed Fe-Zn intermetallic compounds with an aberration-corrected scanning transmission electron microscope, and found that the Fe-centered Zn₁₂ icosahedra indeed seem to behave as if they were large-sized atoms during the plastic deformation.

Electromigration and its effects are widely studied due to their relevance to the semiconductor device industry. In-situ TEM allows investigation into failure mechanisms with high temporal and spatial resolution, as well as control over key parameters including temperature and current density. However, these experiments are complicated by the thermal gradients which result from Joule heating of the metal lines; high current densities in the test samples and ultra-low thermal conduction through the electron transparent silicon nitride support films contribute to relatively large increases in temperature. We have performed finite element modeling to calculate the resulting temperature distributions due to resistive heating of electromigration test structures so that observations can more accurately related to the local temperature.
Calculations have been performed using a range of patterns for metallic test structures, different window geometries, and heat sink structures. Heat generation was volumetric and determined by the current density and temperature dependent resistance of the metal, and loss was a combination of radiative transfer and conduction through the edge of the sample to the TEM holder. The results of these calculations indicate improvements which can be made in sample design to minimize the temperature rise due to Joule heating. Titanium is used as the resistive heating element in these experiments – its high melting point allows for the generation of large thermal gradients, and the phase change from HCP to BCC at 1156K is one indicator of local temperature.
We will also experimentally determine the temperature gradients by observation in-situ of the phase transition upon melting of several transition metals deposited in ~50 nm islands on the reverse side of the support film. Through this study we aim to be able to accurately determine the temperature gradients which result from Joule heating of in-situ electromigration test structures which will allow for the correlation of in-situ observations and a spatially resolved temperature, leading to more quantitative experiments being possible in these systems.
This work was supported by the NYSTAR Focus Center at RPI, C130117, and made extensive use of cleanroom and characterization facilities in the Center for Materials, Devices and integrated Systems (cMDIS) at RPI.

Metal nanocrystals have unique optical, electrochemical, magnetic, and catalytic properties that are strongly dependent on the composition and the crystal size, shape and orientation. Electrochemical deposition is an efficient method for the growth of nanocrystals in liquid solution and synthesis techniques can involve cyclic voltammetry, potential step deposition, or current electrodeposition. However, electrochemical methods for producing nanocrystals in solution are often developed by exploring a wide range of electrochemical driving conditions, chemical concentrations, and additives until a suitable composition and morphology are obtained. In situ liquid cell transmission electron microscopy enables imaging of the crystals as they form. This has allowed systematic exploration of the underlying effects of growth parameters on the crystallization process in liquid solutions.

Here we use liquid cell TEM to establish a quantitative assessment and mapping of the morphology and size distribution of metal nanocrystals formed electrochemically, as a function of liquid environment, in particular the electrolyte layer thickness and the local presence of a metal electrode. The experiments were carried out in a FEI CM30 TEM, operated at 300 kV, using a continuous flow LC-TEM holder. Metal films 20 - 50 nm deposited by electron beam evaporation were patterned into electrodes that partially cover the 50 nm thick silicon nitride membrane that makes up one window of the liquid cell. The cell was filled with an aqueous electrolyte containing either H₂SO₄ or HCl, but with no metal ions present initially. Instead, the metal ions were supplied by applying a voltage between two electrodes that was sufficient to cause corrosion of the counter electrode. We will show movies of this corrosion process as it occurs under a potential scan, correlating the film morphology with voltage and current. After the corrosion process, we find that metal nanocrystals have formed from the released ions, nucleating at locations between the two electrodes. We will present images and movies of the formation of the nanocrystals, particularly under conditions of pulsed current. This allows us to discuss the formation mechanism of these nanocrystals and explore the local crystallization kinetics - the formation rate, morphology, and composition of these electrogenerated nanocrystals as a function of current or voltage, electrolyte composition, electrolyte layer thickness, and spatial configurations around the electrodes - to obtain insights into the processes involved. We show that it is possible to obtain a variety of crystal morphologies by varying the experimental conditions. The experimental approach presented here may be applicable to developing nanocrystal synthesis, as well as making more general measurements of local parameters under an electrochemical stimulus.

Over the past several years there has been a revolution in the spatial resolution achieved in electron microscopy, particularly in the Scanning Transmission Electron Microscopy technique when using a High Angle Annular Dark Field Detector (STEM-HAADF technique). During this time, many materials scientists evolved from relying on High Resolution TEM images to interpreting STEM-HAADF images. While these two techniques provide similar information on a materials crystalline structure, the STEM-HAADF technique arguably offers an order of magnitude more information. As a result, there still remains a question of how to adequately interpret STEM-HAADF images in terms of assigning crystalline structure, identifying atomic defects, or isolating strain, to name a few. We encountered these challenges in our own research when using a Jeol JEM ARM-200F STEM-HAADF (resolution of 0.8Å) instrument to study the atomic resolution images of both noble metal and chalcogenide type nanoparticles. The presentation will discuss our results and findings when using this powerful electron microscope to image nanoparticle samples such as Ag@FeCo@Ag (core@shell@shell) particles, Au@Ag@Au particles, Cu₂S/ZnS Janus structured particles, and other samples. The results will be discussed in terms of the fundamental information observed in the images, how we interpreted the two dimensional images of a three dimensional structure, and our rational for how we can utilize the STEM-HAADF technique to gain new insight into the hidden mysteries of nano-structures in terms of atomic defects, strain, and other phenomena.

In-situ transmission electron microscopy (TEM), Electron Tomography (ET) and Scanning TEM (STEM) diffraction imaging have been some of the most common experiments in the electron microscopy society for several decades. In this work we will present new trends and approaches to these conventional experiments that are now possible using the high speed complementary metal-oxide semiconductor (CMOS) cameras introduced in the recent years.
With the rapid development in nanoscience and technology, more interest has been drawn to in-situ TEM. In most cases reactions being observed at the nanoscale occur at much higher speeds than the charge coupled device (CCD) cameras (with up to 30 frame per second (FPS)) could capture. This limitation was one of the driving forces for development of high speed CMOS cameras that can record data with sub millisecond time resolution. Here experimental results in different materials systems that are uniquely enabled by these high speed cameras will be presented.
Conventional ET consists of acquisition of a series of specimen images at different viewing angles. This is a labour intensive and time consuming process and the final resolution of the 3D reconstruction is directly affected by the number of images, hence tilt range and tilt increment of the experiment. Here application of high speed and high DQE cameras for continuous tomography data acquisition, while the microscope goniometer is continuously tilting will be reported. This approach reduces the total data collection duration from tens of minutes to just a few minutes, and it improves the resolution of the 3D reconstructed volume as a result of reduction of the tilt angle increment to a small fraction of a degree.
In STEM diffraction imaging a parallel or convergent electron beam is scanned on the specimen surface, while an electron diffraction pattern is recorded at each beam position. Such datasets are used to characterize individual nano particles, defects and interfaces and allow accurate measurements of strain and crystal orientation. The frame rate limitation described above causes more challenges in experiments where specimens are beam sensitive and/or drift. We will present STEM diffraction datasets collected with conventional CCD and high speed CMOS cameras and will show how application of CMOS imaging technology can improve the results in such experiments.

Direct observation of coherent collective lattice oscillations known as phonons has remained experimentally challenging owing to their ultrashort native spatiotemporal scales – that is, the combined nanometer (nm), spatial and femtosecond (fs), temporal scales. While advances in ultrafast X-ray and electron probes have been instrumental in revealing the atomic origin of myriad non-equilibrium, photo-excited phenomena, the majority of such studies are conducted in reciprocal space and are subject to ensemble averaging over spot-sizes (and corresponding material volumes) much larger than typical phonon wavelengths. Here, we employ ultrafast electron microscopy (UEM) to image in real-space the evolution, propagation, and scattering of discrete acoustic-phonon wavefronts emanating from crystal defects and material discontinuities in 2H-WSe₂ and 1T-TaS₂ thin films. Probing electron scattering on the image plane, we take advantage of phonon-induced modulations of the local Bragg condition to reveal bright-field (BF) contrast features corresponding to distinct morphologically-dependent phonon modes. From fs time-resolved image scans on a few-layer WSe₂ flake, we extract apparent phonon frequencies and phase velocities of 40 GHz and 5.5 nm/ps, respectively, for a mode evolving in the initial moments after photoexcitation and 3 GHz and 0.9 nm/ps for a large-amplitude mode dominating after a few-hundred picoseconds. Additionally, we obtain auxiliary atomic-scale information from time-resolved selected-area diffraction, from which the resonant behavior is attributed to beating of a c-axis longitudinal acoustic mode. From the ancillary techniques, we piece together the physical mechanisms of photoinduced phonon excitation and the observed BF-imaging contrast. It is expected that these direct manifestations of local elastic properties in the vicinity of material defects and interfaces will aid in the understanding and application of phonon-mediated phenomena in nanostructures.

In situ experiments in the TEM can provide a valuable insight into thermally activated phenomena occurring in materials over a wide range of temperatures. The meaningful interpretation of the data from such studies requires that the temperature control and measurements be accurately calibrated. Advanced micro electro-mechanical system (MEMS) based heating holders (e.g. [1]) exhibit appreciably reduced thermal drift and temperature lag compared to traditional heating holders. These advantages arise from the small sample volume that is heated and the close proximity of the heating element to the sample on the MEMS chip. These MEMS-based heating holders can, therefore, be used for more precisely controlled heating and cooling experiments. However, despite the superior thermal characteristics of these MEMS-based holders, the issue of temperature calibration still remains a significant challenge.
In a recent in situ TEM study the sublimation kinetics of Ag nanoparticles (NP) with a wide particle size distribution were investigated [2]. The observations were interpreted on the basis of the Kelvin equation. This study revealed that the size-dependent sublimation behavior of Ag NPs could be used to identify the temperature of the membrane in contact with the Ag NPs. However, issues such as electron beam heating could potentially result in an overestimation of the temperature, due to poor thermal transfer between the particle and the membrane.
In our work we have studied the sublimation behavior of poly-vinyl pyrolidone (PVP)-capped Ag nanocubes (NCs) with a monomodal size distribution [3]. This particle morphology gives better heat-transfer to the support film, and thus less thermal lag and electron beam heating effects. Moreover, the monomodal size distribution gives a well-defined characteristic sublimation temperature [4]. Thus these observations could enable us to calibrate the temperature in our MEMS based heating stage (Nano Ex- i/V, FEI) accurately, albeit indirectly. The samples were examined in a 200kV TEM/STEM (Talos, F200X, FEI) with a Schottky field emitter. The morphological evolution of single Ag-NCs at different locations on the heating membrane was tracked at specific temperature and time intervals to estimate its sublimation temperature. The role of PVP cap in the electron beam heating effects in these Ag NCs is also investigated.
[1] L. Mele et al., Res. & Technique, Vol. 79, No. 4, 2016, p 239
[2] M.A. Asoro, D. Kovar, & P. J. Ferreira, ACS Nano, Vol. 7, No. 9, 2013, p 7844
[3] A.Tao, P.Sinsermsuksakul, & P.Yang, 45, 2006, p 4597
[4] Y. Ding, F.Fan, Z. Tian, & Z. L. Wang, Small, Vol. 5, No.24, 2009, p 2812
[5] This work was supported in part by a research grant from FEI Company under an FEI-UConn partnership agreement and by the award of an FEI Graduate Fellowship to Sravan Thota. The studies were performed in the UConn/FEI Center for Advanced Microscopy and Materials Analysis (CAMMA).

The scanning transmission electron microscopy (STEM) is a powerful platform for studying materials and their many varied properties (including structural, electronic, magnetic, ferroic etc.) locally and at their fundamental length-scales. In its current standard implementation, STEM imaging is severely restricted by the instrumental linkage in which the information rich electron diffraction pattern (Ronchigram) is reduced to a single value (e.g. through the integrated intensity over a relatively large detector) at each beam location resulting in loss and distortion of information about important but subtle aspects of material properties. Previous theoretical and experimental work suggests that full acquisition of the Ronchigram at each spatial location during a scan can enable super resolution, phase-contrast imaging, imaging of internal fields, and 3D sample reconstruction. Data acquisition and storage has evolved to a point now that it is possible to capture high resolution 4D ptychography data sets rapidly, and the use of large-scale compute facilities enables the processing and mining of these multi-GB data sets to distil the most salient aspects of the data while separating the statistically significant variations (signal) from noise.

Typical scans of 192×192 pixel images containing 384×384 Ronchigrams at every pixel were captured in less than 1 minute. A host of multivariate statistical methods can be brought to bear to extract meaningful information from these large data sets. Statistical methods can be advantageous in that they are parallelizable and can be efficiently implemented on large-scale computing platforms, are model-free and operate with no pre-imposed expectation or bias, and are capable of elucidating inter-pixel correlations unlikely to be noticed by human observation alone.

Though PCA can be used to reduce the dimensionality of information within a data set, it may not always be the most efficient means to compress the and the output of PCA it is often not well suited for finding clusters that exist in subspaces of the data. In order to simultaneously address the difficulty of defining and finding neighbors in extremely high dimensions and the desire to find clusters that are locally defined, we utilize an approach that creates local, patch-based Multidimensional Spectral Hashing (MDSH) codes for dimensionality reduction. MDSH is used to generate small set of dimensions in which similarity among close points is preserved, while allowing distinctions between larger distances to become indistinguishable. This enables us to create a space with many fewer dimensions in which local clustering is preserved. Here, we will explore the use of high-dimensional techniques that automatically select an appropriate number of clusters, to visualize domains in Bismuth Ferrite (BFO), as well as explore how variations in the stride and receptive field of the patches impact the imaging results.

Scanning transmission electron microscopy (STEM) has been a versatile tool for materials characterization with sub-angstrom spatial resolution and sub-second temporal resolution but can be prone to scan and drift distortion during image acquisition. Here, we introduce a new method that enables accurate control of the STEM probe position and scan velocity which minimizes scan distortion at fast scan rates. We show that for any scan path, the actual location of the electron beam always deviates from the assigned beam location as a result of phase lag and inductance, which is related to the beam velocity. Using Archimedean spiral scans with constant angular velocity, we deciphered the characteristic response of the distortion as a function of angular velocity using an aberration-corrected STEM. We demonstrate that our beam control method combined with optimized scan paths can yield less distorted atomic resolution STEM images compared to raster scan for fast frame acquisition. We further present that by accurately controlling the beam path, the beam can selectively interact with designated atomic columns, which enables selective spectroscopy and electron beam induced atomic resolution patterning.
Research supported by Oak Ridge National Laboratory’s (ORNL) Center for Nanophase Materials Sciences (CNMS), which is a U.S. Department of Energy (DOE), Office of Science User Facility (XS, RRU, SK, SJ), by the Division of Materials Sciences and Engineering, Office of Basic Energy Sciences, DOE (ARL and AYB), by ORNL’s Laboratory Directed Research and Development Program, which is managed by UT-Battelle LLC for the U.S. DOE (SJ), and by the Office of Advanced Scientific Computing Research, Applied Mathematics program under the ACUMEN project.(EL and RKA).

The study of capillary-driven phenomena involving metals has vital and wide-ranging technological importance in areas as diverse as sintering, soldering, or general thin-film fabrication. However, capturing these processes in real time is extremely challenging. Such experiments usually involve, and often require, elevated temperatures and strict control of the environment and for a deeper understanding; theoretical modelling has to often take into account the additional complexity of phenomena such as adsorption, diffusion or chemical reactions. The focus of this study has been to observe, in-situ, the interactions between different metallic nanoparticles. This is usually difficult to achieve because the solution-based processes used to synthesize the nanoparticles often require surfactants that cap the surfaces and therefore hamper direct contact. To overcome this we used a modified inert gas aggregation nanoparticle deposition system to deposit silver and gold nanoparticles directly onto a TEM grid. This approach allows the deposition of individual particles, with nominally “clean” surfaces whilst avoiding agglomeration by growing sub-monolayer coverage. The interactions between the particles can be observed in-situ in a high-resolution transmission electron microscope without the need for a “cleaning” step (by either thermal annealing or chemical treatment).
We found that under the beam of the electron microscope these deposited nanoparticles exert attractive forces on each other under vacuum and coalesce upon contact. However, coalescence of two particles with same composition may not be driven by surface diffusion as previously proposed but rather by evaporation-condensation. In contrast when two dissimilar particles meet the silver wets and engulfs the gold. This is due to the unique conditions of our experiments whereby the use of nanoparticles results in relatively high atomic mobility at room temperature whilst volume diffusion remains very slow. As a consequence, it is possible to form a metastable gold-silver interface with a very low associated interfacial energy that drives spreading. It is also important to note that despite the fact that we are manipulating and observing phenomena at the nanoscale (< 10 nm),we describe this using existing concepts derived from continuous analysis. These results provide new fundamental information on the interaction between metal nanoparticles and have critical technological implications. For example, when it comes to the sintering of novel conductive inks and solders our observations suggest that while the combination of two species can promote liquid state sintering and densification (even if they should inter-diffuse in equilibrium) using a single material may result in structural coarsening.

Although transmission electron microscopy (TEM) is an integral tool for materials characterization at high spatial resolution, it is important to account for the interaction of the high energy electrons with the sample. In liquid cell TEM, the sample typically consists of a thin aqueous layer encapsulated between transparent membranes made of silicon nitride or other materials. Such samples have attracted extensive recent interest as they enable the imaging of nanoparticle growth and dynamics, electrochemical reactions and biological materials. However, strong interactions are visible when the electron beam irradiates the water layer. As the beam deposits energy into the water, radiolysis is known to take place, creating reactive species that alter the solution chemistry. Calculations of radiolysis effects for liquid cell TEM of pure water have already quantified the effects that are expected, in particular changes in the pH and the formation of hydrated electrons and hydrogen bubbles. These calculations are helpful in interpretation of liquid cell TEM data, but rely on simplifying assumptions. To calculate energy absorption, the liquid thickness is taken to be small, which is valid only if the thickness of the cell is of the same order as the mean free path of the electrons. Furthermore, the cell windows and interfaces are not included in the calculations. But in a realistic geometry, interfaces can play a key role: for example, the interface between the solution and the silicon nitride windows is a site for nanocrystal nucleation, and the interface between an electrolyte and the metallic surface of an electrode is the site for deposition reactions or gas bubble formation in an electrochemical cell under bias. Metallic electrodes, and to some extent the silicon nitride windows, are strongly scattering materials that generate secondary electrons when the primary beam deposits energy. This results in a higher dose rate near the interfaces - in particular within the liquid that is a few nanometers from the interface, because of the short mean free path of the secondary electrons. The small thickness assumption breaks down in this case, and it becomes important to incorporate the effect of spatially varying dose rate.
We have implemented CASINO simulations to track electron trajectories and hence the dose rate as the function of the cell geometry. The output from the CASINO simulations is used as input to a reaction-diffusion system, simulated in Mathematica, which calculates the solution chemistry by accounting for the creation, diffusion and annihilation rates of the various radiolysis species. This system poses a highly nonlinear set of equations because the time scales for the key reactions and diffusion constants differ by large orders of magnitude. We describe these simulations and discuss the practical implications for liquid cell experiments, especially those that involve materials reactions taking place at liquid/solid interfaces.

The enhanced energy resolution of recently developed monochromators has the greatest impact on the low-loss electron energy-loss spectroscopy (EELS) and it is now possible to probe vibrational as well as electronic excitations in the scanning transmission electron microscope (STEM) [1]. Vibrational spectroscopy in the STEM using focused electron beams opens up many new possibilities for determining material properties at the nanometer level. The delocalized nature of the inelastic scattering makes it possible to use the aloof beam spectral acquisition mode to probe bonding while dramatically reducing electron beam damage. This allows beam sensitive materials to be investigated and we have recently explored the feasibility of OH detection in hydroxides and hydrates with this approach [2]. The surface sensitivity shows that OH species can be detected on surfaces of oxide nanoparticles. The spatial variation of the vibrational EELS signal is being investigated experimentally and theoretically on systems such as model Si/SiO₂ interfaces.
Bandgap determination and interband states can also be mapped in nanoparticles and at nanoparticles surfaces. For example, EELS has been employed to determine bandgap states in Pr_0.1Ce_0.9O_2-δ nanoparticles [3]. Spectral features within the bandgap show a rising onset and extended plateau region. The inter-bandgap feature can be interpreted in terms of transitions of electrons from the O-2p valence band into unoccupied Pr-4f levels lying slightly below the Ce-4f conduction band resulting in broad optical absorption above ~2 eV [3]. Electronic states associated with oxide surfaces may also be probed using aloof beam EELS. Peaks within the bandgap are often observed when the electron probe is positioned just outside an oxide surface (e.g. MgO, TiO₂) [4]. These peaks arise from both the presence of surface states and possibly Cerenkov modes which depend strongly on nanoparticle geometry. The interpretation of these spectral features and the information that can be determined about nanoparticle surfaces will be discussed.

References
[1] O.L. Krivanek et al. Nature 514, 209-212 (2014).
[2] P.A. Crozier et al, Ultramicroscopy (in press)
[3] Bowman, W.J., et al., Ultramicroscopy 2016. 167: p. 5-10.
[4] Q. Liu et al, Ultramicroscopy (in press)
[5] The support from US DOE (DE-SC0004954), NSF grants DMR-1308085 and CHE-1508667 as well as ASU’s John M. Cowley Center for High Resolution Electron Microscopy is gratefully acknowledged.

The development of electron and scanning probe microscopies in the second half of XX century have produced spectacular images of internal structure and functionalities of matter with nanometer and now atomic resolution. Since the beginning of XXI century, the progress in detectors and data acquisition electronics has opened the floodgates of high-veracity information on atomic positions and functionality, often in the form of multidimensional data sets containing partial or full information on atomic positions, bonding, electronic and magnetic functionalities. Harnessing this data flow offers the promise to establish genomic libraries of structure property relationships that can revolutionize materials theory, prediction and synthesis, and enable new sets of tool for direct manipulation of matter. In this presentation, I will discuss the progress towards establishing the corresponding instrumental framework and associated data analytic tools. In Scanning Probe Microscopy, this approach is illustrated via full information capture in SPM based on recording and complete analysis of data stream from photodetector. This general-mode (G-Mode) SPM is illustrated for classical modes such as intermittent contact mode SPM, as well as piezoresponse force microscopy and spectroscopy (PFM) and Kelvin probe microscopy. The analysis of the information contact allows deducing in which cases classical signal processing allows unbiased representation of the tip-surface interactions, and further enables new data rich imaging modes. The big data approaches in STEM are illustrated by streaming of imaging data combined with real-space crystallographic mapping. Here, I briefly discuss the local, Fourier, and semi-local methods for atomic position determination, and the extension of these approaches towards ptychographic data. A deep data approach will allow merging this knowledge with physical models. Here, this concept is illustrated by the analysis of mesoscopic dynamic experiment, namely recovery of reaction and diffusion rates from the observations of particle growth. Finally, a smart data approach will enable algorithms for data identification, expert assessment, and ultimately, control over matter via active feedbacks. Here, I will illustrate this approach via creation of artificial non-equilibrium materials structures via controlled e-beam and He ion irradiation of materials with in-situ atomic resolution imaging.

This research is supported by the by the U.S. Department of Energy, Basic Energy Sciences, Materials Sciences and Engineering Division, and was conducted at the Center for Nanophase Materials Sciences, which is sponsored at Oak Ridge National Laboratory by the Scientific User Facilities Division, BES DOE.

Single walled carbon nanotubes (SWNTs) have shown oustanding capabilities in the realization of new functional devices but are extremely sensitive to any slight changes in their environment, altering their physical properties. A strategy to overcome this difficulty is to use double-walled carbon nanotubes (DWNTs), consisting of two concentric tubes.

In order to better know the basic properties of this kind of tubes in link with their structure, we have developed a systematic and robust procedure using acHR-TEM (aberration corrected Transmitting High Resolution Electron Microscopy), to determine the atomic structure over one hundred DWNTs.

Statistical analyses of their diameters and twist angle between inner and outer tubes show that some configurations are strongly favored whereas some other are never observed. These results reveal the existence of strong coupling between the two concentric tubes in a DWNT for the smaller diameters low than 2 nm.

To go beyond phenomenological approaches, we performed calculations at atomic scale using an empirical model implemented into a Monte Carlo code. Such approach is necessary to understand the nature of the coupling and explain the selectivity of observed patterns.

In order to produce high-efficiency poly-crystalline CdTe solar cells, post-deposition CdCl₂ and Cu diffusion heat treatments (HTs) are necessary. A series of electron beam induced current (EBIC), scanning transmission electron microscopy (STEM) electron energy loss spectroscopy (EELS), and atom probe tomography (APT) measurements have been performed to understand the correlation between carrier separation properties within grain boundaries (GBs) and at the CdTe/CdS interface to elemental segregation and S diffusion at these regions. The EBIC results reveal that the CdCl₂ and Cu HTs convert the CdTe grain boundaries from being non-photoactive to photoactive. The STEM EELS and APT measurements show a Cl segregation and Te depletion within GBs after the CdCl₂ HT, which cause the GBs to be doped opposite of the grain interiors. The band bending at the GBs created by the Cl dopants creates an electric field between the GBs and grains, further increasing the carrier collection. Synergy between APT, STEM-HAADF, and STEM-EELS reveals that Cl also segregates at the CdS/CdTe interface, in which Cl diffuses for a few nms into the CdTe_xS_1-x alloy. To increase the EBIC resolution and determine the location of the pn-junction with respect to the CdTe/CdS interface, a STEM-EBIC holder/experiment was designed for a Nion-UltraSTEM 200. The STEM-EBIC results reveal that the pn-junction is formed several nms inside the CdTe layer, confirming that Cl dopes a thin layer of the CdTe grain interfacing the CdS grain n-type. In addition to the post-deposition heat treatments, recent work has shown that a graded CdTe_xSe_1-x alloy formed by Se diffusion into the CdTe layer can also improve the short-circuit efficiency (Jsc) and overall efficiency of CdTe solar cells. The combination of APT, TEM-selected area diffraction (SAD), and EBIC show the Se content dependence on the CdTe_xSe_1-x structure, and the structural dependence on the photoactivity of the CdTe_xSe_1-x alloy.

This research was supported by the US Department of Energy (DOE) Office of Energy Efficiency and Renewable Energy, Foundational Program to Advance Cell Efficiency (F-PACE), and ORNL’s Center for Nanophase Materials Sciences (CNMS), which is a U.S. DOE Office of Science User Facility.

The semiconductor nanocrystals (NCs) have attracted significant attentions due to their unique photoelectronic properties. The InAs NCs have displayed interesting electronic properties and diffusion patterns that upon doping with metal ions [1-4]. For example, Cu dopants occupy the interstitial sites, the Ag dopants substitute for the In atoms, while the Au atoms form the metal-semiconductor hybrid nanoparticles by segregating to the core of the InAs lattice. The Cu and Ag- doped InAs semiconductor NCs have opposite effects on their optoelectronic properties as a result of doping. However, there is still a lack of understanding of the host-dopant interactions in these three cases. The challenge is the coexistence of multiple length scales that describe different components of the doped NCs. The dimensions of the NCs are of the order of 5-10 nm, the changes in the InAs structure occur in the nm scale, while the changes around dopants occur in the 0.1 nm scale. In our work, multiple experimental techniques, uniquely optimized to probe the system at different length scales, were combined to solve the structures of dopant and host systems as a result of doping and in real time. The probes include x-ray absorption fine-structure spectroscopy (XAFS), transmission electron microscopy (TEM) and x-ray diffractions (XRD). The XAFS is an element specific technique and is sensitive to the local structure surrounding the absorbing atoms. XRD measures the changes in crystal structure of InAs NCs and metallic phase of dopants. By adding electron microscopy for direct imaging and chemical analysis of the elemental distribution in the NCs, we were able to directly observe the changes in morphology and chemical and structural environment of all components in doped InAs NCs. For example, in the case of Au-InAs NCs we observed the co-dependency between the nucleation and growth of Au core in the InAs NCs and the progressive disordering and amorphization of InAs shell. In the case of the case of Ag doped InAs NCs, we observed existence of the solubility limit of Ag in the InAs matrix.

1. T. Mokari, A. Aharoni, I. Popov, and U. Banin, Angew. Chem. Int. Ed. 2006, 45, 8001.
2. D. Mocatta, G. Cohen, J. Schattner, O. Millo, E. Rabani and U. Banin, Science 2011, 332, 77.
3. Y. Amit, H. Eshet, A. Faust, A. Patllola, E. Rabani, U. Banin and A. I. Frenkel, J. Phys. Chem. C 2013, 117, 13688.
4. Y. Amit, Y. Li, A. I. Frenkel and U. Banin, ACS Nano 2015, 9, 10790.

Solid state dewetting^[1] is a topic of current research. Besides targeted patterning, research focuses on the mechanisms and its prevention to avoid degradation of thin film devices. While several studies have addressed solid state dewetting of bare metallic films, we focus on Al thin films covered with a native surface oxide layer. In order to simplify the complexity of the film microstructure, thin epitaxial Al films were grown by molecular beam epitaxy on (0001) single crystalline sapphire (α-Al₂O₃) substrates.
The microstructure and epitaxial orientation relationships of the Al films were analysed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM) methods including electron backscatter diffraction (EBSD), high resolution TEM and C_s-corrected atomic resolved scanning TEM (STEM). The as-deposited Al films form two orientation relationships (ORI and ORII) both containing two twin-related growth variants: {111} Al || (0001) α-Al₂O₃ with ±<-110> Al || <10-10> α-Al₂O₃ (OR I) and {111} Al || (0001) α-Al₂O₃ with ±<2-1-1> Al || <10-10> α-Al₂O₃ (OR II). Atomic resolved high angle annular dark field STEM micrographs of cross-sectional samples indicated strain at the twin boundaries of OR I and a translation in <111> surface normal by 0.91±0.13 Å (high resolution TEM^[2]: 0.84±0.17 Å) compared to an ideal, non-relaxed ∑3 twin boundary.
After annealing for 1 to 45 h at 600°C (T_m(Al): 660°C), instead of Al islands^[3] faceted structures were observed in plan-view SEM micrographs. Site-specific cross-sections prepared by focused ion beam revealed the presence of voids in the Al film with a thin cover layer. The drum-like voids possess distinct facets and reflect the hexagonal symmetry of the basal plane of the sapphire substrate. In contrast to the classical solid state dewetting model^[1][3][4], only volume and grain boundary diffusion is supposed to occur due to the surface oxide layer and void formation can be described by film retraction.^[5] The EBSD investigations indicate that the grain boundaries act as initial points of void formation.
Monochromated electron energy loss spectroscopy of the surface layer revealed a phase transformation from amorphous alumina (as-deposited) to γ-Al₂O₃ (annealed) as proposed in literature from glancing incidence X-ray diffraction measurements.^[4] Although annealing was done at 600°C, faceted epitaxial sapphire ridges form at the Al/Al₂O₃/void triple phase boundary as a consequence of the capillary energy force component acting perpendicular to the film/substrate interface. In situ TEM annealing experiments are planned to investigate the dewetting phenomena and sapphire growth in real time.

References
[1] C. V. Thompson, Annu. Rev. Mater. Res., 42, 399-434 (2012).
[2] G. Dehm et al., Acta Mater., 50, 5021-5032 (2002).
[3] E. Rabkin et al., Acta Mater., 74, 30-38 (2014).
[4] W. Kaplan et al., J. Mater. Sci., 48, 5681-5717 (2013).
[5] S. Dutta et al., J. Am. Ceram. Soc., 95, 823-830 (2012).

Cardiovascular calcification contributes to 17 million deaths globally. It is found in atherosclerotic plaques and remains a major cause of valve stenosis. The only viable treatment for this disease remains the replacement of the tissue via surgical intervention [1]. Recently, we showed the presence of nano/micro calcified particles in calcified heart valves tissues, which are also the first calcified structure that could be detected in vascular tissue. Moreover, cardiovascular calcification is not formed mainly by bone tissue but, surprisingly, by a not previously described, highly crystalline calcium phosphate material [2-4]. By combining state-of-the-art nano-analytical electron microscopy characterisation techniques, here we show the presence of two types of particles within the identified population of calcified particles that are different in their topography, crystallinity, composition and mechanical properties. More importantly, using electron energy loss spectroscopy and atom probe tomography we have directly mapped the location and chemical identity of the atoms and found the presence of organic-mineral interaction within the calcified particles. These results show the presence of new biomineral not described before hence providing new insights in understanding the mineral/material content inside cardiovascular calcification.
References-
Butler, D., 2011. UN targets top killers. Nature, 477(7364), pp.260–1.
Bertazzo, S., Gentleman, E., Cloyd, K.L., Chester, A.H., Yacoub, M.H. and Stevens, M.M., 2013. Nano-analytical electron microscopy reveals fundamental insights into human cardiovascular tissue calcification. Nature materials, 12(6), pp.576–83.
Bertazzo, S., Steele, J.A.M., Chester, A.H., Yacoub, M.H. and Stevens, M.M., 2014. Cardiovascular calcification violet pearl. Lancet, 384(9950), p.1294.
Reznikov, N., Steele, J.A.M., Fratzl, P. and Stevens, MM., 2016. A Materials Science Vision of Extracellular Matrix Mineralization. Nature Materials Reviews, 49(16041).

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.

It is known that, owning to indium fluctuation, GaN-based blue LEDs with high threading dislocation density (TDD) (~10⁸ cm^-2) can still maintain high efficiency. However, GaN-based LEDs with less indium content working from the violet to the deep UV regions exhibit severe efficiency droop due to the high TDD [1]. Therefore, to fully exploit the potential of GaN-based LEDs, it is essential to understand the carrier dynamics around dislocations.
To study the optical properties at the nanoscale, a powerful and widely used technique is cathodoluminescence (CL). Since 1997, it has been confirmed, through a comparison between atomic force microscopy and CL images, that dislocations in GaN act as non-radiative centers [2]. Non-radiative dislocations have been further identified as pure edge or mixed (edge and screw) type using a comparison between plan-view transmission electron microscopy and CL images. In addition, a model based on an effective radius of the dislocation has been proposed to explain the reduced CL efficiency around the dislocation [3]. However, to our best knowledge, the information about the carrier lifetime around the dislocation still remains unknown, because it requires simultaneously picosecond temporal and nanoscale spatial resolutions.
Here, thanks to a new picosecond TR-CL setup combining high temporal (<20 ps) and spatial (~50 nm) resolutions [4], we investigate the dynamics of excitons at T = 10 K around an isolated single edge dislocation in homoepitaxial GaN grown by metal organic vapor phase epitaxy. We obtain the distributions of emission peak energies, effective lifetimes, and CL intensities around such a dislocation. The edge character of this dislocation is identified by observing and simulating the observed ~1.3 meV dipole-like energy shift, due to the strain fields induced by such type of dislocation. By simultaneously recording the variations of both the exciton lifetime and the CL intensity across the dislocation, we directly assess the dynamics of excitons around such a defect. Our observations are well reproduced by a diffusion model [5]. The latter allows us to deduce both an exciton diffusion length of ~24 nm as well as an effective area of the dislocation defined by a radius of ~95 nm, where the recombination can be treated as entirely non-radiative [6]. Therefore, our results pioneer an innovative way to study quantitatively the influence of extended defects on the efficiency of optoelectronic devices.


[1] S. Nakamura, Science. 281, 956 (1998).
[2] S. J. Rosner et al., Appl. Phys. Lett. 70, 420 (1997).
[3] T. Sugahara et al., Jpn. J. Appl. Phys. 37, 398 (1998).
[4] M. Shahmohammadi, et al., Nano Lett. 16, 243 (2016).
[5] W.R. Harding, et al., Electron. Lett. 12, 503 (1976).
[6] W. Liu et al., APL (submitted).

We present a time-resolved cathodoluminescence (CL) imaging microscope based on a 30 keV FEI Quanta 650 SEM, equipped with an ultrafast electrostatic beam blanker, and a Delmic cathodoluminescence analysis system. Electron pulse trains with arbitrary time profiles are achieved with a time resolution down to 5 ns with a repetition rate of 1 MHz. We characterize InGaN/GaN quantum wells emitting at 450 nm using the combined CL spectrum, angular distribution, polarization, and time-resolved CL under 30 keV electron excitation. Using single-photon detection analysis we find a CL lifetime of 20 ns at a spatial resolution less than 50 nm. The resolution is further improved by optimizing the electron optics in the column of the microscope. Two-dimensional lifetime images can be recorded within minutes .

Next, we use a time-correlated single-photon detection system to measure the autocorrelation function of the CL emission g²(τ) of the quantum wells at 450 nm using Hanbury-Brown-Twiss interferometry. The g²(τ) function measures the probability for two photons to be emitted with a certain delay τ, providing direct information on the (single-photon) quantum nature of the emission. We find that a single electron can excite the quantum well more than once, leading to bunching [1] in the autocorrelation function: g⁽²⁾ (0)=5 at a beam current of 140 pA. The amplitude of the bunching signal depends strongly on electron beam current and pulse duration. For a current of 140 pA the bunching amplitude g⁽²⁾(0)=5 for a continuous beam, and increases to g⁽²⁾(0)=9 for a 5-ns pulsed beam. A model to present the electron-induced bunching effect taking into account interacting secondary electron cascades will be presented.

The new time-resolved cathodoluminescence microscope brings together, for the first time, spatial, spectral, angular, polarization, and time-resolved data in one single measurement setting, and enables the study of a wide range of quantum optical phenomena in semiconductor quantum wells, quantum dots, and diamond NV centers, as we will show.

(1) S.Meuret, L.H.G. Tizei, T. Cazimajou, R. Bourrelier, H.C. Chang, F. Treussart and M. Kociak, Phys. Rev. Lett. 114 197401 (2015)

The correlation between a material’s luminescence properties and its morphology, microstructure and local chemistry offers great benefit in the understanding of many technologically important materials and devices. This is particularly true for the novel class of light emitting diode (LED), laser and photovoltaic devices based on semiconductor nanowires (NWs), including disk-in-nanowire heterostructures studied here. This has encouraged a growing interest in performing cathodoluminescence (CL) microscopy at high spatial resolution in a STEM microscope where optical and structural properties can be correlated at the nanoscale [1,2]. Here we use combined HAADF imaging, electron energy loss spectroscopy (EELS) and CL analysis of InGaN quantum disc inserts in a GaN NW proposed for LEDs. We investigate the role of structural and compositional imperfections in the quantum disc and their impact on the optical properties.

For the results of this paper, compositional information was obtained from sub-nanometer scale EELS analysis using means of MLLS fitting to extract the composition and the local luminescence was measured simultaneously using CL. The CL light was acquired using miniature elliptical mirrors (solid angle of about 7.3 sr) integrated into the tip of a conventional cryogenic TEM holder. Light is coupled out of the holder through two optical fibres to an optical spectrometer fitted with a PMT and CCD detectors. This combined EELS / CL system offers the advantage of the best in spectral resolution (up to 4 meV), spatial resolution analysis and sensitivity to microstructural changes. Simultaneous EELS / CL data was collected with the sample at -171C minimizing the influence of the electron beam on the sample and increasing the spatial resolution of the Cl data as result of the enhanced rate of radiative recombination within the QWs.

The analysis was carried out across the InGaN quantum disk active region containing many discs of just a few nanometer in width. Significant variation in the morphology of the quantum discs was observed and could be correlated with the composition and luminescence properties. A defect center was found to emit blue-shifted luminescence with quantum dot-like emission properties.

[1] Kim S. K., Brewster M., Qian F., Li Y., Lieber C. M. and Gradecak S., Nanoletters 9, 3940, 2009
[2] Zagonel L. F., Mazzucco S., Tence’ M., March K., Bernard R., Laslier B., Jacopin G., Tchernycheva M., Rigutti L., Julien F. H., Songmuang R. and Kociak M., Nanoletters 11, 2011, 568

Electronic and optical properties changes induced by defects play an important role in material and devices design. The full picture of these changes must include the understanding of the structure defects along with their spectroscopic signature. Evidently, nano-spectroscopy techniques, such as electron energy loss spectroscopy (EELS) and cathodoluminescence (CL) in electron microscopes (EM) are well adapted to the study of such systems at the nanometer and sub-nanometer scales.
In this contribution, we will discuss different experiments with these two spectroscopic techniques performed in scanning transmission EM (STEM), the main objective being the understanding of the interplay between structural and electronic/optical properties.
The experiments presented have been performed on a VG HB501 (EELS and CL) operated between 60 and 100 kV and on the USTEM Nion200 (EELS) operated between 60 kV and 200 kV. Experiments have been performed at 150 K and 300 K.
We will start by considering the spectroscopic properties of point defects in large band gap materials (diamond and h-BN), which introduce states within the energy bandgap, leading to different absorption/emission signatures. For example, the NV⁰ defect (neutral nitrogen-vacancy complex) leads to emission at 575 nm, while the H3 defect (nitrogen-divacancy complex) leads to emission close to 500 nm [1]. The presence of individual defects was confirmed by light interferometry of the CL emission using a HBT (Hanbury-Brown and Twiss) interferometer [2].
We have used the same technique to identify a structural defect in h-BN, another wide-gap material, which behaves as a single photon source (showing, incidentally, the detection of an individual defect). This defect emits in the 3.5 to 4.2 eV range, with a lifetime of 2.5 ns [3,4]. In the same energy range, another light signal was detected with a longer lifetime (22 – 200 ns). This second emission intensity increases as a function of electron beam irradiation of the h-BN layers, indicating that its origin is linked to defects created by the electron beam, possibly donor-acceptor pairs (B vacancy and interstitial B in-between layers). Different experiments indicate that the first defect is a C_N substitutional atom, but this is yet to be confirmed.
Also in h-BN, we have observed variations of the emission energy of the excitonic line due to extended defects created at folds [5]. The energy changes are caused by the different stacking of the h-BN layers, what was confirmed by simulations.
Finally, we will discuss the use of EELS to characterize the local absorption and electronic properties of defects in thin layered materials.
[1] L. H. G Tizei and M. Kociak, Nanotechnology, 23, 175702 (2012)
[2] L. H. G. Tizei and M. Kociak, Phys. Rev. Lett., 110, 153604 (2013)
[3] R. Bourrellier, Nano Letters, in publication, 10.1021/acs.nanolett.6b01368 (2016)
[4] S. Meuret, Phys. Rev. Lett., 114, 197401 (2015)
[5] R. Bourrellier, ACS Photonics, 1, 857 (2014)

Cathodoluminescence (CL) is an important spectroscopic method for characterizing photovoltaic materials in the electron microscope. When electron-generated free carriers recombine, CL signals are emitted from the luminescent material and provide spectroscopic information that can be used to reveal features of the electronic structure, such as the band gap and defect states near the band edge. These characteristics can be correlated with the microstructure and used to identify microstructural defects that limit the solar cell efficiency. Ultimately, this knowledge can be used to develop processing schemes that optimize material structure and performance. For example, in quantum dot solar cells and polycrystalline thin-film solar cells, defects in the thin absorber layers exist in the bulk and at the surface. Surface defects are important in these systems due to the increased surface to volume ratio compared to conventional Si solar cells. As a result, it is necessary to use CL in high spatial resolution mode so that the effects of bulk and surface defects can be deconvoluted. The spatial resolution of CL in the scanning electron microscope (SEM) is limited by the lateral size of the interaction volume (at least 250 nm). It has to be minimized to break the resolution limit.

We use a home-built CL spectroscopy set-up integrated with a transmission electron microscope (TEM) to measure high-spatial-resolution CL spectra of polycrystalline CdTe. The interaction volume is shown to be about 10 nm in diameter in TEM by Monte Carlo simulations. The improved spatial resolution allows TEM-CL to distinguish CL emissions from grain interiors and grain boundaries. At grain boundaries, we observe both a spectral redshift (≈ 0.01 eV) in the peak location and a reduction in signal intensity compared to that in the grain interiors, suggesting the increased concentration of non-radiative recombination centers. We also use wedge-shaped lamellae with a varying thickness to perform a systematic CL measurement of the grain interiors as a function of thickness and electron beam current. Using a simple model of recombination, we find that the variation of the CL signal with thickness determines the effective bulk lifetime, which has contributions mostly from radiative and surface recombination. The decreased intensity at the grain boundaries in turn determines the reduction of the effective lifetime due to grain boundary defects, enabling an estimate of the grain boundary recombination velocity.

In sum, TEM-CL enables the measurement of CL spectra with high spatial resolution which can be used to compare the radiative recombination rates in grain interiors and at grain boundaries. This can be extended to characterize the emerging photovoltaic materials that consist of nanostructures. The effective lifetime at the grain boundary can be quantified using the simple recombination model. Such information is essential in providing guidelines to improve solar cell efficiency.

In the past 15 years, the use of free electron beams to study the optical properties of nanostructures has proved to be extremely efficient, especially due to the unrivalled spatial resolution attainable in an electron microscope.
Beyond this technical performance, new fields of investigation have therefore been opened. The first relies on the quantitative interpretation of spatially resolved electron based optical experiments – how close are they from optical experiments? The second field relies to the next experimental and conceptual challenges: How can we retrieve the missing information routinely gained in pure optical experiments but not with free electron beams – phase, non-linearity, lifetime measurements, etc …-?
In this presentation, I will present novel interpretations of plasmon mapping experiments, and how they relate to pure optical experimental quantities such as extinction and scattering cross-sections. I will then present preliminary experimental and theoretical results concerning the phase mapping of plasmons. Switching to the problem of lifetimes measurements at the nanometer scale, I’ll show new results proving the measurement of sub-nanosecond lifetimes with sub-15 nm in semiconducting heterostructures. If time allows, I will discuss the recent demonstration of the monitoring of optical non-linearities with an electron beam in such heterostructures.

We have imaged and produced EELS spectra of Nitrogen Vacancy (NV) centers in diamond with our Zeiss Libra TEM with a monochromated electron source and in-column energy filter, With monochromated electron energy loss spectroscopy (EELS) we measure the amount of energy loss that an electron undergoes; this includes optical resonances and inter and intra band transitions but also of course the Cherenkov radiation background. The low acceleration voltage of 40 kV directly reduces the background noise of the Cherenkov radiation. According to the photoluminescence spectra, we find a resonance at 1.9 eV (corresponding to 638 nm wavelength). We have also demonstrated an unexpectedly strong surface-plasmonic absorption at the interface of silver and high-index dielectrics based on electron and photon spectroscopy . The measured bandwidth and intensity of absorption deviate significantly from the classical theory. Our density-functional calculation well predicts the occurrence of this phenomenon. It reveals that due to the low metal-to-dielectric work function at such interfaces, conduction electrons can display a drastic quantum spillover, causing the interfacial electron-hole pair production to dominate the decay of surface plasmons. This finding can be of fundamental importance in understanding and designing quantum nanoplasmonic devices that utilize noble metals and high-index dielectrics.
Depending on the composition, Quantum Materials may act as conductors, insulators, semiconductors or even as superconductors. Combinations of different quantum materials are of high interest to explore new phenomena and act as the foundation for future electronic devices at the nanometer scale. Our quantum materials research is widely spread, reaching from defect formation in graphene to the characterization of hybrid quantum materials. We present our work utilizing Low-Voltage Monochromated EELS and Low-Voltage High-Resolution Electron Microscopy (LV HREM). Together, these often improve the contrast to damage ratio obtained on a large class of samples.

Aberration correction has allowed accelerating voltages in the scanning transmission electron microscope (STEM) to be reduced while maintaining atomic resolution. This facilitates the imaging of two-dimensional materials using annular dark field (ADF) with a resolution of around 1.4 Å, allowing the imaging of individual atoms in materials such as boron nitride and graphene [1,2]. The factor limiting resolution at these incident energies is not only residual coherent lens aberrations, but temporal incoherence leading to extended probe tails. In addition, because many two-dimensional samples are beam sensitive, images typically have a quite low signal to noise ratio. These issues make quantification of images, and especially the unambiguous identification if impurity atoms, difficult.

Deconvolution and band pass filtering potentially offer solutions to both the resolution and noise issues. For a 2D material, the ADF image can be considered a convolution of the probe with the ADF scattering potential. If one could simply deconvolve the probe, the scattering potential should provide unambiguous atomic identification. Deconvolution techniques such as the Weiner filter [3] usually however rely on the use of a regularization parameter to avoid numeric issues. The value chosen for this parameter can lead to quite different results, making true quantification difficult. Parameter free methods such as Richardson-Lucy [4,5] provide robust results, but require the recovered object to be positive. While images are of course positive, the recovered object in this case will be the band width limited ADF potential, whose tails oscillate about zero.

We demonstrate a two stage process, where the effects of temporal incoherence are determined robustly by comparing probes simulated with and without incoherence. This correction is then applied to experimental data in order to remove the “probe tails” and returns a result that can be quantitatively compared with first principles simulations. This method is robust in the presence of noise and also allows the removal of residual probe aberrations.

[1] Krivanek et al., Nature, 462, 7228 (2010).
[2] Zhou et al., Microscopy and Microanalysis, 18, 1342 (2012).
[3] Gonzalez & Wintz, Digital Image Processing, 2^nd Edition, Addison-Wesley, Reading, MA, (1987).
[4] Richardson W.H., Journal of the optical Society of America 62, 55-59, (1972)
[5] Lucy, L.B., Astronomical Journal 79, 745-754, (1974).
[6] This research was supported by the US Department of Energy, Office of Basic Energy Sciences, Materials Sciences and Engineering Division.

Hydroxyapatite (HA) coating has been used to render titanium with excellent osteoconductivity in orthopedic and dental applications. In our previous work, a biomimetic HA coating has been successfully formed on the acid-treated Ti6Al4V discs by soaking them in a modified simulated body fluid at a body temperature and pH. The outermost surface morphology of the coating has been studied extensively using SEM. However, few publications reported the internal microstructure of the biomimetic coating. The adhesive and cohesive bonding of the HA coating to the substrate are determining factors of the success of the implant. Thus, it is critical to develop a good understanding of the cross-sectioning of the heterogeneous, porous coating as well as the bonding between the coating and the substrate.
A dual beam focused ion beam/scanning electron microscopy (FIB/SEM) has been brought together in a single instrument to provide image electron beam, together with a milling ion beam. Ga ions has been the mainstay of the FIB technology over the past years, while the xenon plasma FIB allows for 50x higher removal rates than Ga FIB. As a result, both xenon plasma FIB/SEM (PFIB) and gallium FIB/SEM (GFIB) (FEI Strata 400S DualBeam system) have been used to investigate HA coating on Ti. Cross-sections were milled followed by a cleaning procedure to produce curtaining-free cross-sections. TEM samples were prepared using both PFIB and GFIB for high-resolution interface investigation and chemical analysis. The process and parameters of milling and TEM sample preparation will be compared between PFIB and GFIB.
Cross-sectional analysis is of great interest to visualize the interface between the HA coating and Ti substrate, but the curtaining effect could significantly interfere the resolution of the cross-section image. It has been found that, on one hand, Si cleaning cross-section is more efficient in reducing the curtaining effect of HA coating in PFIB compared to GFIB. On the other hand, Si box patterning is more effective in cleaning the curtaining effect in GFIB compared to PFIB. TEM sample preparation has been simplified with an in situ lift-out method. Due to the higher milling removal rates of PFIB, cracks have been formed in the HA coating, which results in much thicker TEM sample lift-out followed by thinning process on the grid holder. Both the cross-section and TEM images have demonstrated that a tight bonded biomimetic HA coating with a gradient porous structure has been formed on a Ti substrate.
These findings have indicated that even though the PFIB operation is similar to that of the GFIB, there are still significant differences.
Acknowledgements: This work was supported in part by a research grant from FEI Company under an FEI-UConn partnership agreement. The authors would also like to thank the support from NSF grants (CBET-1133883 and CBET-1347130).
Conflict of Interests: Dr. Wei is a co-founder of OrteoPoniX LLC, a biomedical device company.