As we continue to deepen our understanding of material properties and dynamic effects, the visualization of in-situ reactions becomes an increasingly important tool that helps us bridge the gap between the nanoscale performance of atoms and mesoscale properties and effects. While transmission electron microscopy (TEM) is routinely used to correlate static observations with experimental results, the mesoscale is primarily a dynamic arena where the characterization of transient effects is critical. Traditional TEM experiments are able to capture events in the few tens of millisecond time scale, but there is an increasing need to push to sub-ms scales that offer a much finer temporal view of these effects. There is a large divide today in our ability to perform dynamic TEM (DTEM) experiments typically in the pico-second time scales, versus traditional CCD or CMOS cameras which operate at a few tens of frames per second: pushing beyond the millisecond boundary with in-situ TEM experiments will give us a better in-depth understanding of mesoscale reactions in spatial and temporal dimensions.
Extending the visualization to these very small time scales requires simultaneously resolving several challenges including capturing frame rates greater than 1000 / second, improving the signal-to-noise ratio of the hardware image processing chain, and being able to capture all incoming electron signal in both space and time. At the fastest frame rates, there may only be 5 - 10 incident electrons in each sensor pixel to generate signal; losing any incoming signal will significantly degrade the output image quality, and ultimately, our understanding of these sub-ms phenomena. We will discuss some of the challenges associated with extending in-situ TEM to the sub-ms regime as well as present some of the latest results obtained.

Over recent years progress in environmental transmission microscopy (ETEM) has allowed interrogation of certain key problems in material sciences, engineering and catalysis. This technique allows real-time sample observation at elevated temperatures in a gaseous environment. Combined with the ability to conduct electron energy loss spectroscopy it gives ability to gain chemical, structural and morphological information about a sample with high energy and spatial resolution in real-word application conditions. However, it was soon realized that a 30 frames per second image acquisition is a limiting factor. Thus, we have equipped ETEM FEI image corrected Titan 80-300 S/TEM installed in the Center for Functional Nanomaterials at Brookhaven National Laboratory with Gatan K2 in-situ camera capable of providing high frame rate imaging. Back thinned for electron transparency, this direct electron detection CMOS chip has a high resistivity to electron damage and a high sensitivity, leading it to be capable of delivering up to 1600 frames per second. We will present examples of carbon nanotube nucleation and growth as well as oxidation/reduction processes in copper thin films to illustrate the benefits of high frame rate image acquisition. At the same time the ability to record such a high frame rate comes at a cost and leads to substantial challenges associated with efficient data storage, processing, and sharing. Our recent experiments have shown that a data flow of 25TB per week per user could be easily generated. In addition, these huge data sets demonstrate a need for sophisticated data analysis and computational methodology. We will describe how we are enlisting the supercomputing and data storage and transfer capabilities at BNL to deal with these issues, with a focus on providing an optimal utilization in a User Facility environment.

Computational and other related aspects of many structural investigations have grown almost exponentially in the last decade or two. Data sets are enormous, computational analysis has grown both in size and complexity, visualization of results gets more difficult and even things as simple as storage and archiving a typical project have begun to pose serious problems that no-one considered as recently as only a few years ago. One solution is simply to do more of what has always been done: purchase more powerful computers, burn more DVD's, stack data-filled hard-drives on office shelves, etc. There are also solutions to these issues that step outside such conventional approaches and that cause the scientific community to re-think how science could (and maybe should) be done in the future. Some approaches dump such problems into the rubric of "big data" and try to solve the problems in ways that are less dependent on an exact scientific field and more geared towards the common issues present in big data from many different areas. These solutions are beginning to develop in certain academic and commercial settings, and there are national (even some international) efforts to address extremely similar issues that arise in fields as diverse as museum science, nanotechnology, astronomy and bioinformatics. Solutions such as science "gateways" to large data structures and analytical platforms represent generic solutions to these issues that then need to be customized for particular scientific disciplines. There are other systems that can offer massive amounts of computational power to users, but where making the system accomplish exactly what one wants is the major challenge for a user. The Indiana University-Bloomington Electron Microscopy Center (IUB-EMC) has been able to harness a number of these different approaches which are mainly built around IUB's nationally acclaimed cyber-infrastructure. I will present details of particular solutions that are implemented at the IUB-EMC, with the hope that this information will help others in the scientific community think about ways to handle their own problems with Big Data. Many other academic and commercial institutions will have similar solutions in place already, would be capable of creating them without too much effort or at least have been thinking about such issues. It is hoped that as more members of the scientific community change the way they think about big data issues, all our institutions will change as well.

Virtually every domain of science is on the brink of unprecedented eras of discovery, as broad innovations in instrument manufacturing, data acquisition, and computation come online. Each of these domains have sub-fields which are particularly well-positioned to leverage these advances to achieve revolutionary breakthroughs. However, as scientists in these fields make "first contact" with these new capabilities, they are immediately faced with the same underlying challenge: there are very few good practical resources for researchers seeking to apply modern software and hardware technologies to scientific problems.
While this is a source of friction and frustration across all science, for these domains where new data volumes and computing technologies have rapidly become the norm, scientists are faced with the enormous challenge of needing to become experts in supercomputing in order to do basic research. In these areas, proper application of information technology - and not merely the scientific aptitude of the research team - can be the difference between cutting-edge science and millions of wasted dollars.
In this talk, we will approach the scalable scientific data management problem from a principled historical perspective, starting with a discussion of the core reasons why the computing problem has become fundamentally harder. We will then compare and contrast the approaches for "big data" from business computing and the high-performance computing worlds, including an overview of a few of the most popular technologies in the alphabet-soup of commercial and open-source offerings.
Lastly, we will look at the social dynamics of doing science in an era of large data. The challenges of reproducibility, data provenance, and collaboration - already difficult with small datasets - are intractable unless radically different approaches are taken. Alternative "data-centric", thin-client architectures and workflows have myriad benefits, and we will demonstrate how applying various new technologies in this area can catalyze effective research and collaboration, even in the face of unprecedented data sizes.

Data-acquisition software programs for the latest scanning/transmission electron microscopes (S/TEMs) have already been developed in 64-bit platform. Thus, it possible to acquire large-scale datasets such as spectrum images, diffraction images, through-focal image series, tomographic image series, etc. Nowadays, it is straightforward to obtain over a few GB dataset in a single acquisition process, e.g. one spectrum image with 1024 x 1024 pixels and 2048 channels is ~8 GB. Although this trend to acquire such large-scale datasets is desired for many years, it would be more challenging to view the datasets, perform appropriate data analysis, and extract unique features in the datasets. If the datasets are relatively noisy, the feature extraction would be much harder tasks. The large-scale datasets can be efficiently handled by employing advanced statistical approaches such as multivariate statistical analysis (MSA). The several different types of MSA approaches have already been applied to spectrum imaging datasets obtained in electron microscopy as data-mining tools and a quite few software packages to run MSA are available. It is relatively straightforward to apply MSA but some fundamental knowledge of MSA is essential for appropriate analysis of complex spectrum images.
Although the MSA approach is very efficient and useful, it may create unexpected artifacts especially in noise reduction, which may mislead results. These artifacts can be more pronounced when magnitudes of the weak features in the datasets are similar to those of random noise components. In this presentation, advantages/disadvantages of the technique will be discussed and common pitfalls in MSA applications will also be addressed.

Grain boundaries in complex oxides have been shown to readily accommodate non-stoichiometry and impurities, changing the electrostatic potential at the boundary plane and effectively controlling material properties such as capacitance, magnetoresistance and superconductivity. Understanding and quantifying exactly how variations in atomic scale structure and chemistry at the boundary plane extend to the practical mesoscale operating length of the system is therefore critical for improving the material&’s overall properties. In this paper, we show that multivariate statistical analysis (MSA) of aberration corrected scanning transmission electron microscope (STEM) images acquired from meso-scale areas of grain-boundary is an effective routine to understanding the variation in boundary structures that occurs to accommodate non-stoichiometry and vacancies.
MSA of electron micrographs is aimed at achieving simplified representation of structure variations by analyzing correlations across a large set of observations, and was demonstrated in 3D reconstruction of viruses in biological studies [1], phase retrieval using TEM phase contrast imaging [2-3], and more recently point defect analysis using STEM Z-contrast imaging [4]. In our studies, MSA was applied to a series of doped and undoped grain boundaries of perovskite structured SrTiO3 [5-6]. Through effective correction of scanning distortions and drift [7], intensity-based sub-pixel image registration and factor analysis (including principal component analysis), variations in both grain-boundary populations and atomic structures are obtained, which are further attributed to the presence of chemistry variations induced by vacancies and impurities. For instance, in the undoped symmetric tilt grain-boundary of SrTiO3, the presence of Ti-rich non-stoichiometry has been found to play an important role in determining grain-boundary population and energetics [5]. Detailed procedures of MSA and results will be further discussed in the conference presentation. [8]
[1] Vanheel, M. and J. Frank, Ultramicroscopy. 6 187 (1981).
[2] Aebersold, J.F., P.A. Stadelmann, and J.L. Rouvière, Ultramicroscopy. 62 171 (1996).
[3] Trebbia, P. and N. Bonnet, Ultramicroscopy. 34 165 (1990).
[4] Sarahan, M.C., et al., Ultramicroscopy. 111 251 (2011).
[5] Yang, H., et al., Philosophical Magazine. 93 1219 (2013).
[6] Yang, H., et al., Materials Research Letters(In Press).
[7] Jones, L. and P.D. Nellist, Microscopy and Microanalysis. 19 1050 (2013).
[8] This work was supported by the US DoE Grant No. DE-FG02-03ER46057. A portion of the research was performed using EMSL - Pacific Northwest National Laboratory (US DoE contract DE-AC05-76RL01830). Part of this work was also conducted in Research Hub for Advanced NanoCharacterization, The University of Tokyo, supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. Part of the work was supported by the UK EPSRC and the EU ESTEEM2 activity.

Oxygen vacancy plays a major role in controlling the physical properties as in the case of introducing oxygen vacancies by doping Nb, to change intrinsic insulator SrTiO3 (STO) to semi-conductor.
Oxygen vacancies can be indirectly visualized from the defect level luminescence, which is the reason why cathodoluminescence is the main tool for the characterization of defect level in STO. Cathodoluminescence stage with liquid nitrogen cooling capability based on the standard stage for transmission electron microscopy, which can deliver spatial resolution down to 20 nm.
Oxygen vacancies were intentionally introduced in the STO single crystal lattice through heat treatment under controlled environment of temperature and the oxygen partial pressure. Distributions of the hyper-spectral wavelengths were mapped in the regions with dislocations at Liquid nitrogen cooled temperature. Luminescence characteristics from different heat treatment conditions were mapped to be compared from where it was confirmed that the distribution of the oxygen vacancies were not uniform as expected showing locally enhanced emission and fluctuations in spectral wavelengths. Emission wavelength shift by the oxygen vacancy and the lattice defects act as dead emission centers.

In the interest of gaining greater control over the passage of heat in a solid medium, the concept of a metamaterial, with an engineered sub-structure/arrangement of materials has gained much popularity. In this context, we have previously shown that the dual use of the thermal extremum principle - whereby the propagation of heat takes the path of least thermal resistance, along with suitable coordinate transformation techniques - for inducing thermal conductivity anisotropy, yields quantitative criteria for tracing and consequently manipulating the propagation of heat flux [1].
In this talk, we propose ideas related to the bending of heat, through positive/negative refraction at the interface of multilayered anisotropic composites considering the respective thermal conductivity tensors (κ). The positive or negative aspect is considered, according to whether the horizontal and the vertical components of the incident and refracted heat flux vectors point in the same or opposite direction. We will postulate quantitative criteria for whether the angle of refraction ((theta;_r)) should be positive/negative, on the relative magnitude of the components of the conductivity tensor of the materials on either side of the interface. We will propose practical implementation of anisotropic composites to illustrate the sense of refraction through ingenious placement of two isotropic thermal conductive materials, constituted from carbon steel (κ ₁ = 20 W/mK ) and polystyrene (κ ₂ = 0.1 W/mK ) of equal thickness stacked alternately. We consequently show representative positive refraction of theta;= 29 ^o and negative refraction of theta;= -44 ^o at the interface of such composites, when they have been rotated with respect to a horizontal temperature gradient, through rotation angles theta;= 45 ^o , theta;= 60 ^o and theta;= 45 ^o , theta;= -30 ^o respectively.
In summary, we propose that the bending of heat flux at the interface of multilayered composites , that are in turn made from isotropic materials can show negative refraction . Such phenomenon could be used to the directing of thermal energy to useful purpose, e.g., thermal cloaking/insulation , heat concentration and perhaps, a perfect thermal lens.
Reference
[1] Krishna P. Vemuri, P.R.Bandaru, Geometrical considerations in the control and manipulation of conductive heat flux in multilayered thermal metamaterials , Applied Physics letters,103,13311.

Living systems control transport of ions or small molecules across biological membranes using ion channels that form highly efficient and selective pores in lipid bilayers. Although bottom-up synthesis and top-down fabrication could produce pores of comparable size, an unresolved challenge remains to build nanopore scaffolds that fully replicate transport properties of membrane channels. We will show that pores formed by ultra-short carbon nanotubes (CNTs) assembled in the lipid membranes have transport properties that come remarkably close to that goal. These CNT porins can transport water, protons, small ions, and DNA and their ion-rejection properties can be controlled by the charge at the pore mouth. Interestingly, these pores also display the stochastic “gating” behavior common for biological ion channels. Overall, CNT porins represent a simplified biomimetic system that is ideal for studying fundamentals of transport in biological channels, and for building engineered mesoscale structures, such as artificial cells.

The mesoscale architecture of extended protein matrices such as microbial membrane surface layers (S-layers) or collagen scaffolds is critical to their function. Because these systems exhibit a natural propensity for self-assembly, investigating the controls on their assembly and utilizing surface chemical patterns to manipulate the assembly process offers a tool for understanding the principles that underly mesocale organization. Block copolymers, comprised of two dissimilar polymers covalently coupled at one end, can self-assemble into arrays of nanoscopic morphologies, including lamellar, cylindrical, and spherical microdomains, that serve as ideal heterogencous chemical templates for manipulating protein assembly. The size of the microdomains is a function of the polymer size so tuning the copolymer's molecular weight allows for a precise control over the dimension of the block copolymer morphologies. Moreover, the heterogeneous chemical nature of block copolymers allows them to be used as templates for well-defined protein adsorption. Here, we used nanoscopic block copolymer patterns as templates to study the assembly of S-layer proteins SbpA from Lysinibacillus sphaericus (ATCC 4525) by in-situ Atomic Force Microscopy (AFM). Studies were performed on substrates comprised of block copolymers where the microdomains are oriented parallel to and normal to the surface of the film, essentially line and dot patterning, to determine the influence of the type of patterning on the assembly and crystallization of S-layer proteins. To clearly define the crystal orientation on the surface, block copolymer films with well-defined patterns were used. The templates were formed by polystyrene-b-poly(ethylene oxide) block copolymers of various molecular weights after spin coating on faceted sapphire surfaces and subsequent controlled solvent-vapor annealing. Our results show that by controlling the chemical contrast in templates of different geometry and periodicity, protein assemblies could be directed exclusively to the hydrophobic domains of the template. More importantly, our high-resolution AFM measurements indicate that the proteins crystallized in their native lattice while following the structure of the underlying template by preferential adsorption. The results show that block copolymers provide a unique platform for investigating and controlling the mesoscale architecture of protein matrices.

The Guanosine, G-quartet macrocyclic architecture represents a dynamic constitutional system which undergoes reversible conformational exchanges between linear G-ribbons and the G-quadruplex, the columnar system formed by the vertical stacking of four G-quartets. The G-quadruplex plays a very important role in biology, as it is present in nucleic acid telomeres of potential interest to many therapies, and as it seems to play an important role in numerous other cellular processes. The reversible exchanges between the G-ribbons and the G-quadruplex systems may generate distinct self-adaptive mesophases, with the simultaneous formation of mixed polymorphic domains. They may also mutually adapt their distribution based on their own geometrical and interactional internal stability or in response to external factors such as scaffolding confining matrices or ion effectors. Here we use porous silicon (pSi) columnar nanostructured films as a scaffolding matrix to stabilize the G-quadruplexes under confined conditions. The pSi matrix also serves as an optical transducer to monitor in real time the formation of the G-quadruplexes using reflective interferometric Fourier transform spectroscopy (RIFTS). Under confined conditions, we show that the formation of the G quadruplexes induces a significant red shift of the optical signal in the reflectivity spectrum of the pSi host matrix, which is attributed to changes of the refractive index in the pSi film. Furthermore, the formation and stabilization of the Na+, K+ and Ba2+ G-quadruplexes induced by incorporation of the corresponding cations in the pSi matrix is investigated using RIFTS and the optical response correlates well with the expected stability of the cationic complexes. Direct detection of the stabilization of the G-quadruplexes using color-change pSi Rugate filters is also demonstrated. We anticipate this strategy to be a starting point for more sophisticated constitutional reorganization studies under confined conditions and directed at understanding the fundamental aspects of the self-organization of matter, involving in particular controlled structure generation, adaptation and replication processes.
1. G. B. Rusu, F. Cunin, M. Barboiu, Real-Time Optical Detection of Stabilized Artificial G-Quadruplexes Under Confined Conditions, Angew. Chem. Int. Ed. 2013, DOI: 10.1002/anie. 201306230.

Plants are exposed to fluctuations in light and therefore need to balance productive photochemistry and dissipative photoprotection. This is achieved by regulation of the structure and organization of pigment-proteins throughout the thylakoid membrane. In particular, photosystem II (PSII) and its closely associated light-harvesting complex II (LHCII) form supercomplexes within the grana that undergo reversible molecular modifications and large-scale rearrangements to conserve such equilibrium. Photoprotection is achieved by triggering a series of dissipative or repairing reactions denominated non-photochemical quenching, NPQ. A variation of NPQ has been unveiled during the discovery of a new thylakoid protein, a suppressor of quenching, SOQ1.
Using atomic force microscopy, we characterized the structural attributes of grana thylakoids from plants lacking SOQ1 to correlate them with its role in NPQ. We developed a novel image analysis methodology that allowed us to interrogate each detected protein complex and assign its unique identity as part of a densely packed membrane. Our algorithms not only discriminate between crystalline and non-crystalline complexes, they also cluster crystalline particles in different categories. SOQ1 induces protein rearrangements that favor larger separations between photosynthetic complexes in the majority (disordered) phase, and reshapes the PSII crystallization landscape during photoprotection (high-light exposure). It is thought that SOQ1 suppresses quenching directly or indirectly within the LHCII complex. Our structural data indicate that removal of SOQ most likely weakens interactions among light-harvesting antenna complexes, consequently, the separation among supercomplexes increases while the complexes&’ density decreases. The structural light-induced rearrangements that we detect are distinct from known protein organizations associated with the typical heat dissipation mechanisms (qE) providing further support for a role of SOQ1 in a novel NPQ pathway.

Proteins are the molecular machines that carry out the vast array of functions needed for the survival and propagation of all cellular organisms. Many proteins form this machinery by folding into functional building blocks that self-assemble into extended networks to deliver complex functions ranging from photosynthesis, to CO2 separation, selective ion transport, and tissue mineralization. Developing the ability to mimic these in vivo functions by using synthetic molecules would lead to a new class of materials.
In this presentation, we report our progress of using one type of protein-like molecules called peptoids for assembly of two-dimensional (2D) pore-forming networks, exploiting peptoids to mimic the ability of natural proteins to self-assemble. The assembled peptoid-based networks are then further tested for function, such as acting as a matrix for directing crystal formation. Our results show that one acidic peptoid composed of alternating polar and nonpolar residues was able to self-assemble into 2D networks comprising hexagonally-patterned nanofibers in the presence of Ca2+ ions on mica surfaces. Both ex situ and in situ AFM studies showed that peptoid- Ca2+ complexes first formed discrete nanoparticles on mica surfaces, and then these kinetically-trapped intermediates gradually transformed into stable peptoid nanofibers with hexagonal orientations. Although the kinetics of peptoid assembly could be altered through changes in ionic strength or pH, these changes did not stop the formation of 2D networks. Mechanistic studies by AFM-based dynamic force spectroscopy (DFS) indicate that competition between peptoid-mica (PM) and peptoid-fiber (PF) interactions is critical for forming these biomimetic films. The binding free energies extracted from DFS studies show that peptoid-Ca2+ complexes bind much more strongly to freshly-cleaved mica surfaces than to pre-assembled peptoid fibers. Peptoid assembly into extended networks is highly sequence-specific; a slight change of peptoid side chains residues results in a dramatic change in peptoid assembly. CaCO3 mineralization results demonstrate that the assembled peptoid networks are able to direct the formation of CaCO3 crystals with controlled orientations, presumably due to the highly organized presentation of the large number of carboxylic side groups that are contained in this peptoid sequence.

Though progress has been made in artificial photosynthesis, research focuses mainly on the functional imitation of the photosynthetic system while not enough importance is given to structural effects. If we look at a natural leaf, it is a synergy of elaborated hierarchical porous structure and functional components aiming to produce a highly complex machine in which light harvesting, photoinduced charge separation and catalysis modules combine to capture solar light and split water into oxygen and hydrogen efficiently. Evaluating the importance of having a specific structural configuration of the different components is extremely difficult. Nevertheless, this underlines the crucial role of hierarchical macro/meso/micro-porous structures in photocatalysis.[1] The co-existence of light harvesting macrochannels increasing photoabsorption efficiency and allowing efficient diffusion of gas molecules together with a high surface area assured by the presence of microporosity has been found to be the origin of the high photocatalytic performance of TiO2 mesoporous films.[1] In spite of the important promise of hierarchical porosity in light harvesting and mass diffusion, the number of studies in this area is still limited to few compositions and mainly to TiO2.[1] This is due to the lack of efficient and easy synthetic pathways to the desired hierarchical porous materials with a fine control at the three length scales (macro, meso, micro) as well as size, shape and composition of the crystallite grains, equally important in determining the photocatalytic activity.
Herein, we show that assemblying colloidal nanocrystals (NCs) into hierarchical porous structures might be the solution to overcome this lack of control. Microporosity is controlled by the nanocrystal size and shape while macroporosity is attained by using polymeric beads as templates to create photonic structures. We investigated the contribution of each length scale to the photoelectrode activity by performing a series of systematic studies involving change of size, shape and composition as well as electrode architectures. The photocatalytic activity of the hierarchical NC-based electrodes was tested for water oxidation. By offering exquisite tunability of all the critical length scales, our studies contributes to create and understand new structure-properties relationships between NC-based hierarchical porous photoelectrodes and their photocatalytic activity.
[1] Li, Y.; Fu, Z.-Y.; Su, B.-L. Adv. Funct. Mater. 2012, 22, 4634-4667.

The fabrication of 3-D mesoporous materials using building blocks of perovskite niobate nanosheets is demonstrated. Processing includes exfoliation of micron-size powders to form nanosheet suspensions, followed by control of temperature and surface charge to coagulate the sheets in an edge-to-face geometry. Dispersion of the nanosheets is very sensitive to pH and can be adjusted from full dispersion to full agglomeration. The processing parameters can be modified to produce either highly disordered 3-D structures with spherical pores or crumpled layered structures which have 1-D pores. The strong hydrogen bonds between the protonized surfaces of the nanosheets are found to be responsible for the self assembly of the colloidal suspension. Thermal stability is demonstrated for the assemblies to 550C, and the photocatalytic activity for hydrogen yield from methanol solutions are shown for nanosheets with thickness of 2, 3, 4 and 5 NbO6 octahedra. We find very high hydrogen yields without added cocatalysts of up to ~40,000 micromoles of hydrogen per hour per gram.

Colloids are good model systems to study dynamics of condensed matters at an atomic level, because the thermally-induced Brownian motion causes colloids to display the same phase behaviors as atoms or molecules, leading to liquid, crystalline, and glassy phases. An important factor that determines the phase behavior of all kinds of colloidal suspensions is the particle concentration. We demonstrated a simple technique, “dielectrophoretic equilibrium,” called as an “electric bottle,” which is using the electric field gradients for the manipulation of the particle concentration in suspensions of colloidal particles. The electric bottle system provides much better control over the particle concentration than gravity and temperature gradients. Here, we designed the electric bottle with a negative dielectric constant contrast, which drives the particles to collect in the regions of lowest field strength. Confocal scanning laser microscopy was used to follow the individual particles and watch the resulting structures when switching on and off electric fields. The time-dependent changes enable the study of concentration-dependent phase diagram and transformation processes such as crystal nucleation, growth, and melting from a single microscopic sample.

Geometrical frustration arises when a local order preferred by minimization of interaction energy cannot propagate throughout the space because of geometrical constraints. This phenomenon plays a major role in a number of natural and synthetic material systems including water ice, spin ice, spin liquid, Curie-Weiss metals, metallic glasses, and liquid crystal molecules. All of these geometrically frustrated systems are degenerate and tend to form disordered ground-state configurations. Here, we report an analytical, numerical and experimental study on the behavior of buckling-induced geometrically frustrated triangular cellular structures. To our surprise, we find that mechanical instabilities induce the formation of complex ordered patterns which can be tuned by controlling the porosity of structures. In particular, in the case of structures with low porosity, an ordered symmetric pattern emerges, which shows striking correlations with the ideal spin solid. In contrast, for high porosity systems, an ordered chiral pattern forms with a configuration that has not been previously reported. Remarkably, our analysis based on a spin-like model reveals that the connected geometry of the cellular structure plays a crucial role in the generation of ordered states in this frustrated system. Since in our study geometrical frustration is induced by a mechanical instability that is scale-independent, our findings can be extended to different geometries, materials, stimuli, and length scales, outlining a general strategy to effectively control pattern formation over a wide range of systems.

The strongly enhanced optical near-field at the apex of a sharp tip can be used to break the diffraction barrier in scanning near-field optical microscopy, producing optical images with single-digit nanometer-scale resolution and enabling optical investigations of mesoscale hierarchical systems at nearly all relevant length scales. The near-field interactions between a single emitter and the tip can also modify the local density of optical states, which can impart control over various optical properties of the emitter. For example, by spatially scanning the tip in 3D close to the emitter, the emission direction (+1D) and polarization (+1D) can be simultaneously controlled. Furthermore, temporal (+1D) fluctuations in the emission signal from, for example, semiconductor nanocrystal quantum dots, can be suppressed by balancing the near-field excitation enhancement with the fluorescence quenching rate, which varies as the QD blinks. Together, these three +1D controls may enable the realization of a true single photon source where a photon in a particular polarization state can be delivered to a particular location at a particular time.

Traditional optical techniques are inherently surface sensitive due to light-absorption that predominately occurs near the surface, thus making determination of bulk or sub-surface properties of photovoltaic materials difficult. Here we show that two-photon absorption permits sub-surface optical excitation, enabling us to map optoelectronic properties in three-dimensions. For example, carrier lifetimes vary throughout a photovoltaic material due to defects such as grain boundaries, traps and surface defects that act as local recombination centers. Mapping and quantifying the effect of these defects is critical to improving the performance of devices where defects limit efficiencies. We will describe our custom microscope and data acquisition system that allows for the simultaneous acquisition of spectrally- and temporally-resolved photoluminescence and photocurrent as a function of 3D-localized optical excitation. Additionally we will show how we visualize and model these maps to extract relevant semiconductor properties such as bulk minority carrier lifetime and surface recombination velocity.

When it comes to solar cells, the function of photovoltaic conversion of light into electricity stems primarily from the absorbing material to sustain a large electron-hole quasi Fermi level (QFL) splitting in operation conditions. The value of the QFL that can be achieved in given conditions is determined by the amount and nature of defects and can be compared to that of a defectless material (so called radiative limit). This QFL splitting would be therefore the prime quantity to know about a material when investigating its suitability for photovoltaic conversion, but unfortunately it has not been possible up to now to measure it. In this presentation, we will present a technique based on photoluminescence Hyperspectral Imaging (PL-HI) that is able to measure directly the QFL with a spectral resolution of 2 nm and a spatial resolution near the diffraction limit [Delamarre et al., APL 2012]. We will show that the measured QFL is a good approximation of the cell voltage (within tens of meV) and will also show how this technique can be used to measure contactless the local conversion efficiency of an absorber material in thin film solar cells. We will finally discuss various artifacts that have plagued previous attempts at doing the same using exemples from Chalcogenide and III-V solar cells.

Bitumen is the most used material to glue the mineral aggregates in an asphalt concrete mixture. This viscoelastic material consists of a heterogeneous complex mixture of many chemical compounds. Typically, its high sensitivity to temperature and moisture is not desired. The lack of fundamental knowledge regarding physicochemical and self-assembly phenomena in bitumen is recognized as an obstacle to further innovative progress in the field. Therefore, we present studies showing nano/micro-structures formed on bituminous surfaces due to the presence of wax molecules upon temperature annealing procedures. More interestingly, the effect of various factors such as cooling rate, water exposure and recycling was studied. The characterization of such nano/micro-structures was performed using atomic force microscopy under various modes, e.g. tapping mode and peak force quantitative nanomechanical mode, and infrared spectroscopy. The results were linked to the phase behavior of bitumen and the crystalline organization of the wax components. Additionally, the changes in phase behavior and molecular organization at the surface led to changes in the surface tension properties. Further information has been extracted from bulk measurements using for instance x-ray scattering techniques in order to understand the link between surface and bulk structural phenomena.

Understanding preferential solvation of catalysts in solvent mixtures is particularly important to advance catalytic conversion of biomass-derived feedstocks into fuels. During the conversion process both polar and nonpolar liquids (e.g. water and the product fuel molecules) are unavoidably present. If there are preferential interactions of one of the fluid phases with the active catalytic surface or with pore walls, this “epiphilicity”, defined as the preferential segregation near the active area of the catalyst, could lead to inhomogeneities within the reaction medium. This will severely influence the overall catalytic activity or lifetime of the catalyst and therefore the production yield of the biomass-to-fuel conversion.
We report on the application of solvatochromism, observable in UV/visible and fluorescence spectra, to assess and predict previously unconsidered inhomogeneities in a reaction mixture in the immediate vicinity of active centers of novel catclysts for bio-fuel conversion. Our experimental results and corresponding quantum chemical simulations indicate that in solvent mixtures at least one of the components of the liquid phase exhibits epiphilicity". We present data illustrating this phenomenon for homogeneous as well as heterogeneous transition metal catalysts in model toluene-chloroform mixtures and other solvents.

Atomic scale characterization of internal interfaces such as grain boundaries and thin films is needed in order to fully understand the electronic, ionic, mechanical, magnetic, and optical properties of the engineered material. High resolution analytical TEM has given a significant amount of new information about these interfaces, but lacks chemical sensitivity below ~1 at% as well as 3-D information and light element sensitivity. Atom probe tomography in inorganic solids has shown that atomic scale, 3-D characterization is possible with 10 ppm chemical resolution, but a thorough understanding of the laser assisted field evaporation process is needed. Previous studies of inorganic photovoltaic devices have shown that APT is capable of quantifying dopant distributions and interface roughness at resolutions where junction models can be directly correlated. In this work, we show that correlating cathodoluminescence with APT can give a direct correlation between atomic structure and chemistry with energy levels. In ionic conductors, grain boundaries are particularly important as they frequently have conductivities at least two orders of magnitude less than the bulk. Therefore, being able to quantitatively characterize the grain boundary nature to ascertain the reasons behind the decreased conductivity is indispensable for guiding future improvements. In this work an oxygen ion conductor Ce1-xNdxO2-x and a proton conductor BaCe0.2Zr0.7Y0.1O2.95 were analyzed with particular emphasis on analysis of the grain boundary regions. In the Nd-doped ceria, cation and anion segregation at the grain boundary is quantifiable with sub-nm spatial resolution and also found to accurately relate to the ion conducting properties at the grain boundaries as determined with impedance spectroscopy. Correlative TEM and APT illustrates that the structure of the grain boundaries is directly associated with its atomic scale chemistry.

BaTiO3 is the main constituent for the multi-layer ceramic capacitor (MLCC) and is known to have capacitance drop phenomenon, which is still in debate for the degradation mechanism. Structural defects plays major role in controlling the physical properties of the oxides with perovskite structure. Experimental approach to quantify the contribution of defect concentration is not easy because it is not readily visible with any analytical techniques. Cs-corrected transmission electron microscopy made a great contribution to visualize the oxygen vacancies, but it was averaged information through the thickness and difficult to tell the overall average vacancy concentration. Luminescence in BaTiO3 is sensitive to the defect level and defect-level transitions in BaTiO3 studies were typically carried out using cathodoluminescence.
In order to improve the spatial resolution while watching the microstructure, cathodoluminescence stage was developed combined with liquid-nitrogen cooling capability onto the standard transmission electron microscopy stage (TEM-CL). Cathodeluminescence mapping was obtained from the cross-sections of the laboratory-prepared and commercially available MLCC. TEM samples were prepared purely by grinding and Ar+ ion-polishing to avoid shallow Ga implantation when used focused ion beam technique. It was observed that the layered structure of Ba-deficient and Oxygen-vacancy related defect levels are arranged in the layered form in between Ni electrodes.
From these maps, we could visualize distribution of defect at each grain in MLCC. And we could find difference of ionic polarity at each grain and that maybe the one of evidences for capacitance drop mechanism in MLCC.

Since the advent of novel nanostructured materials, the need of new imaging and spectroscopic approaches was made necessary to allow for investigating the opto-electronic properties at nanometer scale. Due to the diffraction limit, traditional optics can only provide information on a length scale that is usually one order of magnitude larger than the dimension of the object under study. Therefore, it is only possible to access spatially averaged quantities that could hide the local properties of such materials. However, near field optics and plasmonics have given the tools to optically probe with true nanometric resolution.
We developed several near field probes, and the last member of this family is represented by the highly performing campanile tips. The campanile tip is a novel nano-optical device that unites optimal near-field properties, including i) highly efficient far-field to near-field coupling, ii) ultralarge field enhancement, iii) nearly background-free imaging, iv) independent from sample requirements, and v) broadband operation. This probe allowed us to locally excite a single Indium phosphate nanowire and locally collect and map the photoluminescence (PL). For the first time, we observed a strong heterogeneity of the photoluminescence along the individual wire due to the local charge recombination originating from trap states - critical information that was unobtainable with previous method.
Increasing the complexity of nanostructured systems on the mesoscale level requires the development of new characterization techniques. To correlate local information with more collective behavior several different but connected properties need to be measured at the same time, generating maps in which different information can be stored in parallel. For this reason, this approach is also known as multidimensional imaging.
For example, understanding the correlation between morphology, local assembly and optoelectronic properties of organic photovoltaics material is instrumental to systematically alter the performance for both, lifetime and efficiency to make OPV materials viable as actual devices. Therefore, we are developing alternative techniques to map topology, photocurrent, and local Raman or PL spectra. Additional local electronic properties, such as surface potential can be easily implemented. Local or global excitation and collection can be chosen to address different properties. Time resolved studies are also possible.
In this paper I will present an overview of new imaging e spectroscopic techniques developed by our group that combine multidimensional imaging approach the sub-diffraction optical resolution.

Imaging materials with mesoscale structure and organization requires molecular and atomic level resolution with a field of view of many microns. Furthermore, these functional materials may have dynamic structures. Scanning probe microscopes have the ability to image these material surfaces in-operando with exquisite resolution and have become one of the most frequently used characterization tools in nanosciences. However, the sequential nature of the scanning limits the speed of data acquisition and conventional instruments typically take several minutes to obtain a high-quality image. We propose new types of non-raster scan paths, which enable high-speed scanning by matching the tip velocities to the mechanical limitations in large-area and high-inertia scanners. Advanced image processing techniques are used to recover high-resolution images from sparse quickly-collected non-gridded position data [1,2]. This allows stretching the speed limits without applying more force or increasing bandwidth requirements on the hardware. By recording several frames per second we visualized chemical reactions such as crystal dissolution and the nucleation and growth of the ZIF-8 metal organic frameworks.
Another limitation of scanning probe microscopy in liquids is the thermo-mechanical noise that dominates the interaction force signal. Compared to operation in air, the smallest measurable forces are more than an order of magnitude higher when imaging in liquids. As a consequence many samples deform under the minimum force required to perform a measurement. This precludes imaging at high temporal and spatial resolution. We developed encased cantilevers to overcome these limits. By fabricating a hydrophobic encasement around the cantilever we keep the cantilever dry, and reduce the fluid viscosity while still allowing the tip to probe the sample in solution. We achieve exceptionally high resonance frequency, Q factor, detection sensitivity, and low force noise (12fN/sqrt(Hz)) enabling gentle high-speed imaging. Furthermore we achieve low-noise interferometric detection of the deflection (6fm/sqrt(Hz)), by using the gap between encasement and cantilever as a Fabry-Pérot optical cavity. These developments expand the frontiers of cantilever based sensing and enable gentle imaging and low noise mass sensing in liquid. Quantitative mass sensing of single nanoparticles and gentle imaging of supported lipid bilayers and high-speed imaging of chemical reactions will be presented.
[1] D. Ziegler, et. al., "Improved accuracy and speed in scanning probe microscopy by image reconstruction from non-gridded position sensor data", Nanotechnology, 24, 335703, 2013.
[2] T.R. Meyer, et al., “Height Drift Correction in Non-Raster Atomic Force Microscopy”, Ultramicroscopy, accepted October 2013.

The use of biomolecules as scaffolds for directed organization of functional inorganic and organic nanomaterials addresses the necessities of assembling multiple functional units with precise control over their spatial arrangement at the molecular level. However, the novel collective behaviors built on these hierarchical functional structures are notable for imperfections and variations as a result of the inherent complexity within the biological structures. Direct correlation of the structural heterogeneities and variations in functions and properties remains a significant challenge for fine-tuning the design and performance of these complex structures.
In our work, we implemented a multimodal single-particle imaging method using scanning confocal fluorescence microscopy together with atomic force microscopy (AFM) topography imaging to study a novel plasmonic heterostructure composed of a fluorophore-modified virus capsid and a gold nanoparticle plasmonic antenna assembled on a DNA origami tile. The high resolution AFM topography imaging enabled us to reveal the anisotropic thickness of the capsid scaffold on a substrate, suggesting an asymmetric planar distribution of dye molecules on the scaffold, which had been overlooked by previous transmission mode electron microscope imaging. Furthermore, the correlative imaging capability enabled us to directly map the fluorescence spectra and intensity of individual heterostructures and determine their correlation with the spatial distribution of the dye molecules and plasmonic antenna on the supramolecular scaffolds. This new knowledge is important for advancing our understanding of the mechanisms that govern the mesoscale assembly of functional molecules on intricate hierarchical biological scaffolds, which provides a promising platform for building future inorganic-biological heterostructures with potential applications in artificial light harvesting and energy transport.

Mastering defect mesostructure and evolution require new methods to find and observe structural and functional defects evolve in 3D and sub-surface systems. X-ray microscopy (XRM) has emerged as a powerful imaging technique that reveals 3D-imaging of interior tomographies. XRM executed both in the laboratory and synchrotron have demonstrated critical analysis and materials characterization on meso-, micro-, and nanoscales, with spatial resolution down to 50nm in laboratory systems. The non-destructive nature of X-rays has made the technique widely appealing, with potential for “4D” characterization, delivering 3D-microstructural information on the same sample as a function of sequential processing or experimental conditions. Understanding volumetric changes, on multiple length scales are fundamental to understanding how materials perform in their local mesoscale architecture. Several XRM applications will be presented including multi-length scale 3D-characterization, in situ testing, quantitative characterization of nonperiodic systems of hierarchical porous polymers and polymer electrolyte fuel cells. XRM extends traditional 2D-microscopy and provides 3D-quantitative datasets for computational modeling when non-destructive evaluation is necessary.
1. He, H. et al. Advanced Functional Materials (2013)
2. Litster, S. et al. Fuel Cells (2013)
3. Sai, H. et al. Science (2013).