Symposium F.MT03—Frontiers of Imaging and Spectroscopy in Electron Microscopy

Block copolymers (BCPs) self assemble into a wide variety of periodic structures and elucidating their nano- and micro-structures has typically been accomplished by a combination of transmission electron microscopy and small angle xray scattering. Two of the more complex microdomain structures in BCPs are the double gyroid (DG) and double diamond (DD). We have developed a 3D reconstruction tool, slice and view scanning electron microscopy to create large volume tomograms of the DG and DD structures enabling detailed investigation of a host of supra- and sub- unit cell morphological characteristics. These include the nature of the curvature of the Intermaterial dividing surface (IMDS), the length and angle distribution of the network struts as well as distortions of the cubic unit cells into triclinic variants. In addition to the periodic structure, we have discovered and classified various point, line and surface defects. Because the minority component in the DG and DD phases are 3D continuous networks, defects lead to changes in topology. We find loop and bridging defects in the networks, as well as dislocation defects and sharp, coherent twin grain boundaries and wide non-crystalline grain boundaries. Such defects generally occur during nucleation and growth of the phase and are important for order-order phase transformations as well as for mechanical and transport properties.

Lightweight aluminum (Al) alloys reinforced with a second phase are widely used in aerospace industry. Nanometer scale precipitates distributed in an aluminum matrix considerably change the mechanical properties of these alloys. The degree of coherency of those precipitates, their density, and material properties (such as Young’s modulus) considerably change the bulk mechanical behavior of the alloys. Therefore, investigation of their properties is of great interest for improving alloy characteristics.
The structure-property relationships between the precipitates and Al matrix are complex and often require advanced characterization techniques for their analysis. The formation of the precipitates results in the outbreak of internal stress fields due to the lattice mismatch between precipitates and Al matrix. Strain mapping techniques using Transmission Electron Microscope (TEM) are proven to be a powerful technique for investigating the mechanisms leading to the formation of stress fields. Commensurately, the technique of Electron Energy Loss Spectroscopy (EELS) provides a complementary information on the Young’s modulus (Y_m) through the analysis of the energies of bulk plasmons.
In this study, different TEM techniques are applied to characterize the influence of the Cu precipitates in an Al alloy. We focused on the characterization of a specific material namely 2024 Al alloy composed of Al-4.35 Cu-1.5 Mg-0.64 Mn-0.5 Si-0.5 Fe (wt.%). The samples are treated at 495°C followed by water quenching. Different aging processes are carried out to control and accelerate the formation of particular precipitates known as θ^”. Strain mapping techniques capable of generating results at different field of views are applied to get insights on the properties of θ^” precipitates. Specimens are prepared with a standard lamella process using Focused Ion Beam (FIB) scanning electron microscope and TEM experiments are performed using aberration corrected microscope of model Titan 80-300 ST from Thermo-Fisher Scientific.
Geometric Phase Analysis (GPA) is applied to high resolution STEM images in order to determine the strain in Al matrix around the θ^” precipitates. This technique allows to determine at a very high spatial resolution the strain at the vicinity of the precipitates and their coherency. However, the field of view is limited to a 20 nm x 20 nm area. Nano beam diffraction (NBD) is used to determine strain at the microscale due to its higher field of view which allows to map mechanical deformation at the grain size scales. Due to the larger field of view (500 nm x 500 nm), this technique, combined with GPA, make the analysis of the strain inside grains possible and show the influence of the dislocations identified and localized with GPA.
Additionally, the determination of plasmon energies was carried out to evaluate the Y_m of the alloy at nanoscale. Moreover, the bulk modulus (B_m), and shear modulus (G_m) are established. Furthermore, the nano-indentation experiments and electron backscatter diffraction (EBSD) mappings are performed to correlate the bulk material properties to the nano and micro scale observations. In this way, a true structure property relationship of 2024 alloys has been investigated at bulk scales as well as at nanoscales.
In conclusion, the presented study shows that the employed techniques of TEM and EELS complement the results obtained by using the techniques of EBSD and nano indentation. Hence their combination enables to gather a deeper understanding of the strain relaxation mechanisms occurring in the alloy. This methodology resulted in the determination of a relation among (i) heat treatment, (ii) precipitates distribution and (iii) mechanical properties. In the end, the obtained results are used as inputs to propose a model for better prediction of mechanical properties of 2024 Al alloy.

Acknowledgements:
The work was funded by ADEK, Abu Dhabi, U.A.E. through the Project No. AARE19-131.

Electron Backscatter Diffraction (EBSD) has become an established crystallographic and microstructural characterization technique, particularly for metallurgical applications. The resolution of the technique is ultimately limited by the practical SEM probe size combined with the overlap of sampling interaction volumes. This spatial resolution is constrained further by the requirement to use high sample tilts to increase the diffraction contrast. As a result, certain applications often require the use of transmission Kikuchi diffraction, or other transmission electron microscopy (TEM) based techniques, which typically demand lengthy foil or lamellae preparation by focused ion beam. By utilizing an array of sensors capable of direct electron detection, mounted on a retractable arm around the microscope’s objective pole piece, we are able gain several advantages compared to EBSD in a standard geometry. This tilt-free sample-to-detector geometry permits the collection of distortion free patterns from a well-defined area and the high sensitivity of the detector allows for pattern collection at lower incident electron energies. This enables a reduction of the sampling volume and has the potential for improved precision and physical spatial resolution with usable signal intensity. Here, we present an adapted methodology to determine the resolution of this tilt-free geometry together with the first experimental results on a molybdenum bi-crystal with a well-defined Σ3 boundary. By sampling perpendicular to the Σ3 boundary with 5nm step and evaluating the variation in the quality of the collected diffraction patterns, the physical resolution can be measured based on the full width at half maximum of the Gaussian curve fit of the pattern quality. Even with 5 nm step, the measured boundary on the indexed map is sharp. The EBSD results are further compared and validated by TEM measurements on lamellae prepared from the same specimen.

There are significant intrinsic challenges to characterizing hybrid (hard-soft) materials due to the difference in properties between components of these structures. Nonetheless, establishing the structure-property relationships of these systems is an essential step in material design. Here we study the recently reported Oleophilic Hydrophobic Magnetic (OHM) sponge, a nanocomposite system that offers the ability to repeatedly adsorb 20-35 times its weight in oil. [1,2] The OHM sponge is synthesized by coating a polyurethane membrane with a nanocomposite, made of nanoparticles, such as Fe₃O₄, and graphite. In addition to the aforementioned challenge of distinguishing both the hard (nanocomposite) and soft (PU sponge) features, this hierarchical material necessitates exploration at multiple length scales. Given the physically and chemically complex characteristics of this membrane, a correlative multimodal approach is needed to study the fundamental underpinnings contributing to the performance of this material. We apply a combination of SEM, S/TEM and Raman spectroscopy to probe the structure, morphology and chemistry of the OHM sponge including with operando techniques.

Preparing thin samples for electron microscopy of hybrid materials can be difficult due to the nonuniform mechanical properties of the system. Here, we embed the OHM sponge in resin at the sub-micron scale and use an ultramicrotome cross section approach to directly assess the nature of the coating. Scanning electron microscopy and spectroscopy techniques provide rich atomic and electronic information about the sponge both on the micro (SEM/Raman) and nano (S/TEM) scale. Using a combination of backscatter electron imaging (compositional contrast) and EDS, the spatial distribution of iron oxide nanoparticles can be explored. Additionally, by exploiting the distinct graphite sp² (π bond) EELS edge and Raman spectral signature, we map the graphite distribution of the sponge, distinguishing it from the other carbon-based materials in the sample. This correlative approach is used to provide a complete assessment of the nature of this material and can be more broadly be applied to other hybrid systems. Further, dynamic electron microscopy is a powerful tool to observe reactions as the progress in real time, and ESEM experiments, in particular, provide a pathway to characterize the solid-fluid interactions that contribute to the OHM sponge's performance.

The presentation will emphasize the holistic understanding of OHM sponge and other hierarchical, hybrid materials that can be attained through correlative and multimodal characterization methods.

[1] Nandwana, V. Ribet, S., et al., Ind. Eng. Chem. Res. 2020, 59, 23, 10945–10954.
[2] V. Dravid. US Patent 62/788,347, Filed: Jan 4, 2019.

Electron energy loss spectroscopy (EELS) nowadays allows us to acquire information on excitations in nanostructures with high spatial and spectral resolution over a broad energy range covering vibrational, optical (valence-electron) and core-electron excitations [1, 2]. The possibility of nanoscale mapping of optical response beyond the diffraction limit has made EELS one of the leading techniques in characterization of nanophotonic structures. However, EEL spectra are typically dominated by linear optical response, as the nonlinear contributions are negligible due to weak interaction between electrons and sample, which prevents the use of EELS in studying the nanoscale behavior of nonlinear optical fields.

The interaction between electrons and optical excitations in a sample can be substantially increased if the sample is excited by external optical pumping. Such a scenario is possible in photon-induced near field electron microscopy (PINEM), where we can reach scattering probabilities of order unity, and obtain multiple energy quanta exchanges between the electron probe and the optical field [3, 4]. Additionally, high light intensities can also trigger substantial nonlinearities in the sample response, which should be reflected in measured PINEM spectra.

In our work [5], we suggest PINEM as a suitable technique for nanoscale mapping of nonlinear optical response. In particular, we show that the interaction of electrons with nonlinear optical fields introduces asymmetries in otherwise symmetric PINEM spectra, which we demonstrate on several realistic examples. First we focus on studying second-harmonic (SH) response of spherical gold nanoparticles. We show that the evanescent SH field, which cannot be probed by conventional nonlinear optical spectroscopies, gives rise to substantially asymmetric electron spectra under achievable illumination intensities. We further calculate SH field in gold nanorods and show that significant asymmetries in the PINEM spectra can be achieved for particular beam positions when the linear-field coupling is reduced. We envision that PINEM performed with variable illumination frequency and intensity could become a powerful technique for characterization of nonlinear optical response with unsurpassed combination of spatial and spectral resolution.

References
[1] P. E. Batson, N. Dellby, and O. L. Krivanek, Nature 418, 617 (2002).
[2] O. L. Krivanek, T. C. Lovejoy, N. Dellby, T. Aoki, R. W. Carpenter, P. Rez, E. Soignard, J. Zhu, P. E. Batson, M. J. Lagos, et al., Nature 514, 209 (2014).
[3] B. Barwick, D. J. Flannigan, and A. H. Zewail, Nature 462, 902 (2009).
[4] A. Feist, K. E. Echternkamp, J. Schauss, S. V. Yalunin, S. Schäfer, and C. Ropers, Nature 521, 200 (2015).
[5] A. Konečná, V. Di Giulio, V. Mkhitaryan, C. Ropers, and F. J. García de Abajo, ACS Photonics 7, 1290 (2020).

The characters of a material depend on the crystal grain size and grain boundaries. Therefore, evaluation of crystals from a viewpoint of crystallography is extremely important. Various methods by using Scanning Electron Microscope (SEM) have been extensively developed for the evaluation of crystals because samples can be handled in a bulk state and observed in the region of μm to nm order. One common method is the Electron Channeling Contrast (ECCI) method¹⁾, which allows to obtain an image contrast corresponding to crystal grain boundaries by using the intensity difference of Back Scattered Electrons (BSEs) emitted from a bulk sample. The other common method is the Electron Backscatter Diffraction (EBSD) pattern method²⁾. The EBSD is possible to analyze crystal orientation distributions using diffraction patterns from a sample detected by a fluorescence screen. Those methods are quite useful for knowing grain distribution in the crystal as well. On the other hand, there are some issues in these methods of measuring a bulk sample. The ECCI is sometimes difficult to characterize the crystal structure and grain boundaries because it detects a mixture of BSEs originating from composition, channeling contrast and a diffraction pattern, while the EBSD has the limitation that a sample should be tilted significantly. Therefore, the lateral spatial resolution is about 100 nm in these methods and it would be difficult to analyze fine crystal structures and grain boundaries.
In order to improve the issues, we applied a multi-segmented BSE detector³⁾. The detector is segmented to 16 parts located above a sample and detects BSEs. The method is similar to the ECCI method. However, in practice it is possible to detect only the diffraction patterns by using the new segmented detector. As a result, it is much more clearly separate the grain boundaries than with the ECCI method. The principle is that the diffraction patterns detected by 16 segments in each element are integrated to calculate the signal intensity distribution and display it as a grain boundary image. Because similar intensities can be obtained from the same diffraction pattern, the grains with similar signal intensities are evaluated as the same crystal orientation with high approximation. The Euclidean distance⁴⁾ was used to compare the signal intensities.
We also tried to apply the decelerating method which is a SEM column technique for improving the electron probe diameter. In the decelerating method, incident electrons keep high energy when passing through an objective lens. Finally, the electron energy is reduced down just before the electron is landed to a sample. The method reduces the spherical aberration coefficient (Cs) and the chromatic aberration coefficient (Cc) related to the electron probe diameter and the spatial resolution can be improved⁵⁾.
In this study, we will report that we tried to quickly acquire the grain boundary image with high spatial resolution using the multi-segmented BSE detector combined with the SEM-deceleration method.

References
1) Martin A. Crimp : Microsc. res. technol., 69, 374-381 (2006).
2) R. P. Goehner and J. R. Michael : J. Res. Natl. Inst. Stand. Technol, 101, 301-308 (1996).
3) T. Otsuka, M. Hara, N. Erdman and S. Kitamura : Proceedings of Microsc. Microanal. 24, 686 (2018).
4) M.M. Deza and E. Deza: Encyclopedia of Distances,（Springer Nature,Berlin, 2009） p. 94.
5) O. Terasaki, S. Asahina, Y. Sakuda, H. Takahashi, K. Tsutsumi, M. Kudo, R. W. Corkery and Y. Ma : JEOL News, 52, 20 (2017).

With the size of electronic devices getting smaller and smaller, it has become important to understand thermal expansion coefficients and thermal properties in sub-nanometer scale [1-3]. However, traditional techniques, such as scanning thermal microscopy or Raman thermometry are limited in spatial resolution due to mechanical constraints or optical diffraction limit [1-3]. In the present work, we will utilize a novel approach of non-contact thermometry based on the combination of low-loss electron energy-loss spectroscopy (EELS) with first principles density functional theory (DFT) modeling to measure the thermal expansion coefficient (TEC) of materials with nano-meter resolution. The approach has been well-tested on 2D materials such as transition metal dichalcogenides (TMDs), graphene and MoS2 by Hu et al [1]. In present work, we will extend the approach to Si nanoparticles to determine its TEC as a function of the temperature. In this contribution, we will also explore the effects of particle size, orientation and composition on the TEC, as well as the limits of measuring the local temperature using this approach.

References:
[1] Hu et al., Phys. Rev. Lett. 120, 05590(2018)
[2] Mecklenberg et al., Sci. 347, 6222(2015)
[3] Mecklenberg et al., Phys. Rev. Applied 9, 014005(2018)
[4] This work was supported by the National Science Foundation (DMR-18314061).

Mn-containing oxides are important materials that have been widely used for industrial applications such as battery, magnetic and superconducting materials. Therefore, a novel material of Mn-containing oxides has been remarkably investigated to improve their capability and quality. They have various physical properties because Mn usually has divalent, trivalent or tetravalent and each valence has different d-electron numbers in Mn-containing oxides. Thus, the determination of valence in Mn-containing oxide is a very important factor to know the characteristics of the material. As an analytical method for Mn valence, X-ray Absorption Fine Structure (XAFS) method is frequently performed in synchrotron radiation facilities [e.g. 1]. X-ray analysis in the synchrotron radiation facilities is quite powerful as it allows in-situ observation. However, it is difficult to conduct an analysis with a finer structure because the X-ray beam size is on the order of microns. For the chemical state analysis in a micro or nano-ordered structure, the method combining with Transmission Electron Microscopy (TEM) and Electron Energy-Loss Spectroscopy (EELS) is popular. Although this method has a high spatial resolution and high energy resolution, EELS cannot directly obtain information of the valence band associated with chemical bonding of the material.
Recently, SXES that can be installed in Scanning Electron Microscope (SEM) and Electron Probe Micro-Analyzer (EPMA), has been developed for the chemical state analysis of a bulk with fine structures [2]. Soft X-ray means the X-ray with energies from 10 eV to a few keV, and includes the X-ray generated by electron transition from valence bands to inner-shell levels. Therefore, we can obtain chemical bonding information by spectral analysis of this X-ray. So far, we reported a valence analysis of Fe oxides by using SEM-SXES system and “Flank” method [3] suggested by Höfer et al. [4]. Flank method proceeds as below:
1) Calculating a spectrum difference between Fe2+ endmember and Fe3+ endmember.
2) Setting Flank positions of Fe Lα and Lβ at the maximum or minimum value position in the difference spectrum.
3) Taking the Lβ/Lα intensity ratio at the Flank position.
This method has higher resolution for valence than the conventional method of peak-intensity and area-intensity ratios. In this study, we attempt to apply this method for Mn-containing oxides.
MnO, Mn2O3 and MnO2 purchased from Hori Mineralogy Ltd. were used for samples. There is no element detected by SXES other than Mn and O. The SXES measurement was performed by using FE-SEM (JEOL JSM-7900F) equipped with SXES (SS-94040SXSER). The soft X-ray spectra were acquired at accelerating voltages of 10 kV, 5 kV and 2 kV with a beam current of 50 nA and acquired for 5 minutes.
In this study, we compared the intensity ratio of Mn Lβ/Lα calculated by the Flank method, peak-intensity ratio method and area-intensity ratio method. As a result, it is found that the Flank method has a higher energy resolution for the valence of Mn than the conventional peak-intensity ratio method and area-intensity ratio method for all accelerating voltages used in this study at 2, 5 and 10 kV. It is expected that this method will be widely applied to the valence analysis of Mn-containing oxides used in various industrial applications.

References:
[1] Kobayashi et al., J. Mater. Chem., 14, 1843-1848 (2004).
[2] Takahashi et al., IOP Conf. Ser.: Mater. Sci. Eng., 109, 012017 (2016).
[3] Kojima, 12th Asia-Pacific Microscopy Conference Abstract (2020).
[4] Höfer et al., Eur. J. Mineral., 6, 407-418 (1994).

Direct electron detectors have revolutionized the field of structural biology and are beginning to find applications in materials science and nanoscience. Here, we show how these detectors can improve the ability of transmission electron microscopy to understand the physical and electronic structure of nanoparticles and thin films. These newly developed direct electron detection cameras reduce the signal-to-noise ratio compared to conventional CMOS / CCD cameras. In this study, electron energy loss spectroscopy (EELS) spectra obtained from several materials systems - Ti2-xNbxC MXenes, Mo4VAlC4 MAX phase and CoPt nanoparticles - are quantified, and the improved signal quality with the direct detection system is demonstrated. We compared the effect of probe size, aperture size, camera length, and imaging filter aperture size on the signal-to-noise ratio. For nanoparticles and thin films, it was shown that probe size and camera length had the strongest influence on the signal intensity arising from inelastically scattered electrons. In contrast, imaging filter aperture size plays a minor role. The optimization of these parameters makes it possible to routinely perform valence-state analysis at an atomic level, with a resolution comparable to bulk approaches such as X-ray photoelectron spectroscopy or X-ray absorption spectroscopy (XAS). The increased Detected Quantum Efficiency (DQE) combined with a sub-pixel detection setting is also beneficial for recording diffraction patterns to perform electron pair-distribution function (ePDF) analysis from these same systems. The improvements in the quality of both spectral and scattering data can be expected to make transmission electron microscopy an even more powerful tool for characterizing nano materials.

Quantitative 3D characterization at high resolution is often required in order to fully exploit structure–property relations of nanomaterials. In recent years, the resolution of electron tomography has reached the atomic scale. However, for a successful 3D reconstruction using electron tomography multiple exposures are required, which is not always feasible when studying, e.g., radiation-sensitive nanostructures. In order to overcome these limitations, an alternative method can be used where the 3D atomic structure is reconstructed from atom counts obtained from a single ADF STEM projection image. These atom counts can be used to create an initial atomic model which serves as an input for an energy minimization to obtain a relaxed 3D reconstruction of a nanostructure [1].

To count atoms with single-atom sensitivity, a minimally required electron dose is necessary, while on the other hand the risk of knock-on damage, induced by the high energy electrons, puts an upper limit on the tolerable dose. An important challenge is therefore to develop experimental strategies to optimize the electron dose by balancing atom-counting fidelity versus the risk of knock-on damage. To achieve this goal, a statistical framework combined with physics-based modelling of the dose-dependent processes is proposed and experimentally verified. This model enables an investigator to theoretically predict, in advance of an experimental measurement, the electron dose resulting in an unambiguous quantification of nanostructures in their native state with the highest attainable precision [2].

For beam-sensitive materials, where the optimal electron dose can be low, the images will exhibit a limited signal-to-noise ratio and additionally, they will show a very weak contrast in the presence of light elements. In order to reliably detect the presence or absence of atomic columns and even single atoms, the maximum a posteriori (MAP) probability rule is proposed as a model-order selection method in which a Bayesian framework is used [3]. By using MAP selection, the probabilities for an increasing number of atomic columns present in an experimental image can be calculated. This method therefore allows us to predict the most probable atomic structure. In addition, in recent work we investigated the use and potential benefit of deep convolutional neural networks to restore low dose, scan noise distorted ADF STEM images.

Our method for atom-counting opens up new opportunities to measure variations of the 3D atomic structure of nanoparticles under the flow of a selected gas [4]. In situ experiments clearly demonstrate that many questions in materials science require a quantification of the material’s dynamics from a series of sequentially recorded images. However, the standard approach for atom-counting, where each frame is analyzed individually, does not take advantage of the time aspect available in a series of ADF STEM images. We therefore extended the atom-counting method using the so-called factorial hidden Markov model. This new method is very promising for revealing atomic scale dynamics [5]. In conclusion, new developments in the field of quantitative STEM will be presented enabling one to reliably quantify dynamic structural changes resulting from adatom dynamics, surface diffusion, beam effects or during in situ experiments.

References
[1] A De Backer et al., Nanoscale 9 (2017), p. 8791.
[2] S. Van Aert et al., Physical Review Letters 122 (2019), 066101.
[3] J. Fatermans et al., Physical Review Letters 121 (2018), 056101.
[4] T. Altantzis et al., Nano Letters 19 (2019), p. 477.
[5] A. De wael et al., Physical Review Letters 124 (2020), 106105.
[6] This work was supported by the European Research Council (Grant 770887 PICOMETRICS to SVA and Grant 815128 REALNANO to SB, Grant 823717 ESTEEM3). The authors acknowledge financial support from the Research Foundation Flanders (FWO, Belgium) through project fundings, a postdoctoral grant to ADB and a doctoral grant to ADw.

The electrical properties of nanoscale electronic devices, such as p-n junctions, transistors and light-emitting diodes, are affected strongly by the presence of defects, dopants, interfaces and surfaces [1]. Just as the performance of field emitters is influenced by geometrical factors such as their shape, curvature and crystallographic orientation [2], in atom probe tomography the shape, chemistry and defects in a needle-shaped specimen can affect the trajectories of field-evaporated ions and the fidelity of the reconstruction of the three-dimensional (3D) atomic positions [3], while tip-enhanced catalytic efficiency is influenced by the shape and species of the tip [4]. A knowledge of the 3D charge density, electric field and electrostatic potential inside and around such nanostructures is therefore required.

Off-axis electron holography provides direct access to the projected electrostatic potential within and around a specimen in the transmission electron microscope (TEM) [5]. Here, we combine off-axis electron holography with electron tomography and a model-based iterative reconstruction algorithm [6] to determine the 3D charge density, electric field and electrostatic potential within and around an electrically-biased C fibre needle. The needle was prepared using a standard focused ion beam preparation procedure. A Nanofactory holder was used for electrical biasing. A µm-sized Au tip was used as a counter-electrode at a distance of 4.5 µm from the needle. The diameter of the apex of the needle was ~60 nm and its length was ~2.5 µm. Two tomographic tilt series of off-axis electron holographic phase images were acquired between -52^o and +48^o with a tilt increment of 4^o. One series was acquired without an electrical bias applied to the needle. This dataset includes only the mean inner potential contribution to the phase. It was used to reconstruct the 3D morphology of the needle and to subtract the mean inner potential contribution to the phase from a second dataset recorded with an electrical bias of 40 V applied to the needle.

The reconstructed 3D charge density in the needle reveals that the charge density is greatest at the apex and has a maximum value of 2.94×10¹⁸ cm^-3. 2D slices extracted from the 3D reconstruction show that the charge density distribution is asymmetrical. The majority of the charge is found to be located close to the surface of the needle. A region with a locally increased charge density may be associated with damage by Ga bombardment during sample preparation. The 3D electric field and 3D electrostatic potential are calculated from the reconstructed 3D charge density on the assumption that image charges are defined by the counter-electrode. The strength of the electric field is greatest close to the apex of the needle and has a maximum value of ~0.25 GV/m [7].

References:
[1] S.M. Sze, Semiconductor devices: physics and technology, Wiley, 2008.
[2] H. Zhang, et.al., Nat. Nanotech. 11 (2016) 273.
[3] F. Vurpillot, et.al., Ultramicroscopy. 159 (2015) 202.
[4] Y. Su, et.al., Nat. Nanotech. 11 (2016) 609.
[5] H. Lichte, et.al., Rep. Prog. Phys. 71 (2007) 016102.
[6] F. Zheng, et.al, J. Elec. Spec. Rel. Phenom. 2020, in press.
[7] The authors acknowledge the European Union for funding through the Marie Curie Initial Training Network Grant No. 606988 under FP7-PEOPLE-2013-ITN).

Supported metal catalysts, consisting of finely dispersed metal nanoparticles on high-surface-area supports, are key players in chemical industry. The interplay between the metal and supporting oxide is an important parameter in modulating the properties of heterogeneous catalysts. To gain fundamental understanding, and therefore increase our control over these metal-support interactions, 3D visualization with nano-scale precision is crucial.

Here, we achieve a quantitative 3D characterization of the local environment of metal nanoparticles embedded in a metal oxide support using dual axis electron tomography. Specifically, we studied raspberry colloid templated (RCT) catalysts – a porous silica support with partially embedded Au₉₇Pd₃ and Au nanoparticles in the pore walls [1] – which are active hydrogenation and oxidation catalysts [2][3]. By performing the tomography analysis over two perpendicular tilt axes, we minimized the missing wedge artifacts around the metal particles, allowing the assessment of the particle shape and geometrical confinement of 400 individual nanoparticles. We complemented the electron tomography analysis with a wet-chemical approach relying on metal overgrowth, to visualize the chemical accessibility of nanoparticles within the support. A combination of silver overgrowth on embedded Au nanoparticles and aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) imaging, and Energy Dispersive X-ray spectroscopy (EDX) mapping unveiled that although the particles were substantially embedded in the silica support they were exposed to the pore network and are readily accessible through the interconnected micropores of the silica matrix in the RCT catalyst. Altogether our work provides a general approach to probe the geometrical confinement and chemical accessibility of a statistically relevant number of metal nanoparticles in supported catalysts in three dimensions and with nano-scale precision.

[1] E. Shirman, T. Shirman, A. V. Shneidman, A. Grinthal, K. R. Phillips, H. Whelan, E. Bulger, M. Abramovitch, J. Patil, R. Nevarez, J. Aizenberg, Adv. Funct. Mat., 28, 1704559 (2018).
[2] M. Luneau, E. Guan, W. Chen, A. C. Foucher, N. Marcella, T. Shirman, D. M. A. Verbart, J. Aizenberg, M. Aizenberg, E. A. Stach, R.J. Madix, A. I. Frenkel, Cynthia M. Friend, Commun Chem 3, 46 (2020).
[3] M. Luneau, T. Shirman, A. Filie, J. Timoshenko, W. Chen, A. Trimpalis, M. Flytzani-Stephanopoulos, E. Kaxiras, A. Frenkel, J. Aizenberg, C. M. Friend and R. J. Madix, Chem. Mat. 31, 5759 (2019).

The trend in miniaturisation of devices is increasing the need for characterisation techniques that are capable of investigating objects in 3D at nanoscale with both structural and chemical information. Microscopy and spectroscopy are correlated together to overcome the boundaries of each technique and to go beyond the limitations of microscopy in data treatment. In this context, we have coupled a Zeiss Orion NanoFab Helium Ion Microscope (HIM) with a specifically designed Secondary Ion Mass Spectrometry (SIMS) system in order to add analytical capabilities to the microscope [1]. In the standard HIM mode, the instrument provides secondary electron (SE) images with a sub-nm lateral resolution. These ultra-high resolution SE images can be correlated in 2D with sub-20 nm resolution SIMS images to combine the nanoscale structure and composition information for a deeper understanding of the investigated samples [2].

This two-dimensional correlative microscopy approach has been extended to three dimensions by creating 3D SE–SIMS representations of nanomaterials to give a more extensive view on the analysed objects [3]. Therefore, photogrammetry is used i.e. the sample stage is tilted and SE images are taken at a specific pattern of rotation angles with a He⁺ primary beam around the region of interest (ROI). The 2D SE images are implemented into a photogrammetry software to obtain a 3D SE model observable at all possible angles and magnifications. The SIMS images of the ROI are acquired in-situ at normal incidence with a Ne⁺ primary beam and projected then onto the 3D photogrammetry representation using correspondence points and by overlaying both image textures. The result displays SE and SIMS information together throughout the 4D reconstruction enhancing specimen visualisation and data interpretation.

In this presentation, we will show step-by-step the workflow of 4D reconstruction for real case studies in nanomaterials applications to illustrate the potential of this method for a variety of many other fields.

The authors gratefully acknowledge support by the National Research Fund Luxembourg (FNR) under grant no INTER/DFG/17/11779689.

[1] T. Wirtz, O. De Castro, J. Audinot, and P. Philipp, “Imaging and Analytics on the Helium Ion Microscope,” Annu. Rev. Anal. Chem., vol. 12, no. 1, pp. 523–543, 2019.
[2] F. Vollnhals et al., “Correlative Microscopy Combining Secondary Ion Mass Spectrometry and Electron Microscopy: Comparison of Intensity-Hue-Saturation and Laplacian Pyramid Methods for Image Fusion,” Anal. Chem., vol. 89, no. 20, pp. 10702–10710, 2017.
[3] F. Vollnhals and T. Wirtz, “Correlative Microscopy in 3D: Helium Ion Microscopy-Based Photogrammetric Topography Reconstruction Combined with in situ Secondary Ion Mass Spectrometry,” Anal. Chem., vol. 90, pp. 11989–11995, 2018.

Traditional scanning transmission electron microscopy (STEM) detectors are large, monolithic regions that integrate a subset of the transmitted electron beam signal scattered from each electron probe position. These STEM imaging experiments record only 1-5 values per probe position, throwing away most of the diffracted signal information. With the introduction of extremely high speed direct electron detectors, we can now record a full image (2D data) of the diffracted electron probe scanned over the sample (2D grid of positions), producing a four-dimensional dataset we refer to as a 4D-STEM experiment. These diffraction images of the electron probe are extremely rich in atomic-scale information, such as the sample structure, orientation, composition, phonon spectra, defect structure and more. The spacing between adjacent STEM probes can be varied from below one Angstrom to hundreds of micrometers, allowing us to probe the functional length scale of materials and devices.

In this talk, I will discuss how 4D-STEM can be used to measure various properties important to materials science. I will show several examples of nanobeam electron diffraction used to measure sample structure, orientation and strain, for samples ranging from metallurgical alloys to conductive polymers. I will also show how 4D-STEM can enable phase contrast imaging methods. Finally, I will also show how the scattering matrix (S-matrix) formalism can be used both to numerically invert sample structure under conditions of multiple scattering of the electron beam, and also for very fast simulation of large 4D-STEM datasets. I will show an updated version of Prismatic, our open source simulation code that includes GUIs for easy application.

Recent progress in aberration correction and monochromation in the scanning transmission electron microscope (STEM) has enabled a wide range of new experiments that greatly enhance the depth of information becoming accessible in materials characterization. Direct detection of electron diffraction patterns and spectra became available just at the right time to push the limits of detectability and resolution of these electron optically advanced instruments even further and make experimental data fully quantifiable, to the point, where the match of simulations and experiment is ultimately only limited by counting statistics.

We will report on the recent development of advanced data reconstruction algorithms for recovering the dielectric function from relativistic EELS spectra [1], a versatile ptychography reconstruction algorithm that includes various regularization options [2], and other ways to process 4D-STEM data [3]. We will present results obtained by applying these algorithms to data having been recorded using a monochromated STEM (Nion HERMES microscope operating between 30 … 200 kV, EELS-resolution < 6 meV, spatial resolution < 0.07 nm) [4] equipped with a Dectris ELA hybrid pixel detector [5] that combines true single electron sensitivity, high speed (> 5000 fps for 514 x 1030 pixels at reduced bit depth, and 2250 full fps at 16 bit) and nearly perfect point spread function (1.3 pixels in integration mode) with the capacity for a high dynamic range (> 10^7). Being able to detect spectra and (zero-loss filtered) diffraction patterns without read-out noise greatly increases the level at which these reconstruction algorithms can extract the underlying dielectric function, local crystal structure or maps of the electrostatic potential.

Example applications that will be presented include the extraction of IR-dispersion information from high-resolution EELS of plasmonic nanostructures and a comparison of various data processing algorithms (ptychography, center-of-mass, etc.) applied to (zero-loss filtered) 4D-STEM data.

[1] A. Eljarrat and C.T. Koch, Ultramicroscopy 206 (2019) 112825
[2] M. Schloz, T.C. Pekin, Z. Chen, W. Van den Broek, D.A. Muller, C.T. Koch, arxiv :2005.01530
[3] B. Haas, M. Schloz, J. Müller, A. Mittelberger, O.L. Krivanek, L. Jones, W. Van den Broek, D. Cooper, J.-L. Rouvière, and C.T. Koch, EMC Abstract (2020)
[4] O.L. Krivanek, N. Dellby, J.A. Hachtel, J.-C. Idrobo, M.T. Hotz, B. Plotkin-Swing, N.J. Bacon, A.L. Bleloch, G.J. Corbin, M.V. Hoffman, C.E. Meyer, T.C. Lovejoy, Ultramicroscopy 203 (2019) 60-67
[5] T.C. Lovejoy, B. Plotkin-Swing, N. Dellby, C.E. Meyer, A. Mittelberger, A. Eljarrat, B. Haas, C. Koch, J. Müller and O.L. Krivanek, M&M Abstract (2020)

Convergent beam electron diffraction (CBED) provides a two-dimensional map of the intensity of scattered beams as a function of angle of incidence. Imprinted on the CBED pattern is a historical record of the multiple interactions of the electron wavefield as it is scattered within the specimen, making the pattern acutely sensitive to the specimen structure.

For more than half a century, this sensitivity has been exploited to determine crystal symmetries, refine crystal structures and measure bonding charge density distributions, to name just a few examples [1,2].

Today, with the advent of new, high speed, high dynamic range detectors, it is possible to read out the whole CBED pattern at each position of an electron beam as it is raster scanned across a specimen (often called “4D-STEM”) [3]. This delivers an unprecedented wealth of specimen information buried in terabytes of data! How can we decode this data to extract the specimen information we want? Various schemes, including ptychography and differential phase contrast, have been demonstrated successfully with fast detectors. These various approaches have enabled enhanced image resolution, mapping of electric and magnetic fields, strain, polarisation domains and octahedral tilts (for a review, see [3]).

In this paper, we will discuss some approaches that exploit the power and sensitivity of dynamical electron scattering that underpins 4D-STEM, to provide alternative imaging contrast mechanisms and interpretative schemes that can reveal targeted specimen information. We illustrate these with applications to metallic nanoparticles, beam-sensitive functional perovskites and semiconducting nanostructures.

[1] Zuo J.M. and Spence J.C.H. (1992) Electron Microdiffraction Springer. ISBN 978-1-4899-2353-0
[2] Tanaka M. (1997) Convergent-Beam Electron Diffraction. In: Dorset D.L., Hovmöller S., Zou X. (eds) Electron Crystallography. ISBN 978-94-015-8971-0
[3] Ophus C. (2019) Microscopy and Microanalysis 25, 563–582.

Confinement of magnetic structures geometrically as well as energetically, leads to novel and unexpected domain behavior. Lorentz transmission electron microscopy (LTEM) is ideally suited for quantitative analysis of magnetic domains and microstructure in functional nanoscale heterostructures. The current state of art LTEM enabled using aberration correctors allows for imaging down to sub-nanometer scale in field-free conditions. In this talk, the study of magnetic domain behavior in strongly interacting magnetic nanostructures namely artificial spin ices, curved 3D magnetic nanostructures, and functional materials that host skyrmions will be presented.
Artificial spin ice lattices consist of lithographically patterned arrays of interacting magnetic islands that exhibit magnetic frustration. We will present the results on emergence of magnetic excitations in such lattices which are topologically non-trivial as well as control of magnetic frustration and emergent ordering such as antiferromagnetic to ferromagnetic phase transition using lattice geometry. Furthermore, we will present a novel approach to understand such frustration behavior based on network topology that can be extended to use of graph neural networks to analyze and predict the behavior of spin ice lattices.
Topologically non-trivial spin textures such as skyrmions present unique opportunities to explore exciting fundamental phenomena, such as the topological Hall effect, as well as novel applications, such as skyrmion-based spintronics. The application of LTEM to visualize and understand the magnetic spin textures in multilayer [Pt/Co]_x thin films that can host skyrmions as well as controlled fabrication of skyrmions using Ga ion implantation will be discussed. In-situ magnetization reversal was used to understand the stability of the skyrmions.
Finally, we will also present a new approach based on machine learning for phase retrieval in LTEM that improves the achievable phase sensitivity and spatial resolution.

The layered oxide cathode materials are promising candidates for high energy density lithium-ion batteries (LIB). However, the Ni-rich layered compounds, especially the end member LiNiO₂ (LNO), suffer from stoichiometric challenges (Li/Ni ratios) and instability issues leading to structural degradation [1,2]. It is challenging to synthesize LNO without defects, which further results in severe capacity loss during electrochemical cycling. The phase transition structure/ mechanism is widely under debate, and obtaining a thorough insight is crucial for enhancing the cathode material performance. Directly visualizing the structure at the atomic scale can give us further insight, nonetheless detecting the light elements in energy materials is not trivial.

All the major studies over imaging lithium are carried out on free-standing single-crystalline nanoparticles [3]. However, the real cathode active materials (CAMs) are polycrystalline particles composed of densely packed primary particles/nanocrystals. This poses challenges in sample preparation to achieve the optimum thickness parameters to visualize lithium. In this report, we have investigated the phase transformed region using aberration-corrected scanning transmission electron microscopy (AC-STEM) combined with the pixelated detector (pnCCD). Selecting specific regions in the convergent beam electron diffraction (CBED) patterns and calculating the so-called virtual enhanced ABF (eABF) images is shown to enhance the contrast of the lighter elements even in the thicker sample of ~ 95 nm [4-6]. Moreover, to verify the dependency of this technique at various thicknesses, final gradational thinning of the focused ion beam (FIB) prepared lamellae is carried out by focused low energy Ar-ion NanoMill. Ultimately, we apply the HAADF (high angle annular dark field), ABF, and eABF imaging techniques simultaneously to image the entire layered-to-rock salt phase transition region.

Interestingly, we have been able to access the phase transformed region in between the layered and the rock-salt structure by collecting four-dimensional datasets (4D STEM). This has enabled us to atomically track Ni as well as the light elements, especially lithium and oxygen. The analysis of the layered-to-rock salt phase transition region suggests that there is no sharp interface, as previously assumed. Instead, the transformation appears to take place gradually [1]. The current findings warrant a re-examination of commonly assumed models on this important class of energy materials. Moreover, our results shine a light on the phase transformation mechanism at high spatial resolution and can influence the future design for stabilizing LNO against structural transitions.

References:

[1] S. Ahmed, M. Bianchini, A. Pokle et al. Visualization of Light Elements using 4D STEM: The Layered-to-Rock Salt Phase Transition in LiNiO2 Cathode Material. Advanced Energy Materials n/a, 2001026.
[2] M. Bianchini, M. Roca Ayats, P. Hartmann, T. Brezesinski, J. Janek, There and Back Again—The Journey of LiNiO2 as a Cathode Active Material, Angewandte Chemie International Edition. 58 (2019)10434–10458. https://doi.org/10.1002/anie.201812472.
[3] Y. Oshima, S. Lee, K. Takayanagi, Visualization of lithium ions by annular bright field imaging, Microscopy (Oxf). 66 (2017) 15–24. https://doi.org/10.1093/jmicro/dfw098.
[4] S.D. Findlay, Y. Kohno, L.A. Cardamone, Y. Ikuhara, N. Shibata, Enhanced light element imaging in atomic resolution scanning transmission electron microscopy, Ultramicroscopy. 136 (2014) 31–41. https://doi.org/10.1016/j.ultramic.2013.07.019.
[5] M. Ohtsuka, T. Yamazaki, Y. Kotaka, I. Hashimoto, K. Watanabe, Imaging of light and heavy atomic columns by spherical aberration corrected middle-angle bright-field STEM, Ultramicroscopy. 120 (2012) 48–55. https://doi.org/10.1016/j.ultramic.2012.06.006.
[6] Rose, H., Phase Contrast in Scanning Transmission Electron Microscopy, Optik 1974, 39, 416–436. (1974).

4D STEM has gained popularity in recent years and is now widely applied to the quantitative phase and orientation mapping[1], strain mapping[2], electromagnetic field measurements[3], and ptychographic reconstructions[4]. The realization of routine 4D STEM is closely related to the development of new detectors. Direct electron detectors, for example, have overcome the limitations faced by CCD and indirect CMOS sensors, providing much higher dynamic range and frame rate, while maintaining single-electron sensitivity. In some cases, e.g. the hybrid pixel array detectors, these detectors are limited in size. (Medipix3 has 256x256 pixels [5], the EMPAD has 128x128 pixels [6]). Therefore, it is of great value to be able to expand the effective size of a small detector.

In this presentation, we demonstrate a method to increase the effective size of a small detector by capturing and combining sub-regions of reciprocal space. We will highlight how to script the microscope to shift the diffraction pattern across the detector. Taking control of the diffraction shift also allows more flexibility in practical application. For example, this technique can be used to capture only certain segments of reciprocal space to save time and storage. It is possible to acquire along an annulus of the diffracted intensity to avoid the central beam and capture only high angle diffuse scattering. In this case, the diffraction pattern can be shifted in a ring pattern instead of a normal grid pattern. Furthermore, we will show how this technique expands a small detector’s application space. Electrons all the way out to the objective lens bore can be captured with a high sampling rate. Limitations and challenges of acquiring such data will also be discussed.

As for reconstruction, we will highlight our automated routine to accurately and precisely align the data post-acquisition. We will also examine the challenges associated with reconstructing the datasets, including rotating the pattern shift axes and lens hysteresis. Consistency of the alignment will also be discussed. We also explore the accuracy of the reconstruction as a function of pattern exposure time and sub-pattern overlap to help guide experimental parameters. Finally, we will also show that being able to capture the Bragg peak intensities at the same time as diffuse scattering provides opportunities for quantitative diffraction analysis.

References:

[1] Fundenberger et al., Ultramicroscopy 96 (2003), P127-137
[2] Muller et al., Appl. Phys. Lett.101(2012), P212110
[3] Ryll et al., Journal of Instrumentation 11(2016), P04006
[4] Lupini et al., Microsc. Microanal 21(2015), P1219-1220
[5] MW Tate et al., Microsc. Microanal 22 (2016), P237
[6] Mir et al., Ultramicroscopy 182(2017), P 44-53
[7] We acknowledge support for this work from the Air Force Office of Scientific Research (FA9550-17-1-0225. This work was performed in part at the Analytical Instrumentation Facility (AIF) at North Carolina State University, which is supported by the State of North Carolina and the National Science Foundation (award number ECCS-1542015). The AIF is a member of the North Carolina Research Triangle Nanotechnology Network (RTNN), a site in the National Nanotechnology Coordinated Infrastructure (NNCI).

We determine the correlation between the structural heterogeneity and unprecedented detailed parameters of medium range ordering (MRO) using 4-dimensional scanning transmission electron microscopy (4-D STEM). The details of MRO constitute nano-scaled structural heterogeneity, such as the type, size, and fraction of MRO structures, which are directly correlated with the important properties, including the ductility and glass forming ability of Zr-Cu-Co-Al metallic glasses. In this study, the quantitative analyses of the data are enabled by our 4-D STEM using the new-generation Electron Microscopy Pixel Array Detector (EMPAD). This new detector can not only support high dynamic range, which is necessary for the quantification, but also take nanodiffraction patterns continuously throughout a large sample area with oversampling. Thus, it allows us to obtain the statistically reliable information on the structural heterogeneity at the nanoscale. We confirm the fact that the important properties of Zr-Cu-Co-Al MGs and Ti-Zr-Ni-Be MGs are directly related to the determined MRO parameters. In the Zr-based MG system, the smaller and more diverse MRO types result in high ductility, and more structurally frustrated MRO type relates to the higher glass forming ability. In the case of Ti-based MGs, the addition of Be in Ti-Zr-Ni can break down icosahedral-MRO structure, which is our hypothesis based on the angular correlation analysis.

The spatial atomic structure of PEN is accessible for the first time with ptychographic reconstruction under low dosage illumination.

Introduction
Organic crystalline structures have been determined by diffraction methods that only generate average structural information from the characterization region. However, crystal structure and orientation in polymers, which depends on e.g. temperature, interfacial interactions and strain, will vary locally. To resolve the detailed microstructure, high resolution imaging is required.
In a case study of annealed thin films of polyethylene naphthalate (PEN) [1], crystalline domains were found to vary with the depth from a surface. We use PEN as a model system to demonstrate high resolution imaging of polymer crystals. High resolution imaging of polymers is limited by the low contrast generated by light elements and beam damage from the high-intensity illumination that is used to get sufficient resolution. Low dose imaging techniques applied here have proved very helpful to acquire data from the pristine structure before it is significantly altered by radiolysis.

Ptychography
A thin polymer specimen can be treated as a weak phase object. When the electron probe scans through a thin specimen consisting of light elements, the amplitude is not expected to change significantly, while the phase contrast will offer more useful information.
Ptychography [2] is a dose efficient imaging technique with phase retrieval capacity by making use of a 4D STEM dataset. It has these advantages: i) Phase can be retrieved by Wigner-Distribution Deconvolution, and lens aberration can be corrected in post-acquisition computation. ii) Depth information can be obtained with optical sectioning, from the 4D dataset recorded in only one scan. This is very useful for beam sensitive materials. iii) Equipped with a fast-pixelated detector, the dataset recorded in low dose illumination can be adjusted to give sufficient signal for reconstruction.

High resolution image of PEN with discrepancy from XRD
Figure. 1 shows a ptychographic image of a grain in PEN. Many of the spatial frequencies present in the image, such as the (200) spacing, agree with previous models derived from X-ray diffraction studies. Noticeable in Figure. 1, however, there are some details between the (200) planes that are not predicted from the previous model. This information can only be uncovered by direct imaging.
This technique will enable the study of crystallography and defects of soft matter, hence assist us understand the property-structure relation of synthetic polymers at atomic level, and potentially the structure of natural macromolecules, e.g. protein, DNA, RNA.

Reference
[1] Shinotsuka, K. & Assender, H. In situ AFM study of near-surface crystallization in PET and PEN. J. Appl. Polym. Sci. 133, (2016).
[2] Yang, H. et al. Simultaneous atomic-resolution electron ptychography and Z-contrast imaging of light and heavy elements in complex nanostructures. Nat. Commun. 7, 12532 (2016).

Label-free spectroscopy and microscopy of live cells offers the capability to examine in-vivo conditions of cellular functions and structure providing valuable insight into biomolecular components of cells [1]. The physicochemical insulation of a samples' biologically relevant environment and instrumentation is critical for the proper understanding of underlying processes and reactions for label-free investigations. The recent development of two dimensional (2D) materials, i.e. graphene, makes it possible to isolate the sample environment from UHV conditions using these materials as molecularly impermeable, electron transparent membranes [2]. The latter opens the application of vacuum-based analytical techniques such as soft x-rays and electrons for in vivo cellular studies. Herein we report the design and fabrication of graphene encapsulated liquid cell platform designed to be suitable for in-situ studies of hydrated biological samples. The encapsulation with graphene allows the sample to be intimately confined and have its interstitial fluid environment preserved while becoming suitable to be studied with multiple analytical techniques [3]. This approach reduces the parasitic adsorption and scattering of the surrounding matrix while still providing nourishment to the cells for an enhanced lifetime [4]. The platform consists of an array of Si3N4 windows on which the samples are deposited allowing for combinatorial batch analysis. The platform has a small cross-section with photon and electrons making it compatible with a wide range of spectroscopies and microscopies. Implementation of sample immobilization with a biocompatible patternable hydrogel media allows for an extended lifetime of realistic conditions and potentially extended cellular lifetimes under measurement. Preliminary results into the viability of graphene encapsulation for biological samples were conducted with fluorescence optical, SEM, and FTIR microscopies.

References:
1. C. W. Freudiger et al, Science, 322 (2008) p. 1857-1861.
2. A. Kolmakov et al, Nature Nanotechnology 6 (2011) p. 651.
3. A. Yulaev et al, Advanced Materials Interfaces, 4, (2016) p.1600734
4. A. Matruglio et al, Journal of Instrumentation, 13, 05 (2018), p. C05016

Solution processing is a common first step in optoelectronic device fabrication, as it is a low-cost method with considerable flexibility. While it has been observed in several systems that solution chemistry affects both the morphology and properties of the deposited material, the manner in which these effects arise is rarely well understood; this is due in large part to the experimental difficulty of characterizing materials in solution. Here we demonstrate how cryo-electron microscopy (cryo-EM) can be used to overcome this challenge, sharing results from two technologically relevant systems: chalcogenide glasses and lead-halide perovskites. We show multiple cryo-EM methods, including single particle analysis, and discuss how an informed choice of method can reveal unique information about each material. Our results shine a light into the black box of solution processing, ultimately enabling better control over film morphology and optoelectronic performance.

Electronic and ionic transport in semiconducting polymers depends strongly on both molecular structure and mesoscale morphology. Transmission electron microscopy (TEM) studies of semiconducting polymers have proven to be challenging due to their inherent beam sensitivity and the interconnectedness of the physical processes occurring at various length scales. The complex, often hierarchical, internal structure of these molecular systems requires characterization tools that allow us to correlate structure from the nanoscale through the mesoscale all the way to their macroscopic properties. We leverage advances in high-resolution TEM (HRTEM) hardware, such as ultra-sensitive detectors and the use of cryogenic temperatures, and parallel computing data analytics tools to characterize the mesoscale organization of semiconducting polymers across areas similar in size to electronic devices (~ 10 μm²). The vast amount of data acquired from HRTEM imaging enables a statistical approach, which is necessary given the inherently disordered nature of semicrystalline materials.

We directly image the structure of a series of semicrystalline conjugated polymers using a fast read-out direct detector to record dose-fractionated exposure series (i.e., “movies”). The dose-fractionated exposure series are then analyzed using Fourier space image processing techniques and computer-based corrections for beam-induced specimen motion. The structure is spatially resolved through mapping of the alkyl (100) or backbone (001) spacing reflections, revealing the rich, peculiar nature of the nanoscale organization of semiconducting polymers. Through analysis of the local orientational correlations, we provide insights into the mesoscale organization of these materials. In addition, constructed orientation maps and director field lines with nanoscale resolution (~3 – 5 nm) allow us to visualize the alignment of molecular chains and understand electronic connectivity. The results show how beam-sensitive organic materials can be visualized and characterized at the nanoscale level using low-dose HRTEM to uncover their long and short-range organization, which is essential for understanding their macroscopic functional properties.

Organic semiconductors are of interest for a variety of applications such as biocompatible electronic devices, transistors, batteries, and solar cells. The morphology and self-assembly of these materials strongly influence charge and ion transport. While our fundamental understanding of structure-function relationships has improved greatly in recent years, significant questions remain about the nature of order at the nanoscale, the role of mesoscale organization, and its connection to the device transport properties. Here I discuss recent work examining the structure of a high-mobility semiconducting polymer previously thought to be amorphous. Thin-film synchrotron scattering, x-ray free-electron laser nano-diffraction, cryo-EM, and 4D-STEM are combined to demonstrate a rich, hierarchal organization with significant local fluctuations in molecular packing and signs of homoepitaxial order. These results combined with molecular modeling suggest a range of intra- and inter-molecular defects that break translational order but allow good long-range electronic coupling necessary for high mobility. This case highlights the need to develop and apply new high-resolution structural probes in heterogenous systems to self-consistently understand structure in disordered materials.

Sub-Ångström-resolution STEM has become a routine tool to study radiation-hard inorganic crystals, but only recently has it been achieved at cryogenic temperatures. This advancement has provided the capability to capture exotic electronic states, such as charge ordering [1-3]. However, precisely tracking the emergence of low-temperature states requires high-resolution and precision STEM across a broad range of temperatures, cryogenic and intermediate, which remains a serious challenge. Here, we demonstrate a new side-entry, dual-tilt cryogenic holder (HennyZ) that combines liquid nitrogen cooling with fast, local heating via a 6-pin MEMS device. This enables intermediate temperature studies at atomic resolution and the flexibility for more complex experiments, such as coupling with in situ biasing [4]. Sub-Å resolution is enabled by consistently low drift rates of 0.3-0.4 Å/s in the full temperature range from ~100-1000 K. Furthermore, the local heating in this continuously variable temperature (CVT) cryo-STEM approach allows the temperature to be cycled repeatedly and rapidly within a single experiment, affording novel dynamic studies. Understanding structural changes in materials at the atomic scale and at temperatures where they occur is critical for manipulating quantum states. We demonstrate CVT cryo-STEM as an effective local probe for revealing unexpected structural phenomena, such as metastable intermediate phases, in Nb₃Br₈—a 2D magnetic material [5,6].

References:
[1] I. El Baggari, et al., Proc. Natl. Acad. Sci. U.S.A. 115 (2018), p. 1445.
[2] B.H. Savitzky, et al., Ultramic. 191 (2018), p. 56.
[3] A.M. Minor, et al., MRS Bull. 44 (2019), p. 961.
[4] B.H. Goodge, et al., Microsc. Microanal. First View (2020).
[5] C. Pasco, et al., ACS Nano 13 (2019), p. 9457.
[6] This work is supported by PARADIM, an NSF-MIP (DMR-1539918), and NSF DMR
1429155 & DMR-1719875.

Oriented attachments are recognized as a common pathway for crystal growth in both natural and synthetic systems, leading interface free or twin related structures. The occurrence of 5-fold twinned nanoparticles leads to unique properties. However, nearly 200 years after their discovery, the formation mechanism is still ambiguous. Using in situ high-resolution transmission electron microscopy combined with molecular dynamics simulations, we demonstrate that 5-fold twinning occurs via repeated oriented attachment of ~3 nm Au nanoparticles. These oriented attachment events create high-energy grain boundaries, which accumulate strain during atomic rearrangements and consequently decompose via nucleation and growth of a special class of twins whose net strain is zero, inducing the 5-fold twin structures. The results provide a quantitative understanding of the 5-fold twinning process; this knowledge provides guidance for interpreting and controlling twin structures and morphologies of a wide range of materials.

Imaging chemical reactions in solution is essential for the understanding complex mechanisms involved in nanomaterial synthesis. Liquid phase electron microscopy (LP-EM) is an emerging technique capable of directly monitoring morphological changes of materials in solution with a unique combination of temporal and spatial resolution. LP-EM has been used to study the reaction mechanisms such as metal nanoparticle formation, polymer self-assembly, wet etching of semiconducting materials, growth of metal−organic frameworks, nucleation of calcium carbonate and Li electrodeposition. These studies provided an unprecedented insight into nanoscale reaction mechanisms that have significantly improved our understanding of nanomaterial synthesis. The primary challenge for the field of LP-EM is understanding the role that the electron beam plays in the observed mechanism. It is common to perform control experiments to establish a critical dose limit with which a process can be observed. A consequence of having to work below a critical dose limit is that LP-EM has mostly been used to monitor reactions over a period of seconds to minutes, despite the fact that many reactions take place over hours to days. Furthermore, even after a critical dose has been established, it is impossible to completely rule out the role of the electron beam using LP-EM experiments alone. Consequently, it is essential to study reaction mechanisms using multiple imaging modalities. Cryo-EM can also provide direct high-resolution information on reaction mechanisms in solution. The advantage of cryo-EM is that electron beam effects are well understood and, using appropriately low doses, are insignificant. The disadvantage of cryo-EM is that it is impossible to know the history or future of any individual particle being imaged. Consequently, the data on reaction mechanisms has to be presented by selecting images that the user determines to be representative of the process occurring in solution. For simple processes that evolve homogenously, it is possible to determine a mechanism with high confidence using cryo-EM alone. However, for complex processes that evolve heterogeneously, the confidence with which mechanisms can be determined by cryo-EM alone is low. Here, a simple process is defined as a process with a single step or single pathway and a complex process is defined as a process with multiple steps or multiple pathways. A homogenous evolution is characterized by a process where only one particle state is observed at each sampled time point, while a heterogeneous evolution is characterized by a process where multiple states are observed at each sampled time point. Consequently, cryo-EM is limited in its ability to resolve complex heterogeneous processes. Despite the obvious benefits of combining LP-EM and cryo-EM, most studies using these techniques have been only conducted with a single EM modality. Here we discuss the application of distributed electron microcopy for studying reaction mechanisms in solution. The term distributed is used because the data required to understand the mechanism is shared between multiple imaging modalities and sample preparation methods. The strength of this approach is that our proposed hypothesis is consistent with data taken from experiments with complementary benefits and limitations.

Electron microscopy has progressed extraordinarly to become a powerful tool for studying nanostructured catalysts at the atomic-scale. Advances in electron optics and detection have made atomic-resolution electron microscopy capable of resolving the three-dimensional surface structure of the complex catalytic nanomaterials. In parallel, the introduction of differentially pumped electron microscopes and micro-electro-mechanical-system devices has enabled in situ observations of catalysts during the exposure to gasses at pressures of up to atmospheric levels and temperatures of up to several hundred centigrade as well as operando studies by concurrent measurements of catalytic functionality. While these developments build a foundation for time-resolved observations of catalysts at the atomic-scale, such observations are challenged by electron-beam-induced alterations of the catalysts and mass-transport phenomena in the fluid phase. In this contribution, I will outline electron microscopy advances addressing the chemical relevance of atomic-scale visualizations. Moreover, I will showcase these “live” visualizations for exploring the dynamical and functional behavior of complex catalytic nanomaterials.

Transmission electron microscopy (TEM) is a powerful materials’ characterization technique for investigating their structure and properties at nanoscale dimensions. It has become an indispensable tool in the present era of nanoscience and nanotechnology. In this study, we employed the energy-filtered TEM (EFTEM) mode to investigate the effect of in-situ heating on the thin-layers of a stimuli-responsive (temperature) poly(N-isopropylacrylamide) (pNIPAm) which has been covalently attached to hollow titania (TiO₂) nanowires. The temperature-responsive behavior was characterized by observing the morphology and composition of samples as a result of heating. The pNIPAm has the swelling to shrinking transition temperature defined as lower critical solution temperature (LCST) of ~32 ^○C at standard conditions. We characterized the morphology and composition of the samples below and above the phase transition temperature of the pNIPAm by applying in-situ heating inside TEM. The EFTEM imaging and elemental mapping were applied at room temperature to determine the conformity of the pNIPAm layer and the elemental distributions of all elements present in samples. In the analysis, the elemental distribution of the element carbon (C) was attributed to the presence of the pNIPAm layer in the samples. The TiO₂ nanowires regions in the samples were identified by the overlapping of generated titanium (Ti) and oxygen (O) elemental maps. The EFTEM analysis was then repeated at the same sample regions by applying incremental in-situ heating, during which the temperature of samples was increased from ambient (23 ^○C) to high (100 ^○C) temperature with an interval of 10 ^○C. The acquired results revealed that the phase transition temperature of the pNIPAm came out to be 95 ^○C inside the TEM, which was higher than the expected value. Similarly, during the cooling of samples inside the TEM, the pNIPAm did not revert to the original morphology once the samples reached room temperature. However, the pNIPAm was found to recover the morphology by cooling the samples in standard atmospheric conditions. In conclusion, our findings show that the pNIPAm may exhibit a different swelling to shrinking behavior inside the TEM than under standard atmospheric conditions. The discrepancy is believed to be due to the rate of heat transfer between the surrounding environment and the samples and the relative humidity at ambient conditions, which evidently play a critical role for the pNIPAm transition from swelling to shrinking as well as from shrinking to swelling phase.

Visualization of the dynamic structure of the liquid, essential for studying the virus and cell activity, has remained elusive for decades due to lack of an effective tool for direct visualization beyond micrometer resolution. Here, we designed a simple liquid-cell for encapsulating the liquid state of sodium for transmission electron microscopic (TEM) observation. The real-time dynamic structure of the liquid-vapor interface was imaged and videoed by TEM on the sample of electron irradiated sodium chloride (NaCl) crystals, a well-studied sample with low melting temperature and quantum super-shells of clusters. The nanometer resolution images exhibit the fine structures of the capillary waves, composed of three zones of structures and features, i.e. flexible nanoscale fibers, nanoparticles/clusters, and a low-pressure area that sucks the nanoparticles from the liquid to the interface. Although the phenomenon was observed based on irradiated NaCl crystals, the similarity of the phenomenon to predictions suggest our first-time videos of the dynamic structure might be useful in validating long-debated theoretical models of the liquid-vapor interface. By implemented the technique to biology, we imaged virus and cell in wet for insights into the 3D structure of the intermediates of viral infection cell.

The oxygen evolution reaction (OER) plays an important role in energy conversion and storage applications, such as metal–air electrochemical cells and the production of hydrogen from water electrolysis. Metal oxide nanoparticles acting as high-quality electro-catalysts allows for anionic redox of oxygen at the particle surface. [1] However, this reaction may lead to considerable dissolution and degradation of the nanoparticle catalyst. Perovskite oxides have been previously demonstrated that, by enhancing the hybridization between the O 2p and transition metal 3d orbitals, the OER activity can be enhanced significantly. [2] KOH electrolyte with Fe impurities have been found to significantly increase the activity of Ni-based perovskite oxides electrocatalysts. [3] In this contribution, we evaluate the role of Fe impurities together with Sr-doping on LaCoO₃ based materials on the activity and stability for OER in alkaline electrolyte using combination of scanning transmission electron microscopy (STEM) imaging, electron energy loss spectrum (EELS) and energy dispersive x-ray spectroscopy (EDXS). Our analysis is conducted on the same particles on pristine and after cycling (Identical-Location TEM; ILTEM), which enables us to track the effect of electrochemical cycling on the particles. In addition to the La_1-xSr_xCoO₃ catalysts, we also studied MnWO particles decorated with Ir catalyst nanoparticles using the ILTEM approach. From the comparison of the same particles on pristine and after cycling, the dissolution of the Ir nanoparticle on the surface of MnWO particles was observed. The results will be combined with first principal calculation to show the effect of surface composition and structural evolution during activity of OER.

References:

[1] E Fabbri et al., Nature Materials 16 (2017), p. 925.
[2] JT Mefford et al., Nat Commun 7 (2016), p. 11053.
[3] L Trotochaud et al., Journal of the American Chemical Society 136 (2014), p. 6744.

Transmission electron microscopy (TEM) is a powerful tool for studying the properties of materials down to the atomic level. In many cases, the quantitative interpretation of images requires simulations based on atomistic structure models. Image simulation for TEM has typically relied on the independent atom model (IAM) that neglects bonding effects, which are measurable and important for understanding many systems. The rearrangement of charge due to bonding is directly detectable in HRTEM images and can be crucial for the interpretation of small differences in the atomic contrast [1]. Moreover, the recent rise in popularity of the four-dimensional (4D) scanning TEM has been partially due to promises of unprecedented sensitivity to electromagnetic fields [2]; however, such data call for rigorous theoretical interpretation through simulations.

Since all electrons and the nuclear cores contribute to the scattering potential, simulations that go beyond the IAM have relied on computationally highly demanding all-electron calculations. We have developed a method to generate ab initio electrostatic potentials when describing the core electrons by projector functions. Combined with an interface to quantitative image simulations, this implementation enables an easy and fast means to model electron scattering. By comparing simulated transmission electron microscopy images and diffraction patterns to experimental data, we can show an accuracy equivalent to earlier all-electron calculations at a much lower computational cost [3].

We have now implemented this method in a freely available open-source program based on Python, dubbed “abTEM” for ab initio Transmission Electron Microscopy. abTEM integrates directly with two popular open-source Python codes: the Atomic Simulation Environment (ASE) [4] for setting up atomistic models, and GPAW for calculating electrostatic potentials based on the real-space projector-augmented wave implementation of density functional theory [5]. A highly efficient first-principles description of van der Waals interactions can be included via the libvdwxc library [6].

abTEM is designed for accuracy and includes up to the second-order expansion of the multislice solution of the high-energy Schrödinger equation and uses the correct integration of the sub-slicing of the three-dimensional potential [7]. Due to integration with ASE, it is straightforward to model temperature in the frozen phonon approximation using realistic configurations from molecular dynamics simulations. abTEM works as a standard full-featured multislice simulation program, and we have additionally implemented features for speeding up simulations at some cost to accuracy.

We demonstrate the capabilities of our method by modeling twisted bilayer graphene, which has recently received considerable attention due to its extraordinary and highly tunable properties [8]. This system requires on the order of a thousand atoms to represent in simulations, which, to our knowledge, is the largest system ever to be used in a quantitative electron microscopy image simulation that includes an accurate representation of bonding effects. The differences are subtle due to the large contribution of the cores; hence a great deal of care needs to be taken when extracting charge rearrangements solely based on (S)TEM measurements.

More information about abTEM is available online [9].


References:

[1] S. Kurasch, Beilstein J. Nanotechnol. 2, 394 (2011)
[2] C. Ophus, Microsc. and Microanalysis 25, 563 (2017)
[3] T. Susi et al., Ultramicroscopy 197, 16 (2019)
[4] A.H. Larsen et al., J. Phys: Condens. Matter 29, 273002 (2017)
[5] J.J. Mortensen et al., Phys. Rev. B 71, 035109 (2005)
[6] A.H. Larsen et al., Model. Simul. Mater. Sci. Eng. 25, 065004 (2017)
[7] I. Lobato and D. van Dyck, Ultramicroscopy 156, 9 (2015)
[8] M. Yankowitz et al., Science 363, 1059 (2018)
[9] https://github.com/jacobma/abTEM

The work was supported by the European Research Council (ERC) Grant No. 756277-ATMEN.

Aberration-corrected scanning transmission electron microscopy (STEM) is a powerful technique to directly observe atomic-scale structures and chemistry inside materials and devices. In the state-of-the-art STEM, a probe size of less than 0.5Å in diameter has been experimentally realized. By combining with new elaborate STEM detectors, we can not only image single atoms, but can also image electric field distribution of atoms through differential phase contrast (DPC) imaging techniques. It then becomes tempting to observe magnetic fields at very high resolution, which should strongly correlate with the properties of magnetic materials and devices. However, high spatial resolution observation of magnetic materials is essentially very difficult, because high magnetic fields (~2T) are always exerted on samples inside the magnetic objective lens. In recent years, we have developed a new magnetic objective lens system that realizes a magnetic field free environment at the sample position. Using this new objective lens system combined with the state-of-the-art higher order aberration corrector, atom-resolved imaging in a magnetic field free environment has been achieved. The details of this magnetic-field-free STEM and the material application results will be presented in the talk.

Nanoscale properties rule the emission and absorption of light by nano-structures, for example the shape of metallic nanoparticles or the atomic design of semiconductor heterostructures. To overcome the diffraction limit of light (≈400 nm) one of the best techniques is electron microscopy. Indeed, the electron beam probes nano-object with sub-nanometer resolution. Both Cathodoluminescence (CL) and electron energy loss spectroscopy (EELS) study the optical properties of nanostructures. They probe, for example, the luminescence properties for CL [1] and the local density of state for EELS [2]. Time-resolved electron microscopy is emerging since a decade, expanding the possibilities of these two techniques. In this presentation, we will use a unique time-resolved transmission electron microscope based on a cold-FEG electron gun [3]. This technology allows, among other things, to reach a spatial resolution of 1 nanometer in pulsed mode. We will discuss the latest results of electron energy gain spectroscopy and the acquisition of the first cathodoluminescence decay traces in a transmission electron microscope.
Electron energy gain spectroscopy is the study of the coupling of a fast-electron with the near field created by the interaction of light with the nanostructure [4]. The group of Zewail in Caltech did the first experimental study in 2009 [5] followed by a few groups [6]. In this presentation, we will discuss electron gain spectroscopy using the first cold-FEG transmission electron gun [7]. This unique TEM allows the highest spatial resolution and highest brightness, leading to a spatial resolution of 1 nm and a strong interaction between the electron beam and the near field.
The development of time-resolved Cathodoluminescence (TR-CL) enabled the measurement of the lifetime of excited states in semiconductors with a sub-wavelength spatial resolution. It was used for example to measure the influence of stacking faults on the GaN exciton [12]. Recently, the first pump-probe cathodoluminescence experiment using diamond revealed the effect of electron excitation on the nitrogen vacancy color center [15]. These results demonstrate that TR-CL is essential to study the correlation between semiconductor optical and structural properties (composition, defects, strain…). While all these pioneering studies were done using a scanning electron microscope, the improvement of the spatial resolution and the combination with other electron-based spectroscopies offered by transmission electron microscopes will be a step forward for TR-CL. In this presentation, we will discuss the first time-resolved cathodoluminescence experiments within a transmission electron microscope.

[1] L. F. Zagonel et al. Nano Lett., vol. 11, pp. 568, 2011.
[2] A. Losquin et al., Nano-Letters, vol. 15, pp. 1229, 2015.
[3] F. Houdellier et al, Ultramicroscopy, vol. 186, pp. 128, 2018.
[4] F. J. G. de Abajo et al, Phys. Rev. Lett., vol. 100, 106804, 2008.
[5] B. Barwicketal, Nature, vol. 462, no. 7275, pp. 902, 2009.
[6] A. Feist et al, Nature, vol. 521, no. 7551, pp. 200, 2015.
[7] F. Houdellier, et al, Ultramicroscopy, vol. 186, pp. 128, 2018.
[8] B. G. Yacobi et al, Springer, 1990.
[9] L. F. Zagonel et al. Nano Lett., vol. 11, pp. 568, Feb. 2011.
[10] L. H. G. Tizei et al, Phys. Rev. Lett., vol. 110, p. 153604, 2013.
[11] S. Meuret et al., Phys. Rev. Lett., vol. 114, no. 19, 2015.
[12] P. Corfdir et al., J. Appl. Phys., vol. 105, 043102, 2009.
[13] R. J. Moerland et al, Opt. Express, vol. 24, 24760, 2016.
[14] X. Fu et al. ACS Nano, vol. 8, 3412, 2014.
[15] M. Solà-Garcia et al, ACS Photonics 7, 232 (2020).

Electrons, used as probes in electron microscopes, are of primal importance to measure material properties with unequalled combined spatial and energy resolution. For instance, inelastically scattered electrons can be collected and energetically separated in order to obtain spectra as a function of their energy loss or, with the same purpose, the radiated light can be collected in the far field and filtered in polarization and frequency. In particular, electron energy loss spectroscopy (EELS) can be considered the richest source of information as it carries the influence from the surface-sustained modes as well as from the bulk response of the system. Additionally, the synchronization of ft laser illumination with electron pulses at the sample opens the routes to study the ultrafast dynamics of nanostructured materials and their influence on optical near-fields. The evanescent components, produced by the excitations in the material, efficiently overcomes the unavoidable energy-momentum mismatch between free photons and free electrons yielding appreciable light-electron couplings. This strong interaction leads to multiple exchanges of quanta between the electron and the optical field, accompanied by a complex sub-fs dynamics. Based on this principle, photon-induced near-field electron microscopy (PINEM) is performed by analysing the resulting multiple gain and loss features in the electron spectra. PINEM experiments have so far relied on coherent light sources (i.e., lasers), for which the measured spectra are well reproduced by assuming sample bosonic excitations that are coherently populated with a large number of quanta. The probability of each electron spectral peaks associated with a net exchange of l quanta is then simply given by squared Bessel function. Additionally, this technique, performed with coherent light, has been combined with an electron free propagation of few millimeters to yield a temporal compression of the electron density of the order of attoseconds. Despite the numerous results obtained in this context, different aspects of the interaction between free electrons and confined light remain unexplored. For instance, the effects induced by the replacement of the coherent laser pulses with a quantum light source on electron spectra, are expected to show interesting physics significantly departing from the one previously studied.

In this work, we use quantum-optics methods in order to describe the quantum nature of the optical field and to understand how the resulting electron spectra strongly depend on the statistics of the sample excitations and their population, whose autocorrelation functions are directly retrieved from the ratios of electron gain intensities. In particular, in the limit of highly populated mode, we recover the usual results of classical PINEM for Fock and coherent states but we find a significant difference in the distribution of the gain and loss peaks of electron spectra in the case of thermal state. Furthermore, we study the role of quantum states of light, e.g. squeezed states, in shaping the longitudinal part of the electron wave function after a macroscopic propagation distance observing a faster compression of the electron density for light states with low phase uncertainty.

In conclusion, we consider this work of fundamental importance in the understanding of quantum many bodies excitations and their statistics as well as in the ability of tailoring the electron wave function via the use of light.

Over last decades, electron microscopy has become very powerful and versatile technique for nano- or atomic-scale imaging and spectroscopy [1]. Major advancements and outstanding capabilities have been made possible thanks better spatial and temporal control over the amplitude and phase of the wave function that characterizes the fast electrons used as sample probes. Control over the beam shape is commonly achieved by means of complex arrangements of magneto- and electrostatic electron lenses that enable sub-Ångstrom focusing and beam scanning, as well as correcting aberrations of electron optics. The phase of the electron wave function can be additionally modified by introducing static phase plates.

We envision an alternative to traditional electron-optics elements emerging by using optically-driven phase plates and enabling dynamical shaping of electron-beam wave functions both in space and time. This approach capitalizes recent experimental demonstrations of wave function control through optical near fields [2-4]. The electron-light interaction can in principle allow us to generate arbitrary electron beam shapes; as an application with high potential for improving the resolution of electron microscopes, we employ it here to correct for aberrations of standard electro- or magneto-static lenses used in current setups.

We demonstrate aberration correction via the interaction of the electron wave function with suitably profiled optical fields. We illustrate this concept by focusing on primary spherical aberration, which due to rotational invariance typically requires complex and expensive aberration correctors. Our work suggests that an optical corrector can help to significantly improve the focal spot and beam profile over the aberrated configuration, and that it can be in principle used to eliminate any undesired beam distortions in both standard and ultrafast (scanning) transmission electron microscopy. Optical aberration correctors could offer better versatility and compactness with respect to traditional corrector designs, and we foresee that they could open a new era of electron microscopy both in aberration correction and in the generation of on-demand electron beams.

[1] P. E. Batson, N. Dellby, and O. L. Krivanek, Nature 418, 617 (2002).
[2] B. Barwick, D. J. Flannigan, and A. H. Zewail, Nature 462, 902 (2009).
[3] A. Feist, K. E. Echternkamp, J. Schauss, S. V. Yalunin, S. Schäfer, and C. Ropers, Nature 521, 200 (2015).
[4] G. M. Vanacore, et al., Nat. Mater. 18, 573 (2019).

Ultrafast electron microscopy allows for the study of nanoscale material dynamics at picosecond time resolution. Electron pulses are typically generated using pulsed-laser induced photoemission from the electron cathode, and often multiple electrons are emitted within a single picosecond pulse. Here we study the dynamics of photon emission after excitation by dense collision cascades generated with controlled electron pulse densities. We use an ultrafast SEM cathodoluminescence (CL) microscope equipped with a pump-probe configuration, which enables a direct comparison between electron and light excitation.
Our pump-probe SEM-CL microscope is based on a 1-30 keV SEM equipped with a Schottky field-emission electron cathode. We use 250-femtosecond laser pulses at a wavelength of 258 nm to excite the W/ZrO electron cathode at a repetition rate from 200 kHz up to 25 MHz and create picosecond electron pulses by photoemission. By tuning the laser power we vary the average number of electrons per pulse from less than 1 up to 500. We record g⁽²⁾(τ) autocorrelation statistics of the CL emission at a wavelength of 450 nm from InGaN/GaN quantum wells using a Hanbury-Brown and Twiss interferometer with 70 ps temporal resolution.
Previous work using continuous and ns-pulsed electron beams has shown that strong bunching (g⁽²⁾(0) up to150) of CL photons can occur when exciting a sample with high-energy electrons, due to the fact that a single collision cascade can generate multiple CL photons [1,2]. Here, we extend this method to study bunching after excitation with ps electron pulses created using photoemission. We find that g⁽²⁾(0)-1 decreases inversely proportional to the number of electrons per pulse, similar to the trend with electron current found in previous work. We introduce an analytical and a statistical model to describe the bunching process and reproduce the measured dependence of g⁽²⁾(τ) on electron density in the ps pulses. Further analysis shows how the number of excitations/photons generated in each collision cascade can be derived from the data. The results suggest that despite the very high density of collision cascades, non-linear effects do not play a role in the plasmon generation, carrier diffusion and carrier recombination effects that follow electron impact on the InGaN/GaN quantum wells.
Next, we use the SEM-CL microscope in pump-probe configuration, by sending part of the femtosecond laser into the SEM chamber. We use this to further study the dynamics of carrier generation and recombination in electron and laser-excited semiconductors [3]. We probe the decay dynamics of bandgap emission in GaAs and CZTS (Cu₂ZnSnS₄) upon ps electron excitation (at 5-30 keV) and 250 fs laser excitation (λ=345 and 517 nm). We study the variations in carrier decay dynamics as a function of laser power and electron pulse density, to deduce the contribution of the different radiative and non-radiative decay mechanisms in each case. These results allow us to compare carrier densities in dense collision cascades and optically excited volumes in these semiconductors.
Finally, we combine CL and PL measurements to perform pump-probe spectroscopy, in which we use a delay stage to control the time between arrival of electron and laser pulses on the sample. We use the electron as a pump and the laser as a probe, and vice versa, aiming to probe non-linear effects in the evolution of dense electron cascades generated by picosecond electron pulses.

[1] S. Meuret, L. H. G. Tizei, T. Cazimajou, R. Bourrellier, H. C. Change, F. Treussart, and M. Kociak. Phys. Rev. Lett 114, 197404 (2015)
[2] S. Meuret, T. Coenen, H. Zeijlemaker, M. Latzel, S. Christiansen, S. Conesa-Boj, and A. Polman, Phys. Rev. B 96, 035308 (2017)
[3] M. Solà-Garcia, S. Meuret, T. Coenen, and A. Polman, ACS Photonics 7(1), 232-240 (2020)

The electrolyte 2.8 M lithium bis(fluorosulfonyl)imide (LiFSI) in dimethoxyethane (DME) has demonstrated good cycling behavior and stable Li morphology upon plating and stripping for high-capacity Li batteries, which is thought to be due to the formation of a stable solid-electrolyte interphase (SEI) layer for good ionic transport. Characterization of this SEI layer on lithium metal has been limited, due to the air-sensitivity and beam sensitivity of these materials, as well as the incompatibility of the electrolyte with the high vacuum environment necessary for nanoscale characterization with electron microscopy techniques. New cryogenic stages and transfer tools have enabled the characterization of frozen solid-liquid interfaces for lithium metal structures. Here, we used cryogenic focused ion beam (FIB) and ultrashort pulse laser ablation to create a cross-section of the lithium metal anode in contact with the electrolyte at various plating/stripping cycles to understand the morphological evolution of the lithium and SEI. Additionally, with cryo-FIB we utilized lift-out procedures under cryogenic temperatures for extraction of a lamella sample for investigation using scanning/transmission electron microscopy with energy dispersive x-ray spectroscopy and electron energy loss spectroscopy. The nanoscale chemical maps have allowed for a greater understanding of the SEI components and how they redistribute with reaction to the lithium metal over the course of many cycles. These results present a step towards visualization of the intact battery stack. We have determined a failure mechanism beyond short circuiting of the lithium metal anodes, due to reaction of the lithium with this LiFSI in DME chemistry.
This work was funded by Sandia National Laboratories’ Laboratory Directed Research and Development program. It was performed, in part, at the Center for Integrated Nanotechnologies, an Office of Science User Facility operated for the U.S. Department of Energy (DOE) Office of Science. Sandia National Laboratories is a multi-mission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. DOE’s National Nuclear Security Administration under contract DE-NA-0003525. The views expressed in the article do not necessarily represent the views of the U.S. DOE or the United States Government.

Transition metal oxide heterostructures host many novel functionalities and have stimulated large interest due to the possibilities of tailoring the functionalities at the atomic layer scale. It is the complex interactions at the interfaces of epitaxial oxide systems that contribute to intriguing physical effects induced by the local variation of electronic and ionic species. In this work, we focus on high-temperature superconductivity (HTSC)[1] of La₂CuO₄-based iso-structural and non–iso-structural systems. In particular, we fabricate (i) La_1.6A_0.4CuO₄–La₂CuO₄ (A = Ca, Sr, Ba) metallic–insulating (M–I) bilayers [2] and (ii) (La,Sr)₂CuO₄–SrMnO₃–LaMnO₃–La₂CuO₄ multilayers [3] using atomic-layer-by layer oxide molecular beam epitaxy (ALL-Oxide MBE) [4].
In order to correlate local structure and superconducting properties, we extensively probe the interfaces using aberration-corrected analytical scanning transmission electron microscopy (STEM) techniques including high-angle annular dark-field (HAADF) and annular bright-field (ABF) imaging, electron energy-loss spectroscopy (EELS), and energy-dispersive X-ray spectroscopy (EDXS). For the atomic-resolution analyses, a JEOL JEM-ARM200F STEM equipped with a cold field-emission electron source, a probe C_s-corrector (DCOR, CEOS GmbH), a Gatan GIF Quantum ERS spectrometer and a large solid-angle JEOL Centurio SDD-type EDXS detector was used. STEM imaging and EELS were performed at probe semi-convergence angles of 20 mrad and 28 mrad, respectively. The collection angles for HAADF and ABF images were 75-310 mrad and 11-23 mrad, respectively. O-O picker tool [5] software has been utilized for STEM image quantification.
The choice of the dopant substantially influences the superconducting mechanisms as a consequence of the dopant distribution near the nominal M–I bilayers interfaces. In the case of Sr- and Ca-doping, sharp interfaces induce striking interface effects, i.e. electronic redistribution. Differently, for Ba-doped bilayer, HTSC is attributed to “classical” homogeneous doping determined by cationic intermixing.[6] Moreover, we demonstrate that sharper Sr-doped La₂CuO₄ interfaces can be achieved by the heteroepitaxial contacts with manganite layers. The dopant distribution in La₂CuO₄ is affected by the elemental intermixing at the first atomic monolayer of the interfacial LaMnO₃ contact. Different superconducting behavior (interface vs filamentary) at the two different interface systems are correlated with the interface sharpness and will be discussed in detail.

References
[1] A. Gozar, G. Logvenov, L. F. Kourkoutis, A. T. Bollinger, L. A. Giannuzzi, D. A. Muller, I. Bozovic, Nature 2008, 455, 782.
[2] Y. E. Suyolcu, Y. Wang, F. Baiutti, A. Al-Temimy, G. Gregori, G. Cristiani, W. Sigle, J. Maier, P. A. van Aken, G. Logvenov, Sci. Rep. 2017, 7, 453.
[3] G. Kim, Y. Khaydukov, M. Bluschke, Y. E. Suyolcu, G. Christiani, K. Son, C. Dietl, T. Keller, E. Weschke, P. A. van Aken, G. Logvenov, B. Keimer, Phys. Rev. Mater. 2019, 3, 084420.
[4] Y. E. Suyolcu, G. Christiani, P. A. van Aken, G. Logvenov, J. Supercond. Nov. Magn. 2020, 33, 107.
[5] Y. Wang, U. Salzberger, W. Sigle, Y. Eren Suyolcu, P. A. van Aken, Ultramicroscopy 2016, 168, 46.
[6] Y. E. Suyolcu, Y. Wang, W. Sigle, F. Baiutti, G. Cristiani, G. Logvenov, J. Maier, P. A. van Aken, Adv. Mater. Interfaces 2017, 4, 1700737.

Domain walls (DWs) in ferroelectrics constitute the two-dimensional (2D) boundary region between domains characterized by different values of the order parameter, i.e. the ferroelectric polarization. One of the most striking properties of DWs is their ability to carry a bound charge, due to the screening by the free carriers of the material. In ferroelectric oxides, such charged DWs behave as 2D conductive pathways embedded in an insulating matrix. This special feature has raised a vast interest in conductive DWs for their potential use in future nanoelectronics. In particular, as the DWs in ferroelectric oxides are often movable under an external electric field, [1,2] they represent a very promising element for novel rewritable circuitry.
Deepening our knowledge about the charge screening processes is of fundamental relevance to achieve a complete understanding of their conduction properties. However, the lack of experimental techniques that are directly sensitive to the localized charges with sub-nanometer resolution makes the identification and the quantification of the screening charges a challenging problem.
In this study, we perform an atomic-scale study of the origin of the enhanced conductivity in charged 109° tail-to-tail (T-T) domain walls in the archetype multiferroic BiFeO₃ (BFO), by combining state-of-the-art scanning probe techniques – such as piezo-response force microscopy (PFM) and conductive atomic force microscopy (cAFM) – with differential phase-contrast (DPC) scanning transmission electron microscopy (STEM). [3]
Our study was conducted on the heterostructure BiFeO₃ (120 nm)/La_0.7Sr_0.3MnO₃ (5 nm)/SrRuO₃ (1 nm)/SrTiO₃ grown by pulsed laser deposition. The ferroelectric domain pattern observed by PFM displays the typical contrast of 71° DWs with head-to-tail DWs, which are oriented along the main directions of the cubic substrate. Interestingly, strongly charged T-T DWs are generated at the merging points of 71° stripes with different orientations. Thanks to their topologically charged state, these charged DWs display enhanced conductivity.
The analysis of the atomic displacements from high-resolution high-angle annular dark-field (HAADF) STEM demonstrates that the charged T-T domain wall presents a mixed Ising-Néel structure and is composed of a strained core with a lateral thickness of 6-8 unit cells. The charged-state of the DW was initially assessed by computing the density of bound charge (ρ_B) starting from the polarization maps. In particular, we image a distribution of positive charges mostly localized at the DW, with localized maxima separated a few nanometers from each other. The ρ_B reaches a maximum value of 1 oxygen vacancy (O_v) every 4 unit cells.
In order to directly image the topological charges at the DW, we employ the differential-DPC (dDPC) imaging technique. The differential form of DPC – thanks to its enhanced sensitivity to low Z species – has been recently been proven to be effective for imaging light elements, such as N, [4] Li, [5] and O. [6] By using dDPC, we successfully image the O columns in the BFO layer and thus map the O_v distribution along the DW core.
Our results elucidate the nature of the conductive DWs in BFO, providing new insights about their structure and charge screening phenomena. In particular, combining atomic displacement mapping and dDPC-STEM imaging we can image and quantify the local charges that are responsible for their high conductivity. This work paves the way for future studies about the structure-properties relationship in such charged DWs.
[1] T. Rojac et al., Adv. Funct. Mater. 25, 2099, 2015.
[2] J. C. Agar et al., Nat. Mater. 15, 549, 2016.
[3] M. Campanini et al., Nanoscale 12, 9186, 2020.
[4] E. Yücelen et al., Sci. Rep. 8, 2676, 2018.
[5] A. Carlsson et al., Microsc. Microanal. 24, 122, 2018.
[6] S. Findlay et al., Microsc. Microanal. 25, 1732, 2019.
[7] The authors acknowledge funding from the Swiss National Science Foundation (project Nr. 200021_175926).

The competition between polar distortions and the mobile charge carriers is a problem of longstanding interest to the development of a quantum theory of ferroelectricity. It is also a key issue for polar superconductors, which are a promising route to unconventional Cooper pairing.
SrTiO₃ is an incipient ferroelectric in unstrained, stoichiometric form, but easily becomes ferroelectric when subjected to small perturbations, including chemical substitution or applied stress. Ferroelectricity precedes superconductivity in doped, strained SrTiO₃, suggesting a complex relationship between two instabilities. Using an atomic-resolution scanning transmission electron microscope, we show atomic-scale evidence of the interplay between itinerant electrons and the correlated polar distortions above Curie temperature.
The undoped, strained SrTiO₃ films show correlated polar nanodomains above Curie temperature, suggesting an order-disorder type ferroelectric phase transition. We observe a systematic suppression of the correlated polar distortions, i.e. nanodomains, and ferroelectricity with introducing Sm dopants. The itinerant electrons screen polar distortions and disrupt nanodomains above Curie temperature. Increasing the dopant density, strain field of the dopants overlap and randomize the polar distortions. The results provide direct evidence that the long-range Coulomb interactions, already present in the paraelectric phase, are driving the ferroelectric transition and are becoming increasingly short-ranged with doping. Moreover, we show that the disorder caused by the dopant atoms themselves presents a second contribution to the destabilization of the ferroelectric state. This disorder completely destroys the nanodomains at high dopant concentrations, thereby. The results thus provide evidence for two distinct mechanisms that suppress the polar transition with doping and the presence of both competing and collaborative microscopic interactions in a ferroelectric superconductor. We also observe an inverse relationship between the critical temperature of superconductivity and the magnitude of polar distortions. Understanding the nature of ferroelectric transition and its interplay with itinerant electrons is of interest for testing the different proposed theoretical models for superconductivity in SrTiO₃.

Realization of precision, close-loop processes requires an integrated suite of integrated hardware [1], software [2] and modeling [3], and materials synthesis/optimization. With spatio-temporal control, deterministic synthesis can be achieved in quantum devices, patterned assemblies, and tailored additive manufacturing of materials. Real time processing multidimensional (hyperspectral) data signals benefit from machine learning approaches presenting an opportunity for implementation of a feedback loop for manufacturing of critical technologies, from quantum computers to 3D printed turbine blades. In order to probe specific phenomena, stimulus and detection must be linked; without real time “learning,” key events are left undetected due to temporal resolution limits or failed temporal triggering alignments. Novel detector designs enable manufacturing processes to be viewed “in operando” at high acquisition speeds (upwards of 4000 fps).The combination of machine learning with these detectors makes automation for atomic-scale manipulation a realistic goal. Here we present the application of machine learning techniques to hyperspectral data to demonstrate a framework for understanding rapidly evolving phenomena using convolutional neural networks trained on >10,000 spectra collected at 400 frames per second during an electron beam manipulation of complex oxide structures. We extend these studies to despite very noisy data collected at 400 fps. unsupervised learning via feature extraction instead of comparing to known labels, which enables the analysis of situations when otherwise inaccessible transformations or metastable states occur. These techniques provide the ability to precisely and predictively tune complex materials toward. To that end, we will highlight extension of these techniiques to surface termination manipulation in 2D materials.

The use of scanning transmission electron microscopes to manipulate substitutional defects in graphene has recently been demonstrated and modeled using ground state molecular dynamics, but the role of electronic excitations induced through inelastic electron scattering in promoting these transformations has so-far remained unexplored. In this work[1,2,3] we probe the effects of electronic excitations on the structural dynamics of localized-defects induced by an electron-beam (a point-charge perturbation), such as substitutional Si/P in graphene quantum-dots as well as the recently observed mobile Si-C₃ defects in graphene nanoflakes. Particularly, the ground and excited state potential energy barriers for the structural transformations are evaluated using time-dependent density functional theory and reveal that the barriers for such localized structural transformations can be effectively lowered upon excitation to the defect-centered excited state through an electronically nonadiabatic pathway. Optically bright excited states in which the barrier is decreased are identified in the low energy region of the electronic spectrum, suggesting that photoexcitation can modulate the reactivity of defects in graphene under electron beam irradiation. Implications of this discovery on directing localized nanoscale transformations using electron-beams will be further discussed, especially in quantum-materials.
[1] “Understanding Beam Induced Electronic Excitations in Materials”, David B. Lingerfelt, P. Ganesh, Jacek Jakowski, Bobby G Sumpter, J. Chem. Theory Comput.,16, 2, 1200–1214 (2020)
[2] “Electronically Nonadiabatic Structural Transformations Promoted by Electron Beams”, David B. Lingerfelt, P. Ganesh, Jacek Jakowski, Bobby G Sumpter, Advanced Functional Materials, 1901901, (2019)
[3] “Localized Electronic Excitations in Silicon-Doped Graphene and their Role in Defect Diffusion”, David B. Lingerfelt, Tao Yu, P. Ganesh, Jacek Jakowski and Bobby G. Sumpter (under review)

The enhanced energy resolution of recently developed monochromators now make it possible to use electron energy-loss spectroscopy (EELS) in the scanning transmission electron microscope (STEM) to probe materials systems in new ways. This powerful spectroscopic capability combined with the small focused electron probe opens the door to explore photonic and phononic/vibrational excitations with high spatial resolution [1-3]. Here we describe recent advances leading to nanoscale and atomic resolution vibrational spectroscopy and novel approaches to probe guided photonic modes in dielectric particles. The fast electron in the electron microscope can excite vibrational modes in materials either through dipole interactions (similar to IR absorption spectroscopy) or impact scattering (similar to inelastic neutron scattering). The spatial localization of these two scattering mechanisms is very different. Impact scattering typically involves higher momentum transfers yielding a spatially localized spectral signal. In a crystal, high momentum transfer interactions are associated with exciting the short wavelength phonon modes at the Brillouin zone boundaries (BZB). Efficient excitation of these modes requires experimental conditions that favor Bragg scattering from the associated Miller planes. This can be easily accomplished in the conventional STEM EELS geometry by tilting the crystal into a suitable zone axis orientation and using a probe convergence angle that spans the Bragg angle associated with the BZB of interest. This convergence angle also permits a small probe to be formed which enables resolution of the associated Miller plane spacing. We have used this approach to demonstrate nanoscale and atomic resolution on a sample of Si, SiO₂ and BN [1]. EELS can also be employed to explore the characteristic photonic modes present in dielectric nanoparticles since the electron beam can provide a source of virtual photons to illuminate the object. We have explored the variation in photonic modes for high refractive index oxide particles such as TiO₂, Ta₂O₅ and CeO₂ [2]. The modes show up as peaks in the bandgap region of the spectrum. The energies of the modes are determined not only by the particle shape and size but also due to coupling between adjacent particles. Coupling between phonons and IR photons can give rise to phonon polaritons [3,4]. Polariton characteristics are also influenced by sample geometry, and so the polaritonic modes can be manipulated by changing the sample shape/size. Here we use STEM EELS to demonstrate variations in the polaritonic response on patterned SiO₂ thin films.

References:
[1] K. Venkatraman et al., Nature Physics 15 (2019) 1237-1241.
[2] Q. Liu et al., Phys. Rev. B 99 16 (2019) 165102.
[3] K. Venkatraman et al., Microscopy 67 (2018) i14 - 23.
[4] A. Konečná et al., Phys. Rev. B 98 20 (2018) 205409.
[5] The support from US National Science Foundation CHE-1508667 and U.S. DOE (DE-SC0004954), the use of (S)TEM at Eyring Materials Center at Arizona State University is gratefully acknowledged.

Advances in electron monochromation have led to a sea change in electron energy loss spectroscopy (EELS) over the last 6 years. Modern monochromators scanning transmission electron microscopes (STEM) have improved EELS energy resolution by an order of magnitude (from ~50 meV down to < 5 meV). This new level of energy resolution enables direct access to a host of new phenomena, such as phonons, vibrational modes, infrared optical excitations, and low-lying states in the electronic structure. Moreover, it also presents many new opportunities to probe difficult and complex systems by exploiting novel interactions between the electron beam and these ultra-low energy excitations. As a result, monochromated EELS is a highly versatile tool that can be applied to a wide range of different disciplines and materials systems and deliver exciting results.

In this talk, I will focus on three different materials systems where monochromated EELS presents a unique opportunity to provide new insights into important topics: biological, optical, and quantum materials. In terms of biological materials, I will demonstrate how we can the efficient long-range coupling of vibrational modes to the electron beam, to provide in depth spectroscopic analysis of ultra-beam sensitive biological materials. Furthermore, I will demonstrate how we can use the combine the natively high spatial and spectral resolution of monochromated EELS with various flavors of momentum-resolution to probe the dispersions of infrared quasiparticles directly at the nanoscale. Lastly, I will discuss how we can apply this technique to coupled networks of colloidal nanoparticles to examine emergent properties in quantum systems.

This work was performed at the Center for Nanophase Materials Sciences, which is a DOE Office of Science User Facility. This research was conducted, in part, using instrumentation within ORNL’s Materials Characterization Core provided by UT-Battelle, LLC, under Contract No. DE-AC05- 00OR22725 with the DOE, and sponsored by the Laboratory Directed Research and Development Program of Oak Ridge National Laboratory, managed by UT-Battelle, LLC, for the U.S. Department of Energy.

Vibrational electron energy loss spectroscopy (EELS) in the scanning transmission electron microscope (STEM) has over the recent years emerged as a versatile tool for characterizing molecular to bulk or surface vibrational modes in a wide range of materials systems, at the nanoscale [1]. The possibility of going beyond the nanoscale, reaching atomic resolution, has been discussed extensively in the community. We recently demonstrated such a capability using a so-called dark-field (DF) experimental geometry, suppressing the relative spectral contribution of elastic and delocalized inelastic dipole scattering by phonons, favoring instead beam electrons having undergone localized impact phonon scattering [2]. This approach allows for acquiring phonon structure images using the acoustic and/ or optical phonon EEL peaks [2] and, is equally applicable to polar [2] and non-polar materials [3]. Moreover, these results verify experimentally that annular dark field (ADF) Z-contrast imaging is based on beam electrons being scattered inelastically by phonons, in agreement with the quantum excitation of phonons model [2]. As a next step, we applied this methodology to probe the vibrational response of a single tri-valent substitutional Si atom in a single layer of suspended graphene. Spatially resolved EEL spectra show a characteristic vibrational signature associated with the Si impurity, localized to the defect atom itself. This is rationalized by means of density functional theory (DFT) calculations, attributing the measured response to the Si atom inducing pseudo-localized phonon modes: i.e. defect modes hybridizing with the graphene bulk continuum, setting up resonant states [3]. Possible applications of the demonstrated atomic resolution and single atom sensitivity, as well as practical aspects of the experimental approach, will be discussed.

[1] O. L. Krivanek et al., Ultramicroscopy 203 (2019) p. 60–67.
[2] F. S. Hage et al., PRL 122 (2019) 016103.
[3] F. S. Hage et al., Science 367 (2020) p. 1124–1127.

Atomic vibrations and phonons are an excellent source of information on nanomaterials that we can access through a variety of methods including Raman scattering, infrared spectroscopy, and electron energy-loss spectroscopy (EELS). In the presence of a plasmon local field, vibrations are strongly modified and, in particular, their dipolar strengths are highly enhanced, thus rendering Raman scattering and infrared spectroscopy extremely sensitive techniques.
In this contribution [1], we experimentally demonstrate that the interaction between a relativistic electron and vibrational modes in nanostructures is fundamentally modified in the presence of plasmons. This could be observed due to the current developments in electron optics, making possible the generation of sub-nanometer wide sub-10 meV electron beams with 60 keV primary energy. We finely tune the energy of surface plasmons in metallic micrometer-long Ag nanowires in the vicinity of hexagonal boron nitride flakes, making it possible to monitor and disentangle both strong phonon–plasmon coupling and plasmon-driven phonon enhacement at the nanometer scale. Because of the near-field character of the electron beam–phonon interaction, optically inactive phonon modes are also observed. Besides increasing our understanding of phonon physics, our results hold great potential for investigating sensing mechanisms and chemistry in complex nanomaterials down to the molecular level.
The strong coupling limit is reached between surface plasmons and phonon modes as the largest coupling constant, g, reached was 52 meV and the full width at half maximum of the two coupled excitations were 23 and 19 meV. The signature of strong coupling is observed along the whole nanowires, including for nanowire regions far away from any h-BN (point to the coherence of the coupled modes along the nanowire). The spatial dependence at nanometer scale of these modes will be described.
More interestingly, the presence of plasmons modify the way fast electrons couple to bulk modes of h-BN, leading to enhancement of their signal and the observation of usually “dark modes”. As an example, the ZO-phonon mode of hBN, usually not observed for electron beams perpendicularly incident on an h-BN flake, can be observed when close to metallic nanowires. We will describe how this enhanced signature changes as a function of the distance of the electron probe to the nanowire surface, varying in the tens-of-nanometers scale. Experiments have been performed with the ChromaTEM microscope (TEMPOS project, ANR-10-EQPX-50), a modified Nion HERMES200 operated at 60 kV. The electron beam was set to have an energy spread in the 6-10 meV range and a nanometer size. Experiments are under way to achieve similar enhancement in molecular systems.
If time allows, recent nano-optics experiments using ChromaTEM will be discussed.

The excitations of quasiparticles govern the physical properties of low-dimensional materials. Furthermore, assessing the irregular stacking sequences and behaviours at imperfect sections (such as boundaries and edges) is undoubtedly important to understand the performance of nanodevices. However, such local information has usually been averaged in conventional inelastic scattering techniques using x-ray, neutron, and light sources because of their diffraction limit. We demonstrate the nanoscale optical and vibrational spectroscopy of 1D or 2D materials by using a monochromatic electron source mounted in a transmission electron microscope (TEM).
All experiments were performed using a TEM (JEOL TripleC#2 at 30–60 kV) equipped with a Schottky field emission gun, a double Wien filter monochromator, and delta correctors. The energy resolution was set to a value less than 50 meV and allowed to access the quasiparticle excitations (i.e. phonon, exciton, and plasmon) of low-dimensional materials by electron energy-loss spectroscopy (EELS). The EEL spectra were collected using a GATAN GIF quantum spectrometer designed for low-voltage operations.
The spatial and momentum resolutions in electron microscopes balance each other and can be finely tuned with magnetic/electrostatic lenses. For instance, an atomically thin probe can be formed with an integration of a large momentum space. This allows us to extract local information from single defects. We have successfully measured the optical gap transitions from a defect of an individual semiconducting carbon nanotube [1,2]. Recently, atomically localized vibrational spectroscopic techniques have also been developed [3,4].
In contrast, an electron probe consisting of a parallel beam has a higher momentum resolution and provides dispersions of quasiparticle excitations. The EEL spectra are collected from each momentum transfer (q) by a pinhole aperture or a q-slit in a diffraction plane. Using this method, we have obtained phonon dispersions of graphene and h-BN from a single layer. Interestingly, the inelastically scattered electrons by phonon excitations provide sufficient signals at a large q even in the second and third Brillouin zone. Such signal enhancement due to a charge modulation at a large q helps draw a phonon dispersion curve of graphene even from a single layer. In addition, the spatial resolution of this method (a few to a few tens of nanometres, depending on its momentum resolution) is more advantageous than other dispersion measurement techniques such as inelastic x-ray spectroscopy. By maintaining the momentum resolution better than ±0.2 Å^-1, each vibrational mode can be assigned at a region of a few tens of nanometres. Therefore, the propagation of each phonon mode at defects such as edges in graphene has been unambiguously visualized [5]. Such local spectroscopy with a large flexibility could help unravel the defect physics of quantum matters.

References
[1] R. Senga, T. Pichler and K. Suenaga Nano Letters 16 (2016), 3661.
[2] R. Senga et al., Nano Letters 18 (2018), 3920.
[3] F. Hage et al., Phys. Rev. Lett. 122 (2019), 016103
[4] K. Venkatraman et al., Nature Physics 15 (2019), 1237
[5] R. Senga et al., Nature 573 (2019), 247

Electron energy loss spectroscopy (EELS) can be used to identify polymers and similar organics, but is limited by the rate at which materials damage when probed with ~100 kV electrons. Any modern scanning transmission electron microscope (STEM) is routinely capable of producing a sub-nanometer probe, but the fluence rate (electrons/nm² s) needed to achieve enough signal at this level of resolution cannot be tolerated by most organics, even at low temperatures. Many polymers show signs of damage well below 10 electrons/Ų, a threshold frequently identified for studying organic material. Even with ongoing detector improvements, “damage-free” EELS fine structure mapping of polymers with nanometer-sized pixels does not seem conventionally attainable. However, the ways in which an electron beam modifies a given polymer—and the resulting changes that can be observed with EELS—are arguably even more unique than the initial EELS fine structure.

The type and extent of beam-induced chemistry for a given polymer depends on at least dose, dose-rate, sample temperature, and vacuum conditions. These dependencies provide opportunities to exploit beam damage as a means for polymer chemistry identification. Beam-induced chemistry significantly precedes substantial mass loss and/or sample distortion in many polymers, particularly at cryogenic temperatures. The simplest option is to probe the fluence dependence by recording a series of STEM-EELS maps with a fixed current and temperature. These 4D fluence-dependent EELS maps can be collected with total fluences normally considered too high for accurate polymer identification, while at lower fluence rates, enabling spatial higher resolution mapping. This will be demonstrated for model polymer blends for <20 nm pixels, using a spectrometer with a CCD-based EELS detector.

Many studies have unintentionally mapped damaged polymers, and/or specifically measured damage evolution with EELS, but resolving and using the beam-induced evolution of sample chemistry to improve mapping is novel. Ultimately, spatial resolution is still limited by beam-induced changes: e.g., the rate of mass loss and the diffusion of generated radicals and gases. While application to polymer EELS will be shown, the approach is quite generic to materials with a measurable chemical evolution preceding significant sample morphology change.

In this study, we perform first-principles calculations to investigate the effect of isotopic substitutions in the vibrational modes of hBN flakes. Using density functional theory calculations the ground state atomic configuration and the dynamical matrix of hBN flake are obtained. The eigenmodes and eigenfrequencies are calculated by diagonalization of the dynamical matrix. Using semiclassical formalism, the interaction between a focused electron beam and vibrations is described. We employ this approach to predict the effect of isotopes on the spatially-resolved electron energy-loss spectra (EELS) of hBN flakes.

We study spontaneous and stimulated interactions between high-energy electrons and photons in nanoscale plasmonic near fields using Electron Energy Loss Spectroscopy (EELS), Cathodoluminescence (CL) spectroscopy and Photon-Induced Near-field Electron Microscopy (PINEM) in a Scanning (Transmission) Electron Microscope (S(T)EM). In recent years, EELS, CL and PINEM have enabled tremendous advances in probing (resonant) optical materials excitations with extreme spatial, temporal and energy resolution [1,2]. In EELS and CL, the time-varying evanescent electric field of the swift electrons serves as an ultra-broadband excitation source that can couple to a wide range of optical modes in a material [3]. The energy transfer during this interaction is a spontaneous process, resulting in a characteristic electron energy loss (EELS) and the subsequent emission of light (CL). Conversely, in PINEM, the electrons interact with an intense laser-driven near-field, giving rise to both electron energy-gain and -loss transitions by stimulated absorption and emission of photons at an energy determined by the laser frequency [4,5].

We demonstrate spatially-resolved EELS, CL and PINEM measurements of highly-localized dipolar optical fields at the sharp tip apexes of a chemically-synthesized Au nanostar [6]. We highlight similarities and differences between the measured EELS, CL and PINEM distributions, and discuss the role of the underlying excitation mechanisms driven either by the electron itself (EELS and CL) or the external laser (PINEM). Supported by electromagnetic boundary element method calculations for a simplified model geometry, we show that in the limit of a single isolated tip mode, spatial variations in the spontaneous and stimulated electron-photon interactions are purely determined by the modal electric near field profile, irrespective of the excitation source.

The interaction strength between free electrons and localized optical fields scales with the Fourier amplitude of the electric near field component along the electron trajectory at a single spatial frequency that is determined by the electron velocity [3,7,8]. As a consequence, the measured EELS, CL and PINEM distributions depend on the initial electron energy, with the maximum interaction strength shifting towards smaller electron energies for increasing field confinement. In general, the full three-dimensional (3D) reconstruction of an electric near-field distribution is therefore only possible by extensive tomographic measurements at multiple electron energies. Yet, using PINEM we show that for an axially-symmetric tip, we can determine the laser-induced surface charge density profile along the tip contour and thus obtain the desired 3D near-field distribution from a single measurement at a fixed electron energy and incident angle.

References:
[1] A. Losquin and T. T. A. Lummen, Front. Phys. 12, 127301 (2017).
[2] A. Polman, M. Kociak, and F. J. Garcia de Abajo, Nature Mater. 18, 1558-1171 (2019).
[3] F. J. Garcia de Abajo, Rev. Mod. Phys. 82, 209-275 (2010).
[4] B. Barwick, D. J. Flannigan, and A. H. Zewail, Nature 462, 902-906 (2009).
[5] A. Feist, K. E. Echternkamp, J. Schauss, S. V. Yalunin, S. Schaefer, and C. Ropers, Nature 521, 200-203 (2015).
[6] N. Pazos-Perez, L. Guerrini, and R. A. Alvarez-Puebla, ACS Omega 3, 17173-17179 (2016).
[7] A. Losquin and M. Kociak, ACS Photonics 2, 1619-1627 (2015).
[8] F. J. Garcia de Abajo and M. Kociak, New J. Phys. 10, 073035 (2008).

Confinement of magnetic structures geometrically as well as energetically, leads to novel and unexpected domain behavior. Lorentz transmission electron microscopy (LTEM) is ideally suited for quantitative analysis of magnetic domains and microstructure in functional nanoscale heterostructures. The current state of art LTEM enabled using aberration correctors allows for imaging down to sub-nanometer scale in field-free conditions. In this talk, the study of magnetic domain behavior in strongly interacting magnetic nanostructures namely artificial spin ices, curved 3D magnetic nanostructures, and functional materials that host skyrmions will be presented.
Artificial spin ice lattices consist of lithographically patterned arrays of interacting magnetic islands that exhibit magnetic frustration. We will present the results on emergence of magnetic excitations in such lattices which are topologically non-trivial as well as control of magnetic frustration and emergent ordering such as antiferromagnetic to ferromagnetic phase transition using lattice geometry. Furthermore, we will present a novel approach to understand such frustration behavior based on network topology that can be extended to use of graph neural networks to analyze and predict the behavior of spin ice lattices.
Topologically non-trivial spin textures such as skyrmions present unique opportunities to explore exciting fundamental phenomena, such as the topological Hall effect, as well as novel applications, such as skyrmion-based spintronics. The application of LTEM to visualize and understand the magnetic spin textures in multilayer [Pt/Co]_x thin films that can host skyrmions as well as controlled fabrication of skyrmions using Ga ion implantation will be discussed. In-situ magnetization reversal was used to understand the stability of the skyrmions.
Finally, we will also present a new approach based on machine learning for phase retrieval in LTEM that improves the achievable phase sensitivity and spatial resolution.

Advances in electron monochromation have led to a sea change in electron energy loss spectroscopy (EELS) over the last 6 years. Modern monochromators scanning transmission electron microscopes (STEM) have improved EELS energy resolution by an order of magnitude (from ~50 meV down to < 5 meV). This new level of energy resolution enables direct access to a host of new phenomena, such as phonons, vibrational modes, infrared optical excitations, and low-lying states in the electronic structure. Moreover, it also presents many new opportunities to probe difficult and complex systems by exploiting novel interactions between the electron beam and these ultra-low energy excitations. As a result, monochromated EELS is a highly versatile tool that can be applied to a wide range of different disciplines and materials systems and deliver exciting results.

In this talk, I will focus on three different materials systems where monochromated EELS presents a unique opportunity to provide new insights into important topics: biological, optical, and quantum materials. In terms of biological materials, I will demonstrate how we can the efficient long-range coupling of vibrational modes to the electron beam, to provide in depth spectroscopic analysis of ultra-beam sensitive biological materials. Furthermore, I will demonstrate how we can use the combine the natively high spatial and spectral resolution of monochromated EELS with various flavors of momentum-resolution to probe the dispersions of infrared quasiparticles directly at the nanoscale. Lastly, I will discuss how we can apply this technique to coupled networks of colloidal nanoparticles to examine emergent properties in quantum systems.

This work was performed at the Center for Nanophase Materials Sciences, which is a DOE Office of Science User Facility. This research was conducted, in part, using instrumentation within ORNL’s Materials Characterization Core provided by UT-Battelle, LLC, under Contract No. DE-AC05- 00OR22725 with the DOE, and sponsored by the Laboratory Directed Research and Development Program of Oak Ridge National Laboratory, managed by UT-Battelle, LLC, for the U.S. Department of Energy.

The enhanced energy resolution of recently developed monochromators now make it possible to use electron energy-loss spectroscopy (EELS) in the scanning transmission electron microscope (STEM) to probe materials systems in new ways. This powerful spectroscopic capability combined with the small focused electron probe opens the door to explore photonic and phononic/vibrational excitations with high spatial resolution [1-3]. Here we describe recent advances leading to nanoscale and atomic resolution vibrational spectroscopy and novel approaches to probe guided photonic modes in dielectric particles. The fast electron in the electron microscope can excite vibrational modes in materials either through dipole interactions (similar to IR absorption spectroscopy) or impact scattering (similar to inelastic neutron scattering). The spatial localization of these two scattering mechanisms is very different. Impact scattering typically involves higher momentum transfers yielding a spatially localized spectral signal. In a crystal, high momentum transfer interactions are associated with exciting the short wavelength phonon modes at the Brillouin zone boundaries (BZB). Efficient excitation of these modes requires experimental conditions that favor Bragg scattering from the associated Miller planes. This can be easily accomplished in the conventional STEM EELS geometry by tilting the crystal into a suitable zone axis orientation and using a probe convergence angle that spans the Bragg angle associated with the BZB of interest. This convergence angle also permits a small probe to be formed which enables resolution of the associated Miller plane spacing. We have used this approach to demonstrate nanoscale and atomic resolution on a sample of Si, SiO₂ and BN [1]. EELS can also be employed to explore the characteristic photonic modes present in dielectric nanoparticles since the electron beam can provide a source of virtual photons to illuminate the object. We have explored the variation in photonic modes for high refractive index oxide particles such as TiO₂, Ta₂O₅ and CeO₂ [2]. The modes show up as peaks in the bandgap region of the spectrum. The energies of the modes are determined not only by the particle shape and size but also due to coupling between adjacent particles. Coupling between phonons and IR photons can give rise to phonon polaritons [3,4]. Polariton characteristics are also influenced by sample geometry, and so the polaritonic modes can be manipulated by changing the sample shape/size. Here we use STEM EELS to demonstrate variations in the polaritonic response on patterned SiO₂ thin films.

References:
[1] K. Venkatraman et al., Nature Physics 15 (2019) 1237-1241.
[2] Q. Liu et al., Phys. Rev. B 99 16 (2019) 165102.
[3] K. Venkatraman et al., Microscopy 67 (2018) i14 - 23.
[4] A. Konečná et al., Phys. Rev. B 98 20 (2018) 205409.
[5] The support from US National Science Foundation CHE-1508667 and U.S. DOE (DE-SC0004954), the use of (S)TEM at Eyring Materials Center at Arizona State University is gratefully acknowledged.

Traditional scanning transmission electron microscopy (STEM) detectors are large, monolithic regions that integrate a subset of the transmitted electron beam signal scattered from each electron probe position. These STEM imaging experiments record only 1-5 values per probe position, throwing away most of the diffracted signal information. With the introduction of extremely high speed direct electron detectors, we can now record a full image (2D data) of the diffracted electron probe scanned over the sample (2D grid of positions), producing a four-dimensional dataset we refer to as a 4D-STEM experiment. These diffraction images of the electron probe are extremely rich in atomic-scale information, such as the sample structure, orientation, composition, phonon spectra, defect structure and more. The spacing between adjacent STEM probes can be varied from below one Angstrom to hundreds of micrometers, allowing us to probe the functional length scale of materials and devices.

In this talk, I will discuss how 4D-STEM can be used to measure various properties important to materials science. I will show several examples of nanobeam electron diffraction used to measure sample structure, orientation and strain, for samples ranging from metallurgical alloys to conductive polymers. I will also show how 4D-STEM can enable phase contrast imaging methods. Finally, I will also show how the scattering matrix (S-matrix) formalism can be used both to numerically invert sample structure under conditions of multiple scattering of the electron beam, and also for very fast simulation of large 4D-STEM datasets. I will show an updated version of Prismatic, our open source simulation code that includes GUIs for easy application.

Imaging chemical reactions in solution is essential for the understanding complex mechanisms involved in nanomaterial synthesis. Liquid phase electron microscopy (LP-EM) is an emerging technique capable of directly monitoring morphological changes of materials in solution with a unique combination of temporal and spatial resolution. LP-EM has been used to study the reaction mechanisms such as metal nanoparticle formation, polymer self-assembly, wet etching of semiconducting materials, growth of metal−organic frameworks, nucleation of calcium carbonate and Li electrodeposition. These studies provided an unprecedented insight into nanoscale reaction mechanisms that have significantly improved our understanding of nanomaterial synthesis. The primary challenge for the field of LP-EM is understanding the role that the electron beam plays in the observed mechanism. It is common to perform control experiments to establish a critical dose limit with which a process can be observed. A consequence of having to work below a critical dose limit is that LP-EM has mostly been used to monitor reactions over a period of seconds to minutes, despite the fact that many reactions take place over hours to days. Furthermore, even after a critical dose has been established, it is impossible to completely rule out the role of the electron beam using LP-EM experiments alone. Consequently, it is essential to study reaction mechanisms using multiple imaging modalities. Cryo-EM can also provide direct high-resolution information on reaction mechanisms in solution. The advantage of cryo-EM is that electron beam effects are well understood and, using appropriately low doses, are insignificant. The disadvantage of cryo-EM is that it is impossible to know the history or future of any individual particle being imaged. Consequently, the data on reaction mechanisms has to be presented by selecting images that the user determines to be representative of the process occurring in solution. For simple processes that evolve homogenously, it is possible to determine a mechanism with high confidence using cryo-EM alone. However, for complex processes that evolve heterogeneously, the confidence with which mechanisms can be determined by cryo-EM alone is low. Here, a simple process is defined as a process with a single step or single pathway and a complex process is defined as a process with multiple steps or multiple pathways. A homogenous evolution is characterized by a process where only one particle state is observed at each sampled time point, while a heterogeneous evolution is characterized by a process where multiple states are observed at each sampled time point. Consequently, cryo-EM is limited in its ability to resolve complex heterogeneous processes. Despite the obvious benefits of combining LP-EM and cryo-EM, most studies using these techniques have been only conducted with a single EM modality. Here we discuss the application of distributed electron microcopy for studying reaction mechanisms in solution. The term distributed is used because the data required to understand the mechanism is shared between multiple imaging modalities and sample preparation methods. The strength of this approach is that our proposed hypothesis is consistent with data taken from experiments with complementary benefits and limitations.

Oriented attachments are recognized as a common pathway for crystal growth in both natural and synthetic systems, leading interface free or twin related structures. The occurrence of 5-fold twinned nanoparticles leads to unique properties. However, nearly 200 years after their discovery, the formation mechanism is still ambiguous. Using in situ high-resolution transmission electron microscopy combined with molecular dynamics simulations, we demonstrate that 5-fold twinning occurs via repeated oriented attachment of ~3 nm Au nanoparticles. These oriented attachment events create high-energy grain boundaries, which accumulate strain during atomic rearrangements and consequently decompose via nucleation and growth of a special class of twins whose net strain is zero, inducing the 5-fold twin structures. The results provide a quantitative understanding of the 5-fold twinning process; this knowledge provides guidance for interpreting and controlling twin structures and morphologies of a wide range of materials.

The electrolyte 2.8 M lithium bis(fluorosulfonyl)imide (LiFSI) in dimethoxyethane (DME) has demonstrated good cycling behavior and stable Li morphology upon plating and stripping for high-capacity Li batteries, which is thought to be due to the formation of a stable solid-electrolyte interphase (SEI) layer for good ionic transport. Characterization of this SEI layer on lithium metal has been limited, due to the air-sensitivity and beam sensitivity of these materials, as well as the incompatibility of the electrolyte with the high vacuum environment necessary for nanoscale characterization with electron microscopy techniques. New cryogenic stages and transfer tools have enabled the characterization of frozen solid-liquid interfaces for lithium metal structures. Here, we used cryogenic focused ion beam (FIB) and ultrashort pulse laser ablation to create a cross-section of the lithium metal anode in contact with the electrolyte at various plating/stripping cycles to understand the morphological evolution of the lithium and SEI. Additionally, with cryo-FIB we utilized lift-out procedures under cryogenic temperatures for extraction of a lamella sample for investigation using scanning/transmission electron microscopy with energy dispersive x-ray spectroscopy and electron energy loss spectroscopy. The nanoscale chemical maps have allowed for a greater understanding of the SEI components and how they redistribute with reaction to the lithium metal over the course of many cycles. These results present a step towards visualization of the intact battery stack. We have determined a failure mechanism beyond short circuiting of the lithium metal anodes, due to reaction of the lithium with this LiFSI in DME chemistry.
This work was funded by Sandia National Laboratories’ Laboratory Directed Research and Development program. It was performed, in part, at the Center for Integrated Nanotechnologies, an Office of Science User Facility operated for the U.S. Department of Energy (DOE) Office of Science. Sandia National Laboratories is a multi-mission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. DOE’s National Nuclear Security Administration under contract DE-NA-0003525. The views expressed in the article do not necessarily represent the views of the U.S. DOE or the United States Government.