Learning heterogeneous reaction kinetics from X-ray videos pixel by pixel

Reaction rates at spatially heterogeneous, unstable interfaces are notoriously difficult to quantify, yet are essential in engineering many chemical systems, such as batteries1 and electrocatalysts2. Experimental characterizations of such materials by operando microscopy produce rich image datasets3-6, but data-driven methods to learn physics from these images are still lacking because of the complex coupling of reaction kinetics, surface chemistry and phase separation7. Here we show that heterogeneous reaction kinetics can be learned from in situ scanning transmission X-ray microscopy (STXM) images of carbon-coated lithium iron phosphate (LFP) nanoparticles. Combining a large dataset of STXM images with a thermodynamically consistent electrochemical phase-field model, partial differential equation (PDE)-constrained optimization and uncertainty quantification, we extract the free-energy landscape and reaction kinetics and verify their consistency with theoretical models. We also simultaneously learn the spatial heterogeneity of the reaction rate, which closely matches the carbon-coating thickness profiles obtained through Auger electron microscopy (AEM). Across 180,000 image pixels, the mean discrepancy with the learned model is remarkably small (<7%) and comparable with experimental noise. Our results open the possibility of learning nonequilibrium material properties beyond the reach of traditional experimental methods and offer a new non-destructive technique for characterizing and optimizing heterogeneous reactive surfaces.

MicroLib: A library of 3D microstructures generated from 2D micrographs using SliceGAN

3D microstructural datasets are commonly used to define the geometrical domains used in finite element modelling. This has proven a useful tool for understanding how complex material systems behave under applied stresses, temperatures and chemical conditions. However, 3D imaging of materials is challenging for a number of reasons, including limited field of view, low resolution and difficult sample preparation. Recently, a machine learning method, SliceGAN, was developed to statistically generate 3D microstructural datasets of arbitrary size using a single 2D input slice as training data. In this paper, we present the results from applying SliceGAN to 87 different microstructures, ranging from biological materials to high-strength steels. To demonstrate the accuracy of the synthetic volumes created by SliceGAN, we compare three microstructural properties between the 2D training data and 3D generations, which show good agreement. This new microstructure library both provides valuable 3D microstructures that can be used in models, and also demonstrates the broad applicability of the SliceGAN algorithm.

X-ray CT observation and characterization of water transformation in heavy objects

Nondestructive observations and characterization of low-density materials composed of low-Z elements, such as water or its related substances, are essential for materials and life sciences. However, visualizing these compounds and their phase changes is still challenging. In this study, an approach to X-ray imaging of water-related substances in heavy objects without the use of contrast agents is proposed. The implementation of the approach is based upon X-ray phase shift, in which the optimal photon energy is simulated for high-contrast X-ray imaging. Proof of concept is provided by observations of resins, water, and clathrate hydrates such as CO₂ hydrate and tetrahydrofuran (THF) hydrate in an aluminum container by diffraction-enhanced X-ray imaging with synchrotron X-rays of 35 keV. These results suggest that the proposed approach is a unique method for visualizing the transformation of these clathrate hydrates and is also applicable to in situ observations of other objects composed of multiphase materials with small density differences.

X-Ray attenuation and image contrast in the X-ray computed tomography of clathrate hydrates depending on guest species

In this study, X-ray imaging of inclusion compounds encapsulating various guest species was investigated based on the calculation of X-ray attenuation coefficients.

5D operando tomographic diffraction imaging of a catalyst bed

We report the results from the first 5D tomographic diffraction imaging experiment of a complex Ni–Pd/CeO₂–ZrO₂/Al₂O₃ catalyst used for methane reforming. This five-dimensional (three spatial, one scattering and one dimension to denote time/imposed state) approach enabled us to track the chemical evolution of many particles across the catalyst bed and relate these changes to the gas environment that the particles experience. Rietveld analysis of some 2 × 10⁶ diffraction patterns allowed us to extract heterogeneities in the catalyst from the Å to the nm and to the μm scale (3D maps corresponding to unit cell lattice parameters, crystallite sizes and phase distribution maps respectively) under different chemical environments. We are able to capture the evolution of the Ni-containing species and gain a more complete insight into the multiple roles of the CeO₂-ZrO₂ promoters and the reasons behind the partial deactivation of the catalyst during partial oxidation of methane.

Multi-scale chemical imaging holds the potential to revolutionize our understanding of the relationships between structure and functionality in complex catalytic materials. Here the authors report the results from the first 5D tomographic diffraction imaging experiment of a complex Ni – Pd/ CeO₂ – ZrO₂/ Al₂O₃ catalyst used for methane reforming.

Secondary-Phase Stochastics in Lithium-Ion Battery Electrodes

Lithium-ion battery electrodes exhibit complex interplay among multiple electrochemically coupled transport processes, which rely on the underlying functionality and relative arrangement of different constituent phases. The electrochemically inactive solid phases (e.g., conductive additive and binder, referred to as the secondary phase), while beneficial for improved electronic conductivity and mechanical integrity, may partially block the electrochemically active sites and introduce additional transport resistances in the pore (electrolyte) phase. In this work, the role of mesoscale interactions and inherent stochasticity in porous electrodes is elucidated in the context of short-range (interface) and long-range (transport) characteristics. The electrode microstructure significantly affects kinetically and transport-limiting scenarios and thereby the cell performance. The secondary-phase morphology is also found to strongly influence the microstructure-transport-kinetics interactions. Apropos, strategies have been proposed for performance improvement via electrode microstructural modifications.

Following lithiation fronts in paramagnetic electrodes with in situ magnetic resonance spectroscopic imaging

Li-ion batteries are invaluable for portable electronics and vehicle electrification. A better knowledge of compositional variations within the electrodes during battery operation is, however, still needed to keep improving their performance. Although essential in the medical field, magnetic resonance imaging of solid paramagnetic battery materials is challenging due to the short lifetime of their signals. Here we develop the scanning image-selected in situ spectroscopy approach, using the strongest commercially available magnetic field gradient. We demonstrate the 7Li magnetic resonance spectroscopic image of a 5 mm-diameter operating battery with a resolution of 100 μm. The time-resolved image-spectra enable the visualization in situ of the displacement of lithiation fronts inside thick paramagnetic electrodes during battery operation. Such observations are critical to identify the key limiting parameters for high-capacity and fast-cycling batteries. This non-invasive technique also offers opportunities to study devices containing paramagnetic materials while operating.

Enhanced Surface Interaction of Water Confined in Hierarchical Porous Polymers Induced by Hydrogen Bonding

Hierarchical porous polymer systems are increasingly applied to catalysis, bioengineering, or separation technology because of the versatility provided by the connection of mesopores with percolating macroporous structures. Nuclear magnetic resonance (NMR) is a suitable technique for the study of such systems as it can detect signals stemming from the confined liquid and translate this information into pore size, molecular mobility, and liquid-surface interactions. We focus on the properties of water confined in macroporous polymers of ethylene glycol dimethacrylate and 2-hydroxyethyl methacrylate [poly(EGDMA-co-HEMA)] with different amounts of cross-linkers, in which a substantial variation of hydroxyl groups is achieved. As soft polymer scaffolds may swell upon saturation with determined liquids, the use of NMR is particularly important as it measures the system in its operational state. This study combines different NMR techniques to obtain information on surface interactions of water with hydrophilic polymer chains. A transition from a surface-induced relaxation in which relaxivity depends on the pore size to a regime where the organic pore surface strongly restricts water diffusion is observed. Surface affinities are defined through the molecular residence times near the network surface.

Larger voids in mechanically stable, loose packings of 1.3μm frictional, cohesive particles: Their reconstruction, statistical analysis, and impact on separation efficiency

Lateral transcolumn heterogeneities and the presence of larger voids in a packing (comparable to the particle size) can limit the preparation of efficient chromatographic columns. Optimizing and understanding the packing process provides keys to better packing structures and column performance. Here, we investigate the slurry-packing process for a set of capillary columns packed with C18-modified, 1.3μm bridged-ethyl hybrid porous silica particles. The slurry concentration used for packing 75μm i.d. fused-silica capillaries was increased gradually from 5 to 50mg/mL. An intermediate concentration (20mg/mL) resulted in the best separation efficiency. Three capillaries from the set representing low, intermediate, and high slurry concentrations were further used for three-dimensional bed reconstruction by confocal laser scanning microscopy and morphological analysis of the bed structure. Previous studies suggest increased slurry concentrations will result in higher column efficiency due to the suppression of transcolumn bed heterogeneities, but only up to a critical concentration. Too concentrated slurries favour the formation of larger packing voids (reaching the size of the average particle diameter). Especially large voids, which can accommodate particles from>90% of the particle size distribution, are responsible for a decrease in column efficiency at high slurry concentrations. Our work illuminates the increasing difficulty of achieving high bed densities with small, frictional, cohesive particles. As particle size decreases interparticle forces become increasingly important and hinder the ease of particle sliding during column packing. While an optimal slurry concentration is identified with respect to bed morphology and separation efficiency under conditions in this work, our results suggest adjustments of this concentration are required with regard to particle size, surface roughness, column dimensions, slurry liquid, and external effects utilized during the packing process (pressure protocol, ultrasound, electric fields).

Topological analysis of non-granular, disordered porous media: determination of pore connectivity, pore coordination, and geometric tortuosity in physically reconstructed silica monoliths

Different approaches are applied and compared, which are universally applicable to quantify pore coordination, pore and pore-throat connectivity, and geometric tortuosity.

Assessing structural correlations and heterogeneity length scales in functional porous polymers from physical reconstructions

A general, model-free, quantitative approach to the key morphological properties of a porous polymer monolith is presented. After 3D reconstruction, image-based analysis delivers detailed spatial and spatially correlated information on the structural heterogeneities in the void space and the polymer skeleton. Identified heterogeneities, which limit the monolith's performance in targeted applications, are traced back to the preparation process.

Morphological Analysis of Physically Reconstructed Silica Monoliths with Submicrometer Macropores: Effect of Decreasing Domain Size on Structural Homogeneity

Silica monoliths are increasingly used as fixed-bed supports in separation and catalysis because their bimodal pore space architecture combines excellent mass transport properties with a large surface area. To optimize their performance, a quantitative relationship between morphology and transport characteristics has to be established, and synthesis conditions that lead to a desired morphology optimized for a targeted application must be identified. However, the effects of specific synthesis parameters on the structural properties of silica monoliths are still poorly understood. An important question is how far the macropore and domain size can be reduced without compromising the structural homogeneity. We address this question with quantitative morphological data derived for a set of eight macroporous-mesoporous silica monoliths with an average macropore size (d(macro)) of between 3.7 and 0.1 μm, prepared following an established route involving the sol-gel transition and phase separation. The macropore space of the silica monolith samples is reconstructed using focused ion beam scanning electron microscopy followed by a quantitative assessment of geometrical and topological properties based on chord length distributions (CLDs) and branch-node analysis of the pore network, respectively. We observe a significant increase in structural heterogeneity, indicated by a decrease in the parameter k derived from fitting a k-gamma function to the CLDs, when d(macro) reaches the submicrometer range. The compromised structural homogeneity of silica monoliths with submicrometer macropores could possibly originate from early structural freezing during the competitive processes of sol-gel transition and phase separation. It is therefore questionable if the common approach of reducing the morphological features of silica monoliths into the submicrometer regime by changing the point of sol-gel transition can be successful. Alternative strategies and a better understanding of the involved competitive processes should be the focus of future research.

Aging of a Pt/Al2O3 exhaust gas catalyst monitored by quasi in situ X-ray micro computed tomography

Catalyst aging effects are analyzed using X-ray absorption micro-computed tomography in combination with conventional characterization methods on various length scales ranging from nm to μm to gain insight into deactivation mechanisms.

7Li in situ 1D NMR imaging of a lithium ion battery

The spatial distribution of charge carriers in lithium ion batteries during current flow is of fundamental interest for a detailed understanding of transport properties and the development of strategies for future improvements of the electrolyte–electrode interface behaviour.

Three-dimensional microstructural imaging of sulfur poisoning-induced degradation in a Ni-YSZ anode of solid oxide fuel cells

Following exposure to ppm-level hydrogen sulfide at elevated temperatures, a section of a solid oxide fuel cell (SOFC) Ni-YSZ anode was examined using a combination of synchrotron-based x-ray nanotomography and x-ray fluorescence techniques. While fluorescence measurements provided elemental identification and coarse spatial mapping, x-ray nanotomography was used to map the detailed 3-D spatial distribution of Ni, YSZ, and a nickel-sulfur poisoning phase. The nickel-sulfur layer was found to form a scale covering most of the exposed nickel surface, blocking most fuel reformation and hydrogen oxidation reaction sites. Although the exposure conditions precluded the ability to develop a detailed kinetic description of the nickel-sulfur phase formation, the results provide strong evidence of the detrimental effects of 100 ppm hydrogen sulfide on typical Ni-YSZ anode materials.

On the roles of graphene oxide doping for enhanced supercurrent in MgB2 based superconductors

Due to their graphene-like properties after oxygen reduction, incorporation of graphene oxide (GO) sheets into correlated-electron materials offers a new pathway for tailoring their properties. Fabricating GO nanocomposites with polycrystalline MgB2 superconductors leads to an order of magnitude enhancement of the supercurrent at 5 K/8 T and 20 K/4 T. Herein, we introduce a novel experimental approach to overcome the formidable challenge of performing quantitative microscopy and microanalysis of such composites, so as to unveil how GO doping influences the structure and hence the material properties. Atom probe microscopy and electron microscopy were used to directly image the GO within the MgB2, and we combined these data with computational simulations to derive the property-enhancing mechanisms. Our results reveal synergetic effects of GO, namely, via localized atomic (carbon and oxygen) doping as well as texturing of the crystals, which provide both inter- and intra-granular flux pinning. This study opens up new insights into how low-dimensional nanostructures can be integrated into composites to modify the overall properties, using a methodology amenable to a wide range of applications.

Understanding Atom Probe Tomography of Oxide-Supported Metal Nanoparticles by Correlation with Atomic-Resolution Electron Microscopy and Field Evaporation Simulation

Oxide-supported metal nanoparticles are widely used in heterogeneous catalysis. The increasingly detailed design of such catalysts necessitates three-dimensional characterization with high spatial resolution and elemental selectivity. Laser-assisted atom probe tomography (APT) is uniquely suited to the task but faces challenges with the evaporation of metal/insulator systems. Correlation of APT with aberration-corrected scanning transmission electron microscopy (STEM), for Au nanoparticles embedded in MgO, reveals preferential evaporation of the MgO and an inaccurate assessment of nanoparticle composition. Finite element field evaporation modeling is used to illustrate the evolution of the evaporation front. Nanoparticle composition is most accurately predicted when the MgO is treated as having a locally variable evaporation field, indicating the importance of considering laser-oxide interactions and the evaporation of various molecular oxide ions. These results demonstrate the viability of APT for analysis of oxide-supported metal nanoparticles, highlighting the need for developing a theoretical framework for the evaporation of heterogeneous materials.