335598 research outputs found
Sort by
Tuning Desolvation Kinetics with Perovskite‐Type Ion‐Conductive Modulators toward Low‐Temperature Zn Metal Batteries
No full textAqueous zinc metal batteries (AZMBs) are regarded as the promising candidates for low-cost, sustainable, but safe energy storage systems. Unfortunately, Zn metal anodes suffer from incomplete desolvation and random dendrite formation, which is attributed to sluggish diffusion kinetics resulted from the strong ion (Zn)-dipole (HO) interactions. Herein, to promote the Zn desolvation and diffusion kinetics, the strategy of constructing perovskite-type ion-conductive kinetic modulators of ZnSn(OH) is initially designed and coated on the Zn metal anode (PIC-ZSH@Zn), regulating ion behaviors against dendrite growth and side reactions of active water. As confirmed by theoretical simulations, COMSOL, time-of-flight second-ionic mass spectroscopy, Raman and various electrochemical analyses, the abundant active sites synergistically weaken Zn-HO interactions to accelerate desolvation to release free Zn, effectively homogenizing the Zn flux distribution to preferentially nucleate and plate metallic Zn. Consequently, the as-fabricated cell maintains reversible stability of 800 h at 10 mA cm with high Coulombic efficiency over 99% under low temperature of 0°C. The paired full cell with PIC-ZSH@Zn presents a high-capacity retention of nearly 80% after 1000 cycles at 1.0 A g at 0°C, reinforcing the operation robustness of AZMBs under low temperature environments
The Influence of Electrolyte Formulation on Gas Evolution in Sodium‐Ion Batteries with NaMnO₂ Cathode
No full textSodium-ion batteries (SIBs) are considered a promising alternative to lithium-ion batteries due to the high availability of sodium resources. Among the various candidates for the positive electrode, layered (O3-type) NaMnO2 has attracted considerable attention. However, understanding of its interfacial stability remains limited. Differential electrochemical mass spectrometry (DEMS) is a powerful tool for monitoring gas evolution and therefore provides valuable insights into side reactions occurring at the interface between anode/cathode and electrolyte. In this work, the gassing behavior of SIB half-cells with NaMnO2 cathode and six representative electrolyte formulations is investigated using DEMS. The results show that electrolytes with fluoroethylene carbonate effectively suppress parasitic reactions and promote the formation of passivating interphases, resulting in improved performance and limited gas release. PC-based electrolytes appear to be more stable than EC-based electrolytes, especially in combination with NaClO4. The use of NaPF6 is associated with increased H2 evolution and possible manganese dissolution, thereby impairing interfacial stability and releasing more lattice oxygen. An increase in the upper cutoff potential enhances gas release, indicating more severe (electro)chemical oxidation of the electrolyte. Overall, this study paves the way for new strategies for tailoring electrolytes to improve the cyclability and safety of SIBs
Training Neural Networks by Optimizing Neuron Positions
No full textThe high computational complexity and increasing parameter counts of deep neural networks pose significant challenges for deployment in resource-constrained environments, such as edge devices or real-time systems. To address this, we propose a parameter-efficient neural architecture where neurons are embedded in Euclidean space. During training, their positions are optimized and synaptic weights are determined as the inverse of the spatial distance between connected neurons. These distance-dependent wiring rules replace traditional learnable weight matrices and significantly reduce the number of parameters while introducing a biologically inspired inductive bias: connection strength decreases with spatial distance, reflecting the brain’s embedding in three-dimensional space where connections tend to minimize wiring length. We validate this approach for both multi-layer perceptrons and spiking neural networks. Through a series of experiments, we demonstrate that these spatially embedded neural networks achieve a performance competitive with conventional architectures on the MNIST dataset. Additionally, the models maintain performance even at pruning rates exceeding 80% sparsity, outperforming traditional networks with the same number of parameters under similar conditions. Finally, the spatial embedding framework offers an intuitive visualization of the network structure
From fine to giant: multi-instrument assessment of the dust particle size distribution at an emission source during the J-WADI field campaign
No full textMineral dust particles emitted from dry, uncovered soil can be transported over vast distances, thereby influencing climate and environment. Its impacts are highly size-dependent, yet large particles with diameters dp>10 µm remain understudied due to their low number concentrations and instrumental limitations. Accurately characterizing the particle size distribution (PSD) at emission is crucial for understanding dust transport and climate interactions.
Here we characterize the dust PSD at an emission source during the Jordan Wind Erosion and Dust Investigation (J-WADI) campaign, conducted in Wadi Rum, Jordan, in September 2022, focusing on super-coarse (10 62.5 µm) particles. This study is the first to continuously cover the full range of diameters from dp=0.4 to 200 µm at an emission source by using a suite of aerosol spectrometers with overlapping size ranges. This overlap enabled a systematic intercomparison and validation across instruments, improving PSD reliability.
Results show significant PSD variability over the course of the campaign. During periods with friction velocities (u) above 0.22 m s (or ∼ 3.3 m s threshold 4 m wind speed), the approximate threshold for local dust emission by saltation, both dust concentrations and the contributions of super-coarse and giant particles typically increased with increasing u, especially under neutral to unstable atmospheric stability conditions. These large particles accounted for about 90 % of the total mass concentration during the campaign. A prominent mass concentration peak was observed near d=60 µm in geometric diameter. While particle concentrations for d<10 µm showed good agreement among most instruments, discrepancies appeared for larger d due to reduced instrument sensitivity at the size range boundaries and sampling inefficiencies. Despite these challenges, physical samples collected using a flat-plate sampler largely confirmed the PSDs derived from the aerosol spectrometers. These findings help to advance our understanding of the dust PSD and the abundance of super-coarse and giant particle at emission sources
Functional morphology of the leg musculature in the marine seal louse: adaptations for high-performance attachment to diving hosts
No full textThe seal louse (Echinophthirius horridus) is a remarkable example of evolutionary adaptation, thriving as an obligate ectoparasite on deep-diving marine mammals under extreme environmental conditions, including high hydrostatic pressure, extreme drag force, salinity, and fluctuating temperatures. To investigate the anatomical and functional specializations enabling this lifestyle, we compared the leg morphology and musculature of E. horridus with its terrestrial relative, the human head louse (Pediculus humanus capitis), using synchrotron-based 3D microtomography and confocal laser scanning microscopy. Our findings reveal that the seal louse has developed a highly compact and robust leg structure with a fused tibiotarsus, an additional set of leg muscles, and a shortened claw tendon—an unprecedented adaptation among insects. These features allow for greater force transmission and reduced metabolic cost during sustained attachment. Behavioral assays further show that E. horridus can only move effectively on hair-like substrates, underscoring its complete reliance on host fur. These findings suggest a highly specialized muscular control system enabling strong, reliable, and reversible attachment in a challenging aquatic environment
Anomaly Detection for Autonomous Driving
No full textWith small fleets of autonomous vehicles of SAE level 4, i.e., such without a safety driver, publicly available, the adoption of autonomous vehicles will only continue to increase. Embedded within shared mobility solutions, this technical advancement can lead to a more sustainable, safe, and comfortable future. Scaling autonomous vehicles more broadly, however, requires handling a wide variety of challenging scenarios, especially those with often rare anomalies. With rising fleet sizes, such scenarios appear with increasing frequency. As many Machine Learning systems follow a closed-world assumption based on a set of known classes, such unknowns remain challenging.
This dissertation addresses anomaly detection for autonomous driving from a holistic perspective, contributing to the generation of scenarios with anomalies, the detection of anomalies, and the handling of anomalies. The first part addresses external anomalies, i.e., such that occur in the environment. Generating scenarios involves providing normal data to train models and creating scenarios with anomalies to evaluate anomaly detection methods. Based on a theoretical systematization of anomalies from the literature, scenarios from all anomaly layers can be created. As generating such external anomalies is often dangerous or infeasible, data is provided through a simulation engine. Based on these scenarios, an anomaly detection method is presented, which is trained on unlabeled sensor data alone. It leverages a world model as a representation of normality, utilizing both camera and LIDAR data. Once detected, anomalies can be integrated into the training process of Neural Networks, removing their status as anomalies. The presented approach handles previously detected anomalies where controlled traffic rule exceptions are required. To achieve this, a situation-aware reward for Reinforcement Learning is introduced.
Next to challenges induced by external anomalies, the driving task can be equally impacted by internal anomalies, such as model failures. This dissertation contributes to the field of internal anomaly detection by detecting model failures without the need for labeled evaluation sets. This is achieved by analyzing the disagreements between two models trained on the same task, but with different learning paradigms. Based on real-world data, the method successfully reveals categorical model failures, most often in seemingly normal situations.
Summarizing, this dissertation presents a holistic set of contributions to the field of anomaly detection for autonomous driving, addressing the generation, detection, and handling of anomalies. This is emphasized by examining both internal and external anomalies
Mutation T9I in Envelope confers autophagy resistance to SARS-CoV-2 Omicron
No full textOmicron has emerged as the most successful variant of SARS-CoV-2. In addition to mutations in Spike that mediate humoral immune escape, the Omicron-specific Envelope (E) T9I mutation has been associated with increased transmission fitness. However, the underlying mechanism remained unclear. Here, we demonstrate that the E T9I mutation confers resistance to autophagy. Rare Omicron patient isolates encoding the ancestral E T9 remain sensitive to autophagy. Conversely, introducing the E T9I mutation in recombinant 2020 SARS-CoV-2 renders it resistant to autophagy. Our data indicate that the E T9I mutation protects virions against lysosomal degradation. At the molecular level, the T9I mutation increases the localization of E at autophagic vesicles and promotes interaction with autophagy-associated proteins SNX12, STX12, TMEM87B, and ABCG2. Our results show that the E T9I mutation renders incoming virions resistant to autophagy, suggesting that evasion of this antiviral mechanism contributes to the efficient spread of Omicron
