Hydrogels are an important material as support matrices for cells to promote growth. These systems have been characterized by electron microscopy, through the application of fixation or sucrose embedding followed by ultramicrotomy. In this way the porosity of the gels can be assessed by transmission electron microscopy (TEM) and/or scanning electron microscopy (SEM). While this approach does give a guide to porosity of samples the absence of water in a hydrogel will have a detrimental effect and its absence distorts the dimensions of the remaining gel. Focused ion beam scanning electron microscopy (FIB-SEM) has been used on dried hydrogels [1], however, we propose a new method of gel porosity characterization by the use of cryogenic-FIB-SEM (cryo-FIB-SEM).  

Cryo-SEM is a long established technique to preserve the water content of a sample and more recently it has been demonstrated that cryo- FIB-SEM can be used for biological and soft matter materials [2]. In this work, the authors have used cryo-FIB-SEM to investigate the porosity and structure of gels whilst in the presence of water. Gels were plunge frozen in slush nitrogen or using a metal mirror freezer and transferred under liquid nitrogen to the sample shuttle of a Cryo-SEM system (Quorum PPT 2000, Quorum Technologies). In the prep-chamber, the sample was coated for 60 seconds using a Pt sputter target. The samples were then loaded into the FIB-SEM (FEI Quanta 3D, FEI). Once in the SEM chamber, the gels were prepared for FIB by deposition (3-4 seconds) of a platinum precursor from the gas injector (set to 27 °C) of the microscope.

The hydrogel samples were milled using an initial current of 1-3 nA to make a rough cut and then by further cuts at lower milling currents (0.3 nA-50 pA), to remove the common milling artefact known as curtaining. SEM micrographs of the visible milled face showed dark patches with largely white areas in between. It was initially postulated that the darker areas were the pores of the gel. In order to test this, the temperature in the SEM chamber was raised to -90 °C, leading to slow sublimation of the water at the FIB milled face (Figure 1). Over approximately 20 minutes images of the slowly subliming gel were acquired. The resulting images show a transition from the black features amongst the majority of lighter contrast through to images with inverse contrast. The final sublimed gel images are clearly interpretable as a porous gel where now the lighter contrast features are identified as the gel strands and the pores are now darker and devoid of water. With a better understanding of the location of these components the original non-sublimed images can be re-examined and the black contrast correlated directly to the polymer and the white to the water.

By inverting the contrast of the original milled face image, it is possible to give an image equivalent to the dehydrated image, but which has all the water bound and is therefore a truer representation of the gel’s morphology (Figure 2). The major advantage to this is that the sample does not undergo shrinkage, and that the process of imaging the milled face hydrated saves time. Additional slices of the freshly milled face may then be acquired to yield a series of slices suitable for 3D rendering (figure 3).

References:

[1] A Al-Abboodi et al., Biotech and Bioengineering, 110 (2013), p. 328.

[2] M Marko et al., J Microsc. 222 (2006), p. 42

Most formulated fine chemical products are complex systems that contain multiple components, with nanoparticles and any incorporated surface coatings interacting with other particles and the dispersant (which can be a liquid or solid). Assessing nanoparticles when suspended in a liquid can be challenging as the particles may disperse individually, agglomerate, aggregate, sediment, chemically-alter or even dissolve and re-precipitate.  With the appropriate sample preparation however, TEM can be used to measure the dispersion and any transformation of nanoparticles suspended in, for example, biological or environmental media [1,2].

Conventional transmission electron microscopy (TEM), with samples prepared by simply drop-casting suspensions onto a thin carbon film, enables imaging and analysis of individual nanoparticles but, because of the drying process, does not capture the particle agglomeration in the dispersion or the surface chemistry when hydrated [3]. To overcome this problem we have prepared thin sections of nanoparticle suspensions for TEM by plunge-freezing a blotted grid into liquid ethane to ensure the aqueous phase vitrifies with no significant redistribution of suspended material. We have used this technique to quantify the dispersion of polymer coated quantum dots, silica and zinc oxide nanoparticles in water and biological cell culture media, identifying the true form in which these nanoparticles are taken up into cells in vitro and thereby providing mechanistic insight to the cellular response at these exposures [3,4,5].

Low dose electron microscopy of nanoparticles suspended in vitreous ice provides opportunity for the analysis of the structure and chemistry of the dispersion, both vital characteristics to understand before any successful biomedical exploitation of nanoparticles. Here, dextran coated iron oxide nanoparticles agglomerated in aqueous suspension and captured in vitreous ice were imaged by bright field TEM and analysed by energy dispersive X-ray (EDX) spectroscopy. Careful control of the illumination conditions (electron dose) permit near native state imaging and confirmation of composition before inducing significant damage to the surrounding ice matrix and subsequent movement of the particles (Figure 1). HAADF STEM imaging was conducted using a 1.3 Å probe and 60 pA probe current, with a resulting EDX map collected in just over one minute showing an iron signal appropriately localised to the nanoparticles (Figure 2).

Going forward, we will use the recently installed FEI Titan Cubed Themis 300 G2 S/TEM at the University of Leeds which is equipped with FEI SuperX EDX spectrometers, a Gatan Quantum ER imaging filter and Gatan OneView CCD to explore the limits of nanoparticle structural analysis (incorporating diffraction and lattice imaging), as well as use of STEM-EDX and electron energy loss spectroscopy for detailed elemental analysis when encased in vitreous ice. In addition to examining the dispersion state of nanoparticles in different suspensions, our goal is to identify and analyse the surface coatings on nanoparticles in the frozen hydrated state, thereby extending the capability of near native state imaging and analysis of nanoparticle suspensions by TEM.

1. N. Hondow, A. Brown and R. Brydson (2015) Frontiers of Nanoscience, 8, 183 – 216.

2. R. Brydson, A. Brown, C. Hodges, P. Abellan and N. Hondow (2015) J. Microscopy, 260, 238 – 247.

3.  N. Hondow, R. Brydson, P. Wang, M.D. Holton, M.R. Brown, P. Rees, H.D. Summers and A. Brown (2012), J. Nanopart. Res., 14, 977.

4. Q. Mu, N.S. Hondow, L. Krzeminski, A.P. Brown, L.J.C. Jeuken and M.N. Routledge (2012) Particle Fibre Toxicol. 9, 1.

5. R. Wallace, A.P. Brown, R. Brydson, S.J. Milne, N. Hondow, P. Wang (2012) J. Phys. Conf. Ser. 371, 012080.

Acknowledgment: We thank FEI for the data shown in the figures which were collected as part of a demonstration at the FEI Nanoport, Eindhoven, and Steve Evans (Swansea University) for the dextran coated iron oxide nanoparticles.

Microstructural investigation of light elements and highly beam sensitive polymer materials using electron microscopy is attractive for elucidating nanostructure but presents numerous challenges. In particular, heavy element staining, often used to obtain image contrast, may obscure or degrade the structure of interest and the acquisition of detailed and spatially resolved information must be balanced with damage of the specimen. Here, scanning electron diffraction (SED) has been used to analyse the crystalline microstructure of unstained polyethylene, overcoming these challenges. SED involves scanning the electron beam across the specimen and recording a diffraction pattern at each position [1] at a high frame rate to enable the data to be acquired before severe degradation of the structure has occurred. In this way, electron diffraction patterns were obtained from an unstained polyethylene sample in 5 nm steps, over areas of a few microns squared and with a 10 ms dwell time. Radiation damage was further minimised by using a high electron acceleration voltage (300kV) to minimise radiolysis and cooling the sample with liquid nitrogen [2]. The diffraction patterns, acquired at every position in the scan, were indexed and analysed by plotting the intensity of a particular reflection as a function of electron probe position to form ‘virtual’ dark field (VDF) images. VDF imaging is much more effective than conventional imaging for visualizing the microstructure of polyethylene. Clear contrast is obtained without staining and the versatile post-facto nature of VDF image formation enables multiple complementary images to to be produced from a single acquisition.
Two novel observations from our SED experiments on polyethylene are highlighted here. The sample of polyethylene was extruded from a melt so as to form ‘shish-kebab’ structures confirmed through BF images of stained microtomed sections. For our experiments, again the samples were microtomed but now unstained to avoid any influence of the stain on the diffraction patterns. The first experiment highlights a lamella-like fragment of polyethylene crystal (likely to be a part of the ‘kebab’ structure). Fig 1 shows a ‘virtual’ BF image and a sample of diffraction patterns that can only be indexed assuming the orthorhombic crystal structure of polyethylene and that the lamella is twisting about a single axis almost parallel to the vertical axis of the image. Moreover, forming consecutive VDF images made it possible to visualize each region of the crystal having a particular orientation in the twisted lamella (Fig 2). In the second experiment, we found that the sample had several micron sized islands of hexagonal polyethylene first seen as a high pressure phase [3]. However, here the patterns reveal a √3 superstructure with weak spots at the 1/3[110]* position (Fig 3). VDF images (Fig 3(b-d)) formed by these supercell reflections revealed domains within which ‘striped’ contrast can be seen; these stripes run at an orientation of approximately 120° to one another. This work demonstrates the applicability of the SED technique to highly beam sensitive materials like polyethylene and the potential for new microstructural insights to be made in this way.

[1] Moeck P. et al., Cryst. Res. Technol., 2011, 46, 586-606
[2] Egerton R. F. et al., Micron, 2004, 35, 399-409
[3] Bassett D. C., et al., Journal of Applied Physics, 1974, 4146-415

PAM and SJK would like to acknowledge funding under ERC Advanced Grant 291522-3DIMAGE. DNJ receives a Vice Chancellor’s award from the University of Cambridge. HJ and HT thank Ms. Makiko Ito for her help in microtoming the polyethylene samples. The authors would like to thanks Anton Jan Bons (ExxonMobil) for initiating this research and stimulating discussions.

Silicon (Si) nanopartcles (NPs) embedded in the insulating or semiconducting matrices has attracted much interest for the third generation of photovoltaics, so called “all-Si” tandem solar cells. In this work, the amorphous Si rich silicon carbide (SRSC) absorber layers with 30% carbon content were deposited using plasma enhanced chemical vapour deposition (PECVD) on quartz substrate at 500 ˚C, and then the SRSC films were annealed at 1100 ˚C in nitrogen for 1 hour ¹. The thermal treatment leads to the SRSC films spinodally decomposed into a Si-SiC nanocomposite. The nanostructures of the phase separated Si and SiC presented in the 15 minutes and 1 hour annealed SRSC films were investigated by two dimensional (2D) energy-filtered transmission electron microscopy (EFTEM). After the thermal treatment, the coexistence of crystalline Si and SiC nanoparticles (NPs) were observed from the high resolution TEM (HRTEM) images and verified by the selected area diffraction (SAD) patterns. After 1 hour annealing, neither Si nor SiC phases are complelely crystallized, the detailed morphologies of Si and SiC nanostructures were studied by electron tomography. For the first time, we make use of EFTEM spectra-imaging (SI) dataset to reveal the three dimensional distributions of Si, a-SiC and c-SiC in sub-volumes. In particular, to obtain more detailed and quantitative information, we have fitted the plasmon spectra with reference plasmon peaks. This enables us to not only to get a quantitative 3D image of all components involved in the materials system in the final tomogram, but also to obtain information about hitherto undetected phases in this system. In such energy resolved plasmon tomograms, the 3D shape of a thin amorphous SiC layer (1) was observed at interface between the crystalline Si network like structure and crystalline SiC NPs. The appearance of the a-SiC interfacial layer is expected from nucleation theory.

Acknowledgements

The authors ackonwledge the support from the EU funded FP7 Project SNAPSUN, the Knut and Alice Wallenberg Foundation and the Swedish Science Council.

References

1.        Perraud, S. et al. Silicon nanocrystals: Novel synthesis routes for photovoltaic applications. Phys. Status Solidi a-Applications Mater. Sci. 210, 649–657 (2013).

In recent years nanoglass materials have attracted a lot of interests due to their special physical properties, which differ significantly from traditional bulk amorphous materials of the same composition [1,2]. For example, a Sc₇₅Fe₂₅ nanoglass exhibited a remarkable plasticity whereas the corresponding ribbon glass was brittle [3]. It has been suggested that these special properties originate from the interfacial regions between amorphous nano domains (analogous to grain boundaries changing the properties of nanocrystalline materials), which have either a different atomic configuration or chemical composition, or both compared to the domain core. Due to the amorphous structure of the materials, the structural variations between the core of a glassy grain and the interfacial region are difficult to distinguish, especially when the composition is similar. By STEM-EDX/EELS spectrum imaging and by EFTEM imaging, such composition variations were confirmed for a number of nanoglass systems synthesized using various methods [2], for example inert gas condensation (IGC), magnetron sputtering and ultra-high vacuum (UHV) cluster deposition. On the other hand, radial distribution function (RDF), which can be extracted from electron diffraction, has proven very sensitive to the small difference in atomic configurations [4], e.g., interatomic distances and coordination numbers. The newly developed STEM-RDF mapping has been shown to be capable of resolving the different amorphous structures at nanoglass core and at the interface, respectively.

In this presentation, we investigated the Sc₇₅Fe₂₅ nanoglass primary particles synthesized by IGC as well as the pellet pressed at 6 GPa. The primary particles were directly collected by a carbon coated TEM grid and the pellet sample was processed by FEI strata focused ion beam (FIB) for TEM observation. From the STEM-EDX mapping on the primary particles (Fig. 1), a Sc-rich shell is clearly resolved. Quantification of the integrated spectra from the out part and from the core of the particle reveals the Sc:Fe atomic ratio being around 4:1 and 2:1, respectively. After consolidation under high pressure, the inhomogeneity in primary particles remains and leads to Sc-rich interface between the areas originated from the particle cores. Additionally, by performing STEM-nanobeam diffraction on the pressed Sc₇₅Fe₂₅ sample, RDF mapping was obtained and two different types of RDFs are distinguished, indicating that there exist two major components, one with higher Fe-Sc coordination number (red curve in Fig.2), and the other one with higher Sc-Sc coordination number (green curve in Fig. 2). By multiple linear least square (MLLS) fitting, corresponding component maps are constructed and shown in Fig.2. In the color-mix map, the green areas represent the interface between particle cores, where Sc-Sc bonding is dominating and the red areas represent the particle cores, where Fe-Sc coordination is considerably higher.

In nanoglass systems synthesized by other method, e.g. NiTi-Cu by magnetron sputtering [5] and NiP by electrochemical deposition, composition and structure fluctuations were also observed. The different amorphous structures constrained locally can be successfully revealed by (S)TEM spectroscopic and nanobeam diffraction methods, which is an important step towards understanding the unique structure in nanoglasses compared to the conventional glasses. The correlation between synthesis and structure of nanoglasses makes it possible to design the amorphous nanomaterials with desired functionalities.

References

1. H. Gleiter, Acta mater. 48 (2000) 1.

2. Gleiter, Beilstein J. Nanotechnol. 4 (2013) 517.

3. J.X. Fang, U. Vainio, W. Puff, R. Würschum, X.L. Wang, D. Wang, M. Ghafari, F. Jiang, J. Sun, H. Hahn, H. Gleiter, Nano Lett. 12 (2012) 5058.

4. X. Mu, S. Neelamraju, W. Sigle, C.T. Koch, N. Totò, J.C. Schön, A. Bach, D. Fischer, M. Jansen, P. van Aken, J. Appl. Cryst. 12 (2013) 1105.

5. Z.Śniadecki, D.Wang, Yu.Ivanisenko, V.S.K.Chakravadhanula, C.Kübel, H.Hahn, H.Gleiter, Materials characterization 113 (2016) 26.

Localized surface plasmon resonances (LSPRs) are the coherent and collective oscillations of conduction band electrons at the surface of metallic nanoparticles. LSPRs are known to localize far-field light to a sub-diffraction-limited length scale, yielding an intense electric field at the particle surface. This effect has been harnessed to dramatically enhance light-matter interactions, leading to a variety of applications such as surface-enhanced Raman spectroscopy (SERS), photothermal cancer therapy and solar energy harvesting. Though a variety of near- and far-field optical methods are used to probe LSPRs, the spatial resolution of these methods is on the order of tens of nanometers, limiting their effectiveness. In contrast, electron energy loss spectroscopy (EELS) performed in a scanning transmission electron microscope (STEM) combines sub-nanometer resolving power with the capability to excite both optical-accessible and –inaccessible plasmon modes and therefore has emerged as one of the leading techniques (Figure 1). In this presentation, I will briefly introduce the STEM/EELS technique and demonstrate the power of STEM/EELS in the characterization of LSPRs. In addition to the traditional use of STEM/EELS for LSPR imaging, we have recently demonstrated that STEM/EELS can also be used to spatially map LSP-semiconductor energy transfer at the nanoscale. The future of STEM/EELS as a window into the nanoscopic world is especially promising, and we expect continued advances in the molecular, optical, materials, information, and energy sciences as a result.

Electron microscopy techniques have been used to probe the optical properties of materials in the subwavelength scale. In particular, it has been shown that using cathodoluminesncence (CL) and electron energy loss spectroscopy (EELS), along with a propitious choice of electron beam energy and target material, resolution below 10 nm is attainable.

In this seminar we will describe recent advances electron spectroscopies in a STEM to probe individual quantum structures. Different experiments attempting to probe their optical properties will be presented (including their lifetimes). Typically, a 1-nm-wide 60 keV electron beam was used to excite the sample. Fundamentally, we have measured spectroscopic signals (CL and EELS) and the second order correlation function (g²(t)) of the emitted light of different materials using an optical spectrometer, an EEL spectrometer and a Hanbury-Brown and Twiss interferometer.

We will start by showing how monochromated EELS and CL experiments are well adapted to probe excitonic excitations in different materials. We will show how EELS is specially adapted to probe excitons in 2-dimensional materials, using MoS₂ and MoSe₂ monolayers as example. For thicker materials, CL is able to more easily produce meaningful information from excitoninc excitations. As examples, we will discuss the luminescence of GaN quantum wells in AlN nanowires (Figure 1) and excitons in hBN flakes.

However, only spectroscopic information does not give a complete picture of the behavior of a specific system. One example are single photon emitters (SPE). To characterize such systems, light intensity interferometry is necessary (a typical experiment in quantum optics). We will demonstrate how a similar experiment using CL is possible, using as example the neutral Nitrogen-Vacancy center in diamond nanoparticles (Figure 2). Also, we will show how this setup allowed the detection of new SPE in hBN.

Finally, we will discuss how a CL setup with a HBT interferometer in a STEM microscope can be used to measure the lifetime of individual quantum emitters with a high spatial resolution (< 15 nm). We name this technique spatially resolved time-correlated cathodoluminescence, SRTC-CL. As an example, we will show that the lifetimes of 8 GaN quantum wells separated by 15 nm in an AlN nanowire can be measured by SRTC-CL.

Tailoring shape of metallic nanoparticles and alignment of nanoparticle assemblies allows controlling the properties of localized surface plasmon resonances, such as peak positions or near field-coupling and enhancement [1]. Electron beam lithography is a versatile tool for nanoparticle manufacturing, but the technique usually suffers from imperfections, surface roughness, and limited spatial resolution, which leads to particle shapes that deviate from design objectives. Similar limitations apply to chemical synthesis, which leads to nanoparticle assemblies with size dispersion and geometry variations. Therefore, to exploit the full potential of plasmonics, full 3D characterization and simulation taking into account the imperfections of real structures become mandatory. Here we present two different tomography-based approaches to understanding complex plasmonic nanoparticles created by electron beam lithography.

In our first approach the precise 3D geometry of a particle dimer fabricated by means of electron beam lithography was reconstructed through electron tomography. This full 3D morphological information was used as an input for simulations of energy-loss spectra and plasmon resonance maps (Figure 1). Here excellent agreement between measured EELS data and theory was found, bringing the comparison between EELS imaging and simulations to a quantitative and correlative level [2].

In our second approach we directly reconstruct particle plasmon fields from a tomographic tilt series of EELS spectrum images. While first approaches and demonstrations of plasmon field tomography were limited to very small particles [3–5] – small enough to neglect retardation – we lift this limitation with our approach making plasmon field tomography generally applicable to nanoparticles of all sizes [6]. Formulation EELS tomography as an inverse problem allows reconstructing the complete dyadic Green tensor for plasmonic particles, which is linked to the photonic local density of states (LDOS). Using this approach we are able to reconstruct the full 3D LDOS for a silver metallic nanoparticle (Figure 2).

This work overcomes the need for geometrical assumptions or symmetry restrictions of the sample in simulations and generalizes plasmon field tomography to particles of all sizes, paving the way for detailed investigations of realistic and complex plasmonic nanostructures.

[1]          S.A. Maier, Springer US, Boston, MA, 2007.

[2]          G. Haberfehlner et al., Nano Lett. 15: 7726–7730 (2015).

[3]          A. Hörl et al., Phys. Rev. Lett. 111:076801 (2013).

[4]          O. Nicoletti et al., Nature. 502: 80–84 (2013).

[5]          S.M. Collins et al., ACS Photonics. 2: 1628–1635 (2015).

[6]          A. Hörl et al., ACS Photonics. 2: 1429–1435 (2015).

We thank Joachim Krenn and Harald Ditlbacher for access to and support with electron beam lithography and helpful discussion. This research has received funding from the European Union within the 7th Framework Program [FP7/2007-2013] under Grant Agreement no. 312483 (ESTEEM2). We acknowledge support by the Austrian Science Fund FWF under project P27299-N27, the SFB F49 NextLite, and NAWI Graz

Recent improvements in energy resolution, enabled by the use of electron energy monochromators, have the potential to turn EELS into a tool able to provide quantitative information of localized surface plasmons (LSPs), such as damping effects in single particles and electron kinetics of single plasmon modes.[1] Crucial to the prospect of quantitative analysis of LSPs is the requirement that the experimental energy resolution must be better than the natural line width of the plasmon resonances, all the while retaining high enough signal-to-noise ratio to enable an accurate determination of the properties of interest.[1] The energy resolution of EELS is customarily determined by the full-width at half-maximum (FWHM) of the zero-loss (ZL) peak. The plasmon resonances, lying in the low-loss regime, often overlap with the broad tail of the ZL peak, blurring many spectral signatures of interest. Until now, several processing techniques had to be applied to overcome these issues, relying for instance on deconvolution algorithms [2] which can introduce artifacts [3] A new generation of electron monochromators now allows for high signal-to-noise ratios while varying the energy resolution controllably, down to the 10meV regime [4].

Here we present recent results aimed at spatially and spectrally resolving the plasmon resonances of individual plasmonic nanostructures and of functional plasmonic devices using a Cs-corrected and monochromated Nion UltraSTEM 100MC (‘Hermes’) microscope with a nominal energy resolution of 10meV. Figure 1 shows spatially resolved LSP modes of two individual Ag particles with different shapes and different number of crystalline domains for the energy ranges indicated in the figure. Both shape and crystallinity appear to affect the plasmonic response. We show how the energy resolution, which also affects the attainable signal-to-noise ratio and dictates the required integration (exposure) time, can be conveniently set and tuned, depending on the inherent properties of the system of interest. Figure 2A shows as-recorded EEL spectra taken from the centre of a Ag nanowire (inset) and showing a narrow bulk plasmon resonance at 3.85eV. We note that the FWHM of the peak does not decrease with decreasing energy resolution from 40meV to 16meV, meaning that 40meV must be below the natural line width of the resonance. In this context, we will discuss the prospect of not only characterizing bare metallic nanostructures, but of also interrogating chemically functionalized plasmonic nanostructures using EELS. We note that the accessible energy ranges (sub-40 meV) also allows us to probe molecules adsorbed onto metal nanostructures. In the raw spectra in Fig. 2B taken from different locations within a sample, spectroscopic signatures of aromatic thiols chemisorbed onto multiple Ag nanoparticles are shown.[5]

References:

[1] M. Bosman et al., Scientific reports, 2013, 3.

[2] R. F. Egerton in “Electron Energy-Loss Spectroscopy in the Electron Microscope“, Springer (New York), 3rd Ed., 2011.

[3] E. P. Bellido et al., Microsc. Microanal., 2014, 20, 767; V. Keast and M. Bosman, Microscopy research and technique, 2007, 70, 211; S. Lazar et al., Ultramicroscopy, 2006, 106, 1091.

[4] O. L. Krivanek et al., Nature, 2014, 514, 209-212

[5] SuperSTEM is the UK EPSRC National Facility for Aberration-Corrected STEM, supported by the Engineering and Physical Science Research Council. PZE acknowledges support from the Laboratory Directed Research and Development Program at Pacific Northwest National Laboratory. WPH is supported by the US Department of Energy, Office of Science, Office of Basic Energy Sciences, Division of Chemical Sciences, Geosciences & Biosciences.

Plasmonic nanostructures enable the concentration of light to the deep subwavelength regime and, thus, are the topic of intense fundamental and application oriented research. Nanoparticles acting as optical antennas are well known for their nanoscale mode volumes, due to the excitation of localized plasmon modes. In this context electron energy-loss spectroscopy (EELS) in a transmission electron microscope (TEM) became a powerful technique as it enables to map the full modal spectrum of plasmon eigenmodes with unprecedented high spatial resolution [1–3]. Beside EELS, Cathodoluminescence (CL) has also recently been used to gain information about the optical response taking advantage of the same high spatial resolution in a TEM [4]. While it is stated that EELS is linked to the full photonic local density of states (LDOS), the CL signal is related to the radiative LDOS [5].

In this work we present a combined EELS/CL study of plasmon eigenmodes on silver nanodisks, using fast electrons in a TEM. In particular we compare the differences of the EELS and CL response using experimental and simulated data. Precise variation of the disk size is achieved by means of electron beam lithography (Figure 1a), enabling a comprehensive study of plasmon excitations on silver nanodisks.

From theoretical considerations it is known, that for certain particle geometries (and therefore for specific surface charge distributions) there exist so called dark modes, which are “invisible” to photons but “visible” to electrons, and therefore can be measured with EELS but not with light [6]. Here we discuss how dark these dark modes are comparing EELS and CL (figure 1b+c). In particular, radial breathing modes (C in figure 1b+c) were predicted to be dark modes [6], although we will show how comparison between EELS and CL can mitigate this statement. Additionally, limitations for the theoretical predictions will be discussed, when the particle size is increased and therefore retardation effects become more important. In this case we show that dark modes are getting brighter. Furthermore symmetry breaking by the excitation source itself, a focused fast electron beam, will be discussed.

[1]          Nelayah et al., Nat. Phys. 2007, 3, 348−353.

[2]          Koh et al., Nano Lett. 2011, 11, 1323−1330.

[3]          Schmidt et al., Nat. Commun. 2014, 5, 4604.

[4]          Yamamoto et al., Phys. Rev. B 2001, 64, 205419.

[5]          Losquin et al., ACS Photonics 2015, 2, 1619–1627.

[6]          Schmidt et al., Nano Lett. 2012, 12, 5780−5783.

This research has received funding from the European Union within the 7th Framework Program [FP7/2007-2013] under Grant Agreement no. 312483 (ESTEEM2). We acknowledge support by the Austrian Science Fund FWF under project P21800-N20, the SFB F49 NextLite, and NAWI Graz.

High energy resolution electron energy-loss spectroscopy (EELS) is now a recognised technique for mapping localized surface plasmon resonances and measuring their resonant energies with a nanometric spatial resolution [1–3]. In contrast to light-based techniques it can further excite and map high order harmonics that are optically forbidden. Here we move on from the study of isolated or tandem plasmonic structures randomly deposited on TEM grids, or more complex structures painstakingly patterned by FIB, to a fast, systematic EELS mapping of precisely patterned Au and Ag films on Si₃N₄ membranes. Following an electron lithography-based preparation, the spatial distributions and mode energies of plasmon resonances and coupling are studied in function of well-controlled variations in structure dimensions.

The measurements are made in STEM-EELS mode using a FEI Titan Themis 60-300 with Gatan Digiscan and GIF Quantum ERS spectrometer. The combination of the X-FEG gun and monochromator gives a sub-nm incident beam with 100–110 meV FWHM of the zero-loss peak and a current of up to 240–250 pA. A fundamental need of the work is that the plasmon excitation can be measured both in the pure Si₃N₄ membrane regions and in the metal film regions. To this end, a high tension of 300 kV is used because, by reducing the relative intensity of bulk plasmon scattering from the metal films, it improves surface plasmon excitation signal to noise. EELS data are normalized by the zero-loss intensity to give the true projected plasmon distribution, without “shadowing” by the metal film [4]. With the fast spectrum imaging mode and high beam current, dwell times are only 0.2–0.25 ms per pixel, allowing us to acquire maps with >10⁵ pixels (e.g. 600 x 600 px) in < 10 minutes per map.

Applying this fast mapping with high spatial sampling to the lithographically-based structures of known layout and dimension gives a highly time-efficient method for studying plasmonic excitations in nanophotonic structures. Figure 1 shows example plasmon resonance low-loss EELS spectra and intensity maps. Parts (a–d) & (e–f) treat the well known plasmonic structures of a silver wire and nano-triangle. Owing to the energy resolution and good signal to noise ratio, there is no need to perform systematic deconvolution of the data to reveal plasmon excitations, even for modes at < 0.5 eV energy loss. High order multipoles are additionally well resolved for both non-penetrating and penetrating trajectories, such as the 2.8 eV breathing mode of the nano-triangle.

Figure 1 (g–h) shows data from more complex coupled gold heptamer apertures. We probe the effect of a nanoscale defect: a small 20 nm size gap in one of the heptamer arms. This feature induces the appearance of an additional low energy mode at 0.84 eV, and affects both the symmetry and intensity of higher order modes. This demonstrates how systematic study of varying geometries allows for in-depth analysis of plasmon resonances.

These experimental studies are combined with modeling of the excitations using a novel “in house” method which aids interpretation of multi-body interactions by simulating EELS spectra using the properties of plasmonic structures, and not the work done on electrons [5]. The simulations and experiments have a symbiotic relationship, with the modeling used to interpret experimental data, and the experimental data used to guide the modeling (e.g. for peak resonance values). Nevertheless, our experimental approach is significantly quicker than the modeling. While so far it is primarily applied to the systematic study of particles and apertures, in the future it will be used to explore the optical excitations of novel nanophotonic structures and materials.

References and Acknowledgements

[1] J. Nelayah et al., Nature Phys. 3 (2007) 348–353.

[2] M. Bosman et al., Nanotechnology 18 (2007) 165505.

[3] F.J. García de Abajo, Rev. Modern. Phys. 82 (2010) 209–275.

[4] N. Le Thomas et al., Phys. Rev. B 87 (2013) 155314.

[5] G.D. Bernasconi et al., J. Opt. Soc. Am. B 33 (2016) 768–779.

We thank the staff of the Center of Micro/Nanotechnology (CMI) of EPFL for support. This research was in part funded by the European Commission (FP7-ICT-2011-7, NANO-VISTA, under Grant Agreement No. 288263).

Cathodoluminescence (CL) and electron-beam-induced current (EBIC) are versatile techniques to characterize semiconductor materials and devices. In this talk, we review our achievement on the study of widegap semiconductors, GaN and SiC.

The major defects in GaN are dislocations. GaN wafers on sapphire include dislocations of 10⁹ to 10⁷ cm^-2 due to lattice mismatch. Fig. 1 shows the secondary electron (SE) and CL images of GaN wafer with different thickness. Such dislocations may agglomerate and form the hexagonal pits of micrometer size. These defects are detrimental for the device performance. Homoepitaxial GaN wafer take over the dislocations of 10⁶ cm^-2 from the seeds. We have performed CL study to distinguish dislocation characters and to clarify the effect of dislocations.

The major defects in SiC are threading screw dislocations (TSD) and stacking faults (SF). TSD act as the killer defects due to the surface roughness at the dislocation core region. SF may be generated and expanded due to e-beam irradiation. EBIC is very effective to characterize these defects in SiC. The e-beam enhanced defect generation of SF (Fig.2) will be reviewed.

At the end, we demonstrate 3D spectra imaging of CL, which is very promising to analyze the details of extended defects.

This work was supported from “GaN project”, Study of Future Semiconductors for Sustainable Society in MEXT, Japan.

Charge carrier lifetime is a key parameter for understanding the physics of electronic or optical excitations. For example the excited state can unveil details of environmental influence, specifically the role of non-radiative transitions. From a practical point of view, lifetimes can largely determine the performances of devices, such as Light Emitting Devices (LEDs) or photovoltaic cells. These usually rely on nanometer scale structures for which small details, such as the presence of single point defects, have to be known with atomic precision. Despite the success of super resolution optical microscopies, they fail as general tools for lifetime measurement at the nanometer scale. In this presentation we will show how we can take advantage of the nanometer probe size formed in a Transmission Electron Microscope and a phenomenon that we recently discovered (referred hereafter as the bunching effect [1]), to study lifetimes of emitter at the nanometer scale without using a pulsed electron gun.

The effect takes its name from the fact that the autocorrelation function g⁽²⁾(τ) of the CL signal coming from quantum emitters (points defects or more generally single photon emitters –SPE-, quantum confined structures…) may exhibit a peak at zero delay – which is a fundamental difference with PL. To measure this effect, we use an intensity interferometry experiment that measures the CL g⁽²⁾(τ). Figure 1 shows that, at low incoming electron currents (I < 100 pA), the g⁽²⁾(τ) of the CL signal intensity I(t) displays a large nanosecond-range peak at zero delay (g⁽²⁾(0) > 35) (bunching), the amplitude of which depends on the incoming electron current. This behavior strongly departs from the PL g⁽²⁾(τ) function which is flat when multiple independent SPE are excited. In this presentation we will show that it occurs because an emitter, like a quantum well, will be excited multiple times by a single electron and will emit a bunch of photons on a time window close to its radiative lifetime. As it will be proved, by simply fitting the experimental curve of the g⁽²⁾(τ) function by an exponential we can retrieve the lifetime of the emitter.

Using this effect we were therefore able to measure very efficiently lifetimes of Gallium Nitride quantum wells (QWs) separated by less than 15 nm, together with their emission energy and atomic structure (Figure 2). Experiments on well separated individual quantum structures shows an excellent agreement with combined time-resolved μ-photoluminescence. We also demonstrate the possibility to measure the lifetimes of emitters of different kinds (defects, QWs, bulk) within a distance of a tenth of nanometers even for spectrally overlapping emissions. This technique is readily applicable to large ensembles of single photon sources and various emitters such as QWs, quantum dots, point defects and extended defects, such as stacking faults (SF).

[1] Meuret et al, PRL 114 197401 (2015)

[2] L. H. G. Tizei and M. Kociak, PRL 110  153604 (2013)

InGaN quantum wells (QWs) show high internal quantum efficiencies over the ultraviolet to green spectrum and in white light emitting diodes (LEDs). However a persistent challenge to the development of higher efficiency devices is the strong polarisation field across the across the QWs along the polar axis. The polarisation induced internal electric fields cause the spatial separation of the electron and hole wavefunctions in the QWs, known as the quantum confined Stark effect (QCSE). It has been proposed that the internal electric field can be suppressed by silicon doping the quantum barriers (QBs) [1]. Moreover, Kim et al. have theoretically shown that the device efficiency may be improved by variations in the silicon dopant concentration through the QWs [2]. To confirm the simulated properties though, it is crucial to resolve the spectral properties of individual QWs.

In this study, nano-cathodoluminescence (nanoCL) reveals for the first time the spectral properties of individual InGaN QWs in high efficiency LEDs and the influence of silicon doping on the emission properties [3]. A silicon doped layer at 5×10¹⁸ cm^-3 is included immediately prior to the growth of the 1^st QW and the QBs between the QWs are subsequently doped to 1×10¹⁸ cm^-3 (sample A). Two further multiple QW InGaN/GaN structures were also grown for reference with QB doping levels of 1×10¹⁸ cm^-3 (sample B) and 1×10¹⁷ cm^-3 or less (sample C). NanoCL reveals variations in the emission wavelength that directly correlate with individual QWs. With QB doping greater than 1×10¹⁸ cm^-3, there is a continuous blue shift in the emission wavelength of each of the subsequently grown QWs. The inclusion of a higher doped layer immediately prior to the growth of the 1^st QW in the LED structure leads to a blue shift unique to the 1^st QW.

The experimental variations in the emission wavelengths were reproduced by Schrödinger-Poisson simulations. The blue shift in emission wavelength through the QWs due to QB doping is found to be caused by screening of the internal electric fields. The reduction in the emission wavelength of the first grown QW due to the higher doped layer is also found to be the result of screening of the internal electric field. The mitigation of the QCSE and consequently stronger overlap of the electron and hole wavefunction, thus should result in an increase in the radiative recombination. NanoCL thus may serve as an experimental approach to study and refine  the design of future optoelectronic nanostructures, including the effects from doping and lead to improvements in device efficiency and functionality.

[1] T. Deguchi, et al., Appl. Phys. Letts. 72, 3329 (1998)

[2] D. Y. Kim, et al., IEEE Photonics. 7, 1 (2015)

[3] J. T. Griffiths, et al., Nano Letts. 15, 7639 (2015)

Due to the specific size-dependent photoluminescence spectra of semiconductor nanocrystals (NCs), their use is promising as building blocks for new electronic and optical nanodevices such as light-emitting diodes, solar cells, lasers and biological sensors.^1,2 In order to design these NCs with tailored properties for specific applications, a high level of control over their synthesis is of key importance. Therefore, it is of great importance to characterize both the shape as the composition of these systems. Here, a range of different colloidal semiconductor NCs are characterized using 2D and 3D electron microscopy techniques.

We will discuss, 2D semiconductor CdSe nanoplatelets (NPLs), both flat as helical shaped³, which are investigated using electron microscopy techniques. The aim is to retrieve structural information using high resolution imaging which enables us to study the growth mechanism of these NPLs. The flat NPLs have mainly {100} edges (Figure 1.A) and only a thickness of 4 to 5 atomic layers (Figure 1.B). The analysis of the helical NPLs shows that they are zinc blende and that the helices are folded uniquely around the ⟨110⟩ axis (Figure 1.D). In order to retrieve the helicity of the ultrathin helical shaped platelets, electron tomography is applied. The three-dimensional tomographic reconstructions confirm that the observed helices fully rotate over a diameter of ∼25 nm and that they are not preferentially left- or right-handed (Figure 1.C).

Furthermore, heteronanocrystals (HNCs) are studied as they improve the stability and, thereby, the surface passivation of the NCs when overgrown with a shell of a second semiconductor with a higher bandgap.  In this manner, the robustness of the system and the photoluminescence quantum yield of the core is increased.⁴  In order to understand the growth process of HNCs, both the 3D structure as the position of the core inside that structure is of key importance. We investigate two types of CdSe/CdS core/shell HNCs, with either a nanorod or bullet shape. High resolution HAADF-STEM microscopy enables us to investigate the crystal structure of the core-shell nanostructure (Figure 2.A,C). Advanced electron tomography based on novel reconstruction algorithms⁵ is used to investigate the 3D shape and to reveal the position of the CdSe core in the CdS shell (Figure 2.B,D). For the CdSe/CdS core/shell bullets, the presence of two types of morphologies was revealed (Figure 2.D). High resolution STEM imaging was used to characterize the surface facets of both morphologies, which enabled us to compare the surface energy of both morphologies. For the CdSe/CdS nanorods, a sequential topotactic cation exchange pathway that yields CuInSe₂/CuInS₂ nanorods with near-infrared luminescence is further investigated⁶.

[1]  Somers, R. C.; Bawendi, M. G.; Nocera, D. G. Chem. Soc. Rev. 2007, 36, 579–591.

[2]  Talapin, D. V.; Lee, J.-S.; Kovalenko, M. V.; Shevchenko, E. V. Chem. Rev. 2010, 110, 389–458.

[3]  Hutter, E. M.; Bladt, E.; Goris, B.; Pietra, F.; van der Bok, J. C.; Boneschanscher, M. P.; de Mello Donegá, C.; Bals, S.; Vanmaekelbergh, D. Nano Lett. 2014, 14, 6257–6262.

[4]  Dabbousi, B. O.; Rodriguez-Viejo, J.; Mikulec, F. V; Heine, J. R.; Mattoussi, H.; Ober, R.; Jensen, K. F.; Bawendi, M. G. J. Phys. Chem. B 1997, 101, 9463–9475.

[5]  Goris, B.; Van den Broek, W.; Batenburg, K. J.; Heidari Mezerji, H.; Bals, S. Ultramicroscopy 2012, 113, 120–130.

[6]  van der Stam, W.; Bladt, E.; Rabouw, F. T.; Bals, S.; de Mello Donega, C. ACS Nano 2015, 9, 11430–11438.

The authors acknowledge financial support from the Research Foundation - Flanders (FWO).

Here, we report on the precise analysis of the atomic structure of inversion domain boundaries (IDBs) in GaN by scanning transmission electron microscopy.  IDBs are a common defect in GaN that traps carriers and leads to a slightly modified luminescence wavelength [1,2].

Our analysis of IDBs in MOCVD grown nanowires confirms recent coherent Bragg imaging (CBI) results [3] stating that the atomic structure of this IDB is different or slightly different from the one determined in 1996 by first-principle calculations (IDB*) [4].  CBI experiments measured a 8 pm shift of the c-planes of the two domains [3], whereas  first-principle calculations predicted no shift. A previous study by STEM [5] found a shift of "ca. 0.6 Å" (60pm), corresponding roughly to the switch of the Ga and N positions without any additional shift. Here in addition to a shift along  c, we show that the interface configuration corresponds qualitatively to the IDB* model (cf. Fig. 1) and that there is a 10 pm dilatation perpendicular to the interface (shown in Fig. 3) in agreement with this model, while CBI did not find a dilatation.

To facilitate the measurement of atom positions across the IDB with picometre-precision, we use HAADF-STEM to avoid coherent effects leading to artefacts. Scanning and drift artefacts are being suppressed by acquiring series of rapid STEM images and aligning them using the newly developed Zorro code. This algorithm is based on calculating estimated drift positions by correlating every frame to multiple frames and minimizing the error of the overdetermined system to obtain a best estimate for the frame positions relative to each other. The sub-pixel aligned frames are then averaged and the peak positions are determined via TeMA (template-matching algorithm).

Our quantitative analysis of experimental and simulated STEM images shows that when atomic columns are very close to each other the measured distance can be slightly different from the real value.  For instance, when the distance between atomic columns becomes smaller than 0.1 nm, the difference between the measured and real values can account for several picometres. This effect can be well observed when an IDB kinks perpendicular to the observation direction leading to closely projected columns in the overlap region of the two domains as seen in Fig. 2. Electron scattering simulations show that the apparently wider distance between atoms is a channeling effect.

These results have provided elements to revisit previous theoretical models of IDBs in GaN.

[1] T. Auzelle et al., Appl. Phys. Lett. 107, 051904 (2015).

[2] R. Kirste et al., J. Appl. Phys., 110, 093503(2011).

[3] S. Labat et al., ACS Nano 9, 9210 (2015).

[4] J. E. Northrup et al., Phys. Rev. Lett. 77, 103 (1996).

[5] F. Liu et al., Adv. Mater. 20, 2162 (2008).

In modern MOS devices, sources and drains are of nanometric dimensions and highly doped (dopant concentration typically > 10²⁰ at.cm^-3). Measuring such dopant concentrations and visualizing their spatial extensions in silicon, although mandatory for the development of the technology, is elusive in practice. Several TEM techniques such as EELS and EDX seem suitable to map dopant concentrations with the required resolution but while they are accurate to measure impurities concentrations, they cannot assess whether these impurities are on interstitial or substitutional sites, what is essential to define doping levels. Moreover, the detection of boron suffers from other physical limitations. Finally, bright-field electron holography has been reported to be suited for such measurements but transforming the electrostatic fields which are measured into doping concentrations is far from straightforward.

In this work, we have explored the possibility to extract boron concentrations from the measurement of changes of the silicon lattice parameter induced by the substitution of boron atoms. For this we use dark-field electron holography (DFEH) on specifically designed samples.

In a first part, we will present the DFEH principle [1]. This is an interferometry technique able to map strain with a precision of the order of 10^-4 and a few nanometers spatial resolution over micrometer fields of view. Two diffracted beams, one passing through an unstrained region of the lattice and acting as a reference, the other one passing through the region where strain has to be measured, are forced to interfere by using an electrostatic biprism and thus create an interference pattern (see figure 1). A phase map is extracted from the pattern by Fourier transform and converted into an atomic displacement field. By using two non-collinear diffraction vectors, all the components of the strain tensor in the observation plane can be obtained.

For our experiment, a sample consisting of five 50 nm-thick doped layers of increasing boron concentrations ranging from 3E¹⁸ at.cm^-3 to 8.5E¹⁹ at.cm^-3 was grown by RP-CVD, under conditions insuring both extremely low concentrations of impurities and the full activation of boron [2]. The sample was further checked by SIMS and ECVP measurements, demonstrating that 100 % of boron atoms are on substitutional sites in all the doped layers. DFEH was used to measure the deformation of the doped layers. We could thus deduce the silicon lattice expansion coefficient (β) resulting from the adding of boron atoms in the crystalline silicon network, from these measurements, as explained below.

The boron atoms being on substitutional sites, the Si:B doped layers can be seen as solid solutions as confirmed by the homogeneity of the deformations imaged by DFEH. These layers are pseudomorphic on the pure silicon lattice as confirmed by the mapping of the in-plane strain by DFEH. Thus, the change of the lattice parameter resulting from the incorporation of boron atoms is solely supported by the out-of-plane strain, through the Poisson’s reaction of the material (figure 2). From the modeling of this sample by FEM and taking into account the relaxation affecting the thin lamella used for DFEH, we are able to retrieve the values of the relaxed Si:B lattice parameter as a function of the substitutional boron concentration. As expected for a solid solution, we find a linear relation between these two parameters. Knowing the boron concentration and the Si:B lattice parameter profiles, we are able to deduce that β coefficient equals -6.5E^-24 cm³ (figure 3). Figure 4 compares the results we have obtained with those found in the literature, often measured by XRD.

Finally, β, the coefficient relating the boron concentration to the lattice parameter, allows us to transform a strain map obtained by DFEH into a “substitutional boron concentration” map with a precision of 3E¹⁹ at.cm^-3 and a spatial resolution of 5 nm. We will illustrate the DFEH effectiveness to measure and image dopant concentrations in “real samples” through few examples, and will discuss the complementarity of the information obtained by this method and by bright-field electron holography.

[1] M.J. Hÿtch, F. Houdellier, F. Hüe and E. Snoeck, Nature 453, pp. 1086-1090, 2008.

[2] F. Gonzatti, J.M. Hartmann, K. Yckache, ECS Transactions 16, pp. 485-493, 2008.

For devices, the junctions between the semiconductors and any metal contacts are crucial for the device performance. A heat treatment is commonly applied to improve the quality of the contacts. For GaAs nanowires with a Au-based contact, annealing can lead to a well-defined metal-GaAs junction within the nanowire [1]. Here, we report an in situ heating, high-angle annular dark-field scanning transmission electron microscopy (HAADF STEM) study on the formation and structural characteristics of such junctions. The nanowires are dispersed on a high-stability TEM heating chip, and local Au contacts are made by lithography before in situ heat treatment within the microscope (Figure 1) [2]. A replacement of GaAs by Au can take place and our study determined key aspects of the reaction mechanism and its kinetics such as the reaction rate and the activation energy. In general, the replacement proceeds one GaAs(111) bi-layer at a time, as demonstrated by lattice resolved HAADF STEM (Figure 2). Ga dissolves in Au and As desorbs, as was previously reported for planar GaAs-Au structures [3]. Using scanning precession electron diffraction (SPED) it was found that there is no fixed epitaxial relation between the newly formed 1D Au-phase and the original GaAs nanowire. The reaction rate and the activation energy for the exchange are accurately determined by tracking the interface between the two phases over relatively long (~0.5 μm) distances.

The morphology of the solid 1D Au-phase is the same as for the original GaAs nanowire. Within it, growth twins are observed. The morphology and growth twins do not alter upon cooling and reheating. For the case where the nanowire is attached to a relatively large Au contact (Au reservoir), the contact acts as a Ga diffusion sink, and only negligible amounts of Ga are found in the formed 1D Au-phase. The in situ STEM and electron diffraction results prove that the replacement reaction takes place in the solid state. For the case where the nanowire is attached to a limited Au supply, as would be the case for a small volume Au deposition onto the nanowire or a Au catalyst droplet used for the nanowire growth, the growing metal segment gradually becomes richer in Ga as the exchange reaction proceeds. The reaction rate is slowed down over time due to the Ga enrichment. Eventually and at sufficiently high temperatures, the Au-Ga segment becomes liquid. Upon cooling of such segments, different Au-Ga intermetallic phases form and the main phases could be identified using a combination of SPED and machine learning (Figure 3).

References:

[1]: M. Orrù, et al, Phys. Rev. Appl., 4, 044010, 2015. DOI: http://dx.doi.org/10.1103/PhysRevApplied.4.044010
[2]: V. T. Fauske, et al, submitted.
[3]: T. Sebestyen, Electronics Lett., 12, 96, 1976. DOI: 10.1049/el:19760075

The authors acknowledge: The Research Council of Norway for the support to the NorFab (197411/V30) and the NORTEM (197405) facilities, and the FRINATEK program (214235), NTNU for support of the initiative “Enabling Technologies” and the EU for support via ERC grant no. 259619 and grant no. 312483 ESTEEM2.

A new generation of electron beam monochromators has recently pushed the energy resolution of (scanning) transmission electron microscopes deep into the sub 20meV range [1]. In addition to the obvious increase in resolution which has made exploring the phonon region of the EELS spectrum possible [1], the increased flexibility of these instruments is proving hugely advantageous for materials science investigations. The energy resolution, beam current and electron optics can be adjusted seamlessly and traded off each other as necessary within a greatly increased range, as will be illustrated on a number of systems studied using a Nion UltraSTEM100MC ‘Hermes’ instrument recently installed at the SuperSTEM Laboratory.

This added flexibility was essential in particular in studying the crystal structure of Li- and Mn-rich transition metal oxides, whose pristine state is not yet fully understood in spite of their great potential as high-capacity cathode materials for Li-ion batteries. Only through complementary electron microscopy and spectroscopy techniques at multi-length scale could the structural make-up of Li_1.2(Ni_0.13Mn_0.54Co_0.13)O₂ crystals be described unambiguously. Systematically observing the entire primary particles along multiple zone axes (Fig. 1) reveals that they are consistently made up of a single phase, save for rare localized defects and a thin surface layer on certain crystallographic facets. More specifically, this careful STEM and EELS study at the atomic scale using revealed that the bulk of the oxides consists of randomly stacked domains that correspond to three variants of monoclinic structure. Crucially, it is also shown that only the very surface is composed of a different Co- and/or Ni-rich spinel with antisite defects [2].

Similarly, atomically-resolved monochromated EELS mapping and imaging were used to study the structure and chemistry of promising complex oxide thermoelectrics such as misfit layered cobaltates [3] or Ca-stabilized A-site deficient titanates [4]. For the latter, careful inspection of images and atomically resolved EELS chemical maps in two orthogonal directions suggests that in the Nd_0.6Ca_0.1□_0.3TiO₃ system, Ca predominantly occupies Nd vacancy shared sites, creating locally a higher occupation of the site and thus promotes short range vacancy-cation ordering in both a and b lattice directions. This in turn results in intricately modulated octahedral tilting distortions of the O sublattice, whose signature can be fingerprinted at the atomic scale through the near-edge fine structure of the Ti (and O) electron energy loss edges [5].

[1] O.L. Krivanek et al., Phil. Trans. Roy. Soc. 367 (2009), pp. 3683-3697; T.Miyata et al., Microscopy 5 (2014), pp. 377-382 ; O.L. Krivanek et al., Nature 514 (2014), pp. 209-212.

[2] A.K. Shukla, Q.M. Ramasse, C. Ophus, H. Duncan, F. Hage and G. Chen, Nature Communications 6 (2015), 8711.

[3] J.D. Baran, M. Molinari, N. Kulwongwit, F. Azough, R. Freer, D. Kepapstoglou, Q.M. Ramasse and S.C. Parker, J. Phys. Chem. C 119 (2015), pp. 21818-21827.

[4] F. Azough, D. Kepaptsoglou, Q.M. Ramasse, B. Schaffer and R. Freer, Chemistry of Materials 27 (2015), pp. 497-507.

[5] The SuperSTEM Laboratory is the U.K. National Facility for Aberration-Corrected STEM, supported by the Engineering and Physical Sciences Research Council (EPSRC). 

The aberration-corrected scanning transmission electron microscope (STEM) can provide real space imaging and spectroscopy at atomic resolution with a new level of sensitivity to structure, bonding, elemental valence and even spin state [1]. Coupled with first-principles theory, this represents an unprecedented opportunity to probe the functionality of complex nanoscale systems. Two case studies will be presented utilizing the Nion UltraSTEM: first, determining the origin of the unexpected ferromagnetism in ultrathin, insulating LaCoO_3-x (LCO) films, which is induced by the spontaneous ordering of oxygen vacancies to relieve misfit strain (see Fig. 1) [2]. Second, the origin of the colossal ionic conductivity in Y-stabilized ZrO₂/SrTiO₃ superlattices [3], which is due to spontaneous disordering of the oxygen sublattice in response to biaxial strain coupled with the incompatibility of the  oxygen sublattices at the interfaces (see Fig. 2) [4-6].

A JEOL ARM 200F has recently been installed in the National University of Singapore, equipped with UHR pole piece, ASCOR hexapole aberration corrector, Gatan Quantum ER and OneView camera and Oxford Aztec EDS system. It is installed within a JEOL isolated room with thermal radiation panels, vibration and magnetic field isolation systems. Fig. 3 shows a Ronchigram at 200 kV with near 70 mrad half angle flat phase region, and initial results from Si [110] show good splitting of the dumbbells with information transfer to sub-Angstrom levels, see Fig. 4. Results on oxides will be presented at the meeting [7].

1. J. Gazquez, et al., Nano Lett, 11, 973 (2011).
2. N. Biškup, et al., Phys. Rev. Lett. 112, 087202 (2014).
3. J. Garcia-Barriocanal et al., Science 321, 676 (2008).
4. T. J. Pennycook et al., Phys. Rev. Lett., 104, 115901 (2010).
5. T. J. Pennycook et al., Eur. Phys. J. Appl. Phys. 54, 33507 (2011).
6. Y. Y. Zhang et al., Adv. Mater. Interfaces 1500344 (2015).
7. S. J. Pennycook, et al., ACS Nano, 9, 9437–9440 (2015).

DOE BES Materials Science and Engineering Division, European Research Council Starting Investigator Award STEMOX # 239739, Fundación BBVA, DOE Grant No. DE-FG02- 09ER46554, National University of Singapore and JEOL Asia.

The introduction of two sorts of cations with different valence and size, such as Ba2+ or Sr2+ and Ln3+, in the A-sites of transition metal perovskite oxides has generated numerous remarkable properties such as high Tc superconductivity,  oxygen storage in cobalt based oxides for the realization of solid oxide fuel cell (SOFC) cathodes , CMR in manganates .

The investigation of the system Sm-Ba-Fe-O in air has allowed an oxygen deficient perovskite Sm2-εBa3+εFe5O15-δ (δ=0.75, ε=0.125) to be synthesized. In contrast to the XRPD pattern which gives a cubic symmetry (ap= 3.934Å), the ED patterns of this phase (Fig.1a) show superstructure spots corresponding to c=5ap”. HRTEM study (Fig1b) revealed that this phase is nanoscale ordered with a quintuple tetragonal cell, “ap ´ ap ´ 5ap”. Bearing in mind that one cubic cell corresponds to the formula Sm0.375Ba0.625FeO2.85, these tetragonal nanostructures can be formulated as Sm2-εBa3+εFe5O15-δ (δ~0.75, ε~0.125). They consist of 5 SmO/BaO layers stacked alternately with 5 FeO2 layers along c. The HAADF-STEM image (Fig.1c) of the Sm2-εBa3+εFe5O15-δ structure along the [100] is clearly established from the contrast segregation that the Ba2+ and Sm3+ cations are ordered in (001) layers along the c-axis. One observes rows of bright dots perpendicular to c, which corresponds to three sorts of Sm or Ba cationic layers, judging from their intensity: pure Sm, mixed Ba/Sm and pure Ba layers. Thus the HAADF-STEM image can be interpreted by the following periodic stacking sequence of the A cationic layers along the c axis: “Sm-Ba-Sm/Ba-Sm/Ba-Ba-Sm”.

It appears from the ABF-STEM images along [100] (Fig.1d) and [110] orientations (Fig.1e) of a single Sm2‑εBa3+εFe5O15-δ domain, that the oxygen positions in all the layers are close to the ideal octahedral positions. However, a closer inspection of the images reveals that the oxygen columns in the equatorial positions close to the Sm layer deviate from their ideal octahedral position, and lie closer to the Sm3+ cations yielding a “zigzag” contrast along [100] and [110].

The “Sm-Ba-Sm/Ba-Sm/Ba-Ba-Sm” chemical ordering is also confirmed by elemental EELS mapping (Fig.2a,b). The spatially resolved EELS data show that the O-K edge spectra corresponding to the “FeO2” planes (labeled A,B,C) exhibit different intensity ratios of the two pre-peaks to the O-K edge, prepeak1/ prepeak2 at approximately 529/531 eV, depending on the nature of the surrounding “Sm,Ba” layers (Fig.2c). The O-K fine structure in the Sm plane is very similar to that of plane A, whereas those of the Ba and Ba/Sm planes are similar to B and C planes respectively. A first observation is that pre-peak 1 at ~529 eV is less intense for oxygen anions close to Sm cations (SmO layers as well as A layers). According to the literature, the height of this pre-peak is generally rather independent of the rare earth element and should be around the same height as pre-peak 2. Pre-peak 2, related to Fe3d eg – O2p hybridized states seems invariant in the structure, apart from the c plane where it is slightly subdued, accompanied by an increase of pre-peak 1 below 530eV. Pre-peak 1 can be attributed to a charge transfer from the eg to the t2g band of Fe (the eg band is usually empty for Fe3+). This increase of pre-peak 1 related to Fe3d t2g – O2p hybridized states is also visible in the Ba/Sm mixed layers, and can be linked to the presence of oxygen vacancies in those planes. This peak is stronger in the C plane suggesting the presence of more vacancies in this plane.The spatially resolved EELS spectra of the Fe-L2,3 edge are plotted in Fig. 2c. The Fe L3 and L2 “white lines” arise from transitions of 2p3/2 → 3d3/23d5/2 (L3) and 2p1/2 → 3d3/2 (L2) and are known to be sensitive to valency and coordination. Our data shows that the A and B FeO2 planes exhibit very similar Fe-L2,3 edges, with an L3 peak maximum at 709.5 eV, and a pre-peak to L3, even if faint at 708 eV. The energy position of the L3 maximum, together with the shape and positions of the L3 and L2 are then indicative of Fe3+ in an octahedral coordination. All the acquired Fe L3 edges are significantly broadened with respect to the plotted references for 6-fold, 5-fold and 4-fold coordinated Fe3+. This broadening can be explained by a change in coordination of the Fe atoms. Bearing in mind that the measured oxygen stoichiometry is 14.25, instead of 15, this suggests that the iron coordination is mainly 6, i.e. octahedral, but may also be mixed with the presence of some FeO5 pyramids in those layers.

The nanoscale ordering of this perovskite explains its peculiar magnetic properties on the basis of antiferromagnetic interactions with spin blockade at the boundary between the nanodomains. The variation of electrical conductivity and oxygen content of this oxide versus temperature suggest potential SOFC applications.

Emergent phenomena at complex oxide interfaces continue to attract attention as the basis for a variety of next-generation devices, including photovoltaics and spintronics. Tremendous progress has been made toward understanding the role of interfacial defects, cation intermixing, and film stoichiometry in single heterojunction systems; however, the techniques commonly used to study these interfaces, such as X-ray photoelectron and absorption spectroscopies, are either sensitive only to near-surface regions or do not offer depth resolution to probe individual interfaces. Here we explore the induced polarization in superlattices of LaCrO₃ (LCO) and SrTiO₃ (STO) using a combination of aberration-corrected scanning transmission electron microscopy (STEM) and monochromated electron energy loss spectroscopy (STEM-EELS). We show that a correlative approach, utilizing an array of local and non-local probes, is necessary to fully understand the defect-mediated origin of the induced polarization in this system.

We have conducted detailed structural characterization of several LCO-STO superlattices, as shown in Figure 1. We employ high-angle annular dark field imaging (STEM-HAADF) to directly measure the induced ferroelectric polarization in the STO layers. We first acquire a relatively high-speed time series of multiple fast frames (0.4 µs px^-1), which are then aligned using both rigid and non-rigid registration to remove both sample drift and scan distortion [1]. Using this procedure we directly measure the induced polarization with picometer precision, as we have demonstrated elsewhere [2]. Our results reveal that the built-in asymmetric potential across the LCO / STO interfaces is sufficient to induce a sizable polarization, on the order of 40-70 µC cm^-2, in good agreement with ab initio calculations [3].

We next perform detailed characterization of chemical intermixing and local electronic fine structure changes to explore how defects affect the induced polarization. Figure 2 shows the result of monochromated EELS measurements of the Ti L₂₃ edge fine structure, overlaid onto the integrated Ti L₂₃ edge signal. An improved energy resolution of better than 0.120 eV allows us to observe significant Ti intermixing through the superlattice, as well as subtle fine structure changes in the vicinity of the LCO layers not apparent in earlier data. Mapping the Ti L₃ t_2g – e_g crystal field splitting across the film, we find evidence consistent with a slight reduction in Ti valence from 4+ to 3+ in the vicinity of the LCO layers, possibly the consequence of La³⁺ substitution for Sr²⁺ or oxygen vacancies. Measurements of the Ti L₃ t_2g / e_g ratio also point toward such a trend: moving from the STO toward the LCO layers the ratio begins to decrease within the intermixed region, indicating a redistribution of electrons from t_2g to e_g states, suggesting a reduction in valence. In light of these results, our experimental STEM-HAADF measurements and accompanying ab initio calculations indicate that the induced polarization is robust against even sizable chemical intermixing and defect formation.

References

[1] Jones, L. et al, “Smart Align – a new tool for robust non-rigid registration of scanning microscope data”, Advanced Structural & Chemical Imaging 1:8 (2015).

[2] Spurgeon, S. R. et al, “Polarization screening-induced magnetic phase gradients at complex oxide interfaces.” Nat. Commun. 6, 1–11 (2015).

[3] Comes, R.B. et al, “Interface-induced Polarization in SrTiO₃-LaCrO₃ Superlattices.” Adv. Mater. Int. (2016). DOI: 10.1002/admi.201500779

A–site defficient perovskites are of particular interest for a range of engineering applications, such as ionic conductors, batteries and thermoelectric materials. The attraction of these systems lies in particular in their chemically stability over a wide range of compositions and operating conditions. Moreover, they can exhibit different degrees structural ordering, such as cation-vacancy or oxygen-octahedral-tilt-domain ordering [1,2] which can significantly affect their macroscopic physical properties.

In this work we combine state-of-the-art monochromated core loss electron energy loss spectroscopy (EELS) measurements with advanced image analysis to investigate the interplay between A-site cation-vacancy ordering and oxygen octahedral tilting domains in a series of Nd_2/3xTiO₃ double perovskites, exhibiting different degrees of structural ordering as a function of their heat treatment. [3,4] As a result they exhibit vastly different thermoelectric properties related to the presence of domain boundaries which  can suppress the thermal conductivity due to increased phonon scattering. [5,6]

 High-angle annular-dark-field (HAADF) imaging and large scale atomically-resolved EELS maps were used to investigate the chemical ordering of cations on the so-called A site of the perovskite structure (Figure 1), while pm-precision annular bright field (ABF) imaging of these compounds reveals the presence of TiO₆ tilting domains, which we show can be correlated with variations in the A-site occupancy (Figure 2).   Furthermore, advanced image analysis of the electron micrographs was used to measure local distortions in the TiO₆ lattice

These local distortions are further investigated using atomically-resolved monochromated EELS measurements with an energy resolution better than 0.1eV, in a Nion UltraSTEM 100MC. In particular, measurements of the of the Ti L_2,3 edge (Figure 3), reveal local changes of the near-edge fine structure, usually not discernible with conventional EELS measurements. Changes in the Ti L_2,3 pre-peak intensity, as well as subtle local variations in the Ti L₃ e_g/t_g intensity ratios, clearly observable thanks the relatively high beam current still available at this energy resolution, are indicative of local distortions in the oxygen octahedral sublattice.

References

[1] D.M. Smyth, Annu. Rev. Mater. Sci. 15, 329 (1985).

[2] M. Labeau, I.E. Grey, J.C. Joubert, H. Vincent, and M.A. Alario-Franco, Acta Crystallogr. Sect. A 38, 753 (1982).

[3] S. Jackson, et al, J. Electron. Mater. 43 (2014) p.2331.

[4] F. Azough, et al, Chem. Mater 27 (2015) p.497.

[5] K. Koumoto, et al, Annu. Rev. Mater. Res. 40 (2010) p.363.

[6] D.J. Voneshen, et a,l Nat Mater 12 (2013) p.1028.

[7] SuperSTEM is the U.K’s national facility for Aberration Corrected STEM funded by EPSRC.  

The study of novel physical properties appearing when two materials are interfaced has become one of the major fields of research in solid state physics over the last decade. For example, in the strive for novel non-silicon based electronics, the discovery of the formation of a conductive layer right at the interface region between 2 insulators (for example LaAlO₃ and SrTiO₃); a so-called two-dimensional electron gas (2DEG) or two dimensional electron liquid appears. Another important example of such emergent phenomena is the appearance of interface magnetism or superconductivity. As the materials involved in those new physical phenomena are often complex oxides, many factors such as strain, oxygen stoichiometry, cation intermixing, oxygen octahedral coupling.…[1,2] have to be considered when discussing their origin. The exact understanding of those phenomena is a key factor in order to turn these research ideas into working devices and in order to search for the most optimal materials.

In parallel, advances in electron microscopy instrumentation and techniques such as the introduction of aberration-correctors of the probe-forming lens have made it possible to achieve sub-angstrom spatial resolution allowing the study of materials on an atom column by atom column basis. High Angle Annular Dark Field (HAADF) combined with Annular Bright Field (ABF) imaging have made it possible to study and understand respectively the cationic and the oxygen sub-lattices in these materials.

As a first example, in the case of a (La,Sr)MnO₃ (LSMO) film grown on a NdGaO₃ (NGO) substrate, the magnetic easy axis of the LSMO film can be reversed by adding a SrTiO₃ (STO) buffer layer between the substrate and the film. Using the statistical analysis of STEM ABF images, we could reveal that even one single STO layer is enough to remove the transferred octahedral tilt of NGO into the LSMO film reversing its magnetic easy axis from the a to the b axis(see ref 1 and figure 1).

In a second part, The appearance of superconductivity in a superlattice of (Sr,Ca)CuO₂ (SCCO) and BaCuO₂ (BCO) will be investigated. Comparing two different heterostructures where 8 layers of SCCO are sandwiched with STO (no superconductivity) and BCO (superconductivity) the effect of the structure of the different layers (orientation of the CuO₂ in the form of planes or chains) on the appearance of superconductivity will be presented.

[1] Z. Liao, M. Huijben, Z. Zhong, N. Gauquelin, S. Macke, R. Green, S. van Aert, J. Verbeeck, G. Van Tendeloo, K. Held, G. A. Sawatzky, G. Koster & G. Rijnders, Nat. Mater. 15 (2016), p.425-431

[2] N. Gauquelin, E. Benckiser, M.K. Kinyanjui, M. Wu, Y. Lu, G. Christiani, G. Logvenov, H. –U. Habermeier, U. Kaiser, B. Keimer & G.A. Botton, Physical Review B 90 (2014), article 195140

J.V., G.V.T. and S.V.A. acknowledge funding from FWO projects G.0044.13N and G.0368.15N.  N.G. and J.V. acknowledge funding from the European Research Council under the 7th Framework Program (FP7), ERC Starting Grant 278510 VORTEX. All authors acknowledge financial support from the European Union under the Seventh Framework Program under a contract for an Integrated Infrastructure Initiative (Reference No. 312483-ESTEEM2).

Although defects may be seen as detrimental, they play essential roles in many materials. In some complex oxides for instance, they present an opportunity to enhance particular properties, or even engineer new ones. An archetypal example is the high-Tc superconductor YBa₂Cu₃O_7−δ (YBCO), where defects are indispensable to immobilize quantized vortices in the presence of magnetic fields. Therefore, determining the atomic structure of defects is critical to unravel and control their effect on the physical properties. In this talk, we will do precisely that with an unforeseen complex point-defect that leads to an unexpected formation of ferromagnetic clusters embedded within a superconductor. The commonest YBCO defect comprises the doubling of Cu-O chains between Ba-O planes, a ubiquitous intergrowth regardless of the deposition technique. Here, we will unclothe the true nature of this defect using a combination of experiments and theory to provide a complete picture of the structure and chemistry of these intergrowths at the atomic-scale and their effect on the electronic and magnetic properties of YBCO. First, we will show, by means of scanning transmission electron microscopy (STEM), how the system solves the local off-stoichiometry induced by the extra Cu-O chain, removing half of the Cu atoms in selected chains, and the distortions induced by the vacancies. Secondly, we will show using density functional theory (DFT) how the complex structure of these intergrowths affects the electronic properties, and yields a magnetization density that extends into the neighboring Cu-O planes. Finally, we will present X-ray magnetic circular dichroism (XMCD) spectroscopy results, which provide evidence of the theoretically predicted Cu magnetic moments and the presence of a dilute network of magnetic defects within the high-Tc superconducting state, see Figure 1. See for more details “Emerging dilute ferromagnetism in high-Tc superconductors driven by point defect clusters”, Adv. Sci. 2016, 1500295.

Highly adaptable crystal structures of complex oxide materials enable changes in composition and provide the opportunity of fabricating them in different forms e.g., thin films and/or heterostructures. Complex oxide heterostructures show noteworthy electronic properties which are absent in bulk forms, and the phenomena occurring at their interfaces are intriguing. The interest attributed to these structures and interfaces is also related to the wide variety of the functionalities such as ferroelectricity, magnetism, superconductivity etc. [1]. One of the most exciting interface effects is high temperature interfacial superconductivity discovered in heterostructures consisting of non-superconducting layers of insulating La₂CuO₄ (LCO) and metallic La_1.55Sr_0.45CuO₄ (LSCO). Moreover, cuprate bilayers are also attractive in which the critical temperature can be enhanced remarkably in comparison to single phase systems [2].

Atomic-layer-by-layer oxide molecular-beam epitaxy (ALL-oxide MBE) with its unique capabilities allowed us to synthesize bilayers consisting of 3 unit cells of metallic La_1.6M_0.4CuO₄ layer and 3 unit cells of the undoped insulating (I) LCO layer neither of which is superconducting by its own, where M represents the dopant, namely Sr²⁺, Ca²⁺, and Ba²⁺ which have the same ionic charge but different ionic radii. The bilayers are superconducting with a critical temperature of ~17.0 K, ~37.0 K and ~38.5 K for Ca, Sr and Ba dopants, respectively.

For STEM investigations, a JEOL-ARM 200F STEM equipped with a cold field-emission electron source, a probe C_s corrector (DCOR, CEOS GmbH), a 100 mm² JEOL Centurio SDD-type EDX detector, and a Gatan GIF Quantum ERS spectrometer was used. High-angle annular dark-field (HAADF) imaging and electron energy-loss spectroscopy (EELS) were performed at probe semi-convergence angles of 20 mrad and 28 mrad, respectively. Collection angles for HAADF and annular bright-field (ABF) images were 75-310 mrad and 11-24 mrad, respectively.

Atomically resolved ABF and HAADF images were simultaneously acquired. HAADF images unveil not only the perfect interfaces between substrate and the bilayers but also the high structural quality (Fig 1a-b). Atomic resolution EELS and energy-dispersive x-ray spectroscopy (EDXS) analyses were conducted to examine elemental distributions across the interfaces. EDXS analyses and EEL spectrum images (SI) reveal different cation redistribution behaviors and lengths depending on the cation size mismatch with respect to the La ionic radius. In Fig.1c the distribution of La and Sr across the interfaces of a Sr-doped bilayer sample is presented, which reveals relatively homogeneous distribution of Sr and also the redistribution of Sr near the interface.

Through ABF imaging, oxygen positions were determined, interatomic distortions (cation – cation and CuO₆ octahedra) were measured and correlated with the dopant distributions obtained via chemical analyses. In Fig. 2a two of the quantitatively studied images (i.e. HAADF and ABF), are presented as an overlay and the arrows indicate the nominal interfaces. Figure 2b shows the measured interatomic distances which are indicated with arrows. Results for the different dopants will be presented and correlations with high-temperature interfacial superconductivity will be discussed [3].

References

[1] H. Y. Hwang et al., Nature Materials, 11 (2012) p.103.

[2] A. Gozar et al., Nature, 455 (2008) p.782.

[3] The research leading to these results has received funding from the European Union Seventh Framework Programme [FP7/2007-2013] under grant agreement n°312483 (ESTEEM2).

Carbon in the sp¹ hybridization (carbyne) is able to form atomic chains that constitute the elementary one-dimensional phase of carbon [1]. Despite many efforts, the synthesis of carbon chains in appreciable quantities remains difficult and even the existence of bulk carbyne is subject of an ongoing controversy. However, the existence of individual carbon chains is undisputed since they have been observed in in-situ TEM studies. The present work allowed, for the first time, to measure the electrical properties of carbon chains by using a dedicated specimen stage for establishing electrical contacts in the TEM.

Carbon chains are unusual conductors that may occur either as metallic cumulene with double bonds or as semiconducting polyyne with alternating single and triple bonds. Now, as it became possible to characterize them electrically, these two electronic configurations can be distinguished. A piezo-driven tip (Nanofactory), integrated into the specimen holder of a TEM, allowed to establish contacts to graphenic material and, by controlled retraction of the electrodes, to unravel chains of carbon atoms (fig. 1a). At the same time, the electrical properties could be measured [2, 3]. By recording current-voltage curves of individual carbon chains, both cumulene and polyyne were identified (fig. 1b). It was found experimentally and by quantum conductance calculations (fig. 1c) that transport through narrow resonant states makes the conductivity much lower than predicted in previous theoretical work. At high applied bias, however, a sudden rise in current occurs, showing the absence of conduction channels at lower energy and their presence at higher energy.

When the 1D system is under strain, the chains exhibit a semiconducting behavior, corresponding to polyyne. Conversely, when the chain is unstrained, an ohmic behavior, corresponding to cumulene, is observed (fig. 1b) [4]. This confirms a recent theoretical prediction, namely that the Peierls distortion, which would stabilize polyyne, is suppressed by zero-point vibrations in an unstrained chain so that cumulene is the stable configuration. In the presence of strain, however, polyyne is favoured by the Peierls instability. Thus, a metal-insulator transition can be induced by adjusting the strain. Furthermore, it is shown that these atomic chains can act as rectifying diodes when they are in a non-symmetric contact configuration, i.e., between a carbon and a metal contact or between two carbon contacts of different type.

Acknowledgements:

Financial support by the LABEX program "Nanostructures in Interaction with their Environment" (NIE) and different projects of the Agence Nationale de Recherche (ANR, France) is gratefully acknowledged.

 References:

1. F. Banhart, Beilstein J. Nanotech. 6, 559 (2015).

2. O. Cretu, A. R. Botello-Mendez, I. Janowska, C. Pham-Huu, J.-C. Charlier, F. Banhart, Nano Lett. 13, 3487 (2013).

3. A. La Torre, F. Ben Romdhane, W. Baaziz, I. Janowska, C. Pham-Huu, S. Begin-Colin, G. Pourroy, F. Banhart, Carbon 77, 906 (2014).

4. A. La Torre, A. Botello-Mendez, W. Baaziz, J.-C. Charlier, F. Banhart, Nature Comm. 6, 6636 (2015).

The characterization of low-dimensional materials such as graphene or transition metal dichalogenides (TMDCs) often requires more than one technique to obtain a thorough understanding of their attributes for specific applications. Graphene and TMDCs both have layered structures and properties that vary significantly with thickness as compared to single layer conformations, making them very interesting for electronics design [1, 2]. Electronic device performance optimization can benefit greatly from knowledge of their crystalline structure and exciton dynamics. The aim of the following analysis is to show how several spectroscopy (Raman/Photoluminescence) and microscopy techniques (AFM/SEM) can, in correlation, provide a more detailed depiction of low-dimensional materials than the constituent measurements could offer in isolation.

Raman spectroscopy, and Raman imaging in particular, has proved to be of great value in differentiating spectra obtained from single, double and multi-layered low-dimensional materials. Raman imaging was also used to evaluate strain, doping, chirality and disorder in graphene and TMDCs [3, 4]. The information acquired with Raman spectroscopy and imaging can be complemented by data from highest resolution microscopy techniques such as Atomic Force Microscopy (AFM), Scanning Near-Field Microscopy (SNOM), or Scanning Electron Microscopy (SEM).

Fig. 1a shows a correlative Raman-SEM image of exfoliated graphene deposited on a non-conductive glass coverslip. The SEM image was acquired in the low vacuum mode using a BSE detector. Different graphene sheets can be clearly visualized. In the center of the SEM image a confocal Raman image was acquired, showing the presence of different number of graphene layers, from one (green) to ten (yellow). The high resolution confocal Raman image (Fig. 1b), acquired from the area marked in Fig. 1a shows the intensity distribution of the G band as a function of layers in the graphene sheet as indicated with 1, 2 and M (multy-) layers. By evaluating the intensity of the D-band, the chirality of the graphene crystal could be determined as highlighted in red color in Fig. 1b. The SNOM image from the same area (Fig. 1c) shows that the transparency of the graphene decreases with increase of number of layers providing direct access to the fine structure of graphene.

In a second example a high resolution SEM image of CVD grown MoS2 is pesented (Fig. 2a). The Raman-SEM image (RISE image)  presented in Fig. 2b shows a correlation of changes in the Raman spectrum at the defects of the crystals.

References:

[1] A. K. Geim , I. V. Grigorieva, Nature 499 (2013), p. 419.

[2] F. H. L. Koppens, T. Nueller, P. Avouris, A.C. Ferrari, M. S. Vitiello, M. Polini, M. Nat. Nano  9 (2014) p. 780.

[3] Y. You, Z. Ni, T.Yu, Z. Shen, Appl. Phys. Letters 39 (2008) p.13112.

[4]  P.K. Chow, R.J. Gedrim, J. Gao, T. M. Lu, B. Yu, H. Terrones, N. Koratka, ACS Nano (in press)

The tunable electrical properties of reduced graphene oxide (RGO) make it an ideal candidate for many applications including energy storage.¹ However, in order to utilize the material in industrially applied systems it is essential to understand the behavior of the material on the nanoscale, especially how naturally occurring phenomenon like wrinkling effects the electronic transport. While there have been many theoretical investigations on the transport behavior, there are much fewer experimental measurements.^{2, 3, 4} Here we used a transmission electron microscope (TEM) with scanning tunneling microscope (STM) probe in-situ holder is used to perform localized electrical measurements on wrinkled, supported RGO flakes. The RGO flakes are deposited onto pre-patterned Au electrodes on SiN membranes. The flakes are 1-3 layers thick and have lateral dimensions on the order of single micrometers. They are deposited through an ammonium laurate (AL) aqueous surfactant. The TEM allows for observation of the local wrinkling structure of the RGO and simultaneously a Nanofactory STM-TEM holder is used to perform localized resistance measurements. The holder allows for manipulation with and positioning of Au STM probes, with diameters less than 100 nm, with sub-nanometer precision. The results show that the overall resistance is low, on the order of single kΩ, if the contact between the probe and the RGO is optimized.  Wrinkles reduce the contact resistance between RGO and the probe. The overall trend for more than 70 measurements is that the total resistance decreases with increasing amount of wrinkles and that the contact resistance dominates the resistance when the probe makes contact with wrinkle free RGO. The sheet resistance and maximal contact resistance for our measurements range from 5-20 kΩ/☐ and 6-20E-7Ω cm², respectively, which falls in line with other scanning probe measurements for graphene and metal contacts.^{5, 6}  Here we use a bottom metal contact for the gold electrodes which eliminates the risk of resist contamination and metal doping of the graphene.⁷ These measurements give evidence that RGO with large amounts of wrinkling on a substrate can transport electrons well and that it is thus not necessary to produce suspended, pristine graphene in order to benefit from its electrical transport properties.

(1) Novoselov, K. S.; Fal'ko, V. I.; Colombo, L.; Gellert, P. R.; Schwab, M. G.; Kim, K. A roadmap for graphene. Nature 2012, 490, 192-200.

(2) Guinea, F.; Horovitz, B.; Le Doussal, P. Gauge fields, ripples and wrinkles in graphene layers. Solid State Communications 2009, 149, 1140-1143.

(3) Guo, Y.; Guo, W. Electronic and Field Emission Properties of Wrinkled Graphene. Journal of Physical Chemistry C 2013, 117, 692-696.

(4) Ladak, S.; Ball, J. M.; Moseley, D.; Eda, G.; Branford, W. R.; Chhowalla, M.; Anthopoulos, T. D.; Cohen, L. F. Observation of wrinkle induced potential drops in biased chemically derived graphene thin film networks. Carbon 2013, 64, 35-44.

(5) Yan, L.; Punckt, C.; Aksay, I. A.; Mertin, W.; Bacher, G. Local Voltage Drop in a Single Functionalized Graphene Sheet Characterized by Kelvin Probe Force Microscopy. Nano Letters 2011, 11, 3543-3549.

(6) Robinson, J. A.; LaBella, M.; Zhu, M.; Hollander, M.; Kasarda, R.; Hughes, Z.; Trumbull, K.; Cavalero, R.; Snyder, D. Contacting graphene. Applied Physics Letters 2011, 98.

(7) Babichev, A. V.; Gasumyants, V. E.; Egorov, A. Y.; Vitusevich, S.; Tchernycheva, M. Contact properties to CVD-graphene on GaAs substrates for optoelectronic applications. Nanotechnology 2014, 25.

We recently demonstrated a controllable and reproducible method to obtain suspended monolayer graphene nanoribbons with atomically defined edge shape [1]. Our method exploits the electron-beam of a Scanning Transmission Electron Microscope (accelerated at 300 kV) to create vacancies in the lattice by knock-on damage and pattern graphene in any designed shape. The small beam spot size (0.1 nm) enables close-to-atomic cutting precision, while heating graphene at 900 K during the patterning process avoids formation of beam-induced Carbon deposition and allows self-repair of the graphene lattice.  Self-repair mechanism is essential to obtain well-defined (zig-zag or armchair) edge shape and, if the electron beam dose is lowered, to perform non-destructive imaging of the graphene nanoribbons.

Drawing the electron-beam path with a software script, we were able to obtain reproducible graphene nanoribbons with sub 10 nm width. Using an in-house built microscopy holder equipped with electrical feedthroughs, we performed 2 and 4 wire measurements on several graphene nanoribbons, with different number of layers. Wide nanoribbons (width> 50 nm) exhibited ohmic behaviour, with conductivity linearly proportional to the width. Narrower ribbons (10 nm < width < 50 nm) displayed non-linear current-voltage (IV) relationship at low temperature (4 K, ex-situ), possibly indicating the opening of a band gap. The narrowest ribbon was 1.5 nm wide, with strong non-linear IV also at room temperature.

Electrical measurements were also performed in the high temperature range (300 K – 900 K), from which we concluded that thermal generated carriers give the main contribution to  electrical conductivity above room temperature.

Ackwnoledgements: This work was supported by ERC funding,  project 267922 - NEMinTEM 

References

[1]        Q.Xu, M. Wu, G. F. Schneider, L. Houben, S.K. Malladi, C. Dekker, E. Yucelen, R.E. Dunin-Borkowski, and H.W. Zandbergen, ACS Nano  7 (2), 1566-1572, 2013

        In this work, we perform an in-depth study of the cyclic telescoping of multi-walled carbon nanotubes (MWCNTs) inside a transmission electron microscope. The nanotubes are observed in real time, while simultaneously measuring their electrical properties. To date, examples of in situ TEM nanotube telescoping have remained few in number due to the highly specialized equipment and technical expertise required.

        Experiments were conducted by first ‘sharpening’ a MWCNT by approaching it with a piezo-controlled tungsten tip and applying a voltage pulse. This results in controlled and localized peeling of the outer walls, giving rise to a telescope structure. The exposed core is then contacted with the tungsten tip and reproducibly pulled out then re-inserted, while measuring conductance behavior at each movement step. Figure 1 shows an example of in situ telescoping of a MWCNT in real time.

        Results demonstrate that an electrical current can be maintained for multiple in-and-out cycles (up to 11 cycles have been tested to date), and for telescoping distances of up to 650 nm. This is exemplified in Figure 2. The systems display complex conductance behavior, with of some MWCNT telescopes demonstrating improved conduction (and current) as a function of the number of cycles, while for most these properties diminish.

        Control experiments have been conducted to investigate the effect of the electron beam on the behavior of the MWCNTs, and have indicated that changes in conductance behavior are not strongly influenced by the imaging process under normal beam conditions.

        One potential application for these structures is use as flexible electrical contacts, which allow for relative motion of the two electrodes, while maintaining electrical conduction.

        The authors would like to acknowledge funding as provided by the National Institute for Materials Science (NIMS), through the International Center for Materials Nanoarchitectonics (MANA).

Controlled manipulation of materials on the atomic scale is a continuing challenge in physics and material science. Doping of two-dimensional (2D) materials via low energy ion implantation could open possibilities for fabrication of nanometre-scale patterned devices, as well as for functionalization compatible with large-scale integrated semiconductor technology. High resolution imaging and spectroscopic analysis of these electronic dopants at the atomic scale is fundamental for understanding sites, retention, bonding and resulting effects on Fermi level shifts. This has been made achievable with the advent of aberration corrected transmission electron microscopy (TEM) and scanning TEM, as well as low-loss electron energy loss spectroscopy (EELS).

We show for the first time directly that 2Ds can be doped via ion implantation. Retention is in good agreement with predictions from calculation-based literature values [1], as are initial results of the sites of dopants and their influence on the band structure of surrounding atoms. We present results of N- and B-implantation in graphene [2,3] as well as current progress with ion-implantation in TMDCs.

1. Ahlgren, E. H.; Kotakovski, J.; Krasheninnikov, A. V. Phys. Rev. B, 2011, 83, 115424
2. U. Bangert, W. Pierce D. M. Kepaptsoglou, Q. Ramasse, R. Zan, M. H. Gass, J. A. Van den Berg, C. B. Boothroyd, J. Amani and H. Hofsäss, Nano Letters, 2013, 13, 4902-4907
3. Demie Kepaptsoglou, Trevor P. Hardcastle, Che R. Seabourne, Ursel Bangert, Recep Zan, Julian Alexander Amani, Hans Hofsaess, Rebecca J. Nicholls, Rik M. D. Brydson,  Andrew J. Scott, and Quentin M. Ramasse, A.C.S., 2015, 9, 11398–11407

Charge density waves (CDW) are periodic modulations of charge density in low-dimensional metals observed as a function of temperature, doping and pressure. Due to electron-phonon coupling CDWs, are also accompanied by a periodic lattice distortion (PLD).¹ Quasi two-dimensional (2D) transition metal dichalcogenides metals like 1T/2H-TaSe₂, 1T/2H-TaS₂, 2H-NbSe₂ are some of the materials that exhibit strong CDW distortions. While CDW can be directly probed using Scanning Tunnelling Microscopy (STM), diffraction and imaging techniques such as High Resolution Transmission electron Microscopy (HRTEM) and selected area electron diffraction (SAED) are sensitive to the structural distortions (PLD) accompanying CDW. In an electric field, and in the absence of crystalline impurities and defects, sliding incommensurate CDW show transport properties similar to a superconducting state.^1,2 It is therefore important to understand the effects of commensuration, defects, and impurities on the static and dynamic properties of the CDW state.^2,3 On a transmission electron microscope (TEM), investigating the effects of  defects on the CDW/PLD state can be done in-situ by generating point defects through electron beam irradiation and at the same time monitoring the structure of the CDW/PLD through electron diffraction and atomic resolved HRTEM imaging. 

Here we report on the interaction of commensurate CDW/PLD with point defects generated by the electron beam in 1T-TaSe₂ and 1T-TaS₂. Due to the CDW/PLD, bulk 1T-TaS₂ and 1T-TaSe₂ are characterized by a commensurate √13 a₀ ×√13 a₀ (a₀= 3.447) superlattice at 180 K and 300 K respectively.¹ Figure 1(a) displays a HRTEM image of 1T-TaSe₂ obtained at 100 K. The HRTEM image was obtained at region consisting of upto 10-15 layers of 1T-TaSe₂ and shows Ta atomic columns. The HRTEM image is also characterized by a bright, dark contrast modulation due to the PLD/CDW. A Fast Fourier Transform (FFT) of the HRTEM image shown in fig 1(b) has spots from the main structure (solid red circles) surrounded by six satellite spots (dotted circle) from the PLD/CDW. HRTEM image analysis involving masking of the satellite spots from the CDW/PLD with a circular mask, followed by inverse FFT helps to visualize the structure of the lattice arising from the commensurate PLD/CDW. The resulting image, shown in fig. 1(c), then displays an intensity image showing a √13 a₀ ×√13 a₀  lattice of the CDW/PLD maxima.

To investigate the effects of electron-beam generated defects on the CDW/PLD, succesive HRTEM images  were obtained from the same sample region over time and at a constant electron dose. Images showing the evolution of PLD/CDW maxima with electron beam irradiation were then obtained from respective HRTEM images using the image analysis procedure elucidated above. Figures 1(d)-(f) show successive HRTEM images obtained over time. The HRTEM images were obtained from a region of upto 15 layers of 1T-TaSe₂ and only show the Ta atomic columns. The PLD/CDW lattices shown in fig.1(g), 1(h), 1(i) correspond to the HRTEM images in figs. 1(d), 1(e), and 1(f) respectively. A radial distribution function (RDF) showing the nearest-neighbor and next-nearest-neighbor periodicities between the CDW maxima can also be calculated. The RDF in figs.1(j), 1(k), 1(l) have been calculated from the CDW/PLD lattices in figs. (g), (h), (i) respectively. The peak with the asterix shows the main periodicity due to the CDW/PLD wave-vector |q_{i =1…6}| =  √13 a₀. Loss of long-range order with increased exposure to the electron beam is characterized by broadening of RDF peaks (see regions marked with dotted rectangle in fig. 1(k) and fig.1 (l)). Based on the characterization of 1T-TaSe₂, and 1T-TaS₂ thin/single layers we show that this loss of long range order is due to interaction of CDW/PLD with S and Se anionic point defects generated by the electron beam.

References and acknowledgments:

^1. J. Wilson, F. D. Salvo, and S. Mahajan, Adv. Phys. 24, 117 (1975).

^2.H. Dai, and C. Lieber, J. Phys. Chem. 97, 2362 (1993)

^3. H. Mutka, Phase Transitions 11, 221 (1988).

Quasi two-dimensional (2D) semiconductor materials are desirable for electronic, photonic, and energy conversion applications and have stimulated extensive research on their synthesis^1-3 and applications^4-6.  One such example is the growth of indium phosphide flag-like nanostructures (Fig. 1) by epitaxial growth on a nanowire template⁷. By the intentional incorporation of defects, control of the nano-flag location along the nanowire was achieved. Long InP nanowires with inherent high defect density in the top area of the flagpole, which pins the catalyst, resulted in top flag geometry (Fig. 1a). Intentional defect creation in the middle section of the flagpole produced middle nanoflags (Fig. 1b). Bottom flags were created by growing defect free primary nanowires under high phosphine flow (Fig. 1c).

Detailed investigation, using aberration corrected high resolution TEM (FEI Titan 80-300), high angle annular dark field STEM and electron diffraction, of the flag-like nano-structures at different stages of the growth provided valuable details about the growth mechanism and the resulting morphology of such nano-structures.

The main conclusion from the TEM observations is that the growth process involves a few stages. The first stage consists of asymmetrical dissolution of the nanowire (NW) tip into the catalyst (Fig. 2a), followed by the Au catalyst unpinning from the top of the nanowire, and its induced migration along the nanowire sidewall. The final position of the catalyst is determined by the position of structural defects (such as stacking faults - Fig. 2b) along the NW (i.e. the "flagpole"). The next stage involves the epitaxial growth of a pure Wurtzite nano-flag on one of the facets of the "flagpole" resulting in a two dimensional nano-membrane on a one dimensional NW sidewall template (Fig. 2c-d).

The comprehensive understanding of the growth process may help improve the synthesis of similar nano-structures. It may allow for better control on the growth process of more complex structures and also help achieving similar nanostructures in a variety of III-V semiconductor material systems with potential applications such as active nanophotonics.

References:

1. Gudiksen, M. S.; Lauhon, L. J.; Wang, J.; Smith, D. C.; Lieber, C. M. Nature 2002, 415 (February), 617–620.
2. Algra, R. E.; Verheijen, M. a; Borgström, M. T.; Feiner, L.-F.; Immink, G.; van Enckevort, W. J. P.; Vlieg, E.; Bakkers, E. P. a M. Nature 2008, 456 (7220), 369–372.
3. Gorji Ghalamestani, S.; Ek, M.; Ganjipour, B.; Thelander, C.; Johansson, J.; Caroff, P.; Dick, K. a. Nano Letters 2012, 12 (111), 4914–4919.
4. Tomioka, K.; Tanaka, T.; Hara, S.; Hiruma, K.; Fukui, T. IEEE J. Sel. Top. Quantum Electron. 2011, 17 (4), 1112–1129.
5. Nakayama, Y.; Pauzauskie, P. J.; Radenovic, A.; Onorato, R. M.; Saykally, R. J.; Liphardt, J.; Yang, P. Nature 2007, 447 (June), 1098–1101.
6. Mourik, V.; Zuo, K.; Frolov, S. M.; Plissard, S. R.; Bakkers, E. P. a. M.; Kouwenhoven, L. P. Science (80-. ). 2012, 336, 1003.
7. A. Kelrich, O. Sorias, Y. Calahorra, Y. Kauffmann, R. Gladstone, S. Cohen, M. Orenstein and D. Ritter, Nano Letters 2016, 10.1021/acs.nanolett.6b00648.

Metallic nanoparticles occupy special places in fundamental research and practical applications, not only because they bridge the gap between atoms and bulk matter but more importantly because some fascinating phenomena appear only at nanoscale level. Research has shown that the properties of nanoparticles are strongly affected by their size, shape, chemical composition as well as their supports.  Hence it is essential to gain full knowledge of their three-dimensional atomic structures, which is still a challenging task for nanoclusters and nanoalloys. 

After a brief overview of the current status of structure research on clusters and nanoalloys, I will discuss what an aberration-corrected scanning transmission electron microscopy (STEM) can offer in this area through an atom counting approach. I will then present two case studies from our recent work: (1) on control and manipulation of the size and atomic structure of Au clusters, formed via gas phase condensation and deposited on amorphous carbon films and MgO(100) surfaces; and (2) on structures of TiO₂(110) supported Au-Rh nanoalloys, prepared through solution based chemical synthesis method, under effects of various thermal treatments.  In both cases, direct evidence of atomistic details of the clusters, nanoalloys as well as nanoparticle-support interfaces will be presented and their possible mechanisms will be discussed, together with computer modeling studies. Our emphasis will be on understanding the relationship between nanoparticle structures and their thermodynamics or kinetics.

Nanoparticle and catalysis research makes extensive use of transmission electron microscopy (TEM). In particular, environmental transmission electron microscopy (ETEM) has attracted considerable attention in recent years. This technique allows us to expose samples to gaseous environments at elevated temperatures in order to investigate local structural changes at the atomic level as the environment changes. Recently, the technique has also been used in nanowire, graphene and electron optical lithography research among others.

Recent developments in TEM instrumentation include monochromation of the electron source and aberration correction of both condenser and objective lenses. These developments have now been introduced onto the ETEM column. The improved spatial resolution and interpretability provided by these additions are beneficial for imaging the surface structure and dynamics of catalyst nanoparticles and provides exciting new possibilities for investigating chemical reactions. In order to take full advantage of this, an understanding of both the interaction of fast electrons with gas molecules and the effect of the presence of gas on high-resolution imaging is necessary.

Using an FEI Titan ETEM equipped with a monochromator and an aberration corrector on the objective lens, we have investigated sintering of supported metal nanoparticles often used in catalysis. A model system consisting of supported gold nanoparticles were prepared by sputter-depositing the metal onto graphene and boron nitride substrates. These samples were imaged under hydrogen at increasing temperatures. Gas was introduced into the environmental cell using digitally controlled mass flow controllers providing accurate and stable control of the pressure in the cell. As the temperature was increased, migrating particles were observed on the support. As they came into contact, a neck was formed between the particles and subsequently, the particles coalesced entirely (Fig. 1). Growth patterns have also been investigated for platinum and palladium nanoparticles supported on silicon oxide substrates. Here, anomalously large particles were observed as the particles were sintered in oxygen atmospheres at temperatures at elevated temparatures. Such large particles have also been observed for industrial catalysts. In this study, we will try to elucidate the mechanisms of metal nanoparticle sintering.

The analytical capabilities of the microscope can be further augmented by adding stimuli such as optical or electrical though the sample holder. Using a holder capable of exposing the sample to light, the redox properties of cuprous oxide have been investigated. Cuprous oxide has been identified as an active catalyst for the water splitting and hydrogen evolution from an ethanol solution. However, Cu2O suffers from photocorrosion. This phenomenon was investigated using controlled atmosphere transmission electron microscopy. Fig. 2 shows how the photoinduced degradation of cuprous oxide to metallic copper under an aqueous atmosphere using bright-field imaging, electron diffraction and electron energy-loss spectroscopy. All three techniques show the transformation from oxide to metal.

Effects of imaging in various elemental as well as di-molecular gases and their effect on imaging and spectroscopy in the environmental transmission electron microscope will also be discussed.

Thin crystalline germanium (c-Ge) and silicon (c-Si) films are widely used in current microelectronic and nanoelectronic devices, such as thin film transistors and highly efficient solar cells. However, Si or Ge films grown from the vapor phase are usually amorphous. Crystallization of such films requires annealing temperatures above 500°C, which are often incompatible with the fabrication processes used in industry. Yet, when a semiconductor film of amorphous germanium (a-Ge) or amorphous silicon (a-Si) is in contact with a metal that forms an eutectic phase diagram (e.g, Au, Ag, Al, Bi, Pd), the crystallization temperature of the amorphous film is reduced significantly. This effect is commonly known as “metal-mediated” or “metal-induced crystallization” (MIC). Despite numerous studies of the MIC effect, the reaction mechanism is still unclear, although the interaction at the interface between metal and semiconductor seems to play a key role in activating the process. In this context, the main motivation of the current work is to understand the mechanism of metal-induced crystallization of semiconductor films in eutectic binaries by studying the interfacial interactions in these systems. Layered films formed by the sequential condensation of components in vacuum are a convenient model system, since the thickness of the film or the size of particles is comparable with the width of the interface boundary in a binary system. By heating such a system in a transmission electron microscope,  the interaction between the metal and the semiconductor film can be observed directly in real time.

We have selected gold-germanium (Au-Ge) films as a model system. Layered Au/Ge films are prepared by the condensation of an amorphous continuous film of Ge of 5 nm, followed by the deposition of a 0.2-0.3 nm Au film on top of the Ge film. KCl crystals covered by an arc-deposited 5 nm carbon layer prior to the deposition of Ge and Au are used as a substrate. The structure and morphology of the films were studied in situ in a FEI TECNAI G2 F20 X-TWIN transmission electron microscope within the 20-400°C range with a Gatan 652 Double Tilt Heating Holder. It has been shown that the process of solid state de-wetting of a 0.2-0.3 nm Au films occurs in the 150-200°C temperature range and results in an a-Ge film uniformly covered with Au nanoparticles with an equivalent diameter of 4.4 nm. The eutectic melting of these Au nanoparticles occurs at temperatures 200-250°C and is accompanied by both an abrupt change of the film morphology and crystallization of the a-Ge film. Long-term annealing of the film at a temperature below the eutectic point does not lead to a change in the mean particle size or to the appearance of c-Ge. Instead, the capillary motion of liquid eutectic particles on Ge surface has been observed (Fig. 1). At higher temperatures, this process intensifies and the motion of the particles becomes visible directly by TEM. This effect leads to the crystallization of a whole surface of amorphous germanium film and causes a drastic change in the Au film morphology. In addition, the formation of metastable fcc Ge phase is found and is shown to drive the capillary motion of eutectic particles. Fig. 2 shows TEM images of the same area of the film taken at 370°C with an interval of 10 seconds between each image. From Fig 2a to Fig 2b, it is evident that the eutectic nanoparticles move, leaving c-Ge behind. However, a more detailed examination of the HRTEM images of the crystalline Ge film shown in Figure 2 exhibit an unexpected feature. In particular, the Fast Fourier Transform (FFT) of region I in Figure 2a, which corresponds to a beam direction of B=[110], shows forbidden reflections for the diamond structure, namely {200} type planes (d=0.283 nm). The appearance of such reflections seems to indicate the crystallization of germanium into a metastable fcc phase, which is then converted into the diamond structure.

In the last decades, radiolytic synthesis routes have exploited the chemical effects of the absorption of high-energy radiation on precursor solutions, to form nanostructures by reproducing a selective reducing/oxidizing environment. Radiation chemical synthesis provides a powerful means to form nuclei which are homogeneously distributed in the whole volume and where the growth rate can be easily controlled.[1] The latter has recently been achieved using liquid cell electron microscopy, with examples of formation kinetics of particles in polar (water)[2] and non-polar (toluene)[3] solvents following a linear relation with applied dose rate. Fine control of particles size with conventional radiation sources typically requires increasing the production of radicals (thus applying “high doses”). On the contrary, the challenge in the scanning transmission electron microscope (STEM), where incident doses are inherently higher by orders of magnitude, is to reduce drastically such production. To illustrate this, we recently proposed the use of non-polar systems (toluene), which are not typically used for radiolytic synthesis outside the STEM, as a solvent that produces much lower amount of radiolytic ionic species upon electron irradiation, as compared to water or other polar solvents such as alcohols (Figure 1 shows an example of

Thus far, most experiments in the liquid cell involved the use of water as a solvent, which explains the large amount of work dedicated to understanding the effects of the electron beam in aqueous conditions. Radiation chemical yields in water are large due to the relatively low bond energies in water molecules, meaning highly reactive oxidizing and reducing radicals and species are created in about an equal amount.[4] Reproducing net reducing conditions for nanoparticle growth can be achieved with the addition of substances that convert primary radicals into free reducing radicals (using OH· scavengers, for instance). The use of organic solvents in the liquid cell allows for tuning the polarity of the medium, giving access to a broader range of synthesis conditions but producing more complex radiolytic products. Here I will discuss, revisiting a number of examples from the literature and presenting our most recent work, general methods for finding more suitable synthesis environments for controlled nanoparticles formation in the liquid cell. Much of the presentation will focus on the solvent radiolysis which is what predominantly dictates the species and yields involved in the chemical processes leading to nanostructure synthesis.[4]

References:

[1] J. Belloni, Catal. Today, 2006, 113, 141-156; S.-H. Choi et al., Colloids Surf., A, 2005, 256, 165-170; J. Belloni et al., New J. Chem., 1998, 22, 1239-1255; M. A. J. Rodgers and Farhataziz, Radiation Chemistry: Principles and Applications, VCH Publishers, New York, N.Y. , 1987.

[2] D. Alloyeau et al., Nano Lett., 2015, 15, 2574-2581; J. H. Park et al., Nano Lett., 2015, 15, 5314-5320.

[3] P. Abellan et al., Langmuir, 2016, 32, 1468-1477.

[4] SuperSTEM is the UK EPSRC National Facility for Aberration-Corrected STEM, supported by the Engineering and Physical Science Research Council (PA). Part of this work was supported by the Chemical Imaging Initiative, under the Laboratory-Directed Research and Development Program at Pacific Northwest National Laboratory (PNNL) and by the US Department of Energy, Office of Science, Office of Basic Energy Sciences, and was performed in part using the facilities located in the William R. Wiley Environmental Molecular Sciences Laboratory, a DOE national scientific user facility sponsored by the DOE’s Office of Biological and Environmental Research and located and the PNNL. PNNL is operated by Battelle for DOE.

Nano assemblies are two- or three-dimensional (3D) collections of nanoparticles. The properties of the assemblies are determined by the number of particles, their position, shape and chemical nature as well as the bonding between them. If we are able to determine these parameters in 3D, we will be able to provide the necessary input for predicting the properties and we can guide the synthesis and development of new assemblies or superstructures. A detailed structural characterization is therefore of utmost importance.

Electron tomography is a versatile and powerful tool that has been increasingly used in the field of materials science. Also for the investigation of nanoassemblies, electron tomography has been of high value as illustrated in Figure 1 [1-3]. However, to extract quantitative information, optimization of the electron tomography experiment is required, especially for large assemblies or assemblies consisting of more than one type of nanoparticle. Here, we will show how advanced electron microscopy,  in combination with novel reconstruction algorithms, enables us to determine the number of particles, their stacking, interparticle distances and outer morphology. This will be illustrated for very large assemblies consisting of spherical nanoparticles [4], assemblies that contain anisotropic particles (Figure 2) and binary assemblies (Figure 3) consisting of particles with different sizes or different compositions [5,6].

Going a step further is the investigation of self assembly and oriented attachment at the atomic scale. Using annular bright field scanning transmission electron microscopy, the possible effect of polarity can be investigated, whereas high angle annular dark field scanning transmission electron microscopy in combination with advanced computational techniques enables the investigation of the interfacial planes in two-dimensional superlattices from nanocrystals [7]. Finally, we will discuss the use of exit wave reconstructions to investigate the effect of surface ligands on self assembly.

[1] S. Bals, B. Goris, L.M. Liz-Marzán, G. Van Tendeloo, Angewandte Chemie 53 (2014) 10600

[2] T. Altantzis, B. Goris, A. Sánchez-Iglesias, M. Grzelczak, L.M. Liz-Marzán, S. Bals, Particle and Particle Systems Characterization 30 (2013) 84

[3] M.P. Boneschanscher, W. Evers, J.J.Geuchies, T. Altantzis, B. Goris, F.T. Rabouw, S.A.P. van Rossum, H.S.J. van der Zant, L.D.A. Siebbeles, G. Van Tendeloo,I. Swart, J. Hilhorst, A. Pethukov, S. Bals, D. Vanmaekelbergh, Science 344 (2014) 1377

[4] D. Zanaga, F. Bleichrodt, T. Altantzis, N. Winckelmans, W.J. Palenstijn, J. Sijbers, B. de Nijs, M.A. van Huis, A. Sánchez-Iglesias, L.M. Liz-Marzán, A. van Blaaderen, K.J. Batenburg, S. Bals, G. Van Tendeloo, Nanoscale 8 (2016) 292

[5] M. Grzelczak, A. Sanchez-Iglesias, L. M. Liz-Marzán, Soft Matter 9 (2013) 9094-9098

[6] T. Altantzis, Z. Yang, S. Bals, G. Van Tendeloo, M.-P. Pileni, Chemistry of Materials 28 (2016) 716

[7] E. Javon, M. Gaceur, W. Dachraoui, O. Margeat, J. Ackermann, M. Ilenia Saba, P. Delugas, A. Mattoni, S.Bals, G. Van Tendeloo, ACS Nano 9 (2015) 3685

Acknowledgements

S.B., D.Z. and N.C. acknowledge financial support from European Research Council (ERC Starting Grant # 335078-COLOURATOMS). The research leading to these results has received funding from the European Union Seventh Framework Programme under Grant Agreements 312483 (ESTEEM2). 

Understanding processes leading to the formation of nanoparticles is of utmost importance for tailoring their size, shape and spatial arrangements, which are crucial for the nanoparticles’ use in heterogeneous catalysis. As the nanoparticle growth involve processes at or across atomic surfaces, observation made in situ at high-spatial resolution would be beneficial for elucidating the nanoparticle growth mechanism. In recent years, transmission electron microscopy (TEM) has become a powerful technique for studying nanoparticles. With TEM, individual nanoparticles can be observed at atomic-resolution and in reactive gas environments [1]. This capability opens up new possibilities for uncovering dynamics of nanoparticles immersed in gas environments. Here, we employ such in situ TEM capabilities to study the growth of Cu nanoparticles supported on SiO₂ by the reduction of a homogeneous precursor consisting of Cu phyllosilicate platelets (Fig. 1) [2, 3]. The homogeneous precursor represents an important class of industrial catalyst precursor material for which local TEM observations can directly be related to the catalyst at a macro-scale.

By monitoring the individual nanoparticles over time, quantitative information about the nucleation time and size evolution of the Cu nanoparticles were obtained from TEM image series (Fig. 2). To employ such quantitative data in a kinetic description of the reduction process, it is imperative to reduce the influence of the electron beam on the process to a negligible level. To achieve that, we developed a low dose-rate TEM imaging procedure in which several regions of the sample was monitored in parallel in a dose-fragmentated way [1,4,5,6].  This strategy allows for a direct evaluating of the role of the electron dose-rate, accumulated dose and dose history, which were all found to play a role on the growth process (Fig. 1C). Based on this evaluation, the role of the electron beam was quantitatively assessed and used to pinpoint illumination conditions that allowed the observation of the inherent thermal reduction process.

In view of this detailed analysis, quantitative data were extracted from the time-resolved TEM image series to describe the thermally induced growth of the Cu nanoparticles supported on SiO₂ (Fig 1B). By comparing the quantitative TEM data of the process with kinetic models, it was found that the growth process was best characterized as an autocatalytic reaction with either diffusion- or reaction-limited growth of the nanoparticles. This finding is significant because the autocatalytic reaction limits probability for secondary nucleation and because the platelike precursor structure restricts the diffusion. This describes a way to synthesize nanoparticles with well-defined sizes, which in turn offers a more stable catalyst[7]. In this way, in situ observations made by electron microscopy provide mechanistic and kinetic insights that can guide the formation of metallic nanoparticles for catalytic applications in a rational way.

References: 

[1] S. Helveg, J. Catal. 328, 102 (2015)

[2] R. van den Berg et. al., J. Am. Chem. Soc. 138, 3433 (2016)

[3] R. van den Berg et. al., Catal. Today (2015)

[4] C. Holse et. al., J. Phys. Chem. 119, 2804 (2015)

[5] J. R. Jinschek and S. Helveg, Micron 43, 1156 (2012)

[6] S. Helveg et. al., Micron 68, 176 (2015)

[7] S. B. Simonsen et. al., J. Am. Chem. Soc. 132, 7968 (2010)

Boron nitride (BN) nanostructures exhibit excellent mechanical properties. For example, the in situ tensile tests on individual BN nanotubes (BNNTs), which were conducted in situ in a transmission electron microscope (TEM) column, demonstrated the strength and Young’s modulus of ~33 GPa and ~1.3 TPa, respectively [1].Such superb mechanical properties of BN nanostructures make them very attractive materials as a reinforcement phase in lightweight composites. Besides nanotubes, nano-BN can be obtained in the form of spherical nanoparticles (BNNPs) with different morphologies – solid or hollow, with smooth or rough surfaces. High electrical and chemical resistance, thermal stability and biocompatibility of BNNPs were reported but their mechanical properties have not been detailed yet.

BNNPs in the present study were synthesized by CVD technique. By changing technological parameters of CVD process particles of various morphologies were obtained. For each type of individual BNNPs in situ compression tests were accomplished in a TEM column. As-synthesized BNNPs were dispersed in acetone and the suspension was placed onto a Si wafer. The truncated cone-like shape indenter with a diameter of top circle of 1 μm was used. Elastic moduli were determined for all tested particles. The obtained results (Fig. 1) show that the high strength enables the hollow spherical BNNPs to withstand a considerable compressive deformation before failure (up to 50% of the hollow particle diameter). The structures also demonstrated a high percentage of elastic recovery.

1. X.L. Wei, M.S. Wang, Y.Bando, D.Golberg, “Tensile tests on individual multiwalled boron nitride nanotubes,” Adv. Mater. 22 (2010), 4895-4899.