Symposium MT06—In Situ Characterization of Dynamic Phenomena During Materials Synthesis

Nucleation plays a critical role in many physical and biological phenomena ranging from crystallization, melting and evaporation to the formation of clouds and the initiation of neurodegenerative diseases^1-3. However, nucleation is a challenging process to study in experiments especially in the early stage when several atoms/molecules start to form a new phase from its parent phase. Here, we advance atomic electron tomography to study early stage nucleation at 4D atomic resolution^4,5. Using FePt nanoparticles as a model system, we reveal that early stage nuclei are irregularly shaped, each has a core of one to few atoms with the maximum order parameter, and the order parameter gradient points from the core to the boundary of the nucleus⁶. We capture the structure and dynamics of the same nuclei undergoing growth, fluctuation, dissolution, merging and/or division, which are regulated by the order parameter distribution and its gradient. These experimental observations differ from classical nucleation theory (CNT) and to explain them we propose the order parameter gradient (OPG) model^6,7. We show the OPG model generalizes CNT and energetically favours diffuse interfaces for small nuclei and sharp interfaces for large nuclei⁷. We further corroborate this model using molecular dynamics simulations of heterogeneous and homogeneous nucleation in liquid-solid phase transitions of Pt⁶. We anticipate that the OPG model is applicable to different nucleation processes and our experimental method opens the door to study the structure and dynamics of materials with 4D atomic resolution.

References
K. F. Kelton and A. L. Greer. Nucleation in Condensed Matter: Applications in Materials and Biology. (Pergamon, 2010).
D. A. Porter, K. E. Easterling and M. Y. Sherif. Phase transformations in metals and alloys. (CRC Press, 2009).
D. Kashchiev. Nucleation: Basic Theory with Applications. (Butterworth-Heinemann, 2000).
J. Miao, P. Ercius and S. J. L. Billinge, “Atomic electron tomography: 3D structures without crystals”, Science 353, aaf2157 (2016).
Y. Yang, C.-C. Chen, M. C. Scott, C. Ophus, R. Xu, A. Pryor Jr, L. Wu, F. Sun, W. Theis, J. Zhou, M. Eisenbach, P. R. C. Kent, R. F. Sabirianov, H. Zeng, P. Ercius and J. Miao, “Deciphering chemical order/disorder and material properties at the single-atom level”, Nature 542, 75-79 (2017).
J. Zhou, Y. Yang, Y. Yang, D. S. Kim, A. Yuan, X. Tian, C. Ophus, F. Sun, A. K. Schmid, M. Nathanson, H. Heinz, Q. An, H. Zeng, P. Ercius & J. Miao. “Observing crystal nucleation in four dimensions using atomic electron tomography”, Nature 570, 500-503 (2019).
J. Zhou, Y. Yang, Y. Yang, D. S. Kim, A. Yuan, X. Tian, C. Ophus, F. Sun, A. K. Schmid, M. Nathanson, H. Heinz, Q. An, H. Zeng, P. Ercius & J. Miao. “Capturing nucleation at 4D atomic resolution”, arxiv.org/abs/1807.10709.

Over the last decade, we have seen a vast development in X-ray imaging techniques, methods and instrumentation, both at synchrotron facilities and in the home laboratory. At present, X-ay imaging experiments can be carried out with spatial resolutions < 10 nm, or with time resolutions down to ~100 ns. It is also becoming more feasible to use other channels than those typical in transmission-based imaging (attenuation, phase contrast), such as diffraction or scattering to form the image contrast, also when studying dynamic systems.
In solidification science, X-ray radiography, topography and tomography are well-established methods for studies solidification dynamics and microstructure formation. The presentation will show recent advances in fast high-resolution quantitative in situ imaging in 3 and 4D, with examples on how experimental output may be used to evaluate and guide computer modelling efforts within the field. Furthermore, some recent and near-future advances within X-ray imaging will be presented, allowing to combine ultra-fast imaging with high spatial resolutions to study governing process dynamics at mesoscopic scales. These advances also include image modalities new to solidification science, that may open for novel studies. Finally, the talk will present some preliminary results from dynamical dark-field imaging studies of dendritic mushy zones and growing eutectics, demonstrating a method that can provide completely new insight into thermomechanical effects in solidification processes.

X-radiography enables the observation of the solidification process in a few hundred microns thin samples of metallic alloys in real-time. Therefore, growth characteristics like solidification velocities or concentration changes can be determined and analyzed. The choice of thin samples avoids a superposition of the solidification structures and exploits the X-ray contrast through a clear distinction between the solid and liquid phases. Another advantage of the thin sample geometry is the possibility to arrange the main surfaces of the sample horizontally in the measurement apparatus, so that large scale fluid flows are reduced. However, radiography is only a 2D representation of a 3D process in which the radiography images display a projection of the 3D sample. Hence, what happens in the thickness direction remains largely hidden. Therefore, the question remains how the confined geometry influences the solidification process and the final microstructure.

We present our results on X-radiography studies of equiaxed dendrite growth in a near-isothermal temperature field with a focus on Al-Ge alloys. We have discovered a dendrite orientation transition (DOT) with increasing Ge concentration in this alloy system. Supplementary post mortem EBSD analysis of the in situ processed samples revealed that the dendrites change their growth directions from the crystallographic <100> to the <110> direction. In addition we found that the thin sample geometry influences the selection of the primary growth directions, so that different growth morphologies develop depending on the orientation of the crystals in the sample. Indeed, confinement in one direction forces the dendrites to select in-plane growth directions. Furthermore, the detailed study of the crystal orientations helped to identify certain 3D growth structures from the transmission contrast in situ images of the dendrites, although the images only represent a projection.

Additional important growth characteristics for dendritic growth are tip growth rates and liquid concentration fields. Both parameters are accessible in our in situ experiments. Tip growth rate measurements typically show a transient, an accelerating and a decelerating regime. However, acceleration occurs only when either the dendrites nucleate far from each other so that they have sufficient time to grow, or when they nucleate in a more undercooled melt. Concentration measurements show that deceleration clearly coincides with the overlap of solute fields from neighboring dendrites. Furthermore, we determined the profile of the solute concentration in front of a dendrite tip. These observations are used to elucidate the influence of confinement on the growing microstructure. To better understand the influence of the thin sample geometry and of important material properties on microstructure formation, a comparison of our experimental data with simulations is also pursued.

In-situ techniques like X-ray radiography at thin samples or tomography in the bulk material has become a powerful tool to investigate solidification process in metals in real time [1-3]

We use laboratory based X-ray radiography and fast synchrotron X-ray tomography to analyze nucleation and the initial growth of the nucleated grain depending on the level of added grain refiner, alloy composition and cooling rate, for the alloy system Al-Ge.

The influence of solute segregation and nucleation undercooling is characterized and parameters for the development of dendritic or globu-dendritic morphology of the grain studied. Fast synchrotron X-ray tomography enables 3D time-resolved observation of dendritic growth. With 50 topographies per second (tps) the destabilization of the solid-liquid interface and the transition from spherical growth towards dendritic morphology, in the early stage of the grain growth, was observed and a critical radius for the transition determined.


[1] M. Becker, S. Klein, F. Kargl, Free dendritic tip growth velocities measured in Al-Ge, Phys. Rev. Materials 2 (2018) 567.
[2] B. Cai, J. Wang, A. Kao, K. Pericleous, A.B. Phillion, R.C. Atwood, P.D. Lee, 4D synchrotron X-ray tomographic quantification of the transition from cellular to dendrite growth during directional solidification, Acta Materialia 117 (2016) 160–169.
[3] A.G. Murphy, W.U. Mirihanage, D.J. Browne, R.H. Mathiesen, Equiaxed dendritic solidification and grain refiner potency characterised through in situ X-radiography, Acta Materialia 95 (2015) 83–89.

Nickel-based superalloys are nowadays still the reference material family to produce blades for advanced turbine engines. Creep-resistant turbine blades and vanes are produced by lost-wax casting and directional solidification, enabling an accurate control of the microstructure and therefore the manufacturing of single-crystal parts. Need for a higher efficiency of turbines motivates the use of materials resistant to higher temperatures. Nevertheless, selecting materials with improved mechanical properties sometimes increases the challenges to process them. For instance, CMSX-4 single crystal superalloy is known to be sensitive to the occurrence of freckles, defects causing the affected parts to be rejected.
Understanding the impact of thermo-solutal convection during alloy solidification is of central importance in the casting industry because convective flow generated from density variations in the liquid is at the origin of many casting defects, such as chemical segregations and freckles. Direct observation of metal alloy solidification by X-ray imaging is an invaluable tool to reveal the dynamics of microstructure formation. Among the various X-ray imaging techniques available using laboratory or synchrotron sources, X-ray radiography is an attractive method that simply consists in illuminating a thin sample with X-ray radiation. The transmitted image is collected on a detector to obtain a projected image of the solidification front, in situ and in real-time.
The present communication reports observations of the solidification of CMSX-4 and binary Ni-11wt.%Al alloys by using fast synchrotron X-radiography. Experiments have been carried out at the ID19 beamline of ESRF (European Synchrotron Radiation Facility) using a new high temperature furnace allowing materials with a high melting temperature to be studied. The use of a high-temperature directional solidification furnace along with the combination of an intense monochromatic beam and a high-speed X-ray sensitive camera allowed us to carry out the direct observation of the development of solutal flow during dendritic growth for Ni-based alloys.
The formation of a dendritic network will be studied for directional solidification experiments carried out at various growth velocities using two solidification configuration : (i) the power-down technique consisting of concomitantly decreasing the temperature of the heating element and (ii) by pulling allowing us to follow the solidification of the whole sample from the bottom to the top. Interaction of the dendritic network with solute plumes will be reported, along with the development of convective flow in the mushy zone. The formation of liquid pockets on the sample side lead to the formation of equiaxed structures during their solidification. The processing of images recorded during pulling experiments allowed us to reveal the segregation pattern forming at the sample scale. In addition, measurements of characteristic parameters (primary, secondary dendrite spacings) will be compared to analytical models.

Since the discovery of quasicrystals (QCs) in 1984 [1], QCs have attracted substantial interest due to their unique structure. For instance, QC patterns exhibit long range order without short range periodicity, and classically forbidden rotational symmetry (e.g. 5-, 8-, 10-, and 12- fold). There exist two classes of QCs with different dimensions of quasiperiodicity: 2D (decagonal, d-) and 3D (icosahedral, i-) QCs. In the former case, the d-QC is stacked periodically along a decagonal (i.e., ten-fold) axis [2]. In contrast, the i-QC shows quasiperiodicity in all directions. While the exceptional structural aspects of QCs have been reasonably well described in the literature, there are still remaining questions in terms of why and how they form in the first place.

In our talk, by leveraging state-of-the-art dynamic imaging techniques, we present our efforts on tracking the nucleation and growth of both classes of QCs near- and far-from-equilibrium, as well as the structural relaxation of QCs upon prolonged annealing. Our research was conducted with the aid of time-resolved synchrotron-based X-ray microscopy (XRM) and dynamic transmission electron microscopy (DTEM). By coupling these two techniques, we have investigated a broad range of length-scales (from tens of nm to hundreds of μm) and solidification pathways (1 K/s to 10⁶ K/s cooling rates). The results provide a wealth of knowledge on the emergence of QCs in deeply undercooled melts; the influence of phasons on QC stability; the growth mode of the QCs from a liquid (solidification) as well as a solid phase (precipitation); and similarities and differences in the crystallization signatures between the QCs and periodic approximant phases. We anticipate that our real-time experiments can provide guidance for simulations on this exotic form of matter.

References
[1] D. Shechtman et al., Phys. Rev. Lett. 53, 1951 (1984).
[2] L. Bendersky, Phys. Rev. Lett., 55, 1461 (1985).

X-ray radiography and tomography imaging studies have been applied widely to study solidifying alloys in order to understand the effect of cooling conditions, alloy composition and presence of external fields such as gravity, electromagnetic pulses, extrinsic additions, etc on crystal formation and growth. However, these studies have predominantly concerned understanding primary phase formation whereas the properties of many engineering alloy castings are more generally (and more dramatically) controlled by the size, morphology, distribution, volume fraction and nature of minority, dispersed secondary phases. These phases may be present in concentrations of only a few volume percent, and are frequently formed at temperatures close to the formation of the majority phase or at lower temperature by eutectic reaction, which makes studying the dynamics of their formation and growth more difficult. These secondary intermetallic compound (IMC) phases are also often ordered structures that exist only over a comparatively narrow range of compositions. In contrast to the disordered primary phases, such as a-Al in aluminium alloys, the way in which the formation and growth dynamics of these ordered IMCs can be affected by changes to casting conditions is much less well-understood.

This investigation applies equipment and software tools developed for the radiographic study of the dynamics of primary phase formation using synchrotron X-rays in aluminium alloys and applies them to model alloys in which normally secondary minority IMCs, known to have a controlling effect on toughness and ductility, are contrived to appear directly from the cooling liquid as primary phase, unconstrained by any prior formation of a conventional primary phase. The formation and growth of two alloy classes are studied in detail: (i) Al-3wt%Fe in order to study the important β-Fe IMCs found in may commercial aluminium alloys, and (ii) an Al-Pt-Er model alloy chosen to have similar morphology IMCs to β-Fe but where the presence of Pt and Er enable the use of a novel pixellated spectrometer detector that spatially resolves the distribution of Pt and Er in the forming IMCs, and in the surrounding liquid. In the former alloy, a technique for estimating the formation temperature of every β-Fe IMC particle directly from the radiographic image sequences is also described.

Real-time observations are combined with detailed post-solidification microscopy including high resolution electron back-scatter diffraction (EBSD) to confirm similarities of the IMCs in both alloy systems, and to identify a common twin plane re-entrant (TPRE) growth mechanism. Extrinsic additions to the alloy such as grain refiners are also investigated to show how IMC formation dynamics and temperatures can be manipulated in a manner similar to that of α-Al dendrites. Examples are given of how the different data from these experiments can be fused to improve our understanding of the factors controlling the formation and growth of IMCs. These insights may ultimately allow better exploitation of the beneficial effects of IMCs on properties such as tensile strength, while minimising their detrimental effects on toughness and ductility, and so enable the development of “tolerant” processing of alloys to enable greater metal recirculation.

The automated assessment of growing dendrites within in situ X-ray videos of alloy solidification can be vulnerable to errors stemming from the underlying traits of the frames provided. Specifically, the high degree of image difference between the start and the end of the observed process, along with factors such as noise and the effective shutter speed used, can lead to a reduction in assessment accuracy. The approach demonstrated here seeks to address and mitigate these issues through a combination of machine learning and multi-stage thresholding, balancing the adaptability of the former with the processing capabilities of the latter. This solution was developed and tested using a video of spatially isothermal, equiaxed solidification of a thin Al-20wt%Cu alloy sample during the 2015 MASER 13 sounding rocket microgravity flight within the on-board XRMON-SOL furnace.

The dynamic transmission electron microscope (DTEM) at Lawrence Livermore National Laboratory (LLNL) was developed to enable imaging of rapidly evolving, irreversible materials processes with nanometer spatial and nanosecond temporal resolutions. DTEM uses electron pulse trains generated by UV-stimulated photoemission to provide multiple (up to 9) image acquisitions across temporal ranges from hundreds of nanoseconds to microseconds, yielding frame rates that are on the order of 10⁶ times higher than conventional in situ TEM frame rates. This pulse train is accomplished with a laser based on an arbitrary waveform generator (AWG) to shape laser pulses to square temporal profiles, producing a user-defined (pulse duration and spacing) UV pulse train that is delivered to the instrument’s photocathode to generate an electron pulse train. Each electron pulse captures an image at a specified delay time, and a high-speed electrostatic deflector directs each image to a different segment of a CCD camera. A multi-frame image (i.e., a movie) is obtained at the end of the experiment. Here, an overview of the DTEM instrument and its operation will be presented.

The majority of DTEM work has focused on imaging of microstructure evolution and precise measurement of kinetics during non-equilibrium phase transformations. As a representative example, studies of rapid alloy solidification (RS), in which solid-liquid interface evolution is monitored and the solidification dynamics are determined, will be discussed. RS occurs in numerous manufacturing processes, such as laser welding and additive manufacturing (AM). Recently, the growth of metals-based AM has led to renewed interest in developing predictive modeling capabilities to describe microstructure evolution under RS conditions (i.e., large deviations from equilibrium solidification due to high interface velocities). Calibration and validation of models to describe these non-equilibrium solidification conditions has proved challenging, largely due to the difficulties associated with acquiring in situ experimental measurements. DTEM results (measured interface velocity and morphology) will be presented in conjunction with efforts to calibrate and validate phase field models to describe RS.

Finally, future directions of research using DTEM will be discussed, with particular focus on efforts to develop an experimental platform to image microstructure evolution and defect formation and propagation during in situ high-strain-rate tensile testing of metals with the goal of understanding the dynamic mechanical properties of materials.

This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract No. DE-AC52-07NA27344. Portions of this work were funded by the Laboratory Directed Research and Development (LDRD) Program under project tracking code 18-SI-003.

Liquid metals can be undercooled significantly below their melting temperatures and bring about formation of a rich variety of metastable microstructure. It is of fundamental and technical interest to understand thermodynamics and kinetics of undercooled solidification. In the past twenty years, solidification velocity-undercooling relationships in pure substances and alloys have been established by means of in situ observations of a traveling thermal front on the surface of the solidifying bulk liquid using a camera or a photodiode. Those relationships have been used to test various theories on rapid solidification and provided insights into metastable microstructure formation under non-equilibrium processing conditions. However, primary solids nucleated from the undercooled liquid may differ from those predicted by an equilibrium phase diagram suggesting metastable or even new solidification pathways. They are more often than never remelted or decomposed in a late stage of solidification or through solid-state cooling leading to difficulties in identification of their crystal structure and changes of solidification pathways. In situ monitoring of undercooled solidification using the time-resolved X-ray diffraction technique can ease these difficulties and provide critical knowledge of solidification pathways, stable or metastable, as well as ripening or decomposition of primary solids. In this talk, I shall present two examples of the use of this technique in in situ monitoring of non-equilibrium solidification of undercooled Nd–Fe–B and Co–Si alloys. In both alloys, metastable solids are formed at medium bulk undercoolings but are decomposed prior to completion of solidification of a bulk liquid. With such knowledge, microstructure formation in as-solidified samples can be well understood. I shall also talk about potential of this technique in advancing current understanding of other non-equilibrium solidification phenomena or in combination with other techniques.

We present an in situ experimental study of crystallographic effects during nonfaceted eutectic growth in thin Al-Al₂Cu samples. Thin (about 10 µm thick) near-eutectic Al-Al₂Cu films were prepared by plasma sputtering on flat sapphire (Al₂O₃) substrates. The thin-sample geometry allows one to observe lamellar-growth patterns in real time without convection in the liquid. We used a reflected-light long distance optical microscope that focuses the contact surface of the metallic film with one of the transparent walls of the sample container during directional solidification.
We recall that in bulk solidification, in brief, the Al-Al₂Cu alloy usually forms lamellar microstructures. The two eutectic phases are also known to present special, or heteroepitaxy crystal orientation relationships (ORs). Marked crystallographic effects, i.e., a dependence of the growth microstructure on the OR, have been reported for a long time in Al-Al₂Cu. More recently, a theory of such crystallographic effects, which takes into account the anisotropy of the surface free energy of the interphase boundaries in the solid, has been proposed, and essentially validated by experimental and numerical studies using model systems [1-3]. This analysis casted light on the distinction that exists between a « floating » dynamics, which is characteristic of an essentially isotropic system, and a « locked » dynamics, during which the lamellae follow a direction that is determined by the OR-dependent anisotropy, and thus possibly grow inclined with respect to the thermal-gradient axis.
Our observations, combined with ex situ crystallographic analysis (X-ray diffraction and EBSD), provide additional evidence for this statement. We studied 8 different eutectic grains of large (typically 4 mm) lateral extension. Most of (7 out of 8) were of the locked type. At low velocity, that is when the lamellar spacing is larger than the sample thickness, the solidification dynamics is essentially 2D. In locked grains, the interphase boundaries appear then as straight lines under the optical microscope, and follow an essentially uniform locking direction. Here are our main findings: 1-Al₂Cu crystals were not arbitrarily oriented with respect to the sample walls, namely, a {001} plane and a {100} plane were both nearly perpendicular to the sapphire plate (within 10°); 2-The apparent trace of the interphase boundaries followed a direction that was systematically belonging to the {100} plane in question; 3-We found some reproducible ORs, but none of them were close to the so-called Beta-6 or Alpha-4 ones that are frequently found in bulk samples [4,5]. It seems therefore that a heuristic density criterion that was successfully invoked for predicting the prevailing coincidence planes in the In-In₂Bi eutectic alloy is of poor help in the present case [6]. Our results also raise new questions about the selection dynamics of the ORs during the early stages of eutectic solidification.

[1] S. Akamatsu, S. Bottin-Rousseau, M. Serefoğlu, G. Faivre, Acta Mat 60.6-7 (2012) 3199-3205.
[2] S. Akamatsu, S. Bottin-Rousseau, M. Serefoğlu, G. Faivre, Acta Mat 60.6-7 (2012) 3206-3214.
[3] S. Akamatsu, S. Bottin-Rousseau, G. Faivre, S. Ghosh, M. Plapp, IOP Conference Series: Mat Sci Eng 84.1(2015) 012083
[4] R. Bonnet, F. Durand, Proceedings of the Conference on In Situ Composites, vol. 1 (1973) 209-223.
[5] U. Hecht et al., IOP Conference Series: Mat Sci Eng 27. 1 (2012) 012029.
[6] S. Bottin-Rousseau, O. Senninger, G. Faivre, S. Akamatsu, Acta Mat 150 (2018) 16–24.

Engineered self-organization presents a future approach to manufacture multi-phase materials with superior properties. In particular, non-equilibrium routes to emergent metastable states could unveil novel functionalities not present in equilibrium states. Herein, we demonstrate the spontaneous formation of metastable, two-phase spiral microstructures during crystallization of a eutectic alloy system. Despite recent reports of spiral self-assembly in both metallic and organic systems, the underlying growth mechanism remains elusive. To reveal the origin of spiral growth, we pursue a multiscale investigation encompassing 3D measurements – X-ray nano-tomography aided by machine learning and electron backscatter diffraction – along with in situ synchrotron X-ray diffraction and atomic-resolution electron microscopy. The results show that crystallographic defects within 'hidden' polytetrahedral phases provide a template for spiral eutectic crystallization. Our findings demonstrate the ubiquity of defect-mediated growth in multi-phase materials and establish a pathway towards bottom-up synthesis of 3D chiral microstructures beyond the scope of existing approaches.

During solidification of a faceted alloy, the growth shape of the crystalline solid is primarily determined by the strongly anisotropic mobility of the interface between the solid and the melt. In a given system, it is not rare to observe facets that grow extremely slowly at finite undercooling, and even appear “frozen” or “blocked” on the spatial scale of the optical microscope, and other ones that grow as fast as if they were atomically rough. During solidification of a dilute alloy, the interplay between such a facet growth kinetics and the diffusion of chemical species in the liquid is then very complex, and achieving a steady-state growth regime, if possible at all, depends critically on the orientation of the crystal and boundary conditions.
Here, we present an experimental study of faceted growth patterns in dilute transparent alloys. We used thin-sample directional solidification techniques that permit to control the orientation of large single crystals with respect to the setup, and observe the evolution of the solid-liquid interface in real time with an optical microscope as a function of the growth rate. We will successively consider two cases. First, in impure biphenyl, the platelike growth shape of the monoclinic crystals includes two low-mobility (001) facets and four high-mobility (110) facets. We found that if a thin crystal is oriented with the (001) facets parallel to the sample walls, steady faceted cells and dendrites form for solidification velocities V above the threshold Vc of the Mullins-Sekerka instability. Both acute-angle [100] and obtuse-angle [010] cells are observed. As previously predicted theoretically, the presence of mobile facets on both sides of the dendritic finger does not strongly affect the regular dendritic scalings. For V<Vc, large single crystals undergo a plastic deformation during growth due to long-range thermal stresses. Second, we considered a cyano-bicyclohexane molecule called CCH4 in the scientific literature. In this system, the interface between a nematic liquid and the so-called smectic B phase (a historical, but misleading name that designates a hexagonal molecular-crystal solid with both orientational and positional long-range orders) presents a single facet plane parallel to the basal plane of the hexagonal stacking. We could bring evidence that the facets have practically no mobility at undercoolings smaller than about 0.1 K. In configurations where the CCH4 molecular layers are perpendicular to the sample walls, but tilted with respect to the main solidification axis, we observed drifting shallow cells (V>Vc) and a type of solitary wave, called faceton (V<Vc). A faceton consists of a localized traveling pattern with an asymmetric faceted shape, which propagates laterally, along the solidification front, in a stationary or oscillatory way. This dynamics casts light on the microscopic growth mechanisms of the facets in this system. It also has some striking similarities with the growth dynamics of silicon polycrystals.

Alloy microstructures formed by directional solidification are often polycrystalline, made up of several large grains of different crystallographic orientations with respect to the temperature gradient. While strongly misoriented grains are generally eliminated at early stages of growth, grains with misorientations less than about 15 degrees can persist to late stages and form subgrain boundaries (SBs), which can influence microstructure selection by mechanisms that are still not well understood. Here we combine experimental, phase-field, and analytical methods to investigate the coupled spatiotemporal evolution of SBs and cellular array structures, including the selection of the primary spacing and the lateral drift velocity induced by the grain misorientation. Experiments were carried out in the DECLIC/DSI experimental set-up aboard the International Space Station where a transparent (Succinonotrile-0.24wt%Camphor) alloy contained within a 3D crucible is solidified under purely diffusive growth regime. The in situ observations exhibit both convergent and divergent SBs, where two grains drift laterally to converge on and diverge from each other, respectively. Near a convergent SB, grains usually inter-penetrate, and the primary spacing of cells decreases in a relatively short range as cells are suppressed and eliminated at the SB. Near a divergent SB, however, a peak accompanied by a long tail is observed in the spacing distribution along the direction normal to the SB. Meanwhile, a divergent SB drifts laterally with a non-trivial velocity comprised between the drifting velocities of individual grains. State-of-the-art phase-field simulations reproduce those novel phenomena observed in the microgravity experiments.

We further develop a discrete nonlinear advection-diffusion equation to describe quantitatively the spatiotemporal evolution of the primary spacing in the presence of SBs. Our results show that advection alone, previously found to describe well the spacing evolution within a single misoriented grain, is insufficient to model quantitatively this evolution in the neighborhood of spacing extrema caused by SBs. We find that incorporation of phase diffusion, which tends to uniformize the spacing, is essential to quantitatively predict both the spacing evolution near extrema and the apparent drifting velocity of SBs that is comprised between the drifting velocities of individual grains. The non-trivial prediction of this apparent velocity directly links phase diffusivity to an experimentally observable quantity. Those insights into polycrystalline directional solidification provide an improved basis for controlling alloy microstructures.

Dendritic microstructures are the most frequent among cast metallic alloys. Understanding their formation has been a problem of long-standing interest because of their profound influence on mechanical properties. During columnar solidification, spatially extended dendritic grains of different misorientations with respect to the temperature gradient growth competitively. This growth competition has been recently investigated using 2D and 3D phase-field simulations. The results have highlighted the limitation of using a traditional minimum undercooling criterion to predict grain selection and the importance of considering details of grain interactions. However, results to date remain limited to moderate growth rates and large temperature gradient that form weakly branched cellular/dendritic array structures. Here we report the results of a study that combines phase-field (PF) and the dendritic needle network (DNN) methods to investigate grain competition in the well-developed dendritic regime relevant for casting of concentrated alloys. The PF model of alloy directional solidification exploits a novel rotationally invariant finite-difference discretization of the anti-trapping current to accurately resolve the dendrite tip operating state of misoriented grains. PF simulations are used to investigate the growth dynamics of secondary and tertiary branches, which play a key role in grain competition. In turn, DNN simulations with branching parameters calibrated by PF simulations, are used to characterize experimentally measurable grain boundary (GB) trajectories on larger length and time scales and to construct GB orientation selection maps as a function of GB bi-crystallography. The results reveal how the grain competition on the macroscale is influenced by the sidebranching dynamics on the dendrite tip scale, which is itself controlled by the magnitude of crystalline anisotropy and noise amplification. Results are compared to available experimental data.

The basis of freeze casting is a phase-separation-induced templating of a second phase (e.g. polymers or particles, or both). The process is complex, a large number of parameters determine the structure and properties of the resulting freeze-cast material: i) the chemistry of the solution or slurry system (e.g. polymer and/or particle composition and properties such charge, size and shape), ii) additives and impurities (and their effect on either the solvent, polymer or particles), iii) processing conditions (e.g. cooling rate, temperature field), and iv) the physics of the strongly anisotropic faceted growth of ice. Empirically over the past 15-20 years, strategies have been developed to control some of these parameters in an attempt to custom-design and manufacture in a single processing step materials over several length-scales. However, while many different materials compositions have been investigated over the years — e.g. ceramic- and metal-based systems, both typically processed with the addition of polymeric dispersants and binders, and polymeric ones, both fibrillated and "smooth" — few have been analyzed with the intent to gain a better understanding of the mechanisms by which the ice crystal growth drives structure formation and self-assembly processes. The latter is what we attempt in this study.
Collating structural features found in the cell walls of freeze-cast ceramics (platelets versus particles), fibrillated biopolymers (collagen, nanocellulose), and "smooth" biopolymers (chitosan, trehalose), we can capture the ice crystal growth process and analyze it "post-mortem." Further assessment of the cell-wall structure using a variety of characterization techniques provides unique insights into the effects and their significance on cell wall assembly of anisotropic ice crystal growth, shear flow generated by the volumetric expansion of ice, and other forces, such as those exerted by the mold, for example. Combining these experimental observations with 3D Phase-field simulations of structure formation in ice-templated materials, detailed in a companion presentation, it becomes possible to gain fundamental science insights and obtain the desired structure-property-processing correlations not only at the more traditional macroscopic, but also at the meso- and microscopic scales.

The observation of the dynamic dislocations-carbon atoms interaction is important in the steel material design. BH steel sheets are bake hardened in the manufacturing process. It is said that solute carbon atoms are segregated in dislocations after bake hardening. Solute carbon atoms are observed by atom probe method [1]. However, there are few examples observed directly including the behavior of dislocations after bake hardening. An electrical resistance measurement method is able to observe both the solute carbon atoms and the state of dislocations [2]. The electrical resistivity measurement can not only be measured on bulk samples, but also the in-situ measurement of the process of deformation after pre-deformation, baking, and hardening after baking. In this study, we developed in-situ measurement methods of electrical resistance in the bake hardening test. And we considered the metallurgy of BH steel sheets in the process of thermomechanical treatment of BH steel sheets.
Cold rolled sheets of Fe-0.02 mass% C-0.1 mass% Mn alloy, Fe-0.002 mass% C-0.1 mass% Mn alloy, and Fe-0.002 mass% C-0.1 mass% Mn-0.03 mass% Ti alloy were prepared. The respective alloys are called carbon steel, low carbon steel and ultra-low carbon steel. These cold rolled sheets were held at 1073 K for 120 seconds and then water cooled. From the test steel, a tensile test specimen with a load axis perpendicular to the rolling direction was cut out. The length of the gauge portion was 50 mm and the width was 2 mm. In tensile deformation, the electrical resistance of each specimen was measured at room temperature. In order to measure the electrical resistance of the test piece only, the test piece was electrically isolated from the device and the electrical resistance was measured in distilled water ice water. When tensile deformation was applied, the stroke speed was 0.25 mm / min. The electrical resistance was measured at a current value of 100 mA by the four-point DC method. In addition, after pre-deformation was applied to some of the test pieces as the bake-hardening treatment, the test pieces were immersed in a bath furnace at 443 K for 20 minutes. After that, the electrical resistance of the same test piece was measured again in tensile deformation.
The structure after water cooling was observed with an optical microscope, and the carbon steel was a composite structure of pearlite and ferrite. Low carbon steel, ultra-low carbon steel has ferrite structure. The average grain sizes of carbon steel, low carbon steel and ultra-low carbon steel were 11 μm, 17 μm and 16 μm respectively.
In carbon steel, the electrical resistance was measured during tensile deformation before bake hardening. As a result, the stress increased with the increase of stroke and the electrical resistance. The gradient of electrical resistance to stress is different in elastic deformation, Lüders deformation, and uniform plastic deformation area. The gradient of the electrical resistance to stress is called a stress-electrical resistance factor. The stress-electric resistance factor is larger in the order of Luders deformation, uniform plastic deformation and elastic deformation. In addition, the electrical resistance during tensile deformation was measured for the sample that had undergone the bake hardening process. As a result, the stress-electric resistance factor after the end of elastic deformation increased more rapidly than the pre-deformation. The increase in electrical resistance may correspond to the increase in lattice defects and the solute carbon concentration.
Thus, we developed a series of methods to measure the electrical resistance in situ in the bake-hardening test, and it became possible to capture microscopic information on the structure of dislocations and solute carbon atoms.

Reference
[1] J. Wilde, A. Cerezo, G. D. W. Smith, Scr. Mater., 43 (2000) 39.
[2] H.Abe, T.Suzuki, H.Hoshina, Tetsu-to-Hagane,65 (1979) 399.

Liquid metal dealloying (LMD) is a novel technique to produce nano-/meso-scale composite and open nanoporous structures with ultra-high interfacial area for diverse applications. This method exploits the selective dissolution of one element of a binary alloy into a metallic melt. Previous experimental and modeling studies have revealed that the kinetics of LMD is diffusion limited and that interfacial pattern formation can be understood in the framework of interfacial spinodal decomposition and diffusion-coupled growth. However, quantitative discrepancies exist between experiment and modeling predictions. Here, we report the results of a phase-field modeling study that reveals a surprisingly strong influence of solid-state diffusion on dealloying morphologies. Even though the solid-state diffusivity is four to five orders of magnitude smaller than the liquid-state diffusivity, it is found to stabilize the diffusion-coupled growth of lamellar structures, thereby favoring the formation of this aligned structure over high-genus topologically connected structures. The results also reveal that solid-state diffusion enables the solid-liquid interface to relax to local chemical equilibrium, thereby influencing the concentrations of alloy elements on the liquid side of the interface and in turn morphological evolution. We also present long-time simulations of dealloying kinetics that shed light on the origin of existing discrepancies between modeling and experiment.

TiO₂-VO₂ alloys are known to undergo spinodal decomposition when cooled below 557°C at a critical composition of 34 mol % Ti.¹ Since the misfit between the rutile-structured TiO₂ and VO₂ lattices is significantly smaller along the a-axis than along the c-axis of the tetragonal unit cell, annealing samples in the spinodal region of the phase diagram results in development of a composition modulation along the [001] direction, with an average quasi-periodicity varying from 15-50 nm, depending on composition and annealing conditions.^1,2 Since the VO₂-rich regions of the resulting nanostructure exhibit a metal-insulator transition and a correspondingly large change in magnetic susceptibility near room temperature, while the TiO₂-rich regions remain insulating at all temperatures and exhibit very different optical behavior than VO₂-rich regions, these alloys are of interest as possible tunable metamaterials.² Full control of the synthesis of such metamaterials requires detailed understanding of their phase separation behavior as functions of temperature, composition, epitaxial strain, and surface orientation. To obtain this information, we recently initiated an x-ray photon correlation spectroscopy (XPCS) study of epitaxial TiO₂-VO₂ alloy thin films. XPCS is a coherent x-ray technique that can provide unique insight into the dynamical behavior of materials. Films with controlled composition, thickness, and surface orientation are grown epitaxially on TiO₂ substrates. Samples are illuminated with a partially coherent monochromatic x-ray beam at Sector 12-ID-D of the Advanced Photon Source. The time- and Q-dependences of the diffuse intensity that develops along [001] directions near Bragg reflections during spinodal decomposition are monitored in situ, while annealing samples. At temperatures in close proximity to the critical temperature for phase separation we are attempting to characterize equilibrium composition fluctuations for the first time in a crystal with a conserved order parameter (Model B in the notation of Hohenberg and Halperin³). At somewhat lower temperatures, e.g., 400°C, XPCS can reveal new information on atomic-level dynamics as the sample undergoes spinodal decomposition. Closer to room temperature, the metal-insulator transition associated with a tetragonal-to-monoclinic phase transition provides another opportunity to use XPCS to observe equilibrium critical fluctuations.
This work is supported by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, Materials Sciences and Engineering Division.

1. Hiroi, Z. et al. Spinodal decomposition in the TiO₂-VO₂ system. Chem. Mater. 25, 2202-2210, doi:10.1021/cm400236p (2013).
2. Chen, Z. et al. Self-Assembled, Nanostructured, Tunable Metamaterials via Spinodal Decomposition. ACS Nano 10, 10237-10244, doi:10.1021/acsnano.6b05736 (2016).
3. Hohenberg, P. C. & Halperin, B. I. Theory of dynamic critical phenomena. Reviews of Modern Physics 49, 436-479, doi:10.1103/RevModPhys.49.435 (1977).

Electroplasticity (EP) is a phenomenon in which applied pulsed electric fields during deformation result in increased formability in metals. This phenomenon has been studied since the 1950’s [1] and researchers have proposed two possible origins of this nearly-universal effect: (1) localized Joule heating at lattice defects and (2) the electron wind effect leading to vacancy migration [2]. However, there is no clear evidence of the true origin of this behavior at the level of individual defects [3]. In this work, in situ Transmission Electron Microscope (TEM) electromechanical tests were performed in order to correlate direct observations of dislocation motion with both mechanical data and applied electrical pulses in an effort to provide clarity on the true origin of this interesting phenomenon.

In this study we have systematically studied the way electrical fields alter the structure and mobility of dislocations in Ni. The controlled electromechanical experiments were performed during in situ testing in a JEOL 3010 TEM, an electrical push-to-pull device (E-PTP) and a Bruker, Inc. Picoindenter with Nanoscale Electrical Contact Resistance (nanoECR). Specific loading and observation directions were selected for the specimens, which were attached via focused ion beam (FIB) lift-out to the E-PTP device. The specimens were deformed uniaxially in tension with the motivation to correlate dislocation behavior and applied current pulses. A sequence of tests was carried out by load control and with variable electrical current pulsing during the mechanical test.

After each test, videos of the dislocation behavior could be directly compared to the electrical and mechanical data. Our initial studies resulted in two types of dislocation motion. The first type of motions happened in less than 0.1sec where pre-existing dislocations escaped from the sample upon pulsing. The other type of motion was slower, with dislocations remaining in the sample for more than 1 second. In the latter case, dislocation motion could be observed through several consecutive steps.

Our observations found that during plastic deformation, relatively large slip traces with dislocations nucleated from the edges could be triggered by the pulses, contributing to large displacement increases in the sample. One example of dislocation motion in Al shows the motion of a dislocation could be correlated directly with the applied current when it was turned on, which can be explained by small changes in the applied electrical current that enabled the dislocation to overcome barriers. In addition, dislocations tend to move with a faster speed and a longer distance during the pulsing period, compared to the purely mechanically-triggered movement. We will also discuss results such as these in the context of local joule heating which can be measured indirectly through expansion of the lattice measured using TEM diffraction.

It is our ultimate goal to quantitatively measure the strain fields and local lattice change around dislocations to understand the external current effects in the plastic deformation of metals and alloys. Comparisons between the lattice strain fields, dislocation behavior and applied electric filed will provide valuable data to further improve atomistic models describing the dislocation motion change with electroplasticity. These experimental and theoretical frameworks will provide a scientific foundation for the design and optimization of alloys with enhanced electroformability for targeted manufacturing applications.

We present the first study by Laue microdiffraction (LMD)of the crystal orientationrelationships in lamellar microstructures obtained by thin-sample directional solidification of theeutectic Al-Al₂Cu alloy. The two eutectic phases are the Al-rich cubic solid solution (α) and the tetragonal intermetallic Al₂Cu (θ). Thin Al-Al₂Cu samples (12 µm thick) were fabricated by plasma sputtering on a clean, optically flat, sapphire single crystal. Eutectic growth was performed in our laboratory by using standard directional solidification (DS) at constant velocity in a fixed temperature gradient, and rotating directional solidification(RDS). The RDS method allows one to rotate the sample with respect to the temperature gradient, and therefore vary continuously the orientation of the eutectic solid during solidification. Both DS and RDS setups are equipped with a long-distance optical microscope for real-time observation of the growth front.
Laue microdiffraction has been carried out ex situ on the BM32 beamline of the EuropeanSynchrotron Radiation Facility(ESRF) in Grenoble, France. In this technique, a thin “white” x-ray beam (typically 5 to 22 keV) of lateral extension in the 100 nm range is sent to the surface of the sample. Selected areas (typically 100X100 µm²) were investigated by using a step-wise scanning motion of the sample holder. At each position (sample pixel) the camera detector records a full Laue pattern, which includes diffracted spots emitted by the metallic film and the sapphire substrate (nondestructive method). The signal coming from the sapphire single crystal was observed to remain unchanged during a scan. The alternation of the aand qlamellae could be clearly seen, and the corresponding sets of diffraction spots univoquely indexed.
We investigated 5 samples containing each 2 to 3 large eutectic grains. Here, we will mainly focus on a reference sample that has also been investigated by x-ray diffraction (XRD pole figures) and EBSD. That sample contained three distinct eutectic grains of the so-called
locked type. Our XRD, EBSD and LMD results are in full quantitative agreement with each other: 1-In the three grains, the Al₂Cu crystals are oriented with a(100)^θplane nearly parallel, and the (001)^θplane nearly perpendicular to the sample wall; 2-The distinct crystal orientation relationships (ORs) that we could identified in the three grains (and in the other samples) are different from those that are most commonly found in bulk Al-Al₂Cu samples; 3-There are strong indications that the interphase boundaries follow a (010)^θplane, independently of the OR. As a main conclusion, we found that the structure of the Al₂Cu phase dominates the determination of the free energy of the interphase boundaries, and selects the lamellar growth direction. Further investigations will be needed for elucidating the OR-selection mechanism at play (nucleation, grain competition). Finally, LMD measurements can be used to measure the mosaicity of the crystals. This work in in progress.

Recent improvements to quantitative phase-field simulations now make it possible to model microstructure formation in three dimensions (3D) on experimentally relevant length and time scales, but quantitative modeling remains so far limited to materials that solidify with atomically rough non-faceted solid-liquid interfaces such as metallic alloys and transparent analogs. We developed a quantitative phase-field formulation for the faceted ice-water system with anisotropic solid-liquid interface energetic and kinetic properties calibrated against existing experimental observations of free dendritic growth of pure water under microgravity conditions. Using 3D phase-field simulations of directional solidification of water with a dilute impurity, we reproduce for the first time salient features of freeze-cast templated structures including the formation of lamellae and undulated ridges. The formation of ridges is shown to be caused by the breakdown of ice platelets, which grow in between lamellae in the basal plane of ice, into linear arrays of cells with rounded tips. The cell spacing determines the ridge spacing and undulations are generated by secondary oscillatory instabilities of cell growth. Our results highlight the crucial role of the anisotropy of both the solid-liquid interface free-energy and the interface kinetics in structure formation. They also demonstrate that other potentially important physical effects neglected in the phase-field model, such as fluid flow generated by the expansion of ice, may play a more secondary role in structure formation. The value of such 3D phase-field simulations is immense since they enable us to investigate little studied and not yet modeled mechanisms of hierarchical microstructure formation and cell wall material self-assembly, both of which define anisotropy, property profile, and overall performance of freeze-cast materials. A companion presentation highlights the experimental results that motivate our work.

Owing to the structural and chemical nature of polycrystalline materials, grain and phase boundaries play a fundamental role in these materials properties by tailoring desired material features (e.g., mechanical, electrical, optical properties). In spite of the evident importance, the basic understanding of these interfaces is still in its beginning, and lacks an essential correlation of its structure and materials properties. In order to provide fundamental insights into the structure and chemical nature of the grain boundary in oxides, grain boundaries of Bi₂O₃-doped ZnO polycrystalline materials in earlier sintering stages were analyzed by Aberration-corrected Transmission Electron Microscopy (FEI Titan Themis Cubed TEM operating at 300kV), where inelastic electrons from the beam were used as a source for mild temperature increasing in the system. Polycrystalline Bi₂O₃-doped ZnO samples were produced by solid-state synthesis route, consolidate by simple uniaxial pressure and sintered at 1100 ^oC for 2 hours. By using a Helios Nanolab Dualbeam 660 (Thermo Scientific), grain boundaries were identified and prepared in thin lamella form, which were studied by TEM. The first feature observed is the Bi-rich (nano)interphase region, where several atomic organizations could be identified. These structures formation was correlated to the complexions (phase-like behavior) [1,2] and their neighboring crystals orientation. Besides, during the analyses, electron beam promoted mild temperature increasing in the sample, which was enough to induce Bi migration, which was possible to observe at atomic resolution and real-time. According to the time-resolved images, the Bi diffusion followed a chemical potential gradient path over the sample, i.e. from higher (bulk) to lower (grain boundary region) chemical potential gradient, inferring it was the driven force for the phenomenon. Moreover, the temperature increasing promoted the Bi-rich phase acting as a liquid-like behavior on the sample surface, following the same chemical potential gradient path over the sample. The interface between ZnO crystals and Bi-rich interfaced also played an important role in the dynamics, where the Bi atoms could crystallize (oxide form) and then incorporated to the grain boundary region. Hence, by following the myriad of phenomena at atomic scale and real-time, this study demonstrates the complexity of grain boundary formation and provides new venues towards the development of advanced materials by grain boundary engineering.

1. Dillon, S.J., et al., Complexion: A new concept for kinetic engineering in materials science. Acta Materialia, 2007. 55(18): p. 6208-6218.
2. Luo, J. and Y. M. Chiang, Wetting and prewetting on ceramics surfaces, in Annual Review of Materials Research. 2008, Annual Reviews: Palo Alto. p. 227-249.

We follow an Al-12%at.pct Cu alloy sample from the liquid state to mechanical failure, using in situ X-ray radiography during directional solidification and tensile testing, as well as three-dimensional computed tomography of the microstructure before and after mechanical testing. The solidification processing stage is simulated with a multi-scale dendritic needle network model, and the micromechanical behavior of the solidified microstructure is simulated using voxelized tomography data and an elasto-viscoplastic fast Fourier transform model. This study demonstrates the feasibility of direct in situ monitoring of a metal alloy microstructure from the liquid processing stage up to its mechanical failure, supported by quantitative simulations of microstructure formation and its mechanical behavior.

Additive manufacturing (AM) are a group of transformative technologies that promise to revolutionize the way we build things. In particular, metal AM have already found many important applications in the field of aerospace, automobile, energy, and defense. Most of metal AM processes involve using high power heat sources (i.e. electron beam and laser) to locally fuse the feedstock to the previous layer. Owing to the small size and fast scanning speed of the heat source, materials experience rapid melting and solidification, as well as multiple heating and cooling cycles during the build. For example, in a laser powder bed fusion process, the heating and cooling rates are generally in the order of 10⁶ K/s, and the local thermal gradient can often reach 10³ K/mm. Given such extreme thermal condition, unique microstructures develop in additively manufactured materials, such as cellular structures, high dislocation density, unusual chemical segregation, and non-equilibrium phases, etc. Some of these structural attributes are favorable for certain applications, while others are deleterious.

At the Advanced Photon Source, we recently developed and applied two different, yet complementary, diffraction techniques to study the microstructure evolution in laser AM processes. One diffraction technique is to use intense pink beam (~24 keV) at the 32-ID beamline for high-speed experiments. The energy bandwidth of a pink beam is generally in the scale of 10^-2, comparing with a highly monochromatic beam with only 10^-4 bandwidth, yet the flux of a pink beam is orders’ higher. With a scintillator-based indirect detection system, pink-beam diffraction patterns can be collected with a rate up to 100 kHz during laser AM of metallic materials. This high-speed diffraction technique offers very high temporal resolution that is needed for studying ultrafast phase transformation. But the dynamic range and reciprocal space resolution of the data are relatively low, limited by the detector and the x-ray energy bandwidth, so it is challenging to use this technique to detect minor phases and sub-grain structures. The second diffraction technique is to use high-energy monochromatic beam (~55 keV) at the 1-ID beamline. A PILATUS3 X CdTe detector is used with a data collection rate up to 500 Hz. Because of the large detection area, high dynamic range and reciprocal space resolution, this diffraction technique allows one to resolve the detailed structures in a diffraction spot and decode important structure information beyond material phase, such as dislocation density, texture, stress field, etc. These two in situ/operando synchrotron diffraction techniques can not only help us understand the unique microstructures of AM materials, but also enable many other exciting research opportunities. These include, but not limited to, the establishment of precise time-temperature-transition diagrams (particularly in the time domain below ms) of structural and functional alloys, and the development of kinetics-driven synthesis routes for novel metastable alloys with hierarchical microstructures and tailored functionalities.

Laser powder bed fusion (LPBF) is a standard technique in additive manufacturing (AM) of metal because of its capability to build complex and precise parts. The laser-based processing is accompanied by rapid heating and cooling, and abrupt changes in phase and structure. Applying the high-speed synchrotron X-ray diffraction of the Advanced Photon Source (APS) provides essential insights into AM processing. In this study, we used high energy monochromatic X-rays and a high-speed Pilatus detector to probe the dynamics of the rapid evolution of the microstructure of AM metals under laser exposure. Various physical phenomena, including phase transformation, thermal effects, and evolving microstructure, are quantified using high-resolution x-ray diffraction. Based on the temperature and phases present, the Bragg rings expand, broaden, vanish, and re-appear.
We have applied this in-situ synchrotron x-ray technique to various materials, including Al-, Ti-, Fe-, and Ni-based alloys. In Ti-6Al-4V, phase transformation between liquid, HCP α phase, and BCC β phase can be clearly identified. Movement of the (110) peak of BCC during the transformation indicates recalescence caused by the exothermic process from BCC to HCP. Although the phase composition is generally as expected from the relevant phase diagram, the kinetics often vary from typical experience. In Inconel 718, e.g., precipitate formation takes place within a few milliseconds after solidification, which conflicts with standard TTT diagrams. The in-situ analysis shows that a few milliseconds are enough for Laves and carbide phases to form. The temperature of specimens is also inferred from the coefficient of thermal expansion (CTE) and the instantaneous lattice parameter(s) value. In addition, high-speed X-ray in-situ imaging technique provides geometric information on vapor cavities, melt pools, and cracking. The coupling of these two in-situ techniques will improve our understanding of the physics underpinning metal LPBF process.

Laser based additive manufacturing allows building layer by layer a designed shape offering versatility and flexibility to many metallurgical sectors. The obtained microstructures depend however strongly on the laser and scanning parameters. Trial-and-error methods to determine the optimal SLM processing parameters for a desired microstructure are time consuming. Numerical modeling approaches can help predicting and understanding the role of individual parameters, but because of the multiple length and time scale character of the physical processes involved, the models become rapidly very complex.

To speed up the selection of parameters, to help validating computational models and to improve our understanding on the link between process parameters and microstructure in SLM, we have developed a miniaturized SLM device (MiniSLM) that allows printing 3D samples during synchrotron X-ray diffraction. The design of the SLM device was driven by specific requirements that are related to its implementation at synchrotron beam lines; the requirement to be portable and to be small and light enough to fit at various beam lines. Furthermore, the X-ray beam needs direct access to the powder bed, minimizing interference with absorbing or diffracting material. The constructed MiniSLM has all basic functionalities of commercial available SLM machines.

We report on Ti-6Al-4V samples that were manufactured in operando during X-ray diffraction at the MicroXAS and MS beamlines of the Swiss Light Source. By varying the laser power, scanning speed and hatch distance, various energy densities were obtained. The dynamics of the alpha and beta phases during fast heating and solidification are tracked with a time resolution of 50µs. We reveal the influence of laser and scanning parameters on the temperatures reached during the process, the cooling rates of α and β phases and the duration the β phase exists during processing. By performing experiments with different scanning strategies including those required for making open lattice structures, we demonstrate that printing of short scan lines, as used in an island scanning strategy, result in coarsening of the prior β grains, lowering of the cooling rates after solidification, and therefore leading to a substantial different microstructure.

The time resolved measurements obtained from in operando xray diffraction provides information averaged over the volume probed by the x-ray beam (varied between 40μ and 200µ). We demonstrate how the in operando results can be used for validation of FEM-based models. Tuning the xray beam profile to the mesh size or vice versa, will optimize the synergetic aspect.

Welding is considered as the most efficient and economical route for joining metals. In fusion welding processes, metals coalesce via extremely rapid melting and subsequent solidification. The melt pool evolution during this melting and solidification sequence determines formation of the final weld joint shape, microstructure and defects. So that scientific insight on weld pool evolution and related phenomena is vital to enhance structural integrity and resilience of the welded structures or components. However, inherent complexity with multi-physics phenomena, associated high temperatures and the rapidness of the welding processes make direct experimental investigations extremely demanding. In this contribution we report the observation of solid-liquid-solid phase transformation and related melt pool dynamics that take place during electric arc welding, obtained through in situ synchrotron imaging at millisecond scale. The analysis flow patterns along with the quantified weld pool surface dynamics revealed us to how different contributing forces dictate the flow conditions over the distinct durations of the relatively short existence of the liquid phase. Our preliminary results suggest the existence of arc, surface tension and gravity dominant regimes during the evaluation of the weld pool. Further, we present our observations on how different welding parameters influence these regimes and develop into different transient conditions.

High energy x-ray diffraction microscopies make possible the in-situ observation of internal microstructural evolution during processing in a wide variety of polycrystalline materials. Using distinct but compatible methods, length scales from millimeters to tens of nanometers can be accessed. Currently, crystalline grain unit cell orientations and strain fields are mapped over cubic millimeter volumes using high energy diffraction microscopy (HEDM). Bragg coherent diffractive imaging (BCDI) is sensitive to Angstrom scale atomic displacement fields that are resolved to a few tens of nanometers inside nano-particles; individual dislocations and their motions have been resolved. Dark field x-ray microscopy (DFM) also is able to image individual defects and their motions. Soon to come upgrades to synchrotron radiation facilities such as the Advanced Photon Source and the European Synchrotron Radiation Facility will accelerate and broaden the applicability of these techniques while also posing significant data volume challenges. Current efforts seek to combine these capabilities to allow mapping macroscopic volumes of a sample at micron resolution and then to zoom in on chosen regions of interest for studies of underlying nanoscale dynamics. Such multiscale measurements have the potential to lead to development of validated multiscale models that are based on fundamental mechanisms and therefore can be predictive rather than being tailored to a specific sample. This talk will present an overview of experimental methods and their capabilities and progress towards integration. Example applications to in-situ studies of mechanically and thermally induced evolution will be demonstrated.

We present our work on the development, demonstration, and application of a new method of quantitative time resolved specular x-ray reflectivity (XRR) for use with monochromatic synchrotron radiation. Our method utilizes a polycapillary x-ray optic to create a convergent beam, illuminating the sample surface with x-rays over a range of incident angles. An area detector collects scattered intensity over a range of exit angles. By altering the geometry, rotating the sample normal out of the plane of the incident beam, we can greatly reduce the effect of diffuse scatter, significantly improving the quality of the measured XRR compared to previous convergent beam XRR methods. We demonstrate our technique with time resolution up to 10 ms and show XRR with intensity over 5 orders of magnitude with q_z resolution sufficient to measure Kiessig fringes from films as thick as 105 nm. Our optic, with a working distance of 22 cm, makes this measurement compatible with many in situ processes. After demonstrating this method, we applied it to the study of interdiffusion in metal-metal multilayers. By measuring the intensity of the first multilayer diffraction peak during rapid sample heating at up to 90 °C/s, we were able to extract the activation energy for interdiffusion in an Al/Ni multilayer system.

Additive manufacturing (AM), also known as 3D-printing, of metals and alloys involves reading a 3D part geometry from a computer-aided design file and building it in a layer-by-layer manner by locally melting the material using a moving heat-source (laser or electron beam). Based on the material feeding approach, metal AM processes can be divided into: (a) direct energy deposition (DED) and (b) powder-bed fusion methods. In DED, the feedstock material in powder/wire form is directly fed into the heat-source as it scans the build surface whereas in PBF, the powder feedstock is deposited onto a powder-bed between two successive layer scans. For both processes, the final as-built part contains surface and volumetric defects, and a heterogeneous non-equilibrium microstructure. These phenomena depend on the AM process parameters, composition of the feedstock, and the desired final part geometry which together determine the material properties, the residual stresses and the subsequent part performance. AM holds the potential to revolutionize the metal and alloy manufacturing sector through its ability to simultaneously synthesize the material from powder/wire form and manufacture the desired part geometry without the need of tooling.

In recent years, increasing number of research efforts are being dedicated to better understand, control and improve the process-microstructure-properties-performance relationship of AM metals and alloys. However, nearly all experimental and modeling efforts in this field have been directed towards studying the heat-matter interactions in the melt-pool and rapid solidification. While these are important topics to study, it is equally important to understand the microstructure evolution in the heat-affected solid zone which undergoes thermal cycling. Consider the layer-by-layer building of a single-track (one heat-source scan per layer) wall and focus on the material at a point X on a deposition layer L (>1). Immediately after the localized deposition, the material at X cools down at a very high temperature rate (ranging from 10³ to 10⁶ °C/s). Addition of more layers (i.e. L+1, L+2, …) results in thermal cycling at X with varying temperature rates and amplitudes. The transient thermal gradients result in the formation of high-amplitude cyclic thermal stresses which must affect the microstructure at X, for example, through dislocation density evolution, recrystallization, precipitation, etc. At later stages of thermal cycling (i.e. during deposition of layer N >> L), a steady-state heat transfer could occur at X resulting in additional microstructural changes. Furthermore, material points at different locations in the as-built part are subjected to different thermal histories resulting in heterogeneous microstructural evolution. To our knowledge, there have been no coherent studies dedicated towards understanding the microstructural evolution during thermal cycling of the material. The main difficulty is tracking the microstructural changes in-situ during the AM process.

In this study, we are particularly interested in understanding the microstructural changes at the intragranular length scale due to thermal cycling during the AM processing of 316L stainless steel. Instead of following the building of this steel inside the AM processing chamber, we thermally cycle TEM lamellae inside a transmission electron microscope and follow their evolution in-situ. The lamellae are prepared from the material synthesized during the building of a wall of 316L stainless steel using a Laser Metal Deposition machine (a DED process). Once inside the TEM, the lamellae are subjected to the temperature vs time history obtained from thermal measurements at a given point-of-interest on the 316L wall during its processing. This thermal cycling results in some very interesting changes in the microstructure which will be presented during this talk.

We describe novel experimental approaches to examine the microstructural evolution of Ti 6%Al 4%V upon cooling from above the beta transus (~995 °C) while imaging in the scanning electron microscope. Ti 6%Al 4%V is a two phase, α+β titanium alloy with high strength and corrosion resistance. The existence of the α/β transformation gives rise to different microstructures and properties through various thermomechanical treatments. Fully lamellar microstructures, bi-modal microstructures, and equiaxed microstructures can each be obtained by accessing different cooling rates upon the final treatment above the beta temperature, each resulting in uniquely enhanced material properties.

Utilizing the capabilities of a heating/ tensile stage developed by Kammrath & Weiss Inc., that also incorporates in-situ nano-indentation capability, we are able to apply real-time imaging techniques in the scanning electron microscope to monitor and control the development of the microstructure. Annealing temperatures up to 1100 °C are attainable, with cooling rates ranging from .01 ° C per second to 4 °C per second. This has allowed us to directly observe the formation of lamellae at different annealing temperature/ cooling rate combinations to determine the lamellar microstructure width, separation, and colony size.

Applying a cooling rate (0.1 °C per second) to samples annealed to 1100 °C showed that lamellar microstructures initiate and form within alpha colonies at temperatures of ~ 920-930°C. When held at the temperature where lamellar formation initiates, the lamellae continue to propagate and lengthen by 50% from their initial length in two minutes time throughout the parent grains. Upon further cooling, the lamellae widen from 3.5- 7 µm to 8-15 µm. Extensive analysis of lamellar microstructure formation and growth will be presented, comparing different temperature regimes, cooling rates and times at temperature.

The ability to further refine thermomechanical processing conditions allows for an assessment of mechanical properties associated with the presence of both phases, and their final microstructure, particularly given the in-situ indentation capabilities available in our system. As such, the ability to deconvolute temperature and time considerations in the formation of lamellar and also bi-modal microstructure is vital, and will be discussed in detail during this presentation.

Acknowledgments: This research is sponsored through grants from the National Science Foundation, Awards CMMI-1647005, and CMMI-1729336. This work was performed with the use of clean room and characterization facilities within the Center for Materials, Devices and Integrated Systems at Rensselaer Polytechnic Institute. We also acknowledge collaborations with Dr. Antoinette Maniatty, Dr. John Wen, Arun Baskaran, Sagar Bhatt, Anant Kekre, and Michael Allahua. Additionally, we acknowledge the help and expertise of Kammrath and Weiss with equipment development.

An approach is proposed to measure grain boundary properties by comparing the evolution of experimentally determined 3D grain structures to that derived from simulation. Grain evolution in pure iron is determined in three dimensions and as a function of time using diffraction contrast tomography at a synchrotron source. The temporal evolution of over 1300 grains is quantified. This provides information on the morphology and kinetics of a large number of grain boundaries in a single experiment. Using a time step from these data as an initial condition in a phase-field simulation, the computed structure is compared to that measured experimentally at a later time. An optimization technique is then used to find the reduced grain boundary mobility that yields the best match of the simulated microstructure. We find a strong variation of the reduced mobility for a given grain boundary misorientation. A strong dependence of the reduced mobility on location in Frank-Rodrigues space is also observed.

Since the intrinsic grain structure of a polycrystal has significant impact on its properties, investigators have focused intently on studying the dynamics of the grain network under various external stimuli, e.g., heat. Upon annealing the material may undergo abnormal grain growth (AGG), in which a few grains grow more rapidly compared to the others. AGG is a common grain coarsening phenomenon that has been observed in a vast array of metallic and ceramic systems. To capture AGG, past experimental studies have relied heavily on two-dimensional (2D) experiments, such as surface characterization via electron backscatter diffraction (EBSD) and optical microscopy. However, such experiments have had limited success in developing a comprehensive model of AGG due to lack of information below the sample surface. Furthermore, these characterization techniques are destructive in nature, and thus it is impossible to follow the individual grain trajectories as a function of time (in the same specimen). As a result, it is impossible to determine the sequence of events that led to AGG in the first place.

Integrating X-ray absorption- and diffraction-contrast tomography (denoted ACT and DCT, respectively) in the laboratory can provide in-depth information on microstructures that are simultaneously polycrystalline and multiphase. Here, we have harnessed this non-destructive method to gain new insights into the problem of AGG in a particle-containing system. We investigate an Al-Cu alloy as a model system. The prepared specimen was imaged through the combination of ACT and DCT on a laboratory X- ray microscope in situ and as AGG proceeds. The 3D data provided an unprecedented opportunity to uncover what determines the expansion and shrinkage of grains — including the abnormal grain — as a function of the initial grain size and topology [5, 6]. However, both of these metrics do not correlate with grain growth rate. Instead, we find a strong correlation between the locations of secondary phase particles (obtained via ACT) and grain boundaries (imaged via DCT). As the specimen is annealed, the distribution of grain boundaries becomes more closely correlated to the distribution of Al₂Cu particles, suggesting that the abnormal grain nucleates in a relatively particle-free region of the sample volume. Meanwhile, the normal grains are pinned by a dense particle cloud. Once it has attained its size advantage, the abnormal grain persists to grow due to an exceptionally large capillary driving force. This study provides new understanding on the origin of AGG with implications for the production of single crystals.

[1] T. Kusama et al., Nat. Commun. 8 (2017), p. 354.
[2] T. Omori, H. Iwaizako, R. Kainuma, Mater. Des. 101 (2016), p. 263
[3] H. Park et al., Mater. Charact. 146 (2018), p. 204
[4] R. Keinan et al., Acta Mater. 148 (2018), p. 225.
[5] M. Hillert, Acta Metall. 13 (1965), p. 227.
[6] W.W. Mullins, Acta Metall. 37 (1989), p. 2979.

Bulk nanocrystalline alloys possess many unique physical properties with tremendous potential for various applications. However, ultra-nano grains of sub-4 nm dimension are extremely difficult to create and stabilize. This research proposes a spontaneous and reversible approach to nanocrystallize Cu-Sn bimetallic alloy by taking advantage of the polymorphic phase transition of Cu₆Sn₅. During temperature increase, and starting from a temperature 275 ^oC lower than the melting point of bulk Cu₆Sn₅, ~10 nm Cu₆Sn₅ nanoparticles were coarsened to tens of nanometers, as observed by in situ TEM. The particle growth continued even during the initial cooling stage. Interestingly, when the cooling process went through the polymorphic phase transition temperature of Cu₆Sn₅, η-Cu₆Sn₅ was transformed to η’-Cu₆Sn₅, and massive grain refinement occurred spontaneously as confirmed by atomic level monitoring in TEM. The final grain size reached a mean value of ~2.5 nm, sufficient small to be in a regime with enhanced ductility, offering highly improved ductility because of inverse Hall-Petch phenomenon, and excellent thermal, electrical and mechanical properties. This process is also reversible, offering an unprecedented method to stabilize extremely small nanograins. Such an ultra-nano material can be technically useful for a variety of applications including as a unique circuit interconnection material which can serve at higher temperatures than its processed temperature. We sucessfully demonstrated using nano-Cu₆Sn₅ as the interconnection mateiral for the packaging of SiC high-power devices. Becasue of the significantly depressed sintering temperature of ultra-fine Cu₆Sn₅ nanoparticles, the SiC can be successfully packaged at a temperature lower than its actural working temperature. Becaue of the enhanced ductility of the inverse Hall-Petch sub-4 nm nanocrystaline joints, nano-Cu₆Sn₅ has a very promissing potential to be used in the packaging of flexible devices. It can offer two attractive advantages simultaneously: processing at low temperature to prevent the damage of the flexible substrate; but working at a relatively high temperature to allow the usage of high-power flexible electronics. Like Cu₆Sn₅, other alloys such as Cu-Al, Fe-Zr, Al-Ni, etc. also possess such a capability of phase transition-induced grain refinement. This work will inspire other researchers to create many additional, sub-4 nm type nanocrystalline alloys, offering unprecedented pathway to realize ultra-nano-grained alloys and more interesting fuctions which were difficult to gain before.

Most technologically useful materials are polycrystalline microstructures composed of a myriad of small monocrystalline cells/grains separated by grain boundaries/interfaces. Grain boundaries play a crucial role in determining the properties of materials across a wide range of scales. A grand challenge problem in engineering of polycrystals is to develop prescriptive process technologies capable of producing an arrangement of grains that provides for a desired set of materials properties. One method by which the grain structure is engineered is through grain growth or coarsening of a starting structure.

Grain growth can be viewed as the evolution of a large metastable network, and can be mathematically modeled by a set of deterministic local evolution laws for the growth of an individual grain combined with stochastic models to describe the interaction between grains. Thus, to develop a predictive and prescriptive theory, investigation of a broad range of statistical measures for microstructure evolution during grain growth in polycrystalline materials are needed and must be obtained using experiments, simulations, data analytics and mathematical modeling. One of the goals of our current research presented here is thus to use data-driven numerical simulations closely integrated with experiments in order to improve upon mathematical models of grain growth and guide design of new experiments.

In this talk, we will report the results of in situ experiments of grain growth in thin films of Pt conducted using a heating holder in a transmission electron microscope (TEM). Our interest here is the dynamics of grain boundary motion. In addition, we will report on ex situ experiments of grain growth in Al, Al-0.2 at% Mn and Cu thin films, with thicknesses in the range of 25-150 nm and annealing temperatures in the range of 150-600 C. A variety of geometric and topological metrics of the grain structure were extracted from bright-field or dark-field (TEM) images or precession electron diffraction crystal orientation maps, including not only the grain size, but also the dihedral angles, the number of neighbors (i.e., side class) and the side class of nearest neighbors of grains.

The experimental grain size distributions for the Al and Cu films were found to be remarkably similar to each other despite the many and significant differences in experimental conditions, which include sputtering target purity, presence of an alloying element, substrate type, film thickness, deposition temperature, actual as well as homologous annealing temperatures, annealing time, absolute grain size, and the twin density within the grains. This similarity argues for a universal experimental grain size distribution, which for grain diameters is lognormal as found previously for thin films at stagnation. Comparison of the experimental grain size distribution with that for two-dimensional, sharp-interface grain growth simulations with isotropic boundary energy shows the distributions to differ in two regions, termed the “ear” and the “tail”. It is shown that the excess small grains in the region of the “ear” are primarily the 3 and 4-sided grains, whereas the excess of large grains in the “tail” region are grains with more than 9 sides. The excesses in the ear and tail regions of the experimental distributions are necessarily balanced by a deficiency in the mid-sized grains with 6-8 sides. We will also present the grain boundary character distributions in the films and compare and contrast our findings with those for bulk samples with grain sizes in the micrometer range.

Central to the advances in functional materials and their processing is the thorough understanding of the fundamentals governing the processing-properties relationships. In additive manufacturing, understanding these relationships is necessary since anisotropy is inherent to the processing techniques, compromising the components quality, performance and reliability. For continuous-flow direct writing (CDW) the materials are formulated as viscoelastic colloidal inks and extruded through a nozzle forming a component. Therefore, these inks are subjected to multiple stimuli during the deposition and post-processing stages, which affect the way these materials self-assemble, store/release stress, and subsequently form and propagate defects.
Seeking to unveil such fundamental relationships, we conduct operando synchrotron x-ray photon correlation spectroscopy (XPCS) studies to investigate the inks’ structural and dynamical evolution during the out-of-equilibrium stages associated with the 3D printing process. During the shearing and relaxation of the inks multiple phenomena take place including loss and recovery of the gel structure, solidification and jamming to name a few. Our setup combines XPCS and CDW at the Coherent Hard X-ray Scattering beamline (CHX-11ID) at the National Synchrotron Light Source II. XPCS is especially suited for investigating the nano-scale spatio-temporal evolution of materials processed by CDW, covering its relevant length and time scales, from tens to hundreds of nanometers, and milliseconds to the hundreds-of-seconds range. We observe the ink’s nanoscale structure and dynamics as a function of the spatial position within the printed filaments, which may further be related to the printed-structures’ solvent evacuation, colloid re-arrangement, and substrate properties (ink/substrate interactions). Additionally, we investigate the dependence on compositional variation, including inks from branched vs. monodisperse colloids and inks containing solvents with different vapor pressures, to expose fundamental differences in the structural and dynamic evolution. While our work serves to enlighten fundamental processes in CDW, it may be extended to other 3D printing techniques. Moreover, it constitutes the basis for the discovery of new pathways for the engineering of materials that exhibit novel properties and unprecedented performance, as well as shedding light onto solving relevant issues in additive manufacturing, including cracking, delamination and poor layer-to-layer adhesion.

Selective laser sintering (SLS) methods are workhorses for the additive manufacturing of polymer-based components. The ease of rapid prototyping and short-run manufacturing by anyone with a laptop and machine access also means it is easy to produce illicit or counterfeit components. It is necessary to have a data-calibrated in-situ physical model of the build process in order to predict expected and defective microstructure characteristics that inform the provenance of these components.

Toward this end, we calibrate 2D and 3D phase-field sintering models for material microstructure evolution and its kinetics during synthesis. We seek to understand and predict microstructure characteristics such as pore structure and layer interfaces in components built by a Formiga P100 SLS printer using PA 2200 Nylon 12 feedstock. This is accomplished by assimilating multiple in-situ data streams, post-hoc imaging analysis, and computational model predictions in an adaptive Bayesian parameter estimation algorithm. In-situ data sources include infrared thermal imaging, built-in printer environmental sensors, and laser pathing guidance, including experimental uncertainty estimates. These are supplemented by post-hoc conical and synchrotron X-ray computed tomography, from which we infer bulk porosity and distributions for local and global pore characteristics. Similar predictions for porosity distributions are provided by the stochastically-driven phase-field sintering model. Randomized feedstock configurations are explored by a stochastic model for Nylon particle deposition and composition, which has been calibrated to bulk particle size analyses from a laser diffractometer. Additional model parameters such as system free energy, interfacial energy, and sintering kinetics are first constrained to physically-relevant search directions by global and local sensitivity analysis, and then explored using an adaptive sampling algorithm to match predictions.

Using this feedback loop, we obtain data-constrained estimates of sintering model parameters along with uncertainty bounds. We conclude by combining these predictions with the infrared thermal maps, and quantitatively relate the in-situ build environment to the resulting microstructure in sintered nylon components.

Solid-state ceramic synthesis is often believed to equilibrium process, as the high temperatures of calcination are expected to minimize kinetic limitations associated with nucleation and growth. Despite this intuition, non-equilibrium phases are often observed during solid-state synthesis, prompting questions regarding the mechanistic origins of these metastable intermediates. Here, we use in situ synchrotron XRD to compare the multistage phase evolution of three important Li-ion cathode materials, LiCoO₂, LiNiO₂ and LiNi_0.8Co_0.2O₂, during a hydrothermal-assisted ceramic synthesis process. While all three systems ultimately crystallize in the stable Rm layered phase after heating, a metastable spinel phase is observed to form prior to LiCoO₂, and disordered rocksalts prior to the Ni-based compounds. We demonstrate that these metastable intermediates are kinetically-facile lithiation products of stable binary oxides that originate during low-temperature synthetic preparation. Our findings highlight newopportunities to engineer low-temperature phases to topotactically template the ceramic synthesis of target crystalline phases.

Electromagnetic (EM) fields can significantly affect crystallization and phase transitions in ceramic materials, leading to low-temperature phase transitions and compositions not observed using conventional furnace-based techniques. Specifically, the application of 2.45 GHz microwave radiation (MWR) during ceramic synthesis has been found to crystallize unexpected phases and influence local atomic ordering. However, the mechanisms underlying how EM fields impact phase stability and promote far-from-equilibrium phases are not well understood. To fully understand the role of EM fields in ceramic nucleation and growth, it is necessary to characterize the relationship between the applied field and the resultant atomic structure. Due to the away from equilibrium nature of EM field-assisted phase transitions, it is necessary to characterize effects on both long range crystalline atomic order and short range atomic order which deviates from the crystalline phase. An understanding of the local atomic distortions induced by MWR and how they impact relative phase stability and the free energy space is especially critical to understanding these processes. An improved understanding of these effects will create the opportunity to utilize EM fields as an additional thermodynamic parameter to access regions of phase space not accessible via conventional synthesis routes.

We utilize a combination of synchrotron X-ray pair distribution function (PDF) analysis, synchrotron X-ray absorption spectroscopy, and density functional theory (DFT) calculations to investigate the impact of MWR on defect formation, phase transitions, and phase stability in ZrO₂ thin films. Through PDF and extended X-ray absorption fine structure (EXAFS) analysis, we demonstrate that the application of MWR promotes defects on the oxygen sublattice, leading to local structural distortions which stabilize a long-range cubic atomic structure. This defective cubic ZrO₂ structure forms at 225°C in the presence of MWR, significantly lower than the temperature conventionally required to synthesize cubic ZrO₂ (>2400 °C). DFT calculations reveal that point defects on the oxygen sublattice can significantly alter the free energy landscape of ZrO₂, and we find that the low-energy defective atomic configurations predicted by DFT correspond well to the experimental ZrO₂ atomic structure observed. This combined experimental and computational study provides direct evidence for the role of EM fields in promoting defect-mediated phase transitions which deviate from the conventional free energy space of ZrO₂, and creates the opportunity to further utilize MWR-assisted synthesis as a means to access away-from-equilibrium phases in ceramic oxide materials.

Electromagnetic (EM) fields have been known to stimulate a range of surprising responses in materials. Microwave radiation is a type of EM field in the frequency range of 0.3 -300 GHz that can lower the synthesis temperature required for ceramics like TiO₂ and induces mixed amorphous-crystalline phase compositions. To better understand the effects of microwave radiation on matter, we study structural changes during microwave heating and microwave radiation-assisted synthesis using in-situ synchrotron X-ray diffraction. We observed anisotropic expansion-contraction of lattice parameters, ‘a’ and ‘c’ of the tetragonal anatase phase, which contradicts the results from conventional thermal heating. When the as-received anatase powders are heated with microwave radiation, an instantaneous decrease in the intensities of diffraction peaks indicates loss of long-range atomic order (decrystallization), which is supported by high-resolution electron microscopy and selected area diffraction patterns. Ex-situ Raman spectroscopy and X-ray photoemission spectroscopy of microwave-heated and microwave-synthesized anatase powders indicate increased defect-generation under microwave exposure. Molecular dynamics simulations on the anatase phase of TiO₂ suggests that introducing an oxygen vacancy can lead to the formation of an interstitial-vacancy pair. So, if microwave radiation can affect defect nucleation in the structure, the generation of defect-pairs away from equilibrium can explain the anisotropic lattice expansion, crystallization at lower temperatures, and/or stabilization of metastable phases under electromagnetic fields. Anisotropic lattice expansion and decrystallization under externally applied fields provide direct evidence for non-thermal coupling between EM fields and matter. Understanding such non-thermal, field-driven processes has implications in engineering low temperature processes for integrating ceramics with delicate polymeric materials for flexible electronics and in tailoring ceramic properties for energy storage, photovoltaic, and catalytic applications.

In situ transmission electron microscopy (TEM) has gained rapid growth in the past decade and attracted tremendous attention in diverse scientific research because it can acquire dynamic information and allow for the mechanistic understanding of various physical and chemical processes in a broad range of materials systems. Specifically, the advancement of in situ TEM in physical and/or chemically reactive environments has enabled the direct observation of structure evolution during materials synthesis. The sintering that transforms the precursor powders into the dense solid phase is the critical fabrication process for ceramic materials. To overcome the limitation of conventional long-holding techniques, recent efforts have been put toward flash sintering, which represents a solid-state synthesis far from equilibrium and shows great technical importance relevant to additive manufacturing. In this regard, to provide an in-depth understanding of the fundamental level, we employ in situ TEM techniques to study the dynamic structure evolution during the sintering of ceramic materials.
Here, we primarily focus on simple transition metal oxides (such as TiO2, Fe2O3, etc.) to investigate their sintering behaviors under various sintering rates. We use a TEM holder capable of in situ heating via MEMS-based devices, on which pristine metal oxide nanoparticles are uniformly dispersed. By tuning the input current density, we can control the power and heating speed that is applied to the nanoparticles, and simultaneously observe their morphological and structural changes in real time. From enormous in situ TEM observations, we found distinct sintering mechanisms relating to sintering rate: the slow sintering include stages of grain rotation, agglomeration, and growth, whereas the flash sintering proceeds directly through abrupt densification and subsequent growth, which is governed by the surface diffusion at different heating load. The effect of grain size and surface orientation also plays an important role in sintering kinetics. Our findings provide valuable insights into understanding the flash sintering mechanism and also show potential implications for developing fast additive manufacturing techniques for ceramic materials.

The growth of high-quality Silicon Carbide (SiC) substrates and nanoparticles is a topic of extreme technological interest due to the importance of this material for current and future technologies. Concurrently, such growth process is challenging due to the meta-stability of different crystal symmetries (polytypes) in the usual growth conditions. Atomic mechanisms ruling the extended defects kinetics in cubic SiC material, manifesting both stacking and anti-phase instabilities, have been investigated by means of cross correlations between simulated evolution and in-situ micro-structural characterizations of crystal growth processes. First we explored the configurational space of cleaved and defective (adatoms and vacancies) surfaces with ab-initio structure predictions algorithms at density functional theory level. Then the atomistic configurations and energetics are input of a Kinetic superlattice Monte Carlo code which simulates the growth kinetics of SiC substrates and nanoparticles. Ab-initio calibrated Kinetic Monte Carlo simulations in super lattices evidence a mutable and complex kinetic behaviour which has been validated by transmission electron microscopy analyses of experimental samples. Both simulations and characterizations show that anti-phase boundaries generate single staking faults with different symmetries or sequences of them. Moreover, boundaries interact with the extended defects of the stacking fault class promoting their annihilation or polarity changes. Defect-defect correlations are mediated by surface instabilities, since a clear interplay between the modifications of the growing surface morphology and the bulk defects occurs and ones cause and influence each others.

Multicrystalline silicon (Mc-Si) ingots for solar cells are manufactured by unidirectional solidification method. In general, since semiconductor materials are not subjected to structure control by a method such as thermo-mechanical treatment after solidification, structure control must be performed in a unidirectional solidification process. However, it has not been well understood how the microstructures in Mc-Si, such as, crystal grains, grain boundaries, twin boundaries, sub-grain boundaries, etc. behave at the crystal-melt interface.
We have attempted to understand the crystal growth mechanisms by directly observing the crystal-melt interface of silicon.
We have developed a lab-scale in situ observation system which is composed with a crystal growth furnace and a digital optical microscope. To date, we have successfully observed crystal-melt interfaces of materials including Si, Ge, Sb, Cu, Si_1-xGe_x, GaSb, and so on.
This presentation will focus on dendrite growth, crystal-melt interface instability, and grain boundary dynamics of silicon. The results of other materials will also be discussed.

A new imaging device for the characterization of silicon during high temperature processing is installed at beamline ID19 at the European Synchrotron (ESRF). We demonstrate the great potential of this imaging setup for the detailed investigation of the interaction and generation of grains and defects that influence the growth process and the final crystalline structure of silicon. Better knowledge of the formation and interaction of defects in silicon is essential in order to improve defect engineering during the manufacturing step of silicon used for photovoltaic applications. In particular, defects like dislocation bunches or sub-grain boundaries reduce the photovoltaic performance and should be therefore analyzed.

In our experiments, two individual camera system that are fully synchronized record the topography and the radiography images in real-time during silicon sample processing in a Bridgman furnace. Compared to previous studies, where the topographs were recorded on X-ray sensitive films, it significantly increases the image acquisition rate allowing a detailed analysis of defect dynamics in the sample.

Topography is the analysis of the fine internal structure of an X-ray diffraction Laue spot in real space that provides information on the crystal quality by the display of lattice distortions. As dislocations and strain fields deform the crystal structure, they can be observed in a topography image by contrast changes. In our experiments, one Laue spot (topograph) is selected and recorded with one of the cameras. The transmitted beam is recorded at the same time by a second camera constituting the radiography images. The radiography contrast originates from the density difference of the solid and the liquid silicon phases and provides information on the morphology and kinetics of the interface. This supplementary information is used to identify new grain nucleation and grain competition events. The in situ studies are complemented by ex situ methods like electron backscatter diffraction and etch pitch measurements, to get information on the grain boundary character and on the dislocation density.

The combination of all techniques allows to analyze the mechanisms responsible for grain and dislocation generation and strain formation. Our work focuses on 300 µm-thin single-crystal silicon float zone samples that contain no visible defects at the beginning of an experiment. During heating, dislocations are generated at the sample edges and their motion intensifies with increasing temperature. We measure their velocities up to the melting point and observe their interactions. In fact, little is known on the nucleation and mobility of dislocations at temperatures beyond 1373 K. When solidification starts, dislocations initially present in the seed propagate in the regrown crystal with the solid-liquid interface. It is important to learn more about dislocation extension from the seed to the regrown crystal, to explore the role of the seed used for example in recent process developments. At the sample edges (solid-liquid-vapor triple point) twin nucleation occurs, which leads to grain competition in the center, which in turn can trigger new grain nucleation. Even dislocation multiplication by the activation of Frank-Read sources can be observed, which underlines the efficiency of the new imaging device for a thorough and comprehensive characterization of silicon processing.

Phase change materials, including Ge₂Sb₂Te₅ (GST), can rapidly and reversibly switch between an amorphous, or glassy, state and a crystalline, or ordered, state via electrical or optical pulses. The amorphous phase and the crystalline phase have very different electrical and optical properties, which means that they can be used in rewritable optical data storage media and in fast, non-volatile electronic memory devices. Using in-situ electron diffraction techniques, we observed that excitation above a threshold fluence with a single femtosecond, 800 nm pulse induces sharp Bragg peaks indicative of large crystalline grains on the order of the probe beam size (80-100 um). This was confirmed by transmission electron microscopy (TEM). Further analysis with the TEM showed that these large single crystalline grains are dendrites located in the center of the laser excitation with a face-centered cubic (fcc) crystal structure. It was determined via energy dispersive X-ray spectroscopy (EDX) that the large single crystalline dendrites are the same composition as the initial amorphous phase. This eliminates constitutional undercooling as a plausible mechanism to explain the formation of the dendrites. The observation of the large grains and the fact that the composition does not change indicates a large crystal growth velocity regime and opens up new opportunities for the growth of single crystal phases of the phase-change materials.

The atomic structure of (001) SrTiO₃ surface has been studied for a long time by various surface sensitive diffraction and spectroscopy techniques. The exotic physical properties exhibited by this surface have been of particular interests since the discovery of a two-dimensional electron gas (2DEG) at the surface [1]. Although atomic-scale surface profile imaging using high-resolution transmission electron microscopy (HRTEM) or scanning transmission electron microscopy (STEM) techniques have been used to visualize the surface structure, assessing the dynamic atomic processes leading to the surface reconstruction still remains challenging.

Here we designed an in-situ HRTEM experiment at elevated temperatures in a Cs-corrected TEM (GRAND ARM 300 CF). TEM samples were prepared from a (001) SrTiO₃ single crystal along the [100] and [110] cross-sectional zone axes by using FIB and attached to a DENSsolutions MEMS heating chip. A negative spherical aberration (Cs) imaging (NCSI) HRTEM condition was applied to observe the surface structure in real-time with all atoms (including strontium, titanium, and oxygen) being clearly resolved with minimal contrast delocalization [2]. In addition to the surface profile imaging at NCSI HRTEM mode, STEM high-angle annular dark-field and annular bright-field imaging have been carried out at room temperature, 300 °C, 500 °C, 700 °C, and 950 °C. In order to determine the chemical composition and the valence state of the outermost surface atoms, atomic-resolution STEM energy-dispersive X-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS) elemental mapping have been performed, respectively.

We observed that the (001) SrTiO₃ surface was terminated predominantly by SrO layer throughout all temperature ranges studied by in-situ TEM. Interestingly, the SrO surface layer showed a unique surface reconstruction largely deviating from the bulk-truncated (1×1) structure with titanium (Ti) ions in the surface SrO layer, particularly occupying the oxygen-oxygen bridge sites (i.e. corresponding to the octahedral interstitial sites in bulk SrTiO₃). The presence of the Ti interstitials was confirmed directly by atomic-scale STEM EDS and EELS elemental mapping. As temperature increases, the interstitial Ti atoms were displaced increasingly upward from the surface plane and caused a pronounced rumpling of the SrO layer. When the temperature is reached 950 °C, a layer-by-layer sublimation occurred from the surface by the formation and migration of atomic step with a unit cell step height. We directly observed that the Ti atoms from the moving step edge migrate and occupy the interstitial sites immediately when the sub-surface SrO layer is exposed as free surface during sublimation. From the Ti-L_2,3 and O-K EELS edges, we observed that the formal valence of the interstitial Ti is mainly +3 and the surface SrO and sub-surface TiO layers contain a significant amount of oxygen vacancies. The observed surface structure of (001) SrTiO₃ with Ti interstitials in the SrO termination has not reported before and we believe this unique surface structure is closely related to the formation mechanism of surface 2DEGs by oxygen vacancies. Our real-time surface profile imaging of the dynamic processes of surface reconstruction provides new insight into the surface science and has the potential to reveal the emergent phenomena such as the formation mechanism of 2DEG at oxide surfaces.

References
[1] Santander-Syro, A. F., et al. Nature 469.7329 (2011): 189.
[2] Jia, C. L., et al. Ultramicroscopy 110.5 (2010): 500-505.

BaTiO₃ (BTO) is a promising candidate for lead-free electronic devices for applications using piezoelectric or ferroelectric thin films. To achieve large-scale integration of BTO with existing semiconductor technology, however, the synthesis of high-quality BTO films at suitably low temperatures is required to reduce the thermal budget during device fabrication. Here, we report a successful synthesis of ultrathin epitaxial BTO films on SrTiO₃ (001) by molecular beam epitaxy with a substrate temperature as low as 280 °C. X-ray diffraction and in-situ reflection high-energy electron diffraction are performed to investigate the structural properties and relaxation of the BTO film during growth. This result paves the way toward the realization of novel nanoelectronic devices using epitaxial BTO films.

Thin films formed by solid phase reaction often exhibit a complex composition of several atomic species. The (re)distribution, diffusive and reactive properties of the individual species as well as their relative motion upon thermal treatment have a distinct influence on the formation properties and thus on overall thin film properties. Rutherford backscattering spectrometry (RBS) and X-ray diffraction (XRD) measurements have proven extremely valuable in the study of thin film formation as a function of thermal treatment. RBS is an ideal technique to unravel the compositional depth profile of thin films. On the other hand, X-ray diffraction provides complementary information on the lattice structure (hence the phase!) and texture of the material.

In conventional studies, several specimens are typically subjected to different heat treatments and subsequently analyzed one by one, in order to obtain a complete overview of the reaction sequence. However, due to the discrete character of this approach, critical stages for the understanding of the development of thin film properties (e.g., formation of transient phases) are easily overlooked. This problem is circumvented by determining the specimen properties in real time, i.e. in situ during annealing/reaction with a high sampling rate. Additionally, real-time measurements drastically decrease the workload, as kinetic parameters, redistribution properties and phase sequence during thin film reaction, for example, can be obtained from a combined ramped RBS or XRD annealing, while numerous specimens have to be analyzed in the conventional approach.

The strength and complementarity of these in situ techniques will be illustrated with a number of typical examples. These will include:
- Growth mechanism of binary and ternary silicide and germanide thin film by thermal annealing. Particular attention will be devoted to the (re)distribution of Pt or Pd during solid phase reaction of Ni_xPt_1-x and Ni_xPd_1-x alloys with Si, and the role of Pt/Pd on the growth kinetics.
- Determining the dominant diffusing species during nickel and palladium germanide formation, two of the most promising candidates for use as ohmic contacts when anticipating germanium-based devices. By continuously monitoring the movement of a thin (inert) marker in situ during phase formation by RBS, the relative contribution of the atomic species to diffusion can be directly deduced.
- Metallization of GeSn, a material which is used to enhance the charge transport properties in the MOSFET channel via strain engineering. Ge_xSn_1-x with Sn concentrations required for device operation (of the order of 6-9 at.%) are metastable. Further treatment such as annealing, implantation or metallization often results in phase separation, i.e. Sn precipitation. Therefore, particular attention has to be paid to the Sn redistribution during metallization, for example Ni/GeSn reaction. A combination of in situ RBS and in situ XRD enables us to follow the (meta)stability and reaction kinetics during reaction.

An additional comment that should be made is that, in order to extract the compositional depth profile embedded in the RBS data, dedicated analysis of each individual spectrum, one by one, is required. Since a single real-time measurement easily contains a hundred or more spectra, analyzing a full data set can be extremely time consuming. To this end, we have implemented a novel analysis technique, i.e. artificial neural networks (ANNs), to tackle the quantitative analysis of large real-time RBS data sets. This approach, which is based on pattern recognition, allows to analyze vast amounts of data without compromising the integrity of the results, thus enabling to conduct systematic real-time RBS studies on large batches of samples or experiments.

Vacuum physical vapor deposition of C₆₀ on a graphene-coated surface is investigated in real-time by utilizing 9.65 keV X-rays from the CHX coherent hard X-ray synchrotron beamline at NSLS-II. X-ray Photon Correlation Spectroscopy is performed in the Grazing Incidence Small Angle X-ray Scattering (GISAXS) mode to achieve surface-sensitive conditions. Local step-flow is monitored through the observation of oscillatory correlations in the later stages of growth after crystalline mounds have formed. An important aspect of the work is that coherent X-rays do not average over complex structures, and this allows us to monitor the growth on polycrystalline surfaces without loss of information. This is important since in conditions where mounds form, the surface is too rough to observe layer-by-layer oscillations; this greatly limits the type of information that can be obtained from experimental techniques that employ low-coherence X-rays or electrons. The experimental results show that the step-flow velocity is nonuniform, and we model the velocity of each step-edge as being a simple function of the lengths of the terraces above and below it. This model predicts that the steps become almost stationary near the edges of the mounds where the local terrace length is very small, and the average slope of the surface is large. It was not previously known that such nonuniform and disordered step arrays as we have observed would follow such a simple growth law. We also find that the sensitivity to local step-flow is increased due to coherent mixing of X-rays scattered from the average mound structure with those scattered from the step array. This effect is a version of heterodyne scattering, where the scattering from the average step array can be considered to be a quasi-static reference signal. This work shows that the use of coherent X-ray scattering provides an approach to better understand surface dynamics and fluctuations during crystal growth. We have also extended these measurements to additional materials systems, including the growth of organic semiconductor small molecules with a lower symmetry than C₆₀, such as diindenoperylene.

The continued demand for power efficient and high-performance microelectronics components is the key driver for the increase in the density and miniaturization of transistors. The performance of these nanoscale transistors depends on their architecture and materials properties. Controlled fabrication of the nanoscale semiconductors materials tailored for microelectronics application requires detailed insight into the chemical and physical processes that take place during their fabrication. Optimization of these nanofabrication processes is currently based on trial-and-error approaches, and these structures are only characterized post processing limiting the insight into how materials evolve during the fabrication.
Here, we describe our development approach, which is based on our in situ liquid phase TEM imaging platform, to directly visualize the semiconductor fabrication related processes occurring in solutions with nanometer resolution in real-time. In my talk, I will show how we can study the etching of Si and SiGe nanowires, which will be an essential element in the next generation gate-all-around field effect transistors (GAAFET). Moreover, I will show an in situ TEM based test platform for studying the effects of anisotropic etching that is often present for many semiconductor materials and describe how different crystallographic Si planes etch.
Our results highlight the importance of visualizing the processes relevant to nanofabrication in order to understand and improve the nanofabrication related processes.

Hafnia based memristive devices are promising candidates as next generation non-volatile memory devices and are appealing because of the compatibility with back-end-of-line processes in the current semiconductor fabrication process and the good chemical and thermal stability. [1-2] One major issue with current devices is the large device-to-device variability [3] which could be resolved by highly textured hafnia where grain boundaries, interconnecting top and bottom electrode, serve as predefined pathways locally constricting the formation of the conductive filament as shown by conductive AFM [4].
In this study, a TiN/HfO₂/Pt stack grown on c-cut sapphire is investigated by monitoring structural changes of the dielectric layer with automated crystal orientation mapping (ACOM) during in situ heating in the TEM.
The 50 nm TiN layer has been grown by using reactive molecular beam epitaxy (RMBE) followed by in situ growing 12 nm amorphous hafnia and ex situ sputter deposition of 100 nm Pt. From one stack, a cross sectional focused ion beam (FIB) lamella has been prepared and transferred onto a Micro Electrical Mechanical System (MEMS) based heating chip. Laterally resolved structural information has been acquired by ACOM in order to track structural changes in the dielectric layer. The structural changes have been compared to another stack which has been ex situ annealed.
A temperature dependant study has been performed to monitor the growth process, thus giving insight into the growth mechanism and formation of grain boundaries in this system. Furthermore, we show, that the transformation from amorphous to the polycrystalline state lowered forming voltages and device-to-device variability serving as basis for a direct structure-property correlation.

References:
[1] Pan, F., et al., Recent progress in resistive random access memories: Materials, switching mechanisms, and performance. Mat. Sci. a. Eng.: R: Reports, 2014. 83: p. 1-59.
[2] Miranda, E.A., et al., Model for the Resistive Switching Effect in HfO2 MIM Structures Based on the Transmission Properties of Narrow Constrictions. IEEE Elec. Dev. Let., 2010. 31(6): p. 609-611.
[3] Ambrogio, S., et al., Statistical Fluctuations in HfO2 Resistive-Switching Memory: Part I - Set/Reset Variability. IEEE Trans. on Elec. Dev., 2014. 61(8): p. 2912-2919.
[4] Lanza, M., et al., Grain boundaries as preferential sites for resistive switching in the HfO2 resistive random access memory structures. APL, 2012. 100(12).

Nanostructured Au films resulting from supersonic cluster beam deposition of bare gold clusters show a non-ohmic electrical behavior and complex and reproducible resistive switching with characteristics suitable for neuromorphic applications [1, 2]. In order to achieve a deeper understanding of the modification induced at the nanoscale by joule heating in a gold cluster-assembled film due to the flow of an electrical current, we performed an in situ HRTEM characterization of the dewetting of the nanostructured films under thermal annealing.
Gold cluster-assembled films were fabricated by supersonic beam deposition on silicon nitride membranes, with nominal thickness of 1, 3, 12 and 18 nm respectively. Their initial morphology depends on the deposited amount of gold: while the samples 1 and 3 nm-thick are constituted by isolated and branched gold nanosized polycrystals, those with the higher thickness shows to be quite more branched, and percolating. The TEM/HRTEM/STEM measurements showed that the samples’ in situ heating leads to the dewetting of the gold polycrystals, and to the depercolation of those initially percolated, via the retraction of all the branches constituting the Au crystals. Such an evolution, common to all samples, was studied in terms of collective behavior, by low-magnification TEM imaging, HAADF-STEM and electron diffraction (ED) pattern collection of large ensembles of gold polycrystals. In particular, the global evolution under thermal in situ stimulus of the gold polycrystals not giving rise to a percolating network was quantified investigating the change of typical features such as their sizes, the latter expressed by their maximum (length) and minimum (width) mean Feret diameter, and the area fraction of the substrate covered by them. With regards to the Feret length and width, as well as to the covered area fraction, both not percolating samples showed that, as a consequence of the branches retraction, these parameters decreased with the thermal increase, bringing to smaller, still polycrystalline but spheroidal nanoparticles, showing a log-normal distribution of their size and with the mean one increasing with the nominal thickness. The same behavior was observed even for the percolating structure, but due to the fact that they were initially connected among them, the determination of the Feret parameters was achievable only after their dewetting. As well, the concomitant thickness increase of the gold nanocrystals during their branches’ retraction was qualitatively observed by HAADF-STEM imaging. The ED diffraction patterns indicated that: i) the as-deposited gold nanostructures did not grow following any preferential crystalline direction, and that ii) while the dewetting gave rise to the size increase of the central parts of the originally branched gold nanostructures, the crystal structure of those zones did not change with the heating and then again no any preferential crystal orientation was found after their thermally-driven evolution’s completion. Finally, the spherical aberration(C_S)-corrected HRTEM allowed determining the atomically-resolved in situ thermal evolution of single gold nanocrystals in the direct space, showing that the branches retraction is accompanied by the most peripheral crystal grains’ boundary movement, bringing to the formation of final, thicker, separate and more regular gold nanocrystals, in agreement with what observed on large ensemble of nanostructures.

[1] C. Minnai, A. Bellacicca, S. A. Brown, P. Milani, Sci. Rep. 2017, 7, 7955
[2] C. Minnai, M. Mirigliano, S. A. Brown, P. Milani Nano Futures 2018, 2, 011002

Direct observation of growing Au-catalyzed III-V nanowires via in-situ transmission electron microscopy can give crucial insights into the theoretical understanding of the growth process, through monitoring the dynamic behaviour of the catalyzing seed particle as well as the resulting nanowire it gives rise to. [1,2] However, the theoretical framework of nanowire growth is to a large extent built on physical quantities which are not directly measureable in a microscope (e.g. the supersaturation of the growth species in the seed particle), often lack the dynamic aspect of the growth process, and rely on material properties and interface energies which are not always well known. This can make it difficult to correlate observations from in-situ experiments with nanowire growth theory.

Here, a stochastic simulation model of III-V nanowire growth is presented, which captures the dynamic processes of nanowire growth in a theoretical framework.[3] The model is based on the material flow of the growth species through the seed particle to follow the composition of the seed over time. Knowing the composition in turn allows for the calculation of nucleation rates using classical nucleation theory. Then, the moment when nucleation occurs, as well as the type (meaning crystal structure/composition), is decided using these nucleation rates and random numbers. Nucleating a new layer in turn affects the composition of the seed which allows the model to capture the dynamics of nanowire growth.

The model is designed to have the same input parameters as those used for experimental growth, such as the flow of growth species and temperature of the system. The output of the simulation include the temporal evolution of the physical quantities used in theoretical works, but also the size and composition of the seed, the crystal structure and composition of the growing nanowire and other properties which can be directly compared with observations from in-situ microscopy. This makes stochastic simulations and in-situ microscopy mutually beneficial. The simulations can provide explanations to experimentally observed trends due to the output of additional physical quantities, while the observations provide real time correlated measurements of the growth dynamics which can be used to improve the model.

An example of this synergistic nature will be presented, where experimental in-situ growth of GaAs nanowires using metalorganic precursors in an environmental transmission electron microscope was combined with simulations.[4] Experiments showed that the two distinct steps in nanowire growth, meaning the incubation period and the step flow growth, occurred on similar time scales and were found to be controlled by different parameters for most of the tested parameter space. Because of this, a step flow process was added to the model, and analyzing the simulations gave insights to how these processes can be limited by one or both of the growth species in different regimes. The experimental values for the step flow times in turn allowed us to tune and improve upon some of the parameters used in the simulations which are not well known, such as the interface energy between the liquid seed particle and the GaAs nanowire.

The flexibility of the simulation framework allows for many uses beyond analyzing step flow times in GaAs nanowire growth. As an example, the input variable corresponding to the flow of growth species can be changed throughout the course of the simulation. This gives the model the freedom to theoretically evaluate growth of heterostructures with different crystal structures for binary nanowires or different compositions for ternary nanowires.

1 - D. Jacobsson et al., Nature, 531, 317–322
2 - J. -C. Harmand et al., Phys. Rev. Lett. 121, 166101
3 - E. K. Mårtensson et al., Nano Lett., 19, 2, 1197-1203
4 - C. B. Maliakkal et al., arXiv:1905.08225

One-dimensional platinum-based nanowires (NWs) have been demonstrated to possess improved activity and stability towards oxygen reduction reaction (ORR) for proton-exchange membrane fuel cells. Sufficient understanding on the growth mechanism at atomic level and seeking approaches for facile synthesis of platinum-based NWs are emerging. Here we report a novel scalable synthesis of surfactant-free thin platinum-based nanowires network electrocatalysts based on a generic solid state method. The growth stages and formation mechanism are investigated by in situ transmission electron microscopy (TEM). It reveals that the enhanced surface diffusion of Pt atoms due to the H₂ adsorbate, especially on {100} planes, can establish a liquid-like behavior which facilitates the diffusion. Surface coalescence then prefers to occur on this specific plane, until the exposed Pt (100) surface is attached by other nanoparticles. In situ atomic observation demonstrated that four stages are included among this solid state orientated attachment: 1) H₂ adsorption, preferentially on Pt {100} facets, 2) higher local diffusion on Pt {100}, 3) attachment and formation of necking area between Pt nanoparticles, and 4) coalescence to form a short NW. The resulting catalysts demonstrated enhanced ORR performance: the as-prepared platinum-based NWs delivered the mass activity of 0.71 A/mg_Pt for Pt NWs and 1.52 A/mg_Pt for Pt-Ni alloy NWs, and the specific activity of 2.20 mA/cm²_Pt for Pt NWs and 3.58 mA/cm²_Pt for Pt-Ni alloy NWs, showing promising improvement over the commercial Pt/C catalyst. This newly developed solid state H₂-assisted method with the atomic understanding from in situ TEM observation provides an applicable path for scalable production of surfactant-free Pt-based NWs with excellent activity towards ORR. It is prone to better understand the surface adsorbate-promoted growth mechanism in gas at the atomic scale thus presents the possibility to synthesize flexible shape-controlled nanostructures in gas, similarly as in liquid phase.
[1]. Ma, Y. L.; Gao, W. P.; Shan, H.; Chen, W. L.; Shang, W.; Tao, P.; Song, C. Y.; Addiego, C.; Deng, T.; Pan, X. Q.; Wu, J. B. Adv. Mater. 2017, 29, 1703460.

A novel process, the formation and growth of metallic Ag nanowires (NWs) and nanoparticles (NPs) on a-Ag₂WO₄ upon electron beam irradiation has been discovered and was extensively investigated by different experimental techniques including transmission electron microscopy (TEM) [1-3].A combination of long time atomistic molecular dynamics (MD) simulations [4] and in situ transmission electron microscopy (TEM), was used to probe the coalescence of Ag nanoparticles (NPs). In the experiments we were able to observe and analyze rarely seen in situ Ag NPs growth produced by irradiation of an α–Ag₂WO₄ crystal by electrons from a TEM electron beam in vacuum, which were observed in real-time. This phenomenon can be considered as a clear example of surface plasmon resonances (SPR), in which the electron beam creates Ag NPs that become SPR electric dipoles. Nearby Ag NPs undergo coalescence processes that are controlled by dipole-dipole interactions, forming larger clusters [4]. The welding and sintering of Ag NPs analyzed with computer simulations, provided meaningful insights at atomistic level and show the role of the crystallinity and shape of the Ag NPs, lattice plane matching, and also the effect of stacking faults during these processes helping the understanding of the experiments. They uncovered an interface evolution process displaying a disorder to order transition [5] which was also verified experimentally, supporting the conclusions of this study.
Work is supported by CNPq (304073/2015-6), CAPES and FAPESP (2013/07296-2, 2016/23891-6). SDUMONT Computer at LNCC, CENAPAD-SP and IFGW are acknowledged for computer time.

[1] E. Longo, L. S. Cavalcante, D. P. Volanti, A. F. Gouveia,V. M. Longo, J. A. Varela, M. O. Orlandi and J. Andres, Sci.Rep., 3, 1676(2013).
[2] J. Andrés, L. Gracia, P. Gonzalez-Navarrete, V. M. Longo,W. Avansi Jr., D. P. Volanti, M.M. Ferrer, P. S. Lemos, F. A. L.Porta, A. C. Hernandes and E. Longo, Sci. Rep.5, 5391(2014).
[3] W. da Silva Pereira, J. Andrés, L. Gracia, M. A. San-Miguel, E. Z. da Silva, E. Longo and V. M. Longo, Phys. Chem. Chem .Phys.17, 5352 (2015).
[4] Giovani M Faccin, Miguel A San-Miguel, Juan Andrés, Elson Longo e Edison Zacarias da Silva.J. Phys. Chem. C, 121 (12), pp 7030–7036 (2017)
[5] E. Z. da Silva, G. M. Faccin, T. R. Machado, M. de Assis, Nadia G. Macedo, J. C. Sczancoski, S. Maya-Johnson, Juan Andrés, Elson Longo, and M. A. San-Miguel. J. Phys. Chem. C, 123(17), 10703-11328(2019)

We investigate several mechanisms involved in GaAs nanowire formation, using an aberration-corrected transmission electron microscope (TEM) equipped with beam sources of Ga and As₄ molecules. The molecular beams are supplied to a heated SiC membrane and the GaAs nanowires form by vapor-liquid-solid growth, with or without Au catalyst particles. Growth is observed in real-time with the atomic resolution.
We adjust the As₄/Ga beam flux ratio to control the geometry of the catalyst droplet. The contact angle of the droplet on the nanowire top facet has a strong influence on the stacking of the GaAs atomic layers: wurtzite (WZ) or zinc-blende (ZB) crystal structures can be obtained [1]. Both structures form layer by layer, but specific features are associated with the growth of each of these crystalline phases. WZ phase stacking systematically corresponds to nucleation events occurring at the periphery of the interface between the droplet and the nanowire. In such a case, this interface is perfectly flat and the step flow is relatively slow [2]. ZB stacking is obtained mainly at large droplet contact angles, the top facet being truncated and the step flow much faster (instantaneous at a camera acquisition rate of 10 frames per s). In this situation, the truncation volume varies periodically, at the frequency of monolayer formation as already reported in other systems [3].
We show that the previous features are not intrinsically related to the crystal structure which grows. Indeed, at contact angle below 90°, ZB monolayers can form with a slow step flow and no facet truncation at the liquid/solid interface. Moreover, we use an artifice to produce a truncation of the top facet while the droplet contact angle leads to WZ formation. In this situation, the WZ stacking is maintained but the monolayer step flow becomes fast. This shows that the truncation of the top facet acts as a reservoir of atomic species which can accelerate the completion of a monolayer, whatever the crystal structure being formed.
Finally, we will show that a fine control of the beam sources allows us to switch abruptly from one crystalline phase to another and thereby, to fabricate crystal phase superlattices.
We acknowledge the French National Research Agency (ANR) for funding the TEMPOS/NanoMAX project within the EQUIPEX "Investissements d'Avenir" programme.
[1] D. Jacobsson, F. Panciera, J. Tersoff, M. C. Reuter, S. Lehmann, S. Hofmann, K. A. Dick, and F.M. Ross, Nature 531, 317 (2016).
[2] J.-C. Harmand, G. Patriarche, F. Glas, F. Panciera, I. Florea, J.-L. Maurice, L. Travers, and Y. Ollivier, Phys. Rev. Lett. 121, 166101 (2018).
[3] A. D. Gamalski, C. Ducati, and S. Hofmann, J. Phys. Chem. C 115, 4413 (2011)

Crystal growth of III-V semiconductor nanowires assisted by a liquid droplet requires, apart from successful nucleation, that the particle continuously wet the top of the nanowire. Recent TEM observations have shown that with increasing droplet volume, the interface between the nanowire and the liquid may become truncated [1,2], and eventually the droplet may be displaced from the top facet [3]. While this has been observed and discussed from the nucleation point of view, it has not been addressed from the droplet dynamics perspective.

Here we compare in-situ transmission electron microscopy observations of the droplet-nanowire interface during growth of Au-assisted GaAs nanowires, with calculations of the stability of the droplet on the nanowire top facet by comparing the capillary forces. By expanding the Young’s equation for droplet on a surface to include a truncated facet, we find limits for the solid-vapor and liquid-vapor surface energy ratio with respect to the contact angle of the droplet. These limits confine an interval of surface energy rations between which the droplet remain on top of the nanowire. In addition, it is predicted that formation of a truncation of the solid-liquid interface increases the stability for having a droplet on the top of a nanowire.

From the experimental observations, based on real-time imaging of the crystal growth at ~20 frames/s, we observe the dynamic droplet-nanowire interface as the droplet is filled with Ga which results in a truncated interface from time to time. Along with these observations, we estimate the surface tension and contact angle of the droplet based on the projected droplet volume. These experimental data and observations are then combined with our theoretical calculations in order to experimentally estimate the surface energy of the GaAs wurtzite {10-10} solid-vapor interface. We get an experimental prediction of the surface energy for the crystal growth conditions of GaAs (pressures and temperatures) in comparison to the existing theoretical predictions made for ideal environments and reconstructions [4, 5]. The comparison can further be used to hypothesize if, or how, the surface of interest has been reconstructed.

To conclude, using high frame-rate in-situ TEM imaging in conjunction with theoretical estimations we will present a reasoning for the formation of a truncation at the droplet-nanowire interface. In addition, we will show the observed droplet dynamics and how they are used to experimentally estimate the surface energy of the wurtzite GaAs nanowire sidewall.

References
[1] Wen, et al, Phys. Rev. Lett. 107, 025503 (2011)
[2] Jacobsson, et al, Nature 531, 7594, 317 (2016)
[3] Tornberg et al, J. Phys. Chem. C 121, 21678 (2017)
[4] Sibirev et al, Phys. Solid State 52, 7, 1531 (2010)
[5] Leitsmann et al, J. Appl. Phys. 102, 1 (2007)

Complex metal oxides with a variety of crystalline nanostructures have been the subjects of extensive research thank to their peculiar properties, such as multicolor emission and multiferroics, as well as potential applications in photonics and renewable energy. One of the main challenges is to have an effective protocol to rationally design the materials, especially one-dimensional (1-D) binary and ternary oxide heterostructures with new functionalities. The challenge begs the understanding of the material growth mechanism at the atomic level. In-situ transmission electron microscopy (TEM) is a unique tool to investigate the dynamics at the growth front in the real time with high spatial resolution.
In my talk, I will present our ex- and in-situ TEM studies of zinc oxide (ZnO) nanowire synthesis using germanium (Ge) catalyst via vapor-solid-liquid (VLS) mechanism as a model system. The Ge catalyst particle has an unusual geometry and a large volume compared to the growing nanowires. This is very different from the comparable dimensions of catalysts and nanowires in the conventional VLS synthesis, for instance gold-group IV (silicon) and gold- group III-V (GaAs) nanowire systems. This may result in different interface dynamics and, consequently, growth kinetics in the Ge-ZnO system. More interestingly, under our ex-situ synthesis parameters by thermal evaporation, ZnO grows into twisted nanowires and tubular nanotubes, in addition to the well-reported straight nanowires and nanoribbons. The observation implies the coexistence of both VLS (layer-by-layer growth at the catalyst-nanowire interface) and screw-dislocation driven (self-perpetuating atom assembly around the screw dislocation core) pathways in the Ge-ZnO system. The understanding obtained from this study will serve as a guideline to extend the knowledge of 1-D material synthesis and expand the library of complex oxides realized by VLS.

In situ X-ray absorption near-edge structure (XANES) measurement is a powerful tool to investigate local chemical bonds of nanomaterials. However, only a few studies have been reported to clarify the growth process of carbon nanotubes (CNTs) by XANES. Here, we demonstrate in situ XANES study for CNT growth process by thermal decomposition of SiC(000-1) face [1] and that for single-walled carbon nanotube (SWCNT) growth by alcohol catalytic chemical vapor deposition (ACCVD) using Co and Ni catalysts. C K-edge, and Co and Ni L-edge XANES measurements were performed at BL-7A in KEK PF and BL-5S1 at Aichi SR, respectively.
In CNT growth by thermal decomposition, CNTs are formed on SiC(000-1) surface by heating above 1200°C in a vacuum, due to the desorption of Si atoms from SiC surface. C K-edge XANES results showed that the orientation of C-C bond changed with the increase of the sample temperature. When the temperature was 1100°C, C-C bonds were parallel to the surface. However, they were aligned to perpendicular to the surface above 1300°C. Taking into account the results of first-principles density functional theory/molecular dynamics (DFT/MD) simulation [2], at the beginning, the graphene flakes were formed at ~1200°C, which were parallel to the SiC surface. As the temperature rose, they were curved by the formation of bonds with the dangling-bonds on SiC(000-1), and hemispherical structures (“carbon nanocaps”) were formed. When the temperature was above 1300°C, CNT growth was proceeded perpendicular to the surface [3]. Our results showed that the size and structure of graphene flakes formed at the initial stage would determine the structure (diameter, chirality and the number of walls) of CNTs.
Next, we performed Co and Ni XANES measurements to clarify the chemical states of these metal catalyst particles during SWCNT growth by ACCVD. Co and Ni L-edge XANES spectra showed that “as deposited” Co and Ni catalysts were both partially oxidized and they were reduced to metal particles during heating. However, after SWCNT growth started, Co catalysts were transformed to carbides, but Ni catalyst remained metallic. This indicates that, during SWCNT growth, carbon atoms dissolve into the Co catalyst particles, while carbon atoms diffused only on the surface of Ni catalyst particles. Taking into account the similarity in the bulk phase diagrams of C-Co and C-Ni, the difference in the chemical states of Co and Ni catalyst particles is due to the specific property of nano-sized particles, which is consistent with the DFT simulation [4].
This work was partially supported by the Meijo University Research Branding Project for Cultivation and Invention of New Nanomaterials under the Ministry of Education, Culture, Sports, Science and Technology (MEXT) Private University Research Branding Project (Global Development Category). The NEXAFS measurements were performed under the approval of Photon Factory Program Advisory Committee (Proposal No. 2011G539).
[1] M. Kusunoki et al. Appl. Phys. Lett. 77, 531 (2000).
[2] N. Ogasawara et al. Chem. Phys. Lett. 595, 266 (2014).
[3] T. Maruyama et al. J. Phys. Chem. C 119, 26698 (2015).
[4] Y. Okamoto et al. Theor. Nanosci. 8, 1755 (2011).

Limited understanding of the factors influencing the yield of CNTs relative to the number of catalyst particles remains an important barrier to their large-scale production with high quality and to tailoring CNT properties for applications. This lack of understanding is evident in the frequent use of Edisonian approaches for achieving high-yield CNT growth, and the sometimes confusing influence of trace residues on the reactor walls. In order to grow CNTs reproducibly and enable large-scale growth, it is imperative to understand, fundamentally, how these common practices impact catalytic activity and thus nucleation success rate. Herein, we use ambient pressure-X-ray photoelectron spectroscopy (AP-XPS) and in situ Raman spectroscopy to study the influence of carbon and hydrogen on the coupling between catalyst reduction and CNT nucleation, from an iron catalyst film. The XPS results reveal a positive correlation between the degree of catalyst reduction and the density and quality of the CNTs, verifying that effective catalyst reduction is critical to CNT nucleation, and to the resulting CNT growth yield. We also show that hydrocarbon by-products from consecutive growths can facilitate catalyst reduction and increase nucleation success rates significantly. The in situ Raman studies reveal that the addition of carbon decreases the incubation time, thereby enabling a higher CNT nucleation density. Our findings suggest that common practices used in the field, such as reactor preconditioning, aid in the reduction of the catalyst population, thus improving nucleation success rates and enabling growth of dense forests.

Tuning the optoelectronic responses of metal-oxide semiconductors in a stable, but reversible manner is identified as the limiting factor for their deployment in real-world flexible electronic devices. Currently, external stimuli (e.g. pressure, temperature) are applied on these inorganic semiconductors to modify their optoelectronic response, ultimately reverting to their initial state upon relaxation. Therefore, this work demonstrates the effect of polarized molecules for reversibly manipulating the optical, electrical, and magnetic transport properties of transition metal (TM) doped metal-oxides and is validated computationally. As a proof-of-concept, the electronic properties of TM dopants and their coupling with rare earths (RE) are probed using time-resolved optical spectroscopy. While these materials have applications in bio-sensors and flexible displays, the results of which have potential for quantum computing.
Functionalized Ni doped TiO₂ thin films were characterized to identify effect of surface dipoles (para-substituted benzoic acid ligands) on their optoelectronic and structural properties. The ligand-induced local structural and electronic changes around Ni²⁺ were investigated via neutron scattering (NS) and soft x-ray absorption spectroscopy (XAS). Upon functionalization, a distinct shift in the t_2g:e_g filling and energy levels were observed in the Ni L₂ edge XAS spectra and was attributed to a local geometry distortion. Our real time (RT) time-dependent density functional theory (TDDFT) calculations further confirm that the applied ligand field changes the 3d metal orbitals hybridization, resulting in the observed variation in the Ni L₂ edge XAS. Additionally, in situ ultraviolet photoelectron spectroscopy (UPS) measurements on functionalized TiO₂:Ni films reveal that the direction of charge transfer and band bending is reversible and proportional to the dipole moment. Next, TiO₂:Ni²⁺ shell layer with controlled thickness, is coated on the core β-NaYF₄:Er³⁺ nanoparticles to facilitate Ni²⁺-Er³⁺ energy transfer for dynamic upconversion luminescence via adaptive Ni²⁺ absorption. UV-Vis and in situ photoluminescence (PL) measurements on these surface-modified RE core-TM shell phosphors indicate a clear shift in the optical absorption (~ 50 nm) and emission spectra, demonstrating the ligand-dependent luminescence of core-shell phosphor. This direct design method of combining experiments and theory will provide a robust approach for designing weak field dependent optoelectronic materials. More importantly, this ability to reversibility manipulate interfacial charge transfer processes of doped semiconductors has direct implications in the development of quantum computing type materials and memory-storage devices.

DNA origami strategy has greatly promoted the development of DNA nanotechnology as it can construct delicate 2D/3D nanostructures. The current method to prepare DNA origami is through a woven approach to fix a long strand DNA into certain shapes. This article describes a novel strategy to fold the same 2D DNA sheet into multiple complex structures driven by introducing hydrophobic interaction. The pathway of the folding process can be adjusted by tuning the distribution and chronic order of cholesterol, resulting in different complicated structures. This method has been proved efficient for constructing variable nanostructures based on the same DNA origami sheet, and termed as origami+. This work represents a new strategy in DNA nanotechnology and will improve the ability to manipulate objects at nano-scale.

Reference:
[1] Y. Zhang†, C. Wang†, Y. Dong*, D. Wang, T. Cao, S. Wang and D. Liu *, Adv. Funct. Mater. 2019, 1809097

As an alternative computing architecture, neuromorphic computing has drawn enormous attention worldwide. Various analog switching devices have been demonstrated as synapses for neuromorphic computing. A particularly promising design involves an analog neural network accelerator based on a cross-point array of non-volatile memory elements. However, these devices often show non-ideal switching characteristics that will limit the accuracy in neural network training. Improved control of analog switching can be achieved by replacing each 2-terminal memory element with a 3-terminal memristor structure that has a channel with a “read” terminal on each end, plus a control electrode that is analogous to the gate in a transistor. We are particularly interested in memristor structures that do not use lithium as the diffusing species, due to its sensitivity to air and humidity. We have therefore developed a fast-switching memory element based on the use of a copper superionic conductor, Cu₁₆Rb₄Cl₁₃I₇. This unusual compound has the highest reported room temperature conductivity of any solid electrolyte, 0.34 S/cm. Ex situ measurements suggest that a voltage at the control electrode causes Cu⁺ to migrate through the solid electrolyte and deposit onto the channel, changing its resistivity, as measured using the two read terminals. However, ex situ measurements cannot directly address failure modes and the physical factors limiting reliability, even when structural and electrical characteristics are correlated before and after operation of the device. We therefore describe the use of transmission electron microscopy, with structural and analytical (EDS/EELS) capabilities, to investigate the operation mechanism and limitations of copper-based 3-terminal synaptic cells. We have fabricated electron-transparent thin film samples on commercial TEM window chips with a copper control electrode and various materials for the channel. The chips are accommodated in a Hummingbird TEM holder with biasing capability. We will describe the characteristics of these devices and correlation of their electrical performance with microstructural evolution. We will also compare the device operation with liquid cell TEM observations of the cyclic pulse deposition of copper from a liquid electrolyte, which has been proposed as an alternative synaptic cell design. For both liquid and solid electrolytes, we believe that in situ TEM techniques have a vital role to play in the material and architecture design of memristor-based 3-terminal devices.

Corrosion is the interaction between service materials and environments that involves chemical or electrochemical reactions at interfaces. The corrosion-induced deteriorate of proper material functions is the most common cause of the failure during service. Alloyed nanoparticles (NPs), an emerging material compound involving several elements in the form of disordered solid solutions, is attracting increased attentions since their novel properties in applications of catalysis and superior structural alloys. The corresponding service environment for such materials usually involves super critical conditions such as exposure to chloride-containing solutions, acidic, high-temperature, high oxidative gases, which requires in depth understanding of corrosion behaviors at both structural and chemical level. In contrast to alloys in bulk phase, the corrosion-resistant properties of nanoparticle phase is poorly studied due to the relatively fresh discovery of such materials system and the challenge of obtaining nanoscale information regarding to the typical few nanometer sizes of NPs. With the reduced size of alloys from bulk to NPs, it is crucial to utilize advanced characterization techniques like TEM to gain atomic resolution information during oxidation process. In situ TEM, with combined sub-nanometer resolution and real time observation capabilities, is essential for such study.

The oxidation behavior is specifically essential to be investigated for alloyed NPs with working conditions expose to atmospheric pressure gases at elevated temperature. To do so, an in situ gas-cell TEM holder is utilized to allow the alloyed NPs expose to various of gases environment. The distance between two Si microchips can be adjusted as small as 50 nm or as large as 2 μm, accommodating the major size range of synthesized NPs. Air, H₂, and water vapor gases are used to flow into the cell with a constant volumetric flow rate. A temperature controller is used to set the temperature from RT to 800 °C. The in situ gas-cell allows atomic resolution imaging at both room and elevated temperatures, and the structural evolution during oxidation is observed in real time. The oxidation states for various elements are evaluated by electron energy loss spectroscopy (EELS) at different temperatures. The observation is further investigated by density functional theory (DFT) calculation to gain fundamental understanding of oxidation behavior. The present study provides insights into how oxidation initiated at the surface of alloyed NPs, and help gain deeper understanding of designing anti-oxidation composition for next generation functional alloys.

Melanins are ubiquitous biopigments responsible of manifold functionalities in living organisms¹. Eumelanin is a hygroscopic, black-brown biomacromolecule belonging to the family of melanin biopigments . It is synthesized by the oxidative polymerization of (5,6)-dihydroxindole (DHI) and (5,6)-dihydroxindole 2-carboxyl acid (DHICA) building blocks. Different available polymerization sites, a wide variety of supramolecular arrangements and co-existing redox states, bring chemical and physical disorder to the biopigment². There is a significant interest in eumelanin physicochemical properties such as broadband optical absorption, free radical scavenging, metal-ion chelation and hydration dependent electrical response, this last mainly attributed to protonic transport^3,4. Owing to the complex structural organization, a universal theory correlating supramolecular aggregation to fundamental optical and electrical properties of eumelanin does not exist yet. Furthermore, the charge transport properties of the biopigment, object of numerous studies, are still elusive^5–9.
Our group is currently investigating the solid-state polymerization of spin coated DHI and DHICA thin films at surfaces. We observe that the monomers aggregate, possibly eventually polymerizing in a time scale of 48hr at ambient condition. Images obtained by Atomic Force Microscopy (AFM) show that after a few days at ambient condition DHI thin films grow into well defined fern-like structures on certain areas of the surface. This process, attributable to a diffusion limited aggregation (DLA), is followed by the growth of crystal-like agglomerations about 400 nm-high, on the fern structures.
On the other hand, DHICA thin films do not form crystal-like agglomerations. We observe, instead, that regularly shaped nano-pillar structures, uniformly coating the substrate surface, grow in about 1hr at ambient condition. The pillars gradually undergo an in-plane and out-of-plane growth, reaching average width and height of 100 nm and 300 nm. UV-Vis spectra of DHI acquired over time, show an absorption peak at 305 nm that gradually shifts to 322 nm after one month at ambient condition. We tentatively interpret this red shift as an evidence of solid-state polymerization. Interestingly, both in DHI and DHICA spectra, we observe an increase of the absorbance both in the UV and Visible regions, this latter likely due to an improved -electron delocalization of oligomeric structures over time.
Being non-toxic, abundant, solution processable and quite biodegradable, eumelanin is a promising candidate for green electronics¹⁰: as such our results are of paramount importance to exploit the full technological potential of the biopigment.

References
1. d’Ischia, M. et al. Pigment Cell and Melanoma Research, 28(5), 520-544 (2015).
2. Panzella, L. et al. Angew. Chemie - Int. Ed. 52, 12684–12687 (2013).
3. Micillo, R. et al. Sci. Rep., 7 (41532) (2017).
4. Di Mauro, E., Xu, R., Soliveri, G. & Santato, C. MRS Commun., 7, 141-151 (2017).
5. Sheliakina, M., Mostert, A. B. & Meredith, P. Adv. Funct. Mater., 1805514 (2018).
6. Mostert, A. B. et al. Proc. Natl. Acad. Sci., 109(23), 8543-8547 (2012).
7. McGinness, J., Corry, P. & Proctor, P. Science (80). (1974).
8. Wünsche, J. et al. Chem. Mater. 27, 436–442 (2015).
9. Migliaccio, L. et al. Front. Chem. 7, 1–8 (2019).
10. Irimia-Vladu, M. Chem. Soc. Rev. 43, 588–610 (2014).

Crystallization, a process of arranging atoms or molecules into a highly organized structure, is widely used to modify physical properties of materials in macroscale. Although various crystallization methods have already been developed, crystalline control with nanoscale precision remains a challenge. Here, we report a scheme that uses electron irradiation to induce nanoscale localized crystallization in amorphous aluminum oxide (Al₂O₃). The crystallization mechanism is based on the electron irradiation which induces thermal heating in the absorbing material, supplying energy to rearrange the atoms into an organized crystal form. As for the capability of precise crystalline control, it is originated from the real-time imaging function of scanning electron microscopy (SEM) and transmission electron microscopy (TEM), which allows localized crystallization (nanoscale phase change) with sub-10 nm precision. In addition, the crystallization leads to volume change, which generates stress in the thin film and triggers self-assembly of 3D nanostructures. Various complex three-dimensional (3D) functional nanostructures can be realized based on this scheme.

The most common method for controlling structure through processing is through simple thermal excursions, as temperature causes atomic diffusion, and thus allows structural re-ordering. One of the primary experimental method used to determine the local internal structure of nanomaterials is that of transmission electron microscopy (TEM). Therefore, the most common dynamic in situ microscopy experiments revolve around the relatively simple act of heating nanomaterial samples.

Thin-film technology has allowed much more local heating of the specimen in TEM heating holders if one is studying nanomaterials, enabling more stable imaging during in-situ TEM heating experiments. More recently, direct temperature measurements on these thin-film heating systems have made temperature measurements at the sample become closer to becoming reality. Although double-tilt TEM heating holders have been available, they suffer from some of the same limitations as standard double-tilt TEM holders do. Specifically, backlash in the tilting mechanism and lack of repeatability of the tilt makes it difficult to get the sample in exactly the right orientation and to know the exact angle of rotation. In this work we will use an optimized double-tilt thin-film heating platform that minimizes mechanical artefacts in the tilting mechanism and provides very stable image performance when heated to temperatures up to 1000°C, where it can be run more than 150 hours. Outside of standard temperature characterization of these thin-film heating TEM systems using different high resolution optical thermal imaging techniques, we also used melting standards to confirm accurate temperature response of the on-chip temperature sensor to indicate the temperature at the sample. We can identify very local and discrete melting events of samples with known melting points and correlate this to the sensor response, where local melting happens at very discrete moments (typically within a few frames), and surface diffusion occurs in matters of seconds following that.

In addition to this, we used this new in-situ TEM sample heating system to characterize model nanomaterial systems. Specifically, we studied the behavior of gold nanoparticles and nanorods on SiN_x substrates as function of temperature. Starting with characterizing of particle motion and stability under the beam and then the phase transformation and (rapid) shape change at the melting point. We will present rapid frame-to-frame phase transformation process of gold nanorods as we ramp the sample temperature to 1000°C and show that we can use the system to track phenomena in nanostructures over the full temperature range of the TEM heating system with good image stability. These results illustrate the optimized overall mechanical stability and temperature performance of this newly developed in-situ TEM double-tilt heating platform.

In micromachining and nanoelectronic field, etching is an important method to selectively construct two-dimensional or three-dimensional nanostructure semiconductors by introducing etchants or sacrificing the surface with powerful force [1]. ZnO nanostructure has been one of the most popular nanomaterials being studied in the past decades due to its widespread potential applications in photovoltaics, catalysis, flexible electronics, and solid state lighting, etc. [2,3]. To improve the ZnO nanostructure based device performance, some research groups used acidic or basic wet etching to help to expose the high surface energy crystal planes such as the semipolar {10-11} family surfaces [4]. However, it is hard to control the surface exposure regularity of the etched ZnO nanostructures. In this work, in situ liquid cell transmission electron microscope (LCTEM) is employed to help resolve the crystal shape change in the acid etching process and reveal the morphology transformation process. To mitigate and clarify the well-known electron beam effect in LCTEM [5], optical microscopy (OM) and scanning electron microscope (SEM) has been used in combination of the LCTEM method. By using such a combination method, the etching process of ZnO nanostructures can be robustly and confidently analyzed from atomic to nanometer scales and the etching mechanism can be simulated. Controlling pH=1~5 and temperature from 20 ^oC to 60 ^oC, the nanostructure will be changed from pencil shape to crown shape, as well as the etching rate will be changed. And both {10-11} and {10-1-1} can be preferentially exposed. This etching study provides some insight into controllably surface etching for nanoscale crystal surfaces.


[1] P. Pal, K. Sato, Micro and Nano Systems Lett. 3, 6 (2015)
[2] C.B. Ong, L.Y. Ng, A.W. Mohammad, Renew. Sust. Energ. Rev. 81, 536 (2018)
[3] R. Kumar, O. Al-Dossary, G. Kumar, A. Umar, Nano-Micro Lett. 7, 97 (2015)
[4] A. Ostapenko, T. Klöffel, J. Eußner, K. Harms, S. Dehnen, B. Meyer, G. Witte, ACS Appl. Mater. Interfaces 8, 13472 (2016)
[5] N.M. Schneider, M.M. Norton, B.J. Mendel, J.M. Grogan, F.M. Ross, H.H. Bau, J. Phys. Chem. C 118, 359 (2014).

The adoption of in situ and operando techniques for imaging dynamic processes within the transmission electron microscope (TEM) has opened new avenues of research in both the fields of materials and biology [1]. What was once considered an unorthodox methodology, requiring labs to develop their own home-built solutions, has achieved mainstream acceptance within the microscopy community. Multiple companies now provide commercial solutions for microchip based closed cell sample holders that can safely encapsulate liquids between electron transparent membranes and deliver capabilities far beyond the basic systems introduced a decade ago [2]. Not only can researchers control the liquid environment, but they can simultaneously deliver external stimuli directly to the sample with specially fabricated microchips, enabling benchtop experiments to be replicated at the nanoscale inside the TEM [3,4]. Optimized holder designs facilitate fluid control and enable direct line of site to EDS detectors [5]. State-of-the art, low sensitivity cameras and new diffraction techniques permit enable the low electron doses necessary to expand in-situ liquid microscopy performance and achieve better resolution of fragile and low contrast materials.
LC-TEM has enabled researchers to obtain new insights into the mechanistic pathways of nanomaterial growth and degradation as well as biological mechanisms such as RNA transcription and real-time protein aggregation [6,7]. Here we review developments over the last decade of in-situ liquid cell electron microscopy as a standalone technique, as well as the challenges that must be addressed in order to adapt experimental parameters to the unique environment of liquid based samples.
References:

[1] Demmert et al., in (2016) ‘Visualizing Macromolecules in Liquid at the Nanoscale’, in Ross, F.M. (ed.) Liquid Cell Electron Microscopy. Cambridge: Cambridge University Press, pp. 334–355.
[2] de Jonge et al, PNAS, 106 (2009): 2159-2164
[3] Karakulina et al, Nano Letters (2018) 18 (10) pp 6286-6291 [4]
[4] Smith A1:T106 al, ACS Central Science (2017), 3(1) 58-65
[5] Zaluzec et al, Microscopy Microanalysis, 20.2 (2014): 323-329
[6] Varano et al., Chem. Commun., 2015, 51, 16176-16179.
[7] DiMemmo et al., Lab Chip, 17 (2017) p. 315

Unexpected high pressure phenomena have been reported in boron nitride, Na-Cl and water, which are not observed at ambient conditions, indicating pressure can determine the stability of crystal structure. [1-3] It is widely accepted that palladium and palladium hydride (Pd-H) have a face-centered cubic (FCC) crystal structure under environmental conditions, even at high temperature in the case of pure Pd. In this study, we firstly report hexagonal close-packed (HCP) Pd-H which is a new palladium hydride phase previously suggested by density functional theory (DFT) calculation in high pressure. [4]
We synthesized hexagonal Pd-H by irradiating high energy electron beam on aqueous solution of Pd in a transmission electron microscope (TEM). We adopted graphene liquid cell (GLC) to acquire high resolution images in liquid. [5] The electron beam induced nanoparticle growth in liquid was traced by FEI Titan TEM equipped with Gatan OneView CCD camera, which can acquire a single image within 0.04 seconds.
Crystal structure analysis of the Pd nanoparticles was performed by analyzing diffraction patterns generated by fast Fourier transform (FFT) of high resolution TEM images acquired during particle growth. Through the diffraction patterns obtained from various zone axis, it was revealed that synthesized Pd nanoparticles have HCP structure which is different from the structure of Pd and Pd-H at ambient conditions, FCC. From energy dispersive X-ray spectroscopy (EDS) analysis conducted with a TEM equipped with Bruker Super X-EDS system, no element except for Pd was detected, suggesting oxygen doesn’t influence on the formation of hexagonal structure. Given that EDS cannot detect elements lighter than boron and inter-planar distance of the particle is greater than that of an imaginary pure HCP Pd which has same interatomic distance with FCC Pd, created particle is assumed to be Pd-H.
Thermodynamic stability of HCP structure of Pd-H was verified with a computational study. A simulation show that Pd-H is likely to have hexagonal structure when pressure is applied and concentration of hydrogen in palladium is high. For PdH_0.5, HCP is more stable than FCC only when it’s under pressure higher than a few GPa, while HCP phase is more stable even at ambient pressure in the case of PdH₁. Considering the previous reports asserting that materials lying between two graphene layers experience high pressure up to a few GPa and that radiolysis of water produces hydrogen radicals, formation of hexagonal Pd-H is attributed to not only hydrogen radicals but high pressure applied by water trapped between graphene layers. [3,6] We tried to synthesize the HCP particles outside a TEM by irradiating electron on liter scale vessel under ambient atmosphere, but resulting products were mostly FCC particles, which might be due to the absence of high pressure.
Also, it was confirmed that the nanoparticle has hexagonal structure even at the beginning of the growth stage and inter-planar spacing of the particles does not change remarkably during in-situ observation, suggesting hydrogen radicals participate in both nucleation and growth reactions. It’s quite different from FCC Pd-H formation process where hydrogen is absorbed by an existing Pd matrix. In addition, we investigated thermal stability of hexagonal Pd-H by an in-situ TEM heating holder (Gatan heating holder model 652). It is known that hydrogen desorbs easily from FCC Pd-H at around 100°C even in hydrogen atmosphere. [7] Interestingly, hexagonal Pd-H maintained its crystal structure even when it’s annealed over 300°C under vacuum without noticeable lattice constant change.

References
1. Lijun Zhang et al., Nat. Rev. Mater. 2 (2017)
2. Weiwei Zhang et al., Science 342 (2013)
3. G. Algara-Siller et al., Nature 519 (2015)
4. Xiao Yang et al., Sci. Rep. 7 (2017)
5. Jong Min Yuk et al., Science 336 (2012)
6. Nicholas M. Schneider et al., J. Phys. Chem. C 118 (2014)
7. Yiming Niu et al., Angew. Chem. Int. Ed. 58 (2019)

FePt nanoparticles (NPs) with face-centered tetragonal (fct) structure are of great interest for applications in data storage and for electrocatalysis due to a combination of properties including high magneto-crystalline anisotropy, high Curie temperature, small superparamagnetic critical size and high chemical stability. Solution phase techniques provide an excellent strategy for preparing monodisperse FePt NPs with a narrow size distribution, good compositional control and well-defined shape. Traditional ex situ experiments using TEM and XRD analytic methods show that fct-FePt NPs pass through several intermediate stages during solution synthesis from Pt and Fe precursors. First, fcc-Pt and crystalline Fe₃O₄ (magnetite) phases appear, presumably formed by reduction of the Pt and Fe precursors, respectively, at lower temperature. Subsequently, the Fe₃O₄ phase disappears and the fcc-FePt phase appears. Finally, the fcc-FePt phase transforms into fct-FePt when the sample is heated to a higher temperature. Controlling these reactions is important for precise control of the resulting nanoparticles, but the mechanism of the formation of fcc-FePt from the fcc-Pt and Fe₃O₄ phases is difficult to characterize using only ex situ data. We have therefore used in situ liquid cell electron microscopy to explore the mechanism of the reaction between crystalline Fe₃O₄ NPs and Pt solution in an aqueous medium. We find that microstructural changes take place in crystalline Fe₃O₄ NPs upon flowing a Pt precursor solution. The size of the crystalline Fe₃O₄ NPs gradually decreases and they eventually disappear, followed by re-formation of small NPs. Under certain conditions we also observe the formation of unusual structures where the NPs develop hollow cores. We consider the effect of initial Fe₃O₄ NP size, Pt concentration and temperature on the morphological changes, and discuss a model in which the crystalline Fe₃O₄ NPs are reduced and alloy with Pt atoms to form fcc-FePt NPs. Understanding the intermediate reaction steps can contribute to a better control of the formation of fct-FePt NPs.

The interfacial structure that forms at the junction between a 2D layered material and a 3D crystal can dominate the physical phenomena exhibited by the combination of materials. In particular, epitaxial or low defect density 2D/3D interfaces minimize electrical contact resistance and allow efficient charge transfer, and are essential for achieving proximity-induced superconductivity for spintronics applications and for developing catalysts and field effect or optoelectronic devices. In situ TEM provides a unique tool for examining the structure, formation and evolution of 2D/3D interfaces by following along as 3D nanocrystals are grown on 2D substrates. The experiments require deposition capabilities as well as ultrahigh vacuum (UHV) to control the sample surface cleanliness and avoid extraneous reactions such as oxidation. We therefore use a custom TEM with base pressure in the 10^-10 Torr range and an attached series of vacuum chambers with heating and deposition capabilities. We first transfer 2D layers or heterostructures onto substrates that can be cleaned by direct-current heating in the vacuum system. Without breaking vacuum we then deposit metals or semiconductors using evaporation or reactive gas precursors. Imaging can be carried out at the same time as, or immediately after deposition, to follow the evolution of the interfacial structure. We will illustrate the capability of TEM to analyze nanocrystal nucleation and evolution, orientation relations, interface dislocations, local strain fields and defects in several metals and semiconductors on 2D materials. We will discuss how parameters such as substrate temperature and defect density affect the morphology of deposited metals, including reactive systems such as Nb on transition metal dichalcogenides. We will show a strategy for achieving epitaxial growth of materials that show poor epitaxy, such as Ag, Ge or Si, by using a template made of another, epitaxial material such as Au. We can also measure the catalytic properties of individual metal nanocrystals as we use them to grow semiconductors using chemical vapor deposition. We finally describe the use of sequential deposition to form heterostructures and the effects of patterning the 2D material. We conclude by discussing the importance of recent developments in microscopy instrumentation in analyzing 2D/3D materials growth to help design functional nanoscale materials.

Group V 2D materials are a new class of materials [1] which includes various allotropes of phosphorus, arsenic, antimony [2, 3] and bismuth. Amongst them, we find the α and β phases of antimonene, which display thickness dependent electronic and topological properties, including a single-layer semi-conducting phase and a quantum spin-hall state at multilayer thickness. Using Transmission Electron Microscopy (TEM) and real-time Low-Energy Electron Microscopy (LEEM) observations of the van der Waals growth of antimonene on graphene, we discuss the growth modes of α-antimonene and its transition to β-antimonene occurring at a critical thickness of ~5nm. Moreover, using synchrotron-based X-ray Photoemission Electron Microscopy (XPEEM), we discuss the environmental stability and bonding nature of both phases. These results are an important step towards developing large-scale epitaxial growth methods for group V 2D materials.

[1] D. Geng and Y. Y. Hui, Adv. Mater. 2018, 30(45), 1800865.
[2] M. Fortin-Deschênes, O. Waller, T. O. Mentes, A. Locatelli, S. Mukherjee, F. Genuzio, P. L. Levesque, A. Hebert, R. Martel, O. Moutanabbir., Nano Lett. 2017, 17(8), 4970–4975.
[3] M. Fortin-Deschênes, R. M. Jacobberger, C. A. Deslauriers, O. Waller, É. Bouthillier, M. S. Arnold, O. Moutanabbir., Advanced Materials. 2019, 31(21), 1900569.

Benefited from a large density of 2D layer edges exposed on the surface, vertically-aligned 2D MoS₂ layers have recently harvested excellent performances in the field of electrochemical catalysis and chemical sensing. With the increasing versatility for high-temperature demanding applications, it is vital to identify their thermal-driven structural and chemical stability as well as clarify its underlying principle. Despite various ex-situ and in-situ characterizations on horizontally-aligned 2D MoS₂ layers, the direct in-situ heating of vertically-aligned 2D MoS₂ layers and the real-time observation of their dynamics at the near-atomic scale have never been approached, leaving their thermal stability poorly understood. Moreover, the geometrical advantage of the surface-exposed vertically-aligned 2D MoS₂ layers is anticipated to unveil the structural dynamics of interlayer van der Waals gaps and its correlation with thermal energy, unattainable with 2D MoS₂ layers of any other structural forms. Herein, we report a comprehensive in-situ heating TEM study on cleanly transferred vertically- aligned 2D MoS₂ layers up to 1000^oC. Several previously unknown striking phenomena were observed in the course of heating; (1) formation and propagation of voids between the domains of vertical MoS₂ layers with distinct grain orientations starting at about 900^oC. (2) subsequent melting of 2D MoS₂ layers accompanying a formation of Mo nanoparticles at about 945^oC, much lower than the melting temperature of their bulk counterpart. (3) initiation of melting from the surface-exposed 2D layer vertical edge sites, congruently supported by density functional theory (DFT) prediction. These new findings obtained with vertically-aligned 2D MoS₂ layers will offer critical insights into better understand the thermodynamic principle that governs the structural stability of general van der Waals 2D crystals as well as provides technological guidance for materials design and optimization in their potential high-temperature applications.

It is now well understood that careful in situ transmission electron microscope (TEM) experiments are required in order to reveal, at high resolution, materials phenomena which are representative of the true natural behavior. One such study, of the crystallization of high-k dielectric tantalum oxide (Ta₂O₅) thin films on a silicon substrate, has shown how the individual mechanisms of nucleation and growth contribute to the overall transformation and processing behavior, and even allow in situ capacitance measurements on individual crystals. The determination of the reaction activation energy is an important component of this type of work ^[1]. This approach has been extended to describe the kinetics of oxygen ion.vacancy ordering in epitaxial thin films of ceria (CeO₂) grown on yttria stabilized zirconia single crystal substrates, as revealed by in situ aberration-corrected TEM^[2].

[1] Sinclair, R., MRS Bulletin, 38(12), pp. 1065-1071 (2013).
[2] Sinclair, R., Lee, S. C., Shi, Y. and Chueh, W. C., Ultramicroscopy, 176, pp. 200-211 (2017).

A better understanding of the formation mechanism of nanoparticles is needed to control growth and fine-tune their properties for applications. SAXS, WAXS and absorption measurements are commonly used to obtain insight in the formation mechanism of nanocrystalline systems. However, these measurements are typically performed in capillaries or by taking aliquots, therefore represent particle growth under conditions that are different than the actual reaction conditions. We developed a reactor to reproduce the conditions under which nanoparticles are synthesized by using specially designed glassware. In this reactor it is possible to do in-situ SAXS, WAXS and absorption measurements in oxygen and water free environment, at high temperatures and with a time resolution on the order of milliseconds. Furthermore, we were also able to add solid reactants or inject a liquid at any time during the synthesis, while measuring.

Here we report in-situ measurements aimed at giving more insight into one of the most debated nanoparticle formation mechanisms: the formation of CdSe nanoplatelets. These platelets have a cubic crystal structure, which makes the non-isotropic growth in only two dimensions counterintuitive. Several seemingly contradicting mechanisms for the formation have been suggested, e.g. oriented attachment [1], ligand-induced two-dimensional growth [2] and nucleation on crystal facets [3]. There is a consensus that the platelets only form in solution when a mixture of short and long chain carboxylates is present. SAXS measurements show that the NPLs form after the short chain cadmium carboxylate has melted (240°C). The in-situ absorption measurements reveal a large effect of the temperature when the short chain carboxylate is introduced in the reaction mixture (190-240°C). This triggered us to investigate the effect of addition temperature and the melting point of the short chain carboxylates in more detail. The SAXS and WAXS measurements provide information on the size, shape, crystals structure and melting of the carboxylate, whereas the absorption measurements made it easier to distinguish between a mixture of quantum dots and nanoplatelets.

[1] Yiya Chen and X. Peng, J. Am. Chem. Soc., 139 (29), 10009-10019 (2017)
[2] S. Ithurria, G. Bousquet, and B. Dubertret, J. Am. Chem. Soc., 133 (9), 3070-3077 (2011)
[3] Andreas Riedinger and David J. Norris, Nature Materials 16, 743–748 (2017)

In contrast to the classical monomer-by-monomer addition models, crystallization by addition of particles, ranging from multi-ion complexes to fully formed nanocrystals, is now recognized as a common phenomenon. Although experimental observations clearly demonstrated crystallization by particle attachment, especially oriented attachments, many fundamental aspects remain unknown, such as the interplay of solution structure, interfacial forces, and particle motion. Here we use in situ transmission electron microscopy and atomic force microscopy to directly observe and measure the particle motion and interfacial forces, respectively. Integrating theoretical calculations and simulations, we revealed the orientation dependent attachments of particles, interfacial structures, and the factors that affect them, such as temperature, pH, ionic strength, particle shapes, and solvents, etc. In addition, we also demonstrated that strain relaxation at the grain boundaries, induced via oriented attachments, forming twin structures. The specific systems that we investigate are metal (Ag and Au), oxides (TiO₂), and minerals (mica), etc. The work provides a deep understanding of mechanisms of oriented attachments and factors control the process and thus provide a guide to design and fabricate the morphologies and structures, which are closely related with material properties and functions.

Advanced synchrotron sources of high flux and detector systems capable of collecting, in just a few seconds, full X-ray diffraction (XRD) patterns of powder samples create opportunities for in situ studies of a variety of processes such as catalysis, energy storage and conversion, and crystallization of nanoparticles from amorphous precursors. There will be thousands of XRD patterns containing information on structural changes along the processes. Particularly in the case of crystallization, information on crystal phase, lattice strain, and particle size distribution (PSD) as a function of time and temperature, covering the whole process from amorphous to fully crystalline phases.

Besides position of the diffraction peaks that can provide information on thermal expansion coefficients from nanoscale to large particles with bulk properties, width and area of the diffraction peaks allow a detailed investigation of the growth kinetics, including nucleation and coarsening processes in the time-dependent PSDs. Particle size from diffraction peak widths stand for just one value that can be produced by an infinity of PSDs. In situ measures of integrated intensities during growth of crystalline particles provide a direct route to access fluctuation in the population N(t) of particles as a function of time t. Nucleation (dN/dt>0) and coarsening (dN/dt<0) processes can be evidenced from the experimental integrated intensities when compared to theoretical values for PSDs with constant number of particles and with imposed constraints to the determined particle sizes from diffraction peak widths.

Coarsening (grains growing in size) under constant volume of diffraction is possible at the cost of small grains that dissolve providing material to larger grains in a process widely known as the Ostwald ripening phenomenon. Since first described by Wilhelm Ostwald in 1896, the phenomenon has been observed in a number of general nanocrystal growth systems where broad spectrum of particle sizes are present [1]. This process play a fundamental role in determining the evolution in time of the particle population, and it is crucial for controlled synthesis of nanocrystals.

For PSDs with broader distributions, particle sizes from diffraction peak widths has complex behavior with the PSD parameters such as mode (most probable value) and standard deviation, and where dynamical diffraction corrections can be indispensable. For strong Bragg reflections, neglecting dynamical corrections in broad PSDs with many particles of sizes above a few hundred nanometers leads to an aberration regarding the actual mode value. Without accounting for dynamical effects to suppress the kinematical intensity contribution of large particles, the position and width of the PSDs would have to be much smaller than they actual are to reproduce the experimental peak widths. In this work we developed a systematic procedure to solve time-dependent PSDs based on in situ XRD experiments. It is based on the kinematical approach of x-ray diffraction, with proper integrated reflectivity corrections from dynamical diffraction calculation [2]. As practical demonstration we applied this analytical procedure to the single-phase crystallization process of bismuth ferrite [3]. In situ XRD data collection where carried out with synchrotron x-rays of 12keV, acquisition time of 38 s per XRD pattern, and enough flux to monitor diffracting particles as small as 20 nm.

[1]Z. Zhang, Z. Wang, S. He, C. Wang, M. Jin, and Y. Yin,
Chem. Sci. 6, 5197 (2015).
[2] S. L. Morelhão, Computer Simulation Tools for X-ray Analysis, Graduate Texts in Physics (Springer, Cham, 2016).
[3] G. Clarke, A. Rogov, S. McCarthy, L. Bonacina, Y. Gun’ko, C. Galez, R. Le Dantec, Y. Volkov, Y. Mugnier, and A. Prina-Mello, Scientific Reports 8, 10473 (2018).

The potential of environmental scanning electron microscopy (ESEM) as a versatile, in-operando technique that can accommodate many length scales, substrate geometries and reactive environments has been highlighted in the context of recent in-situ studies of catalytic graphene and hexagonal boron nitride (h-BN) chemical vapour deposition [1-3]. We have set up a dedicated new ESEM system based on a modified adoption of the ZEISS NanoVP aperture technology to facilitate improved high-resolution imaging, combined with heating capabilities of up to 1000°C and a customised local gas injector system. We present here our latest results using this new in-operando capability to foster an understanding of crystal growth and phase transformations on the nano-scale specifically for 1D and 2D materials. We show in-operando ESEM of the Ni silicidation of top-down patterned Si nanochannels revealing unexpected time-dependent phenomena that contribute to inhomogeneous silicidation lengths. This silicidation process plays an important role in VLSI transistor fabrication as it allows the formation of highly conductive metal silicide contacts very close to the gate [4,5]. However, while bulk silicidation is well understood, the altered kinetics for small diameters, which are increasingly applied in highly scaled transistors, are not fully understood [6]. We introduce strategies how such silicidation on the nano-scale can be much better controlled, offering new insights for scaled contact formation. We also show an in-situ ESEM study of localised etching of 2D materials addressing key questions on the nucleation and etch kinetics specifically of graphene, h-BN and TMD materials such as WS₂.

References
[1] P. Kidambi et al., Nano Lett 13, 4769 (2013)
[2] R. Weatherup et al Nano Lett. 16, 6196 (2016)
[3] R. Wang et al. ACS Nano 13, 2114 (2019)
[4] S. P. Murarka, Silicides for VLSI Applications. Academic Press, 2012
[5] H. Iwai et al., Microelectronic Engineering, 60(1), 157–169 (2002)
[6] W. Tang et al., Semicond. Sci. Technol., 29(5), 054004 (2014)

Understanding and directing electrochemical reactions below the micrometer is a long-standing challenge in electrochemistry. In this respect, electrodeposition using scanning probe microscopy provides an attractive platform, not only for additive nanoscale manufacturing, but also for the investigation of local dynamics such as mass transport and reaction kinetics.

We report on the direct writing of copper nanostructures using an oscillating nanoelectrode operated with an Atomic Force Microscope (AFM) in aqueous solutions. The nanoelectrode is used to locally initiate electrochemistry, while the AFM allows for in-situ probing of the deposited material. We demonstrate the writing of arbitrary structures beyond cluster size with lateral dimensions in the order of the tip radius.
Strikingly, this local writing is only possible in highly dilute electrolytes (~ 1 µM CuSO₄). Moreover, we find a profound effect of the dynamic operation of the nanoelectrode on the deposition kinetics.

We explain our results by considering the extended screening length and slow charge dynamics of the dilute electrolyte, which allows the nanoelectrode to operate inside, and disturb, the electrical double layer (EDL) at the solid-liquid interface. We visualize these (dis)charging dynamics and their effect on the reaction directly through the electrodeposited structures, and experimentally observe the Debye time of the electrolyte as a threshold value for growth. Furthermore, we show that this threshold time increases as a function of applied potential.

Our results lead us to propose a reaction mechanism based on cyclic charging and discharging of the EDL, which facilitates local electrodeposition through direct electrostatic interaction between ions in the liquid and the nanoelectrode.

One of the factors limiting catalytic activity of nanoparticles is the lack of control over formation and evolution processes (i.e. coalescence and coarsening) and over size and shape selectivity at high temperature. Here, we show that zeolite nanosheets of the MFI type (MFI-nanosheets or MFI-Ns), formally known as the Zeolite Socony Mobil–five (ZSM-5), can be used to keep cobalt catalyst nanoparticles docked to the support external-surface and to help retain their small size when exposed to high-temperature reducing conditions for a prolonged time. In situ environmental transmission electron microscopy (ETEM) shows that the MFI-Ns can influence not only the size of nanoparticles but also their shape via catalyst/support registry. ETEM image analysis revealed the preferential docking of nanoparticles to energetically favored configurations –namely to four or less pores consequently regulating the resulting population of nanoparticles. We ascribe the observed registry to catalyst/support interactions and to the periodic surface of the MFI-Ns with local minima, which helped stabilize nanoparticles and suppress evolution. This study elucidates how a nano-porous support surface may provide a path forward for controlling nanoparticle formation and evolution kinetics and for achieving size- and shape-selectivity at high temperature.

The resolution of optical microscopy is generally limited by diffraction to approximately half a wavelength, which prevents close monitoring of the motion of nanoparticles with optical techniques. Moreover, the intensity of light scattered from a single nanoparticle scales as the square of the particle volume, making it difficult to detect the scattered light from small particles with sufficient speed to characterize particle dynamics. Various superresolution techniques exist that overcome these difficulties, but these generally require careful fluorescent labeling or else are applicable only in rather limited circumstances, and are often too slow to be of use in following the dynamics of self-assembly.

We introduce an interferometric optical technique that overcomes many of these difficulties, and therefore will allow the detailed characterization of the dynamics of self-assembly at the nanoscale. This technique measures the optical anisotropy in the scattering from individual nanoparticles or small assemblies of nanoparticles, which makes it possible to determine the three-dimensional alignment between pairs of particles or of a single anisotropic particle with time resolution on the order of a few microseconds. We will present a theoretical development of this technique, as well as experimental demonstrations on monitoring the rotational motion of individual gold nanorods and the assembly of superparamagnetic nanoclusters.

Looking forward, the technique could be straightforwardly combined with existing single-particle tracking techniques, and since it is interferometric in nature, particles as small as a few nanometers could potentially be tracked, both in position (x,y,z) and orientation (θ,φ). In addition, each measurement results in an optical spectrum, which has the added advantage that for the assembly of metal nanoparticles, the plasmon nanoruler effect could be used to determine the separation between particles with high accuracy. Taken together, these capabilities have the potential to afford in situ monitoring of nanoscale self-assembly with unprecedent detail in a straightforward, non-invasive, and affordable setup.

Since the 1930s, molybdenum oxide (MoO₃) is in the scope of knowledge to be a very selective catalyst for the conversion of methanol to formaldehyde by oxidation [1]. It also has promising gas sensing properties and the utilization of MoO₃ nanorods in gas sensors exhibits a highly selective response to NO₂ and NH₃ [2]. Catalysis as well as sensing requires high surface area materials and thus a controlled formation of materials on the nanoscale. Gas phase synthesis is an established technology to produce small particles in high purity without any template.
We investigated the nucleation and growth of molybdenum oxide nanoparticles in flames using time-of-flight mass spectrometry (TOF-MS) and particle mass spectrometry in combination with molecular-beam sampling. This sampling technique instantaneously interrupts all reactive gas-phase processes and thus allows a spatially resolved measurement of gas and particle-phase. TOF-MS can be used to identify the composition as well as the transient formation and growth of intermediate species such as clusters from various locations in a reactor.
MoO₃ nanoparticles are synthesized by the decomposition of molybdenum hexafluoride (MoF₆) in a premixed hydrogen/oxygen flame. Due to low pressure (20–30 mbar), an almost one-dimensional, spatially extended flat flame is formed enabling spatially-resolved measurements along the reactor axis. The progress from the initial decomposition of the gaseous precursor to the formation and growth of (MoO₃)_n clusters and nanoparticles has been studied by varying the distance between the burner head and the sampling nozzle position. The measurements show a sequential cluster formation followed by a subsequent particle nucleation. The number concentrations of differently sized clusters strongly vary indicating “magic” cluster sizes. In contrast to previously studied WO₃ nanoparticle formation [3], (MoO₃)_n clusters with excess of oxygen are observed. Moreover, (MoO₃)_n cluster and nanoparticle formation is much more prone to the flame chemistry during synthesis, which can be used to tune the nanoparticle growth process.

References:
[1] H. Adkins, W.R. Peterson, J. Am. Chem. Soc. 53 (1931) 4, 1512-1520.
[2] Taurino, A. Ferleo, L. Francioso, P. Siciliano, Appl. Phys. Lett. 88 (2006), 152111.
[3] S. Kluge, H. Wiggers, C. Schulz, Proceedings of the Combustion Institute 36 (2017) 1, 1037-1044.

In-situ pump probe electron scattering to study the nucleation and growth of thin films is still a technique in the early stages of development. Unlike in 4-D microscopy, where a single electron scattering event is observed 10⁶ times where reversible phenomena is interrogated, most material self-assembly at the nanoscale evolves and requires a single packet of 10⁶ electrons to probe. We will present a different approach to ultrafast pump probe electron scattering by using a laser-pumped electron gun in a reflection geometry to collect the specular beam rather than a diffracted beam which allows for lower electron loss in the material.

We have developed this technique to directly measure the ultrafast laser assisted diffusion of self-interstitials that are generated by irradiating GaAs with a 150 fs laser operating at 780nm using a fluence below the melt threshold but above the High Spatial Fluence Laser Induced Periodic Surface Structure (HSFL) formation threshold (0.04 J/cm²). We have described the defect generation mechanism in the literature [1]. Self interstitials are generated via an ultrafast, athermal process in less than 600 fs after the laser pulse irradiation. Subsequent 1 kHz irradiation for one second causes the formation a self-assembled grating that eventually has a wavelength of 180 nm consisting of ~90 x 200 nm walls and ~90 x 200 nm trenches (measured with Transmission Electron Microscopy). Preliminary pump probe experiments show that the growth does not continue after the last pulse. The growth rate of the material above the original surface is at least 2000 Å/sec assuming the process takes the full millisecond between pulses. However, we have much evidence that the ultrafast laser assisted diffusion stops somewhere between 100 ps to 50 ns, hence the actual growth rate is between 10⁶ and 10⁹ Å per second.

We will present results from a nonadiabatic quantum molecular dynamics calculation where we have studied the enhancement in mass transport of a self-interstitial atom during and after irradiation. The model couples the vector potential of the laser field directly to the Hamiltonian and allows us to directly monitor the bonding and electron-ion coupling. The method, which requires 40 attosecond time steps, has revealed additional athermal, ultrafast atomic transport mechanisms that are consistent with the extremely large growth rates we observe.

This talk will present our electron gun design and performance as well as results from our pump probe electron scattering study of the nucleation and growth of the first islands that appear after ~300 to 400 exposures. These islands are ~0.5 nm tall and ~20 nm wide and require a probe with a commensurate wavelength. Our electron pulses have a temporal width of ~2 ps after acceleration to 30KV. Results will be shown for multiple experiments where the electron beam will be delayed from 50 ps after the 400th exposure to 50 ns after the 400th exposure. By observing the reduction in the specular beam intensity we will be able to extract an arrival rate of our subsurface interstitials assisted by the optical field.

References:
[1] M. J. Abere, B. Torralva, and S. M. Yalisove, Appl. Phys. Lett. 108, 153110 (2016).

Colloidal quantum dots have recently entered the consumer market as phosphor material for television screens. Over the past 30 years of research, an enormous variety of synthesis procedures has been developed for quantum dots of different chemical compositions, sizes, and shapes. In most cases, the metal and anion precursors are mixed at temperatures of 200–350°C in the presence of coordinating ligands. The outcome of the synthesis is determined by the chemical nature and concentrations of the precursors, the reaction temperature, the presence of additives, and other parameters. While for most synthesis procedures we have a reasonable model of the underlying formation mechanisms, this model is usually supported only by indirect evidence. For example, intermediate stages in the reaction may be studied by taking aliquots from the reaction mixture and analysing them ex situ using electron microscopy. Alternatively, optical spectroscopy has allowed the study of quantum dot formation mechanisms in situ, but this method relies on reference measurements that relate properties to structure. Recently, X-ray scattering experiments have been explored as a promising route to obtain in situ data on quantum dot formation. However, X-rays penetrate poorly through the reaction mixture, so the experiments could only be performed using thin capillaries as reaction vessels, which severely hinder reactant mixing compared to conventional reaction flasks.

Here, we follow the formation of colloidal CdSe quantum dots using in situ X-ray scattering under reaction conditions comparable to typical synthesis procedures used in academic labs. To this end, we have developed a reaction-flask design that allows for proper reactant mixing while minimizing attenuation of the X-ray probe beam. We can thus track the size, size dispersion, and concentration of quantum dots during growth, without the use of calibration curves. Even for the intensively studied system of CdSe quantum dots, we show that the conventional model of separated nucleation and growth fails to provide an accurate description of the quantum dot formation. Instead, we find that the nucleation phase extends for minutes, continuing while existing quantum dots in the ensemble are growing. Although nucleation and growth are not separated, we nevertheless achieve a narrow size distribution of the final product. Based on a comparison to simulations, we propose that this seemingly counterintuitive finding is evidence for a radius-dependent quantum dot growth rate. This allows smaller quantum dots to grow faster than bigger ones, resulting in a narrow size distribution. The origin of a radius depended growth rate has most likely to do with the different faceting of larger quantum dots compared to smaller ones.

Our method has provided unexpected insights into a synthesis procedure that has been studied in depth for over 25 years. This highlights the importance of in situ measurements of nanocrystal growth under realistic reaction environments. With follow-up experiments, we will investigate the synthesis of other types of nanocrystals, including NaYF₄ upconversion nanocrystals and CdSe nanoplatelets.

Broad ion beam bombardment of solid targets leads to the spontaneous formation of nanoscale structures, such as ripples or ultrasmoothened surfaces, but complete fundamental understanding of the phenomenon is still elusive. In this work, the surface sensitive X-ray technique of Grazing-Incidence Small-Angle X-ray Scattering (GISAXS) is used to study the kinetics and dynamics of nanoscale surface evolution. While non-coherent X-rays can measure the averaged kinetics on a surface, coherent X-rays make it possible to differentiate the dynamics of fluctuations around the average kinetics. Real-time studies of ion bombarded Si, Ge, and SiO₂ surfaces have been conducted. The target materials were kept either below or above their recrystallization temperatures. In the early stage of room temperature ion bombardment, the evolution of averaged kinetics is well fit with a linear theory formalism employing a viscous relaxation term. The X-ray Photon Correlation Spectroscopy (XPCS) two-time correlation function shows a novel behavior in the early stage, with memory stretching back to the beginning of the bombardment. During the late stage where the average surface structure on a given length-scale is no longer evolving, dynamical processes exhibit compressed exponential correlation behavior. Ion bombardment experiments above the target materials’ recrystallization temperatures are ongoing and their results will also be reported.

Three-dimensional (3D) electrostatic capacitors have the potential to outperform their conventional two-dimensional (2D) counterparts. In this talk, I will present our work on the use of real-time X-ray scattering techniques to aid the development of a novel class of 3D tricontinuous bulk conductor/insulator/conductor nanocapacitors using nanoporous gold (NP-Au), a 3D bicontinuous network synthesized by selective alloy corrosion (dealloying), as the scaffold and primary building block. These X-ray scattering experiments are performed at the University of Pennsylvania’s Dual-Source and Environmental X-ray Scattering Facility, an instrument capable of simultaneously probing the Ultra Small-Angle X-ray Scattering (USAXS) and Wide-Angle X-ray Scattering (WAXS) regimes in various experimental environments. In the first part of this talk, I will present a fundamental study on the synthesis of 3D NP-Au, the scaffold of our 3D nanocapacitor. In particular, we use in situ USAXS/WAXS to probe the evolution of microstructure, morphology, and crystal structure during electrochemical dealloying (synthesis) performed under potential control in neutral electrolyte. Our studies have revealed that both zero-dimensional (0D) Au nanoparticles and 3D NP-Au form during the dealloying process, an important and previously unknown finding. In the second part of my talk, I will present our studies on post-synthesis thermal coarsening in combination with X-ray scattering (in situ USAXS) to tailor the ligament/pore size distribution in the 3D NP-Au scaffold, which is used as the first conductor in the bulk conductor/insulator/conductor nanocapacitor. In connecting these in situ USAXS/coarsening results with in situ transmission electron microscope (TEM) observations, we have found that the abovementioned 0D Au nanoparticles play an important role in the initial stages of the coarsening process. Following the post-synthesis thermal coarsening, the insulator (alumina, Al2O3) and second conductor (aluminum-doped zinc oxide, AZO) are grown on the pore walls of the starting 3D NP-Au scaffold using atomic layer deposition (ALD), resulting in a 3D tricontinuous bulk NP-Au/Al2O3/AZO nanocapacitor. In the third part of this talk, I will discuss our use of operando Grazing Incidence SAXS (GISAXS) methods to probe the microstructural and morphological changes of the tricontinuous bulk nanocomposite during ultra-fast charge/discharge operation.

High-resolution imaging of electron beam-sensitive crystalline materials is one of the most difficult applications of transmission electron microscopy (TEM). The challenges are manifold, including the acquisition of images with an extremely low beam dose, the time-constrained search for crystal zone axes, the precise alignment of successive images, and the accurate determination of the defocus value.
We reported that using a direct-detection electron-counting camera, it is possible to acquire useful high-resolution TEM images with electron dose as low as a few electrons per square angstrom to ensure that the intact structure was captured before damage occurred [1]. In this talk, we will present a suite of new methods that we recently developed to address the rest challenges mentioned above. Our methods advance the HRTEM of extremely beam-sensitive materials from “occasionally possible” to “routine”. We demonstrate the effectiveness of our methodology by capturing atomic-resolution (~ 1.5 Å) TEM images of several metal organic frameworks (MOFs) that are generally recognized as highly sensitive to electron beams. In the case of MOF UiO-66, individual metal atomic columns, various types of surface termination, and benzene rings in the organic linkers, are clearly identified. We also successfully apply our methods to other electron beam-sensitive materials, and achieve atomic-resolution TEM imaging of the organic-inorganic hybrid perovskite CH₃NH₃PbBr₃ for the first time [2]. More recently, we applied this new technology to prove the successful encapsulation of single molecule magnets in MOF NU-1000 [3], and to investigate the evolution and transformation of various defects in MOF UiO-66 [4]. We also found a new TEM specimen preparation technique particularly useful for beam-sensitive materials.

[1] Y. Zhu, M. Pan, Y. Han et al., Nature Materials, 2017, 16, 532-536
[2] D. Zhang, Y. Han et al., Science, 2018, 359, 675-679
[3] M. Wriedt, Y. Han et al., JACS, 2019 (accepted)
[4] L, Liu, D. Zhang, Y. Zhu, Y. Han Nature Chemistry, 2019 (accepted)

With the continued development of scanning probe microscopy techniques, manipulation of single molecules has become possible. Thereby, the manipulation path can be chosen at will and energy barriers between potential minima on that pathway can be quantified, as can the energy landscape around the molecule before and after manipulation. To explore the practicality of this novel pathway to catalysis research, we selected benzene molecules on a Cu (100) surface as a model system. We first choose a specific manipulation path and then move the tip at constant but continuously reduced heights while recording the oscillation amplitude Aand phase phi with the microscope operated in our recently developed tuned-oscillator (TO) detection scheme [1]. To preserve the accuracy of recovered tip-sample interaction potentials and forces, we use oscillation amplitudes significantly larger than the decay length of the tip-sample interaction potential [2,3]. Analyzing the full (x, z,A, phi) data array then allows recovery of the potential energy U(x, z) acting between the tip and the sample, the force on the tip normal to the surface vertical tip-sample force F_n(x, z), and the force F_l(x, z) that acts on the tip along the manipulation path (i.e., lateral) with meV, pN, and pm resolution. In 54 distinct manipulation events, the molecules were either pushed, pulled, jumped to the tip, or did not move depending on the chemical surroundings of the molecule and the chemical identity of the tip. For further insight, we compared the experimentally measured energy landscapes and manipulation outcomes with computational results obtained using a Bayesian Optimization Structure Search (BOSS) protocol [4], which also delivers that the diffusion barrier evolves with the chemical identity of the tip and the surrounding of the molecule. This machine learning technique developed at Aalto University delivers the minimum energy pathways and entire energy landscapes with minimal computational effort.
References:
[1] O. E. Dagdeviren et al., Nanotechnology 27, 065703 (2016).
[2] O. E. Dagdeviren et al., Physical Review Applied 9,044040 (2018).
[3] O. E. Dagdeviren et al., Review of Scientific Instruments 90,033707 (2019).
[4] M. Todorović et al., NPJ Computational Materials 5, 35 (2019).

Pt nanoparticles (NPs) have been widely used as electrocatalysts for oxygen reduction reactions (ORR) in fuel cells of interest in the automotive industry¹ and hydrogen evolution reaction (HER) in alkaline electrolyzers.² However, the high cost as well as the high d-band center of Pt hinder its application as a cost-effective fuel cell material. Instead, Pt is combined with other metals. The resulting Pt-based bimetallic NPs are potential candidates to overcome the issues associated with pure Pt without compromising the electrochemical performance. Since the underlying electrocatalytic reactions, such as adsorption and desorption of reactants, intermediates and products, take place on the surface of the electrocatalysts, the performance is determined by the surface composition of the bimetallic NPs.³ Tremendous effort has been made in synthesizing Pt-based bimetallic NPs with precisely controlled surface structure and chemistry. In particular, chemical and electrochemical processes have both been used to tune the surface composition of bimetallic NPs.⁴ Chemical processing is often difficult to control due to random site etching, but electrochemical processing is relatively versatile and it can be straightforward to etch the non-noble metal away. However, important questions regarding this electrochemical tuning remain unresolved, regarding the interplay of kinetic etching phenomena and external parameters such as the NP shape, initial composition, surfactants, temperature and etching potential.

Here, we use in situ liquid cell electron microscopy as a direct probe of the electrochemical etching of oleylamine-capped Pt-based bimetallic NPs in aqueous sulfuric acid electrolyte solution at elevated temperatures with high temporal and spatial resolution. We have modified liquid cell chips in order to perform simultaneous electrochemical biasing and heating. We synthesize Pt-based bimetallic NPs in the shape of rhombic dodecahedra with well-defined facets.⁵ The particles contain a Pt skeleton covered by the non-noble metal, for example Ni, as well as a surface layer of Pt surrounding the entire NP. Our real-time observations reveal that the etching kinetics of these Pt-based bimetallic NPs upon application of a bias is dependent on the reaction temperature. Furthermore, the applied bias required to etch the NPs is also dependent on the reaction temperature. We will discuss the etching mechanism, taking account of the effects of the electron beam on the observations. Understanding the electrochemical behavior of these Pt-based bimetallic NPs at elevated temperatures potentially lays the foundation for rational design of desired structures for application as electrocatalysts.

References
[1] Stamenkovic, V. R.; Fowler, B.; Mun, B. S.; Wang, G.; Ross, P. N.; Lucas, C. A.; Marković, N. M. Science 2007, 315, 493-497.
[2] Chen, C.; Kang, Y.; Huo, Z.; Zhu, Z.; Huang, W.; Xin, H. L.; Snyder, J. D.; Li, D.; Herron, J. A.; Mavrikakis, M.; Chi, M.; More, K. L.; Li, Y.; Markovic, N. M.; Somorjai, G. A.; Yang, P.; Stamenkovic, V. R. Science 2014, 343, 1339-1343.
[3] Liao, H.; Fisher, A.; Xu, Z. J., Small 2015, 11 (27), 3221-3246.
[4] Zhang, B.-W.; Yang, H.-L.; Wang, Y.-X.; Dou, S.-X.; Liu, H.-K., Adv. Energy Mater.2018, 8 (20), 1703597.
[5] Niu, Z.; Becknell, N.; Yu, Y.; Kim, D.; Chen, C.; Kornienko, N.; Somorjai, G. A.; Yang, P., Nat. Mater. 2016, 15, 1188.

Gas-solid reaction plays an important role in catalysis, gas sensing and gas storage. Therefore, reaction process on metal surface with gas has been extensively studied at the atomic scale. However, the reaction process on metal surface with gas under electric field remains poorly understood. In our study, we investigated the reaction process on metal surface under electric field by means of atomic-resolution in situ environmental transmission electron microscopy (ETEM), and found that an unexpected metal (Pd)-gas (N₂) compound was formed on the metal surface under an electric field.^[1]

We studied Pd surface by using a pair of Pd electrode tips with nanometer separation. The reaction process on Pd surface in gases, such as O₂ and H₂, is well understood. Here, we performed systematic study on Pd surface in four typical environments (vacuum, O₂ (100 Pa), H₂ (100 Pa) and N₂ (100 Pa)) with and without an electric field. First, we studied Pd surface without an electric field.^[2] We found that the surface oxidation of Pd surface in O₂ has a dependence of previous exposure to H₂: after prolonged exposure to H₂, the oxidation of Pd surface is suppressed. Based on the ETEM observation and electron energy loss spectroscopy (EELS) measurement, we concluded that subsurface adsorption of H atoms suppresses the oxidation of Pd surface. Then, we studied Pd surface pretreated in H₂ under an electric field. In vacuum, a tunneling current flowed across the nanogap between electrically biased Pd electrode tips. In O₂ and H₂, Pd atoms on the positive electrode tip were evaporated and then deposited to the negative electrode tip without any changes in the crystal structure.^[1] In N₂, we found that a peculiar compound was formed on the positive electrode tip.^[1] A PdN compound, which is predicted and confirmed only at extremely high pressure and temperature, was formed at low pressure and room temperature under an electric field. The gas-solid reaction was reversible with the change of the magnitude of an electric field. It is most likely that electrons passing through the nanogap induce the gas-solid phase transition. We will show the in situ movie of the dynamic phenomena including the gas-solid phase transition. In addition, we will show the result of ETEM observation of Pt surface. These findings obtained only by means of in situ ETEM observation have opened new route toward synthesizing nanomaterials.

References
[1] T. Tamaoka, R. Aso, H. Yoshida and S. Takeda, Nanoscale, 2019, 11, 8715-8717
[2] T. Tamaoka, H. Yoshida and S. Takeda, RSC Advances, 2019, 9, 9113-9116

Cross-correlative techniques using transmission electron microscopy (TEM) and synchrotron transmission X-ray microscopy (STXM) are becoming more common and are used to study a wide range of electrochemical reaction-systems at the nanoscale. This approach has already started to produce new insights on the dynamics and structural changes during processes as lithium ion insertion/extraction, dendrite formation, metal nucleation, and metal corrosion. Despite the power of this approach, the challenges associated with replicating bulk-scale electrochemistry data in the environmental cell microscopy platform are well-known. First, the hardware components are not best optimized to perform in the reduced scale environment. Second, the chips configurations such as sizes and aspect ratios of different electrodes suitable for various electroanalytical measurements are poorly designed. Here, we use operando liquid cell TEM/STXM microscopy platform, validating cross-correlative electroanalytical bulk-scale measurements with model systems, to study the oxygen evolution reaction (OER) in the electrodeposited thin films of Ni(OH)₂/Co(OH) for charge storage applications such as batteries and supercapacitors.

The studies presented here are performed using electrochemical cells, which consist of two microfabricated chips sandwiched with electron transparent SiNx membranes for viewing in TEM/STXM. Using a new developed hardware system with a specialized configuration of working, counter and reference electrode, we performed cyclic voltammetry studies in 0.1M CuSO₄ and 20 mM K₃Fe(CN)₆/20 mM K₄Fe(CN)₆ in 0.1M KCl solutions to validate the bulk-scale correlation between TEM and STXM. Next, we study nucleation mechanism during the electrodeposition of Ni(OH)₂/Co(OH)₂ film growth on the substrate electrode using TEM/STXM. We compare oxygen evolution on 50 μg cm-2geom Co(OH)₂ on glassy carbon in a macroscopic rotating disk electrode (RDE) cell at a rotation rate of ω = 1600 rpm and a scan rate of ν = 10 mV s-1 and in the operando STXM cell with Co(OH)₂ on Pt at defined voltage scan rate and electrolyte flow rate. The current-potential relationship in the operando STXM cell shows similar Tafel slopes at the same potentials to the RDE tests validating the high electrochemical fidelity of the operando STXM cell. The liquid-electrochemistry hardware developed for this work can be used to study range of materials using cross-correlative imaging in TEM/SEM and X-ray microscopes.

We study the oxidation of clean suspended mono- and few-layer graphene in real-time by in situ environmental transmission electron microscopy [1]. In order to study the oxidative etching of suspended graphene, we have developed a cleaning scheme that results in the near complete removal of hydrocarbon residues over the entire visible sample area (~50 µm²) [1, 2]. This etching process allows the top-down fabrication of graphene nanoribbons with lower edge roughness than can be achieved by electron beam lithography.

At pressures below 0.1 mbar we observe anisotropic oxidation and the formation of hexagonal holes with armchair-oriented edges, with an edge roughness below 1 nm. We see increasingly isotropic oxidation at higher pressures, eventually leading to irregular holes at a pressure of 6 mbar. We find that the growth rate of the initial defect is more stable than that of an extended hole. We show that this is due to re-arrangement of the edge atoms, leading to a stable defect which over time is broken by consecutive discrete and stochastic atomic removal events.

Finally, we find that few-layer suspended flakes are stable against oxidation at temperatures up to at least 1000 °C in the absence of impurities and electron beam-induced defects. These findings show (i) that the oxidation behavior of mono- and few-layer graphene depends critically on the intrinsic roughness, cleanliness and any imposed roughness or additional reactivity from a supporting substrate and (ii) that the activation energy for oxidation of pristine suspended few-layer graphene is up to 43% higher than previously reported for graphite. These results have implications for applications of graphene where edge roughness can critically affect the performance of devices, and more generally highlights the surprising (meta)stability of the basal plane of suspended bilayer and thicker graphene towards oxidative environments at high temperature.

[1] J. D. Thomsen, J. Kling, D. M. A. Mackenzie, P. Bøggild, T. J. Booth. "Oxidation of Suspended Graphene: Etch Dynamics and Stability Beyond 1000 °C". ACS Nano, 13, 2, 2281-2288 (2019)
[2] J. D. Thomsen, T. Gunst, S. S. Gregersen, L. Gammelgaard, B. S. Jessen, D. M. A. Mackenzie, K. Watanabe, T. Taniguchi, P. Bøggild, T. J. Booth. "Suppression of Intrinsic Roughness in Encapsulated Graphene". Physical Review B, 96, 014101 (2017).

The core-shell catalysts with atomic layered surface Pt through liquid phase atomic layer deposition has been proven to be one of the most promising oxygen reduction (ORR) electrocatalysts with high activity and low consume of Pt.^1-5 It is important to directly observe how the atomic layers grow via in situ technique. On the other hand, corrosion is a major and unavoidable challenge for the degradation of properties of multimetallic core-shell electrocatalysts during electrocatalysis, which restricts the practical application of this type electrocatalysts severely.^6-7 Therefore, it has been an urgent issue to understand the evolution and mechanism of corrosion of Pt-based core-shell catalysts during the reaction with an aim to the effective utilization of the catalysts. ⁸
In this presentation, we will first report a 3D growth mechanism for the formation of core−shell nanostructures involving a hybrid process with initial island growth and surface diffusion. The desired surface structure can be achieved by adjusting the competition between island deposition and surface diffusion.⁹ Then, we will show the corrosion behavior of Pd@Pt core-shell electrocatalysts in the etching solution, which depends on the effects of geometry and surface defect.^10-11 Finally, we can demonstrate the capability of electrochemistry inside of TEM, showing both the oxidative etching and the re-deposition behaviors during the cyclic voltammetry in TEM.

References:
1. Wu, J.; Yang, H., Platinum-based oxygen reduction electrocatalysts. Acc. Chem. Res. 2013, 46, 1848-1857.
2. Bian, J.; Zhang, H.; Jiang, Y. Y.; Jin, C. H.; Wu, J. B.; Yang, H.; Yang, D. R., Nano Lett. 2015, 15, 7808-7815.
3. Yan, Y. C.; Shan, H.; Li, G.; Xiao, F.; Jiang, Y. Y.; Yan, Y. Y.; Jin, C. H.; Zhang, H.; Wu, J. B.; Yang, D. R., Nano Lett. 2016, 16, 7999-8004.
4. Xiong, Y. L.; Shan, H.; Zhou, Z. N.; Yan, Y. C.; Liu, Y. F.; Tian, H.; Wu, J. B.; Zhang, H.; Yang, D. R., Small, 2017, 13, 1603423.
5. Wang, X.; Choi, S.; Roling, L.; Luo, M.; Ma, C.; Zhang, L.; Chi, M.; Liu, J.; Xie, Z.; Herron, J.; Mavrikakis, M.; Xia, Y. Nat. Commun. 2015, 6, 7594.
6. Lim, B.; Jiang, M.; Camargo, P. H.; Cho, E. C.; Tao, J.; Lu, X.; Zhu, Y.; Xia, Y., Science 2009, 324, 1302-1305.
7. Zhang, L.; Roling, L. T.; Wang, X.; Vara, M.; Chi, M.; Liu, J.; Choi, S.-I.; Park, J.; Herron, J. A.; Xie, Y., Science 2015, 349, 412-416.
8. Wu, J.; Shao, H.; Chen, W; Gu, X.; Tao, P.; Song, C.; Shang, W.; Deng, T, Adv. Mater., 2016, 28, 9686-9712.
9. Wu, J.; Gao, W.; Wen, J.; Miller, D. J.; Lu, P.; Zuo, J. M.; Yang, H., Nano Lett., 2015, 15, 2711–2715.
10. Wu, J.; Gao, W.; Yang, H.; Zuo, J. M., ACS Nano, 2017, 11, 1696-1703.
11. Shan, H.; Gao, W.; Xiong, Y.; Shi, F.; Yan, Y.; Ma, Y.; Shang, W.; Tao, P.; Song, C.; Deng, T.; Zhang, H.; Yang, D. R.; Pan, X. P.; Wu, J., Nat. Commun., 2018, 9, 1011

Since the first discovery of supercooling in 1724, the study of supercooled matter has been mainly limited to varying temperature or pressure. Early this year we have demonstrated an electrochemical approach to generate supercooled sulfur and observe the dynamic process in situ (see Publication 1 below). Our methodology combines dark-field optical microscopy, a transparent electrochemical cell, and a fast camera to visualize the process at single microdroplet with millisecond time resolution. This platform may open up opportunities for studying supercooled liquids as the droplets approach either homogeneous nucleation to the crystalline state or enter into the glass transition. Relevant to understanding lithium-sulfur battery chemistry, liquid sulfur is observed to form in the electrochemical cell and elucidates a long-debated reaction pathway for sulfur redox reaction in this environment. In a follow-up work (see Publication 2 below), we applied the above methodology to image in situ the electrochemical formation of polybromide droplets, which is used in zinc-bromine flow batteries, a promising technology for stationary energy storage. Using our unique platform, we reveal the liquid nature, chemical composition, and strong interaction with the electrode, which cannot be done using conventional analysis tools. The results provide insights into the role of complexing agents and guide the future design of zinc-bromine flow batteries.
This presentation is based on the following two recent publications from my lab:
1. Proceedings of the National Academy of Sciences, 2019, 116 (3), 765-770
https://doi.org/10.1073/pnas.1817286116
2. Angewandte Chemie International Edition, 2019, 131, 1-7
https://doi.org/10.1002/ange.201906980

Lithium/Sodium ion batteries have high energy density and are extremely light and inexpensive to produce, thus the success of Li/Na ion batteries may help to achieve a major leap in green energy storage technology. Advancing our understanding necessitates the development of experimental tools capable of operando characterization that can discern mechanisms of operation and degradation in the native operating environment. The most illuminating and useful characterization methods are those capable of providing detailed mechanistic information of charge/discharge dynamics of individual grains and particles. TEM investigations specialize in revealing structural and compositional information with nanoscale spatial resolution and subsecond temporal resolution. By utilizing our newly designed electrochemical liquid cells, we devoted our effort to the in situ and operando multimodal study of the dynamic Electrode-Electrolyte Interface reactions.
For lithium ion batteries, we captured dynamic electrochemical lithiation and delithiation of Au anodes in a commercial LiPF₆/EC/DEC electrolyte using transmission electron microscopy (TEM). The inhomogeneous lithiation, lithium metal dendritic growth, electrolyte decomposition, and solid-electrolyte interface (SEI) formation were observed in situ. In addition, the dynamic lithiation/delithiation of MoS₂ nanosheets in a LiPF₆/EC/DEC commercial electrolyte were also observed. Upon discharge in a voltage range of 1.8-1.2 V, MoS₂ on the Ti electrode underwent irreversible decomposition resulting in fragmentation of the MoS₂ nanosheets into 5-10 nm MoS₂ nanoparticles. For sodium ion batteries, we report an in situ study of Na deposition with the development of electrochemical liquid cells, in which the electrodes within a cell are patterned in different configurations to produce a flat or sharp curvature. The Na deposition shows granular growth. Variations in the SEI thickness on a Na nano-grain introduce differences in the local growth rate and disparities in the morphology. The newly formed Na grains preferentially deposit at the base of existing grains close to the electrode.