119092 research outputs found
Sort by
Responsible Evaluation of AI for Mental Health
No full textAlthough artificial intelligence (AI) shows growing promise for mental health care, current approaches to evaluating AI tools in this domain remain fragmented and poorly aligned with clinical practice, social context, and first-hand user experience. This paper argues for a rethinking of responsible evaluation -- what is measured, by whom, and for what purpose -- by introducing an interdisciplinary framework that integrates clinical soundness, social context, and equity, providing a structured basis for evaluation. Through an analysis of 135 recent *CL publications, we identify recurring limitations, including over-reliance on generic metrics that do not capture clinical validity, therapeutic appropriateness, or user experience, limited participation from mental health professionals, and insufficient attention to safety and equity. To address these gaps, we propose a taxonomy of AI mental health support types -- assessment-, intervention-, and information synthesis-oriented -- each with distinct risks and evaluative requirements, and illustrate its use through case studies
The drastic influence of the TiO₂ polymorph on the formation of core–shell structures in Na₀.₅ Bi₀.₅ TiO₃ –25SrTiO₃
No full textThe formation of Na₀.₅ Bi₀.₅TiO₃–25SrTiO₃ solid solutions is highly sensitive to variations in the oxygen vacancy concentration, which strongly influence the Sr²⁺ and O²⁻ interdiffusion during calcination. In addition to conventional approaches such as adjusting the Bi₂O₃ content or doping, selecting the TiO₂ polymorph provides an additional strategy to control oxygen vacancies and with that microstructure and functional properties of ceramics. XRD revealed that rutile TiO₂ promotes the formation of intermediate Bi₁₂ TiO₂₀ , thereby suppressing Bi₂O₃ sublimation, and reducing the final oxygen vacancy concentration. This yields pure Na₀.₅Bi₀.₅TiO₃ core fractions of up to 16.9%, surrounded by SrTiO₃ -rich shells. The core–shell formation correlates with an increased positive strain from 0.21% to 0.27%. Electrical properties, however, are more dependent on the oxygen vacancy content than core–shell presence. Thus, anatase-based samples, benefiting from faster oxygen diffusion, exhibit more homogeneous microstructures with fewer core–shell structures and up to three orders of magnitude higher oxygen ion conductivity than rutile-based ceramics. Interestingly, the opposite effect was found for pure NBT, highlighting the dependence on the solid solution. Overall, the results demonstrate that the TiO₂ polymorph selection is a powerful strategy for effectively controlling reaction pathways, defect chemistry, and the resulting functional properties
Synthesis and thermal properties of dense Si─Al─C─N‐based polymer‐derived ceramics
No full textIn the present study, aluminum‐modified preceramic silicon polymers were synthesized via chemical modification of a commercially available organopolysilazane using an aluminum amido complex. The incorporation of aluminum into the polymer structure and its effect on ceramization behavior and processability as well as the thermal and mechanical properties of the obtained Si─Al─C─N ceramic materials were systematically investigated. The microstructure and chemical composition of the dense, predominantly amorphous monolithic Si─Al─C─N were characterized using solid‐state nuclear magnetic resonance, scanning electron microscopy, and transmission electron microscopy. Aluminum incorporation led to enhanced densification of Si─C─N, resulting in monoliths with a porosity fraction as low as 2 vol% achieved for the composition with the highest aluminum content. Al incorporation was shown to also result in a significant reduction in thermal conductivity, thus, Si─Al─C─N formulations exhibited values as low as 0.6 W m −1 K −1 . The hardness and Young's modulus remained nearly unchanged upon aluminum incorporation, with values of ca. 14.5 and 156 GPa, respectively, for the high‐aluminum‐content sample. To the best of our knowledge, the present study reports for the first time on the thermal and mechanical properties of dense, mainly amorphous Si─Al─C─N ceramics, highlighting the suitability of Si─Al─C─N as excellent material for possible thermal insulation at temperatures beyond 1000°C
Balancing climate impact, resource use, and environmental safety: assessment of treatment options for lithium-ion battery recycling wastewater
No full textRecycling of lithium-ion batteries (LIBs) generates wastewater containing persistent ions, such as hexafluorophosphate (PF6-) and bis(fluorosulfonyl)imide (FSI-). Without targeted treatment, these salts can enter surface waters and pose environmental risks. To address this, a process combining nanofiltration, electrochemical oxidation, and ion exchange was developed. In addition to pollutant removal, the process enables recovery of potassium hexafluorophosphate (KPF6), a salt used in glass production, metallurgy, and the pharmaceutical industry.
Experimental investigations determined key design parameters for ion exchange process, including kinetics, resin stability, and suitable regenerants. The combined treatment achieved compliance with current discharge limits and produced analytical grade KPF6. Life Cycle Assessment (LCA) results for global warming (GWP) and abiotic depletion potential (ADP) highlighted clear benefits of recovery, primarily from reducing hazardous-waste incineration.
Economically, levelized treatment costs (LTC) amounted to 183 €/m³ without recovery, and decreased to 95 €/m³ when recovery was included, largely by lowering disposal needs. Marketing the recovered salt generated revenue, yielding a positive gross profit and making the process competitive with the benchmark evaporation, costing 88 €/m³ (without benefits) and 50€/m³ (with benefits).
Overall, integrating resource recovery into wastewater treatment improves both environmental and economic outcomes. Strengthening the regulatory framework by introducing discharge limits for total fluorine would enable more comprehensive monitoring and reduce the release of persistent salts from LIB-recycling effluents into surface waters
Development and analysis of a battery degradation model for use in mixed-integer linear programming
No full textThe transition towards electromobility in conjunction with increasing bidirectional charging is also placing greater focus on battery aging. Given the popularity of mixed-integer linear programming (MILP) in analyzing energy systems, there is a significant need for suitable battery degradation models. However, the complicated battery aging processes are often highly simplified in the literature. Therefore, we develop and evaluate a more suitable and accurate model. For this purpose, an existing non-linear continuous battery degradation model is discretized and piecewise linearized. The discretization error is kept below , while the additional linearization causes a maximum deviation of for calendar and for cyclic degradation. To further evaluate the model and its variants, it is implemented in an optimization problem representing a residential bidirectional charging use case. Distinct State-of-Charge (SOC) modeling for calendar degradation results in solving times that vary significantly (minutes to days), while this variation is less severe for different cyclic temperature variants. Although the effect on the optimal operating strategy is only limited, the general neglect leads to considerable differences, especially regarding the dominating calendar degradation. The concluding comparison with a frequently used energy throughput method shows an increase in solving time (seconds vs. minutes), but also significant differences in operating strategy. The developed model leads to lower SOCs due to calendar degradation, and provides a more accurate cyclic degradation and economic evaluation of bidirectional charging, offering the possibility of a more adequate integration of battery degradation and its effects into MILP models
Implementation of a digital twin in additive manufacturing of copper — methodology, implications, and future prospects
No full textDigital twins are increasingly being used to visualize, analyze, or control physical processes and systems. Implementation currently poses challenges for users due to the cross-domain complexity of digital twins. In this study, the authors utilize a self-developed method to support the implementation of a digital twin (DT) for a powder bed fusion additive manufacturing system (PBF-LB/M) for copper components, utilizing a green laser. The study highlights the support offered by the developed approach and the implications of using DTs for PBF of copper. The DT focuses in particular on monitoring maintenance requirements, assisting in the selection of correct process parameters, and alerting plant operators in the event of problems. In addition, a process model focused on lack of fusion is implemented, based on earlier studies. In the human–machine system, DTs thus represent a further building block towards an improved process stability in PBF-LB/M of copper, and the method used lowers the barrier to entry for widespread use of DTs
Soft and Rigid Wetting at Small Scales
No full textThe present work aims to extend the understanding of the physics of wetting phenomena at small scales for statics and dynamic processes. The cumulative thesis consists of three chapters. In the first chapter, we give a brief overview of the physics of wetting and the mathematical machinery it employs. Then, in the second chapter, we summarize the research findings and provide explanations regarding how research works included in the third chapter are connected and amalgamated to answer the research question. Finally, the third chapter contains the research works.
The main question addressed in the thesis is how droplets and films/rivulets spread and evaporate on complex solids and imbibe in the elements of their surface structures. A particular focus of the thesis is in gaining the further comprehension of the effect of the surface forces – considered in the frame of the disjoining pressure concept in the present work – on droplet and film/rivulet statics and dynamics. For that, we study behavior of five systems: (1) nanodroplet statically wetting a rigid solid; (2) nanodroplet spreading over a rigid solid; (3) nanodroplet resting and spreading on a deformable solid; (4) liquid nanorivulet in a steady-state in rigid and soft wedges; (5) contact line of a volatile droplet or rivulet pinned to a protruding wedge-shaped non-uniformity of the substrate topography.
We showed that contact angle of a droplet drastically depends on the droplet size as long as the height of the droplet is comparable with the range of the surface forces. We analyzed the spreading behavior of nanodroplets over thick wetting films and found that at the nanoscale, neither Tanner, nor Cox-Voinov laws hold. In the course of spreading, the ripples are shown to be formed at the advancing contact line, the depth of which was also shown to be affected by the surface forces in the wetting film.
When the substrate is soft, it responds to the presence of a liquid by the deformation. We investigated how surface forces influence the shape of a wetting ridge formed in the vicinity of the contact line of a nanodroplet and how softness of the substrate affects the dynamics of spreading. We demonstrated that the evolution of the wetting ridge height while spreading is neither self-similar, nor monotonic.
The Young-Laplace equation is known to have unbounded solutions or to not have any solutions for a rivulet in a solid wedge in the case when the Concus-Finn condition on a liquid contact angle and a wedge opening angle is satisfied. We showed that when the surface forces are taken into account and, hence, the Young-Laplace equation is augmented with the disjoining pressure term, the steady-state solutions exist. The shape of the rivulet was shown to depend on the wedge geometry as well as on the surface force parameters. Besides, we have considered the case when the walls of the wedge are deformable and demonstrated how wedge softness, geometry as well as the thickness of a wetting film on the wedge's walls affect its deformation.
In many cases, when spreading over a substrate with a complex surface morphology, the contact line of a liquid droplet or rivulet pins to sharp peculiarities thereof. A straight wedge makes a simple portrait of such physical solid's non-homogeneity. We considered a contact line region of a volatile liquid to built an analytical model for the local flows in both liquid and gas phases. The flow in the gas phase is the Stefan flow. We performed parameter study to understand how different factors such a wedge opening angle and a ratio of fluid and gas viscosities affect the flow topology
AICD Bench: A Challenging Benchmark for AI-Generated Code Detection
No full textLarge language models (LLMs) are increasingly capable of generating functional source code, raising concerns about authorship, accountability, and security. While detecting AI-generated code is critical, existing datasets and benchmarks are narrow, typically limited to binary human-machine classification under in-distribution settings. To bridge this gap, we introduce , the most comprehensive benchmark for AI-generated code detection. It spans , across , and , including recent reasoning models. Beyond scale, AICD Bench introduces three realistic detection tasks: ()~ under distribution shifts in language and domain, ()~, grouping generators by architectural lineage, and ()~ across human, machine, hybrid, and adversarial code. Extensive evaluation on neural and classical detectors shows that performance remains far below practical usability, particularly under distribution shift and for hybrid or adversarial code. We release AICD Bench as a to drive the next generation of robust approaches for AI-generated code detection. The data and the code are available at this https URL}