Ion Beam Scanning Electron Microscopy Research Articles

Effect of cell pressure on the electrochemical performance of all‐solid‐state lithium batteries with zero‐excess Li metal anode

AbstractAll‐solid‐state cells (ASSCs) typically operate at a specific pressure to ensure good contact between the solid electrolyte and the electrode‐active materials. However, establishing the ideal cell pressure is challenging because of the various cell structures, the mechanical characteristics of solid electrolytes, and the extent to which the volume of the electrodes changes during cycling. In this study, we propose a specially designed cell assembly that adjusts to the changes in volume that occur during cycling while maintaining a constant cell pressure. The evaluations indicate that the spring in the cell assembly effectively reduces the stress incurred from the volume expansion that occurs in the electrode during charging (lithiation) and the volume contraction that occurs during discharging (delithiation) while maintaining the prescribed cell pressure. The capacity fading—as a function of the cycle number—decreases when operating ASSCs comprising a cell assembly that include a spring, compared with those that exclude a spring. Focused ion beam–scanning electron microscope reveals no cracks and delamination in the LiNi0.8Co0.1Mn0.1O3 (NCM811) composite cathode of the ASSCs, operated at 25 MPa, with a spring‐equipped assembly. The Ag nanolayer that deposits on the Cu foil is an effective collector metal, forming a dense lithium plating layer on the Ag/Cu foil anode.

3D Multiscale Lithium‐Ion Cell Modeling for LiFePO4 Freeze‐Casted Electrode Structures Using Synchrotron X‐Ray and FIB/SEM Tomography

AbstractThe performance of batteries and the associated operating areas depend, among other things, on the 3D microstructures of the electrode materials, and thus fundamental research is required in the field of electrode design. A multiscale microstructure‐resolved 3D model is developed that investigates two different LiFePO4 freeze‐casted electrode structures, that is, cellular and lamellar. The microstructure is simulated directly from the X‐ray computed tomography data and the nanostructure is combined with the pseudo‐2D simulation approach, where the morphological parameters and the distribution of the binder, carbon, and LiFePO4 are obtained from ex situ focused ion beam scanning electron microscopy measurements. The discharge performance shows that the lamellar structure exhibits a lower ohmic overvoltage and achieves a higher gravimetric capacity compared to the cellular structure, even though both electrode materials have the same porosity and amount of active material. The simulation reveals that the performance is not only directly influenced by the lithium‐ion transport through the porous structure but also by the current distribution through the active material. Based on these insights, lamellar electrode structures should be considered for next‐generation battery electrodes. The modeling approach can assist in electrode fabrication by identifying defects or suggesting better structural parameters.

Three-dimensional images reveal the impact of the endosymbiont Midichloria mitochondrii on the host mitochondria

The hard tick, Ixodes ricinus, a main Lyme disease vector, harbors an intracellular bacterial endosymbiont. Midichloria mitochondrii is maternally inherited and resides in the mitochondria of I. ricinus oocytes, but the consequences of this endosymbiosis are not well understood. Here, we provide 3D images of wild-type and aposymbiotic I. ricinus oocytes generated with focused ion beam-scanning electron microscopy. Quantitative image analyses of endosymbionts and oocyte mitochondria at different maturation stages show that the populations of both mitochondrion-associated bacteria and bacterium-hosting mitochondria increase upon vitellogenisation, and that mitochondria can host multiple bacteria in later stages. Three-dimensional reconstructions show symbiosis-dependent morphologies of mitochondria and demonstrate complete M. mitochondrii inclusion inside a mitochondrion. Cytoplasmic endosymbiont located close to mitochondria are not oriented towards the mitochondria, suggesting that bacterial recolonization is unlikely. We further demonstrate individual globular-shaped mitochondria in the wild type oocytes, while aposymbiotic oocytes only contain a mitochondrial network. In summary, our study suggests that M. mitochondrii modulates mitochondrial fragmentation in oogenesis possibly affecting organelle function and ensuring its presence over generations.

Interactive Machine Learning Improves Accuracy of Coal Porosity Segmentation in Focused Ion Beam–Scanning Electron Microscopy Images

Interactive Machine Learning Improves Accuracy of Coal Porosity Segmentation in Focused Ion Beam–Scanning Electron Microscopy Images

Growth hormone remodels the 3D-structure of the mitochondria of inflammatory macrophages and promotes metabolic reprogramming.

Macrophages are a heterogeneous population of innate immune cells that support tissue homeostasis through their involvement in tissue development and repair, and pathogen defense. Emerging data reveal that metabolism may control macrophage polarization and function and, conversely, phenotypic polarization may drive metabolic reprogramming. Here we use biochemical analysis, correlative cryogenic fluorescence microscopy and cryo-focused ion-beam scanning electron microscopy. We demonstrate that growth hormone (GH) reprograms inflammatory GM-CSF-primed monocyte-derived macrophages (GM-MØ) by functioning as a metabolic modulator. We found that exogenous treatment of GM-MØ with recombinant human GH reduced glycolysis and lactate production to levels similar to those found in anti-inflammatory M-MØ. Moreover, GH treatment of GM-MØ augmented mitochondrial volume and altered mitochondrial dynamics, including the remodeling of the inner membrane to increase the density of cristae. Our data demonstrate that GH likely serves a modulatory role in the metabolism of inflammatory macrophages and suggest that metabolic reprogramming of macrophages should be considered as a new target to intervene in inflammatory diseases.

3D-reconstructions of zygospores in Zygnema vaginatum (Charophyta) reveal details of cell wall formation, suggesting adaptations to extreme habitats.

The streptophyte green algal class Zygnematophyceae is the immediate sister lineage to land plants. Their special form of sexual reproduction via conjugation might have played a key role during terrestrialization. Thus, studying Zygnematophyceae and conjugation is crucial for understanding the conquest of land. Moreover, sexual reproduction features are important for species determination. We present a phylogenetic analysis of a field-sampled Zygnema strain and analyze its conjugation process and zygospore morphology, both at the micro- and nanoscale, including 3D-reconstructions of the zygospore architecture. Vegetative filament size (26.18 ± 1.07 μm) and reproductive features allowed morphological determination of Zygnema vaginatum, which was combined with molecular analyses based on rbcL sequencing. Transmission electron microscopy (TEM) depicted a thin cell wall in young zygospores, while mature cells exhibited a tripartite wall, including a massive and sculptured mesospore. During development, cytological reorganizations were visualized by focused ion beam scanning electron microscopy (FIB-SEM). Pyrenoids were reorganized, and the gyroid cubic central thylakoid membranes, as well as the surrounding starch granules, degraded (starch granule volume: 3.58 ± 2.35 μm3 in young cells; 0.68 ± 0.74 μm3 at an intermediate stage of zygospore maturation). Additionally, lipid droplets (LDs) changed drastically in shape and abundance during zygospore maturation (LD/cell volume: 11.77% in young cells; 8.79% in intermediate cells, 19.45% in old cells). In summary, we provide the first TEM images and 3D-reconstructions of Zygnema zygospores, giving insights into the physiological processes involved in their maturation. These observations help to understand mechanisms that facilitated the transition from water to land in Zygnematophyceae.

Cobalt chromium alloys in fixed prosthodontics: Investigations of mechanical properties and microstructure

Statement of problemCobalt chromium (Co-Cr) alloys possess beneficial mechanical properties because alloys, even in thin sections, can resist high mastication forces and exhibit an acceptable bond to the surface porcelain layer. Traditional manufacturing techniques of Co-Cr alloys such as casting have been replaced with newer fabrication techniques, such as milling, laser melting, and presintered milling. Despite scarce documentation, these new manufacturing techniques are being used to fabricate dental and implant constructions. PurposeThis in vitro study investigates the hardness, yield strength, elastic modulus, and microstructure of the most commonly used Co-Cr alloys for fixed prosthodontics based on manufacturing technique. In addition, this study investigates the effect of heat treatment on the mechanical properties and microstructure of these materials. Material and methodsFive Co-Cr alloys were included (dumbbell and rectangular shaped) based on four manufacturing techniques: cast, milled, laser melted, and presintered milled. Commercially pure titanium grade 4 and titanium-6 aluminum-4 vanadium ELI (extra low interstitial) were included for comparison, and yield strength and elongation after fracture were evaluated. The specimens were tested for hardness using the Vickers test and for elastic modulus using a nondestructive impulse excitation technique. The microstructure of selected specimens was analyzed using focused ion beam-scanning electron microscopy (FIB-SEM) and energy dispersive X-ray spectroscopy (EDS). ResultsThe mechanical properties depend on the manufacturing technique used; the laser-melted and presintered Co-Cr specimens demonstrated the highest mechanical properties, followed by the milled and cast groups. Both the laser-melted and the presintered milled Co-Cr specimens showed smaller grain size compared with the cast and milled Co-Cr specimens. The titanium-6 aluminum-4 vanadium ELI demonstrated higher hardness and yield strength compared to commercially pure titanium grade 4. No major differences were observed for the selected materials regarding the mechanical properties and microstructural appearance after heat treatment. ConclusionsThe laser melting and presintered milling techniques produced higher mechanical properties compared with the cast and milled Co-Cr. These findings were confirmed through microstructural analysis with respect to the grain size, precipitation, and number of pores.

Toward Operando Structural, Chemical, and Electrochemical Analyses of Solid-State Batteries Using Correlative Secondary Ion Mass Spectrometry Imaging.

The global transition from fossil fuels to green energy underpins the need for efficient and reliable energy storage systems. Advanced analysis and characterization of battery materials is not only important to understand fundamental battery properties but also crucial for their continued development. A deep understanding of these systems is often difficult to obtain through only pre- and/or post-mortem analyses, with the full complexity of a battery being hidden in its operational state. Thus, we have developed an operando methodology to analyze solid-state batteries (SSBs) structurally as well as chemically before, during, and after cycling. The approach is based on a specially designed sample holder, which enables a variety of electrochemical experiments. Since the entire workflow is performed within a single focused ion beam scanning electron microscope equipped with an in-house developed magnetic sector secondary ion mass spectrometer, we are able to pause the cycling at any time, perform analysis, and then continue cycling. Microstructural analysis is performed via secondary electron imaging, and the chemical mapping is performed using the secondary ion mass spectrometer. In this proof-of-concept study, we were able to identify dendrites in a short-circuited symmetric cell and to chemically map dendritic structures. While this methodology focuses on SSBs, the approach can directly be adapted to different battery systems and beyond. Our technique clearly has an advantage over many alternatives for battery analysis as no transfer of samples between instruments is needed and a correlation between the microstructure, chemical composition, and electrochemical performance is obtained directly.

264-LB: Aerobic Exercise Training Reorganizes Skeletal Muscle Mitochondrial and Lipid Droplet Architecture

Exercise is a known intervention to alleviate insulin resistance. This study investigated how a 12-week endurance exercise intervention changes skeletal muscle mitochondrial and lipid droplet architecture in patients with type 2 diabetes. Participants completed 12 weeks of endurance exercise at 70% of their VO2 max five days a week at one hour a day. Muscle biopsies were taken before and 72 hours after the last exercise session. A subset of these muscle biopsies from two participants (58±0 years, 42.3±3.3 BMI, 2.5±0.7 medications, 8.2±0.1% HbA1c) was imaged using focused ion-beam scanning electron microscopy. Due to these results being preliminary, general trends are discussed. A deep-learning model segmented 14,197 mitochondria and 4,400 lipid droplets to assess individual mitochondrial and lipid droplet 3D structures, and mitochondria-to-lipid interactions. Surprisingly, exercise did not change individual mitochondrial and lipid droplet volume, surface area to volume ratio, or sphericity. There was an increase in mitochondria-to-mitochondria and mitochondria-to-lipid droplet interactions after endurance training. These preliminary data suggest that exercise may relieve insulin-resistant skeletal muscle by increasing the connectivity among mitochondria and lipid droplets prior to changing individual organelle morphology. Further analyses are ongoing to confirm these findings. Disclosure H. A. Parry: None. C. L. Axelrod: None. V. Baena: None. W. S. Dantas: None. J. P. Kirwan: None. B. Glancy: None.

Kinetic Monte Carlo (KMC) simulation of sintering of nickel oxide-yttria stabilized zirconia composites: Model, parameter calibration and validation

This study proposes a kinetic Monte Carlo (KMC) model for the sintering of two-phase composites using nickel oxide-yttria stabilized zirconia (NiO-YSZ) as a case study. An artificial neural network (ANN) assisted method is used to calibrate the KMC model parameters by comparing the simulated microstructures and the real microstructure reconstructed by focused ion beam scanning electron microscopy (FIB-SEM). It is demonstrated that the ANN calibrated KMC model can predict quantitatively the microstructure evolution of NiO-YSZ composite. The microstructural parameters such as volume fraction, specific surface area, and triple-phase boundary length density of sintered NiO-YSZ are well predicted within acceptable errors. The proposed framework can be adapted for the simulation of the microstructure evolution of any other composite electrodes for solid oxide fuel cells and other composites fabricated by sintering.

Three-dimensional propagation behavior of hydrogen-related intergranular cracks in high-strength martensitic steel

Using X-ray computed tomography and focused ion beam-scanning electron microscopy (FIB-SEM) serial sectioning, this study investigated the three-dimensional propagation behavior of hydrogen-related intergranular cracks in martensitic steel with a tensile strength of 1.2 GPa, focusing on segments of prior austenite grain boundaries (PAGB segments). X-ray computed tomography revealed that the crack morphology was more continuous in the hydrogen-charged specimens. Through FIB-SEM serial sectioning, we found that crack-tip blunting and ductile rupture of un-cracked ligaments were associated with specific PAGB segments in the uncharged specimen. In the case of hydrogen-related intergranular crack propagation, even very fine, low-angle PAGB segments (sub-micrometer size) can act as obstacles for crack propagation. Based on these results, we propose that the misorientation of each PAGB segment has a large influence on the local arrestability of intergranular crack propagation.

Insight into mechanism of boosted oxygen reduction reaction in mixed-conducting composite cathode of solid oxide fuel cell via a novel open-source pore-scale model

Composite engineering is one of the most practical strategies for mitigating thermo-mechanical instability in state-of-the-art cobalt-based mixed ionic-electronic conducting electrodes for solid oxide cells. Nevertheless, the mechanism underlying improved performance and durability of composite electrodes is not substantially examined and understood. Pore-scale modeling can effectively bridge the gap between material property, microstructure discovery, and performance evaluation. However, most of the previous pore-scale models are built on in-house codes or commercial software that is closed-source to the public, limiting customization and community involvement. Here we report for the first time a microstructure-resolved 3D pore-scale model based on an open-source Lattice Boltzmann library, implementing a well-validated electrochemical module that considers gas, ion and electron transport, as well as electrochemical reaction at triple-phase-boundary (TPB) and double-phase-boundary (DPB). The pure La0.6Sr0.4Co0.2Fe0.8O3-δ (LSCF) and 50/50 vol% composite LSCF/GDC (Gd-doped ceria) cathodes are reconstructed using focused ion beam-scanning electron microscopy and then are investigated under various current densities, temperatures, oxygen concentrations, and thicknesses. Results reveal the advantages of boosted TPB density and ionic conductivity outweigh the disadvantages of lower DPB density and electronic conductivity for the composite cathode. Furthermore, the composite cathode performs better at low temperatures and low oxygen concentrations due to the presence of more macropores, the stable bulk ion conductivity of GDC, and higher TPB density. Local current distribution at active sites indicates a thickness of 20 to 40 μm is favorable for reduced overpotential and moderate reaction homogeneity.

Application of immuno- and affinity labeling with fluorescent dyes to in-resin CLEM of Epon-embedded cells

In-resin CLEM (Correlative Light and Electron Microscopy) of Epon-embedded cells involves correlating fluorescence microscopy with electron microscopy in the same Epon-embedded ultrathin section. This method offers the advantage of high positional accuracy compared to standard CLEM. However, it requires the expression of recombinant proteins. In order to detect the localization of endogenous target(s) and their localized ultrastructures of Epon-embedded samples using in-resin CLEM, we investigated whether immunological and affinity-labeling using fluorescent dyes applied to in-resin CLEM of Epon-embedded cells. The orange fluorescent (λem ∼550 nm) and far-red (λem ∼650 nm) fluorescent dyes examined maintained a sufficient level of fluorescent intensity after staining with osmium tetroxide and subsequent dehydration treatment with ethanol. Immunological in-resin CLEM of mitochondria and the Golgi apparatus was achieved using anti-TOM20, anti-GM130 antibodies, and fluorescent dyes. Two-color in-resin CLEM revealed that wheat germ agglutinin-puncta showed the ultrastructures of multivesicular body-like structures. Finally, taking the advantage of high positional accuracy, volume in-resin CLEM of mitochondria in the semi-thin section (2 μm thick) of Epon-embedded cells was performed by focused ion beam scanning electron microscopy. These results suggested that the application of immunological reaction and affinity-labeling with fluorescent dyes to in-resin CLEM of Epon-embedded cells is suitable for analyzing the localization of endogenous targets and their ultrastructures by scanning and transmission electron microscopy.

To be or not to be a conodont. The controversial story of Pseudooneotodus and Eurytholia

The genus Pseudooneotodus (Drygant, 1974) is a genus of small and conical elements widely distributed from the Middle Ordovician to the Early Devonian throughout the world. Because of its unusual shape, Pseudooneotodus has long been considered enigmatic, and only in the late nineties of the last century the genus has been finally placed within conodonts according to histological data.This study investigates possible similarities between Pseudooneotodus and Eurytholia (Sutton et al., 2001), an incertae sedis genus of enigmatic plates with a phosphate composition. An association of over one hundred specimens of Pseudooneotodus beckmanni and Eurytholia bohemica was analyzed from conodont residues in two distinct geographical areas: the Prague Basin (Požáry and Mušlovka sections, Bohemia, Czech Republic) and the Carnic Alps (Rauchkofel Boden section, Austria). Through an investigation that combines the use of optical and electron microscopy (including focused ion beam scanning electron microscopy), X-ray microdiffraction, and trace element (HFSE) analysis by mass spectrometry, differences between these fossil groups were observed and compared with data resulting from typical conodonts (Dapsilodus obliquicostatus, Panderodus unicostatus and Wurmiella excavata) recovered from the same samples.

Investigating structural attributes of drug encapsulated microspheres with quantitative X-ray imaging.

The intra-sphere and inter-sphere structural attributes of controlled release microsphere drug products can greatly impact their release profile and clinical performance. In developing a robust and efficient method to characterize the structure of microsphere drug products, this paper proposes X-ray microscopy (XRM) combined with artificial intelligence (AI)-based image analytics. Eight minocycline loaded poly(lactic-co-glycolic acid) (PLGA) microsphere batches were produced with controlled variations in manufacturing parameters, leading to differences in their underlying microstructures and their final release performances. A representative number of microspheres samples from each batch were imaged using high resolution, non-invasive XRM. Reconstructed images and AI-assisted segmentation were used to determine the size distribution, XRM signal intensity, and intensity variation of thousands of microspheres per sample. The signal intensity within the eight batches was nearly constant over the range of microsphere diameters, indicating high structural similarity of spheres within the same batch. Observed differences in the variation of signal intensity between different batches suggests inter-batch non-uniformity arising from differences in the underlying microstructures associated with different manufacturing parameters. These intensity variations were correlated with the structures observed from higher resolution focused ion beam scanning electron microscopy (FIB-SEM) and the in vitro release performance for the batches. The potential for this method for rapid at-line and offline product quality assessment, quality control, and quality assurance is discussed.

Comparative connectomics and escape behavior in larvae of closely related Drosophila species

Evolution has generated an enormous variety of morphological, physiological, and behavioral traits in animals. How do behaviors evolve in different directions in species equipped with similar neurons and molecular components? Here we adopted a comparative approach to investigate the similarities and differences of escape behaviors in response to noxious stimuli and their underlying neural circuits between closely related drosophilid species. Drosophilids show a wide range of escape behaviors in response to noxious cues, including escape crawling, stopping, head casting, and rolling. Here we find that D.santomea, compared with its close relative D.melanogaster, shows a higher probability of rolling in response to noxious stimulation. To assess whether this behavioral difference could be attributed to differences in neural circuitry, wegenerated focused ion beam-scanning electron microscope volumes of the ventral nerve cord of D.santomea to reconstruct the downstream partners of mdIV, a nociceptive sensory neuron in D.melanogaster. Along with partner interneurons of mdVI (including Basin-2, a multisensory integration neuron necessary for rolling) previously identified in D.melanogaster, we identified two additional partners of mdVI in D.santomea. Finally, we showed that joint activation of one of the partners (Basin-1) and a common partner (Basin-2) in D.melanogaster increased rolling probability, suggesting that the high rolling probability in D.santomea is mediated by the additional activation of Basin-1 by mdIV. These results provide a plausible mechanistic explanation for how closely related species exhibit quantitative differences in the likelihood of expressing the same behavior.

The Caudofoveata (Mollusca) Spicule as a Biomineralization Model: Unique Features Revealed by Combined Microscopy Methods

Caudofoveates are benthic organisms that reside in the deep waters of continental slopes in the world. They are considered to be a group that is of phylogenetic and ecological importance for the phylum Mollusca. However, they remain poorly studied. In this work, we revealed the structure of the spicules of Caudofoveatan mollusks Falcidens sp. The spicules presented a hierarchical organization pattern across different length scales. Various imaging and analytical methods related to light and electron microscopy were employed to characterize the samples. The primary imaging methods utilized included: low voltage field emission scanning electron microscopy (FEG-SEM), focused ion beam-scanning electron microscopy (FIB-SEM), high-resolution transmission electron microscopy (HRTEM), atomic force microscopy (AFM), and electron diffraction. In addition, we performed a physicochemical analysis by electron energy loss spectroscopy (EELS) and energy dispersive X-ray spectroscopy (EDS). A crucial factor for successfully obtaining the results was the preparation of lamellae from the spicules without damaging the original structures, achieved using FIB-SEM. This allowed us to obtain diffraction patterns of significant areas of well-preserved sections (lamellae) of the spicules in specific directions and demonstrate for the first time that the bulk of these structures is organized as a single crystal of calcium carbonate aragonite. On the other hand, AFM imaging of the spicules’ dorsal surface revealed a wavy appearance, composed of myriads of small, pointed crystallites oriented along the spicules’ longer axis (i.e., the c-axis of the aragonite). The organization pattern of these small crystallites, the possible presence of twins, the relationship between confinement conditions and accessory ions in the preference for mineral polymorphs, and the single crystalline appearance of the entire spicule, along with the observation of growth lines, provide support for further studies employing Caudofoveata spicules as a model for biomineralization studies.

FIB-SEM combined with proteomics and modification omics clarified mechanisms of lipids synthesis in organelles of Chlorella pyrenoidosa cells with high CO2 concentration

In order to explain reasons why flue-gas CO2 (normally containing high CO2) enhanced carbon fixation and lipids synthesis with increased photochemical electron production in microalgae cells. Focused ion beam scanning electron microscopy (FIB-SEM) was combined with proteomics and phosphorylation modification mics to clarify mechanisms of lipids synthesis at protein and organelle levels in Chlorella pyrenoidosa cells cultivated with high CO2 concentration (15 % v/v). The volumes of chloroplast and endoplasmic reticulum in subcellular organelles increased by 47 % and 306 %, respectively, compared with the control, which improved conversion efficiency of starch grains to lipids (lipid content increased by 57 %). Proteomics and modifications omics revealed that protein translation and ribosome structure and biogenesis-related enzymes were significantly modified by phosphorylation, which regulated protein biological functions. Glycolysis, pentose phosphate pathway and other carbohydrate metabolic pathways were markedly enriched and promoted the expression of lipid synthase, which was consistent with enhanced carbon fixation in photosynthesis, expansion of subcellular organelles and improved lipids synthesis.

Logistics of Bone Mineralization in the Chick Embryo Studied by 3D Cryo FIB-SEM Imaging.

During skeletal development, bone growth and mineralization require transport of substantial amounts of calcium, while maintaining very low concentration. How an organism overcomes this major logistical challenge remains mostly unexplained. To shed some light on the dynamics of this process, cryogenic focused ion beam-scanning electron microscopy (cryo-FIB/SEM) is used to image forming bone tissue at day 13 of a chick embryo femur. Both cells and matrix in 3D are visualized and observed as calcium-rich intracellular vesicular structures. Counting the number of these vesicles per unit volume and measuring their calcium content based on the electron back-scattering signal, the intracellular velocity at which these vesicles need to travel to transport all the calcium required for the mineral deposited in one day within the collagenous tissue can be estimated. This velocity at 0.27 µm s-1 is estimated, which is too large for a diffusion process and rather suggests active transport through the cellular network. It is concluded that calcium logistics is hierarchical and based on several transport mechanisms: first through the vasculature using calcium-binding proteins and the blood flow, then active transport over tens of micrometers through the network of osteoblasts and osteocytes, and finally diffusive transport over the last one or two microns.

Phase Segmentation of Ni/YSZ Anode for Solid Oxide Fuel Cells by Deep Learning

A deep learning network is implemented for the phase segmentation of Ni/YSZ anode of solid oxide fuel cells (SOFCs), as manual segmentation of focused ion beam-scanning electron microscopy (FIB-SEM) images is time-consuming. Segmentation is performed on two samples with different volume ratios of Ni to YSZ. The mean intersection over union (mIoU) reaches 0.9363, indicating good agreement between the predicted images and the manually segmented images. Furthermore, the impact of image augmentation on segmentation accuracy is investigated, and it is found that augmentation is necessary when the number of images in the training set is small or the training and testing sets come from different samples. Furthermore, mixing a sixth of manually segmented images from a new sample into the training set can significantly improve the accuracy of automated segmentation for the new sample. This strategy reduces the number of images requiring manual segmentation and improves processing efficiency.