Microscopy Conference 2023

Session chairs

Prof. Dr.

This session focusses on functional thin films and advanced structured materials targeted for applications in a wide range of energy and information technologies, including intriguing physical effects at interfaces of epitaxial complex oxide systems, electron and ion conductors, photovoltaics, thermo-electrics, dielectrics, and resistive switching. The optimization of these devices, from materials’ design to applications, requires a thorough knowledge of the structural, physical, and chemical properties, its interactions with the environment and the role of interfacial effects, for which recent technical innovations in electron microscopy hardware and software have provided a fertile ground for their analysis with unprecedented spatial and spectral resolution. We welcome contributions to latest developments at the forefront of scientific achievements and impact in the design, synthesis, characterization, and applications of functional thin films.

Planar defects in non-stoichiometric SrFeO3-δ perovskite type thin films studied by analytical scanning transmission electron microscopy

Abstract text (incl. figure legends and references) Strontium-ferrite (SFO), SrnFenO3n-1, provides a fascinating system in which stoichiometric changes can be explored in order to study their impact on the material's crystallographic structure, its properties, and, in particular, on defect formation. For n = 2, SFO adopts a brownmillerite-type structure of alternating layers of FeO6 octahedra and FeO4 tetrahedra with iron exclusively in +3 oxidation state. Increasing the oxygen content, with n = 4 and n = 8, SFO reveals tetragonal and orthorhombic structures with iron in +3 and +4 oxidation states (see, e.g., [2]). For the limiting case that n → ∞, SFO is found in the stoichiometric, cubic perovskite structure as SrFeO3, where Fe is uniquely in +4 oxidation state, forming the well-known FeO6 octahedra as building blocks of the perovskite structure. From a properties point-of-view, SFO is known to be a mixed ionic-electronic conductor and consequently, tuning the structure and defects provides a mean to control its conductivity. We present an experimental study of non-stoichiometric SrFeO3-δ, where the loss of oxygen (δ &gt; 0) must lead to oxygen vacancies, possibly structural defects, and thus resulting in changes in the oxidation state of Fe as well. We used HAADF- and ABF-STEM to develop a structural model of planar defects, which run "packman"-style along {001} planes in the SFO film grown on SrTiO3, and which are formed due to the oxygen deficiency [2] (see Fig. 1). Atomic-resolution EDX spectroscopy reveals that the defective bands are formed by Fe2O2 double layers. The defects strongly resemble the Sr4Fe6O12+δ structure viewed along the pseudocubic [107] direction, which consists of bands of corner-sharing FeO6 octahedra separated by bands formed of periodically arranged FeO5 polyhedra identified as trigonal bipyramids and tetragonal pyramids, respectively. To assess the impact of these chemistry-driven structural defects on the physical properties, EELS was used to measure the Fe L3,2 and the O K excitation edges. Subtle but significant differences in the fine structure of the O K edge at the defect planes compared to bulk spectra were identified and interpreted by complementing the experimental data with simulations using FEFF9. This analysis suggests an increased electron doping of the Fe 3d eg bands in the Fe2O2 defect layers compared to the bulk SrFeO3, which as a consequence must affect SFO's electron-hole conductivity. Moreover, analysis of the Fe L3,2 excitation edge uncovers the expected change of the Fe oxidation state in the defective planes from +4 towards +3. In summary, STEM imaging combined with EDX and EELS measurements revealed the nature of planar defects present in SrFeO3-δ thin films grown on SrTiO3 and their effect on the electron-hole conductivity. Controlling the amount of oxygen vacancies in these perovskite-based SFO thin films thus allows the electronic conductivity of these films to be adjusted. Figure 1 HAADF STEM: overview (a) and detail (b) of the atomic structure of the planar defects formed in oxygen deficient, perovskite-type SrFeO3. References [1] M. D. Rossell, A. M. Abakumov, Q. M. Ramasse, R. Erni, ACS Nano 7 (2013) 3078-3085. [2] M. D. Rossell, P. Agrawal, M. Campanini, D. Passerone, R. Erni, Phys. Rev. Mater. 4 (2020) 075001. [3] P. Agrawal and D. Passerone are kindly acknowledged for providing support for interpretation of the FEFF9 data, and we would like to thank M. Campanini providing the script for peak-pair analysis.

MS 4: Functional thin films

Invited talk

Tuning interface sharpness and interface superconductivity at La2CuO4 heterostructures

Abstract text (incl. figure legends and references) Heterostructures of transition-metal oxides enable emerging unique physical properties at their interfaces [1]. The structural adaptability of La2CuO4 allows for designing different heterostructures, and superconductivity takes place at the interface between overdoped (metallic) and undoped (insulating) La2CuO4 layers grown by oxide molecular beam epitaxy (MBE)[2]. In addition to homo-epitaxial systems[3], heterogeneous multilayers of La2CuO4 with, for example, perovskite 113-type and 214-type lanthanum nickelate layers revealed the impact of the interface sharpness on thermoelectric and superconducting properties, respectively [4,5]. In this work, we fabricate (La,Sr)2CuO4–SrMnO3&shy;–LaMnO3–La2CuO4 superlattices [6] using oxide MBE and focus on the interface sharpness and the related superconducting mechanism compared to cuprate–cuprate interfaces. We use 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) to examine the interfaces. A JEOL JEM-ARM200F STEM equipped with a cold field-emission electron source, a probe Cs-corrector (DCOR, CEOS GmbH), a Gatan GIF Quantum ERS spectrometer and a large solid-angle JEOL Centurio SDD-type EDXS detector was used for atomic-resolution analyses. Our findings demonstrate that hetero-epitaxial cuprate–manganite layers can achieve sharper Sr-doped La2CuO4 interfaces. The dopant distribution in La2CuO4 is affected by the elemental intermixing in the first atomic LaMnO3 monolayer. We further underline that structurally sharp interfaces can be chemically rough (Fig. 1), and the chemical intermixing dominates the physical properties. Importantly, different superconducting behaviour (e.g., interface vs filamentary) can be customized with interfacial design [7]. References [1] Y. E. Suyolcu, G. Christiani, P. A. van Aken, G. Logvenov, J. Supercond. Nov. Magn. 2020, 33, 107. [2] A. Gozar, G. Logvenov, L. F. Kourkoutis, A. T. Bollinger, L. A. Giannuzzi, D. A. Muller, I. Bozovic, Nature 2008, 455, 782. [3] 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. [4] F. Baiutti, G. Gregori, Y. E. Suyolcu, Y. Wang, G. Cristiani, W. Sigle, P. A. van Aken, G. Logvenov, J. Maier, Nanoscale 2018, 10, 8712. [5] P. Kaya, G. Gregori, F. Baiutti, P. Yordanov, Y. E. Suyolcu, G. Cristiani, F. Wrobel, E. Benckiser, B. Keimer, P. A. van Aken, H.-U. Habermeier, G. Logvenov, J. Maier, ACS Appl. Mater. Interfaces 2018. [6] 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. [7] This work has received funding the European Union's Horizon 2020 research and innovation programme under grant agreement No. 823717 – ESTEEM3. Figure 1. (a) STEM-HAADF image of a La2CuO4–SrMnO3–LaMnO3–La2CuO4 superlattice demonstrating coherent "A" and "B" interfaces. (b) EDX line-scan profiles reveal Cu (red) and Mn (green) intermixing at the interfaces. Sr (orange) and La (blue) profiles are given as guides indicating the individual layers. The blue arrow depicts the region of the acquired EDX line scan profiles.

IM 2: Spectroscopy

Abstract talk

Spatially localized electronic states at the LaAlO3/TiO2 interface: mapping of a 2D electron gas

Abstract text (incl. figure legends and references) Perovskite heterostructures play an increasingly important role in optoelectronics applications, such as solar cells, LEDs and photo detectors [1]. The combination of two different semiconducting materials often leads to drastic changes of the electrical, magnetic or optical properties in the vicinity of the interface. Thus, full characterization of the localized electronic states is of utmost importance. We approach this goal with a combined ab-initio and experimental study of the electronic states localized at the LaAlO3/TiO2 interface. The material of interest is a heterostructure consisting of the cubic perovskite LaAlO3 and TiO2 in anatase phase. Bulk anatase is typically a semiconductor with a few eV wide direct band gap. In combination with LaAlO3, however, unoccupied states appear directly above the Fermi energy for 2-3 TiO2 layers closest to the interface. When a small energy window is chosen for imaging, similar to elemental mapping albeit with a window of 0.5 eV or even smaller, the localized electronic states can be mapped in real space [2]. We compare the experimental results to inelastic channeling simulations of a Ti-La terminated interface based on [3]. The simulated spectrum image for an energy loss of 456.4 eV can be seen in Figure 1(c). Ti atoms at least 1 unit cell from the interface show negligible intensity or no intensity at all and can therefore be considered &quot;bulk-like&quot; in this context. Ti atoms close to the Ti-La terminated interface, however, show significant intensity at this energy loss, which is especially apparent in the second atomic column from the interface. The orbital there resembles the shape of a dxy-orbital with x the horizontal axis and y the vertical axis in the figure. The experimental map of the Ti L3-edge onset pre-peaks lies in close agreement with the simulations (see Figure 1(b)). Due to the difficult mapping conditions presented by the low signal-to-noise ratio and due to elastic channeling, the original shape of the orbitals is not exactly reproduced. More importantly, however, the spatial confinement of the relevant electronic orbitals and the resulting local increase in intensity is clearly visible. Electrons occupying these states can, thus, be identified as a two-dimensional electron gas. The direct mapping of such an electron gas opens up new ways of investigating heterostructures in electronics with future application of understanding and designing new and improved optoelectronic materials. [4] Figure 1. (a) HAADF image of a 12 nm thick heterostructure oriented in the [100] crystallographic directions of both anatase TiO2 and LaTiO3 recorded at a high tension of 300 kV. (b) STEM-EELS spectrum image for a collection semi-angle of 24 mrad and an energy window at the Ti L3-edge onset. (c) Simulated spectrum image for an energy loss of 456.4 eV and a 1 unit cell thick sample with the same parameters as in (a) and (b). [1] Cheng et al., Adv. Optical Mater. 10, 2102224 (2022) [2] Löffler et al., Ultramicroscopy 177, 26 (2017) [3] Wang et al., Journal of Applied Physics 108, 113701 (2010) [4] The authors gratefully acknowledge funding from FWF under grant nr. I4309-N36.

MS 7: Ceramics and composites

Prof.

Probing the interplay of atomic structure and properties in perovskite oxide thin films

Abstract text (incl. figure legends and references) Perovskite-structured oxide thin films provide a fascinating playground for the study of condensed matter physics, owing to the strong coupling of lattice composition and structural distortions to functional properties. Here, we illustrate how recent developments in electron microscopy are vital for study of their structure–property relationships. Measurements are made in scanning transmission electron microscopy (STEM) mode using a double aberration-corrected FEI Titan Themis 60-300 at 200/300 kV. For measuring atomic structure, high angle annular dark field (HAADF) and annular dark field (ABF) 90° rotation image series are recorded, and then corrected for linear and non-linear scan distortions. Electron energy-loss spectroscopy (EELS) is carried out using a Gatan GIF Quantum ERS. Samples are prepared via mechanical polishing and broad ion beam milling, or by focused ion beam and lift out, and cleaned with an IBSS MCA. First, we look at NdNiO3/SmNiO3 superlattices, where macroscopic measurements of a metal-to-insulator transition (MIT) show an unexpected transition from two electronic phases to a single one as layer thickness decreases below a critical value. Using monochromated EELS to measure subtle changes in the Ni L3 and O K ionization edges, we not only confirm this behavior, but also map the electronic phases with a sub-nm spatial resolution [1]. STEM structure measurements show, however, that each layer retains the composition and lattice distortions of its parent lattice; a result that helps motivate a new Landau theory of electronic phase boundary energy [2]. To further study this behavior, we test decoupling the layers by incorporating LaAlO3 interlayers into the superlattices. Here, atomic resolution STEM is essential for diagnosing deposition problems and for measuring the coupling of structural distortions across interfaces; see Fig. 1. Lattice distortion connections next come to the fore in the study of LaVO3 films. STEM helps show how epitaxial strain plays a key role in determining the film's orientation, giving an out-of-plane orthorhombic long axis under the tensile strain set by a (110) DyScO3 substrate [3]. Since the DyScO3 instead has an orthorhombic long axis that is in-plane, such a film inherently forces an interface that couples the incommensurate oxygen octahedra rotations of the two orthorhombic orientations. Remarkably, STEM reveals this rotation-coupled interface forms within film, rather than at the substrate/film interface, creating a sharp 90° domain boundary within a single compound. Justified by second principles energetic simulations, this finding promises new avenues for creating novel functional properties [4]. In summary, we illustrate myriad ways that aberration-corrected STEM is critical for uncovering the interplay between atomic structure and physical properties in functional thin films. Looking towards the future, in order to achieve further insights, our aim is to couple these precision measurements with variable sample conditions. [1] B. Mundet et al., Nano Lett. 21, 2437 (2021). [2] C. Domínguez et al., Nat. Mater. 19, 1182 (2020). [3] H. Meley et al., APL Mater. 6, 046102 (2018). [4] D.T.L. Alexander et al., under review. Fig. 1. (a) HAADF STEM image of a NdNiO3/SmNiO3 superlattice with LaAlO3 interlayers. (b) depth profile of average Δy displacements of A-site cations from their "ideal" cubic positions (right) shows how each layer recovers the distortions of its parent compound.

MS 3: Low-dimensional and quantum materials

Microstructure-functionality relationships of pulsed laser deposited CuI thin films

Abstract text (incl. figure legends and references) In the field of transparent p-type semiconductors, copper iodide (CuI) holds an outstanding position. It combines high transparency in the visible spectral range with excellent hole transport properties [1]. To make CuI thin films suitable for complementary electronics, i. e. to develop CuI into an equivalent partner for well investigated transparent n-type semiconductors, control of electrical and morphological properties is required. Pulsed laser deposition (PLD) showed recently successful outcomes regarding film quality and systematic functionality tuning [2]. The use of powder-based targets allows a high flexibility regarding composition and can provide lateral and vertical composition gradients as well as controlled doping. With the possibility to change source targets in-situ, template layers can be grown to improve lattice match and surface smoothness due to the decrease of grain boundaries. This works focuses on the microstructural investigation of (doped) CuI films employing EBSD, SEM, FIB, TEM and STEM-EDX. The obtained results are used to explore microstructure-functionality relationships by linking XRD findings and electrical properties with the results gained by the aforementioned electron microscopy techniques. The first part of this work examines the impact of intentional p-doping with selenium (Se) on the film morphology. CuI with a lateral Se gradient was PLD grown on a two-inch diameter sapphire wafer. Site-specific TEM samples were prepared by focused ion beam (FIB) machining corresponding to a certain Se content. Whereas for each Se value, growth starts with the ´default´ zinc-blende γ-phase, high resolution TEM showed that a relatively high Se content seems to favor the CuI beta phase (trigonal crystal system) within the ongoing growth process. For a lower Se content, a new CuI phase, presumably with an up to now unpublished wurtzite 4H structure, was discovered with high-resolution TEM and related diffraction analysis (Fig. 1). This crystalline phase change also matches with results regarding XRD and electrical properties, where considerable changes occur for a selenium content above x(Se) = 1 at% [3]. The second part investigates the possibility to suppress rotational domains by employing PLD grown template layers such as sodium bromide (NaBr). By growing NaBr on cubic SrF2 (111) substrates before depositing the CuI thin film, a heterostructure was constructed where the moisture sensitivity of NaBr could be employed to obtain free standing CuI thin films. XRD phi-scans showed single-crystallinity and no occurrence of rotational domains up to 1-2 µm film thickness. Nevertheless, films need to be sufficiently thicker to be utilized as, for example, front-back device processing. Samples out of reasonable thick films could be prepared for EBSD to visualize domain orientation, SEM to depict surface roughness and TEM to analyze grain boundaries to examine the cross-section as well as the front- and backside of the CuI film. It was shown that for thicknesses above 1-2 µm, the single crystallinity gets lost, probably due to accumulation of defects during the PLD growth [4]. [1] M. Grundmann et al., phys. stat. sol. (a) 210, 1671-1703 (2013). [2] P. Storm et al., APL Mater. 8, 091115 (2020). [3] P. Storm et al., Phys. Status Solidi RRL 15(8), 202100214:1-6 (2021). [4] P. Storm et al., J. Mater. Chem. C 10, 4124-4127 (2022). Fig. 1: CuI:Se film with lateral Se composition gradient on 2-inch diameter c-sapphire wafer with marked position for FIB sample extraction (left). High-resolution TEM images 1,2, and 3 close to the interface substrate-film (right).

MS 2: Metals and alloys

Prof. Dr. rer. nat. habil.

Seeing structural evolution of organic molecular nano-crystallites using 4D-scanning confocal electron diffraction (4D-SCED)

Abstract text (incl. figure legends and references) Direct observation of organic molecular nanocrystals using electron microscopy is highly challenging, due to their radiation sensitivity and complex structure. 4D-STEM nano-beam diffraction (NBD) using small convergence α has been demonstrated to reveal the rich structural information in soft materials [1]. In this context, homogeneous beam-specimen interaction is desired to maximize dose/information efficiency before total structural damage, and spot diffraction signals are preferred for sensitivity and accuracy of Bragg peaks detection in dense reciprocal lattices typical in organic crystals. In NBD, the beam illuminates the sample with an Airy disk profile, and diffraction disks size is coupled to α, not optimized for dose efficiency and Bragg peaks detection due to spread signals and possible disks overlapping. Here, we introduce 4D-scanning confocal electron diffraction (4D-SCED) which combines high dose efficiency and high angular resolution. 4D-SCED applies defocused pencil beam illumination (Fig. 1b) on the sample and combines a confocal electron optic setup with a camera to record spot diffraction patterns (Fig. 1c). The defocused illumination reduces the dose with a homogenous beam sample interaction, and the confocal optics transfer spot-like diffraction signals, optimized for Bragg peaks detection. The spatial and angular resolution can be estimated using geometric considerations (Fig. 1d), which is largely decoupled from α. We first compare SCED and NBD with a single crystal thin film of 2D molecular α,ω-DH6T bilayer (fig. 1e-h). An order-of-magnitude higher diffraction signal is seen in SCED compared to NBD, enabling accurate Bragg peak detection for structural studies, e.g. layer rotation relation. We apply 4D-SCED to study an active layer in organic solar cells, DRCN5T:PC71BM BHJ thin films after solvent vapor annealing (Fig. 2a-c). With careful balancing spatial and angular resolution for the given dose budget, structural details of DRCN5T nano-crystallites oriented both in- and out-of-plane are imaged at ~5 nm resolution and dose budget of ~5 e-/Å2. Finally, we use 4D-SCED to study the structural evolution during thermal annealing by in situ heating the thin film in the TEM (Fig. 2d). The evolution of crystallite size (coarsening) and texture, as well as the progressive enrichment of PC71BM at interfaces are directly revealed [2]. The unique combination of high dose efficiency and high angular resolution makes 4D-SCED an ideal technique for studying beam-sensitive soft materials. The new possibilities of the technique are currently employed to explore further soft materials. Refs: [1] Bustillo, K., et. al., Acc. Chem. Res. 11 (2021) 2543 [2] Wu, M., et. al. Nat. Commun. 13 (2022) 2911 The authors acknowledge financial support from DFG via projects GRK1896 and SFB953. Work at the Molecular Foundry has been supported by U.S. DoE under contract No. DE-AC02-05CH11231. Fig 1. (a) STEM setup. (b) Defocusing the probe mitigates dose for beam-sensitive samples. (c) Scheme of SCED. (d) A geometric scheme for considering spatial and angular resolution. (e-h) Comparison of NBD and SCED applied to an α,ω-DH6T bilayer. Fig 2. Visualizing (a) edge-on (with color wheel) domain orientation and (b) face-on (grayscale) domain location of the donor crystals in DRCN5T:PCBM blends. (c) Mapping the whole area with 4D-SCED dataset. (d) The structural evolution of donor nano-crystallites during an in situ annealing experiment in the TEM.

MS 4
Functional thin films

Appointment

Session chairs

Description

Programme

MS 4Functional thin films

Appointment

Session chairs

Description

Programme

MS 4
Functional thin films