Publikationen

Browse

Now showing 1 - 20 of 5607

Flexible process variant binding in information systems with software product line engineering
(Universität Ulm, 2025-06-19) Hehnle, Philipp; Reichert, Manfred
Different organisations often run similar digitised business processes to achieve their business goals. However, organisations often need to slightly adapt the business processes implemented in an information system in order to adopt them. Various approaches have been proposed to manage variants in process models. While these approaches mainly deal with control flow variability, in previous work we introduced an approach to manage implementation variants of digitised business processes. In this context Software Product Line (SPL) Engineering was applied to manage a set of common core artefacts including a process model from which Process-Aware Information Systems (PAIS) can be derived, which differ in the implementation of their process activities. When deriving a PAIS, implementations are selected for each process activity and then included in the PAIS at compilation time. One challenge that has not yet been solved is giving users of digitised business processes the option of selecting multiple implementations at runtime. This paper extends our previous work by not only allowing for the selection of activity implementations at compile time, but also at start time and runtime. Consequently, it becomes possible to defer the decision as to which implementation should be selected to start time and runtime. Furthermore, multiple implementations of a particular activity may be selected and executed concurrently. The presented approach also allows customising the input and output data of activities. Data from expert interviews with German municipalities suggests digitising business processes with varying implementations is a widespread challenge and our approach is a way to mitigate it.
Investigation of expansion and potential of Si-dominant anodes with binder modification in full cells
(Universität Ulm, 2025-06-16) Kloker, Gabriele; Anjass, Montaha; Brauchle, Felix; Rutz, Daniel; Vrankovic, Dragoljub
Silicon as anode material in lithium-ion batteries shows immense volume change during (de-)lithiation. This swelling is an important factor to consider when developing Si anodes for practical applications, but also the utilization and behavior of both electrodes. Using Si capacity only partially results in less volume change, thus mitigating degradation. The impact of cell balancing and voltage window for limitation of capacity is investigated in full pouch cells via reference electrode and expansion measurement setup. Therefore, this work shows Si-dominant anodes with to different extends partially neutralized polyacrylic acid (PAA), cycled in normal and mild voltage ranges to reveal the impact on anode behavior and expansion. PAA binder is beneficial for silicon-containing anodes and neutralization of PAA improves the performance of this binder. In this work, it is shown that binder modification also has a large impact on anode coating expansion and silicon utilization in a voltage range of 3.0–4.2 V. Differences in performance between PAA neutralization degrees are not as prominent in case of 3.3–4.2 V voltage window. The reduced utilization of Si in this case leads to an over five times higher cycle life for all binder variations, at the same time sacrificing capacity.
Low-Loss Frequency Selective Surface for Sub-Millimeter Wave Radiometer Applications
(Universität Ulm, 2024-10-31) Kienle, Veronika; Ettorre, Mauro; de Sagazan, Olivier; Sauleau, Ronan; Waldschmidt, Christian; Chaloun, Tobias
This contribution provides the analysis and design of a low loss, multi-layer frequency selective surface (FSS) on a fused quartz substrate for radiometer applications at sub-millimeter wave frequencies. The multi-layer FSS generates a bandpass filter response and consists of three metallic FSS layers and two fused quartz substrates. The design parameters like substrate heights and sheet impedances of the FSS layers are deduced by modeling the proposed structure as a second-order Butterworth bandpass filter (275 GHz center frequency and bandwidth of 20 GHz) with two serial half-wavelength resonators and three K-inverters. The FSS employs an inductive metal mesh with a unit cell size of 200 μm constructed by a micro-machining lift-off process. Transmission measurements conducted in the 220 GHz-330 GHz range with a quasi-optical setup verified an in-band insertion loss lower than 0.2 dB.
A Study on Molding Broadband Integrated Antennas for a G-Band Radar
(Universität Ulm, 2025-12-30) Kienle, Veronika; Weißer, Marina; Hitzler, Martin; Matt, Felix; Bord, Robin; Waldschmidt, Christian
This paper examines the impact of Glob Top encapsulation on the radiation performance and mutual coupling of closely spaced transceiver MMICs with a hybrid antenna system for a G-band radar. This hybrid antenna system consists of an integrated patch antenna, a parasitic patch resonator on a glass superstrate and an external dielectric rod (DR). The study reveals that while mechanical protection is provided by the encapsulation, its effect on electromagnetic characteristics is significant. Simulations and measurements reveal that radiation efficiency decreases with an encapsulation. Additionally, ripples in the gain pattern get stronger as encapsulation height increases, primarily due to multi-path interference in the near field of the external DR. Despite these effects, encapsulation has hardly any influence on the coupling between the MMICs, which is approximately -30 dB. These results provide insight into the electromagnetic impact of protective coatings, which is critical for optimizing the performance of mmWave and sub-THz radar systems.
Wall Analysis of Buildings Using a Compact Multi-Band Radar System with a Common Phase-Center Antenna
(Universität Ulm, 2026-07-29) Riese, Nico; Kienle, Veronika; Hitzler, Martin; Waldschmidt, Christian
A compact multi-band radar system operating from 24.05–26.5GHz, 76–84GHz and 167–182GHz using three FMCW radar chips is presented. This system concept is enabled by a PCB-based common phase-center antenna, which illuminates a single dielectric lens. The antenna performance is evaluated by means of gain and radiation pattern measurements. A multitude of wall structures are analyzed using the proposed multi-band radar system. This demonstrates the versatility and robustness of the multi-band approach for the application of building analysis.
Compensating thickness effects in micro X-ray fluorescence spectroscopy using integrated optical microscopy for thickness determination of soft matter block copolymer membranes
(Universität Ulm, 2025-12-19) Müller, Riccarda; Weckenmann, Leon; Aslanova, Nigar; Gupta, Yesleen; Schacher, Felix H.; Streb, Carsten; Leopold, Kerstin
A non-invasive method using the integrated optical microscope of a laboratory-based 2-dimensional micro X-ray fluorescence spectroscopy (2D µXRF) instrument to determine the thickness of soft matter samples has been successfully developed, validated, and applied. This easy-to-use method is applicable to soft matrices in a thickness range from 25 to 1000 µm. The main advantage of this method is that thickness determination is directly related to the physical thickness of the sample, rather than its optical thickness, and it does not affect the sample structure, i.e., it does not require any drying or embedding. Elemental composition and distribution analysis by 2D µXRF can be performed on the same sample specimen, using the same setup directly before or after thickness determination. Knowledge of the thickness of different samples can be used to normalize elemental intensities to compensate for the mass-thickness effect, enabling reliable comparison of samples of different thicknesses. By systematically probing several points on a sample surface, this approach can also be used to correct the elemental intensity of unevenly thick soft matter samples. The aspect of correcting element intensities based on the determined thickness should enable direct quantification of element contents using µXRF and external calibration in future work.
Cryo-EM observation of AA amyloid fibrils in mouse model of systemic AApoAII amyloidosis
(Universität Ulm, 2025-09-11) Andreotti, Giada; Higuchi, Keichii; Schmidt, Matthias; Fändrich, Marcus
The co-deposition of amyloid fibrils from different precursor proteins is a topic of increasing relevance for protein misfolding diseases. Using cryo-electron microscopy (cryo-EM), we here determined the structures of two serum amyloid A (SAA) protein-derived amyloid fibril morphologies that were extracted from a mouse strain that is primarily known to be associated with apolipoprotein A-II-derived amyloid fibrils. The two fibril morphologies show the same protomer conformation as in previously reported ex vivo amyloid fibrils from SAA protein but a different relative arrangement of fibril protein stacks. These data establish that serum amyloid A-derived amyloid fibrils share the same fibril protein fold in different mouse strains and disease contexts.
Advancing ionic liquid-based electrolytes for rechargeable magnesium batteries
(Universität Ulm, 2026-08-25) Elkhafif, Omar; Jacob, Timo; Krause, Simon
Rechargeable magnesium batteries (RMBs) are emerging as a promising alternative to Li-ion batteries due to their high theoretical capacity and being one of the most abundant elements in Earth’s crust, leading to low raw material cost and wide availability.1 However, the development and commercialization of RMBs require addressing several challenges related to their anode, cathodes, and electrolytes. Mg anode is prone to surface passivation and poor reversibility; cathodes suffer from slow Mg2+ sluggish diffusion as well as interfacial instability; and electrolytes show a narrow electrochemical window, high overpotentials, and incompatibility with both electrodes.2,3 Since the electrolyte plays a vital role in the development of RMBs, finding a suitable electrolyte is essential to address most of the previously mentioned challenges. However, several obstacles related to the proper selection of Mg salts as well as a suitable solvent hinder their development. For instance, electrolytes containing salts with anions such as hexafluorophosphate, perchlorates, trifloromethanesulfonate often show poor compatibility with magnesium metal. These anions either react with the Mg electrode or decompose on its surface, forming a passivation layer. Furthermore, non-aqueous solvents such as nitriles, esters, amides, carbonates, and sulfones tend to decompose on the highly reactive Mg anode, resulting in its passivation.4 The products of electrolyte decomposition tend to form Mg dendrites.5 Herein, key parameters such as Mg2+ diffusion kinetics, safety, chemical and electrochemical stability require further consideration. Ionic liquids (ILs) are emerging in battery applications due to their unique properties, including a wide electrochemical window, thermal stability, low vapour pressure, and high ionic conductivity.6 Nevertheless, the deployment of ILs in Mg batteries is accompanied by other challenges, primarily arising from impurities such as water residues introduced by the ionic liquids’ hygroscopic nature or contaminants resulting from their synthesis process. From this perspective, hydrophobic ionic liquids (1-methyl-1-propylpiperidinium bis(trifluoromethylsulfonyl)imide (MPPip-TFSI) and 1-butyl-1methylpyrrolidinium bis(trifluoromethylsulfonyl)imide (BMP-TFSI) have been systematically studied. Their drying process has been studied over different molecular sieves of different pore diameters. Then their electrochemical behavior towards Au(111) was characterized through cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS). Moreover, the MPPip-TFSI/Tetraglyme system with 0.1 M Mg(TFSI)2+ 0.01 M Mg(BH4)2 probed the Mg deposition/dissolution in the presence and absence of a trace amount of water. Nevertheless, even rigorously dried ILs, Mg deposition from pure IL and 0.1 M Mg(TFSI)2 takes place at high overpotential and is accompanied by electrolyte decomposition. Therefore, a systematic three-stage electrolyte optimization was carried out. Firstly, different co-solvents with different ratios were explored, followed by the addition of 0.01 M Mg(BH4)2. The subsequent step involved the optimization of IL:co-solvent ratios based on different electrochemical parameters, including Mg deposition overpotential, coulombic efficiency, cathodic and anodic peak currents. Thereafter, Mg deposition/dissolution was confirmed through structural characterization using SEM, EDS, and XRD. The optimized electrolyte was further validated in a full cell using WS2-PANI cathode. In addition to the electrolyte optimization, the interfacial stability of Mg anodes remains a critical challenge in the context of dendrite formation. Since Mg dendrites have been considered unlikely to form compared to Li dendrites, recent studies indicate that Mg dendrites can indeed be formed under certain conditions. To resolve these discrepancies, Mg|Mg symmetric cell and Mg|TiS2 asymmetric cell with the optimized electrolyte, abbreviated as MIDS electrolyte, were systematically investigated under different current densities. Its electrochemical behavior was studied by cycling voltammograms, galvanostatic cycling, and electrochemical impedance spectroscopy. Furthermore, ex-situ SEM/EDS complemented by operando optical microscopy exhibited the impact of low and high current densities, particularly the growth mechanism and kinetics. In the Mg|TiS2 full cell, the influence of low and high current rates on the cyclic stability, capacity retention, and coulombic efficiency was also tested. Detailed mechanisms of Mg dendrite formation under a certain current range, as well as the cell self-healing, are provided in this study. Taken together, this cumulative work enhances the understanding of electrolyte design, electrode-electrolyte interface, and dendrite formation and healing mechanisms. By elucidating the coupled effect of impurities, electrolyte composition, and current density, this thesis presents deep insights into the fundamental challenges in Mg batteries and paves the road towards the design of a stable and efficient Mg battery.
Population sizing and optimal restart strategy for evolution strategies in highly multimodal landscapes
(Universität Ulm, 2026-08-25) Schönenberger, Lisa; Beyer, Hans-Georg; Kestler, Hans; Meyer-Nieberg, Silja
Evolution Strategies (ES) are an optimization technique inspired by natural evolution. The evolutionary process of ES is based on the principles of mutation, selection, and recombination, whereby a population of potential solutions is adapted in each generation. ES can be applied to a wide range of optimization problems. In particular, ES has demonstrated a remarkable efficacy in continuous, highly multimodal optimization problems. However, in the context of high multimodality, ES typically requires a sufficiently large population size to successfully locate the global optimum. Conversely, excessively large population sizes result in increased computational costs. This raises the question of the minimum population size required to achieve a satisfactory success probability. A straightforward method to execute ES without knowledge of the unknown optimal population size is to implement restart strategies. In the context of restart strategies, the current search process is regularly stopped, and the optimization algorithm is restarted. In the case of ES, the population size changes following each restart. Since each restart requires resources, too many restarts should be avoided. However, the required resources increase with the population size. It is therefore also important to ensure that the optimal population size is not exceeded by a margin too large during the final restart. The objective of this research is to determine the optimal rule for adjusting the population size λ. To this end, a loss function is introduced to quantify the wasted computational cost relative to the optimal strategy. A crucial requirement for any efficient restart strategy is that its loss, relative to the optimal λ, remains bounded. It will be demonstrated that not all strategy types are bounded. However, for a particular strategy type, in which λ is increased multiplicatively by a constant factor ρ, the relative loss function is bounded. Furthermore, it will be demonstrated that within this strategy type, there exists an optimal value ρ = 2 that minimizes the maximum relative loss. This doubling strategy has already been used in many implementations of evolutionary algorithms to control the population size. Acceptable performance has been demonstrated in several applications. However, a formal proof of optimality is presented and the underlying conditions are discussed for the first time within the scope of this work. Another question that arises concerns the minimization of the average relative loss of a multiplicative restart strategy. To this end, the optimal restart parameter that minimizes the maximum average relative loss will be calculated, resulting in an optimal ρ that lies in the interval [2.5,3.5] and is larger than the commonly chosen value of ρ = 2. In the context of examining restart strategies, the knowledge of the optimal population size was not a prerequisite. A subsequent inquiry concerns the value of the optimal population size and how this value scales with the dimension. A model will be presented that simplifies the complex structure of the examined multimodal functions by interpreting local landscape oscillations as frozen noise. This approach will allow conclusions to be drawn from the established theory of noisy landscapes to the more complex original functions. The model will be implemented on a selected number of well-known multimodal test functions, namely Rastrigin, Bohachevsky, Ackley, and Griewank. For the Rastrigin and Bohachevsky function, the population size scales with O(√N ln(N)), where N denotes the problem dimension. Another key finding is that for the Ackley function, the scaling behavior depends strongly on the initial values. If the algorithm starts in a certain vicinity of the global optimizer, the dependence on the dimension N is rather weak. However, if the initial value exceeds a certain distance R to the optimizer, the population size scales exponentially with R. The Griewank function is known for its counterintuitive behavior of getting simpler to be optimized with increasing dimension, although the number of local minima increases with the problem dimension. This fact can also be explained by applying the frozen noise approach. Overall, the study results in the observation that in many instances the population size scales sublinearly with the problem dimension. This contrasts with gradient-based optimization techniques, which require multi-start strategies in the presence of multimodality, resulting in exponential cost increases with the problem dimension. Although this fact is already known from experiments, this thesis presents an initial approach to derive it analytically. The aforementioned study demonstrated that the optimal population size is contingent upon the underlying function. However, in general, the objective function is not known. One method for achieving success with moderate population sizes in highly multimodal landscapes is to use rescaled mutations. In this method, the search space is explored with a large mutation strength, but the step toward the new parental centroid is reduced by a factor κ. The population sizing equation previously mentioned will be expanded to include ES with rescaled mutations. The result of this analysis indicates that the optimal population size scales inverse linearly with κ. Consequently, even with a given moderate population size, a successful outcome can be achieved in highly multimodal landscapes by reducing the step toward the new parental centroid according to this scaling law.
Numerik zwischen Wissenschaftssoziologie und Fachdidaktik : Eine Interviewstudie zu den Charakteristiken der Disziplin und zu den Kompetenzen von Studierenden in Numerik-Veranstaltungen
(Universität Ulm, 2026-08-25) Burr, Laura; Urban, Karsten; Hochmuth, Reinhard
Substantielle und sinnvolle Fachdidaktik kann nur mit einem fundierten Verständnis der Fachwissenschaft einhergehen. In der vorliegenden Arbeit steht die Numerik deshalb zunächst selbst im Fokus und es wird untersucht, welche Merkmale und Aspekte die Disziplin charakterisieren und welche Aufgaben und Ziele sie verfolgt. Im fachdidaktischen Teil dieser Arbeit werden anschließend die Kompetenzen von Studierenden in Numerik-Veranstaltungen herausgearbeitet. Im Rahmen einer empirischen Untersuchung wurden insgesamt 17 Interviews mit Numerikern durchgeführt, aus denen die entsprechenden Charakteristiken und Kompetenzen rekonstruiert werden konnten. Die Interviews wurden mit der qualitativen Inhaltsanalyse nach Gläser und Laudel (2010) ausgewertet. Es konnten insgesamt neun charakteristische Merkmale und Aspekte in den Ergebnissen identifiziert werden. So wurde in den Interviews beispielsweise die Interdisziplinarität und die Anwendungsorientierung der Numerik beschrieben. Die befragten Numeriker haben außerdem zwölf unterschiedliche Kompetenzen der Studierenden in Numerik-Veranstaltungen hervorgehoben, zum Beispiel das Algorithmisieren und das Implementieren. Drei der identifizierten Kompetenzdimensionen werden am Ende dieser Arbeit exemplarisch anhand von Übungsaufgaben der Universität Ulm und der Humboldt-Universität zu Berlin rekonstruiert.
Crystallographic and electronic structure of tungsten diselenide and its heterostructures
(Universität Ulm, 2026-08-25) Leiter, Robert; Kaiser, Ute; Beránek, Radim
This dissertation deals with monolayers and vertical heterostructures based on tungsten diselenide (WSe2), which are studied using High-Resolution Transmission Electron Microscopy (HRTEM). Even under low-voltage conditions, Se atoms are continuously removed from the lattice and agglomerate to form increasingly complex defect structures. The whole "life cycle" of a WSe2 monolayer is reported under electron irradiation, revealing the formation and evolution of various defect structures. When stacking two layers of WSe2 under a controlled low twist angle, it was found that the two lattices undergo atomic reconstruction, leading to enlarged domains of homogeneous stacking. This was even found in some cases with heterostructures constructed with various combinations of WSe2 and other TMDs, despite large discrepancies in lattice parameters. Subsequently, exfoliated layers of Ca2Nb3O10 are studied, where atomic defects in freestanding monolayers are observed, and electron energy-loss spectroscopy also shows the changes in the excitons of tungsten diselenide when a vertical heterostructure is constructed with a monolayer of the perovskite on top.
Single-molecule tracking and kinetic analysis in living cells and multicellular organisms
(Universität Ulm, 2025-06-25) Gebhardt, J. Christof M.
In a living organism, the interplay of stochastically interacting molecules brings forth structures and processes robustly organized in space and time. Single-molecule localization microscopy and tracking emerged as important techniques to visualize the super-resolved spatial distribution and real-time motion of individual fluorescently labeled molecules in a living cell or multicellular organism. Thereby, single-molecule tracking (SMT) enables quantifying kinetic mechanisms underlying vital organismal processes. This review covers the methodology of SMT in living cells and multicellular organisms, including labeling approaches and microscopy techniques, with a focus on recent developments in temporal excitation patterns facilitating extracting kinetic properties and the analysis of molecular kinetic parameters accessible by SMT. Emphasis lies on the application of this methodology to quantifying the kinetics of proteins such as transcription factors, and some recent examples are highlighted. The review concludes by envisioning future perspectives of SMT in living systems.
Nanoscale electrochemical sensing of tramadol with a polyoxometalate-modified electrode
(Universität Ulm, 2025-10-09) Njiajo Talla, Donald Eric; Djuffo Yemene, Astree Lottie; Zambou Jiokeng, Sherman Lesly; Schüttler, Konstantin M.; Mbomekalle, Israel; Mizaikoff, Boris; Kenfack Tonlé, Ignas
Ensuring the accuracy and reliability of analytical methods is essential to ensure the efficacy and safety of pharmaceutical treatments. This study describes a rapid and cost-effective electroanalytical nanosensing platform for tramadol detection, an opioid analgesic widely used in clinical settings but prone to misuse. The sensing material is based on a polyoxometalate, [H4PW17V(IV)O62]9− (POM), which was successfully synthesized, then characterized by Fourier-transform infrared (FT-IR) spectroscopy, scanning electron microscopy (SEM), powder X-ray diffraction (PXRD), and X-ray photoelectron spectroscopy (XPS) which together allow to confirm its morphology, composition, crystalline structure, and surface chemical states. A glassy carbon electrode (GCE) modified with this POM (POM/GCE) was evaluated by cyclic voltammetry using [Fe(CN)6]3− and [Ru(NH3)6]3+ as redox probes, and square wave voltammetry was employed for tramadol detection in aqueous solution. Key experimental parameters including pH, suspension volume, frequency, and accumulation time were optimized to enhance the sensor sensitivity. Under optimal conditions (pH 3, 3 μL modifier, 25 Hz, no accumulation time), the sensor exhibited a linear response from 0.06 to 0.78 μM (R2 = 0.998), with a detection limit of 4.03 nM and quantification limit of 13.4 nM. The method showed a good reproducibility (RSD 2.2 %) and stability, and was successfully applied to the analysis of pharmaceutical formulations, yielding recovery rates between 97 and 101 %. In comparison to UV–Visible Spectroscopy, the proposed method demonstrated higher sensitivity, a lower detection limit, and improved selectivity for tramadol determination in pharmaceutical formulations. These advantages highlight its potential as a reliable alternative for routine analysis and drug quality control.
Investigations on electrochemical magnesium deposition and dissolution with ether-based electrolytes
(Universität Ulm, 2026-08-20) Schick, Benjamin W.; Jacob, Timo; Bresser, Dominic
Rechargeable Mg batteries are seen as a promising future battery technology. To develop electrolytes with improved properties for the electrochemical Mg deposition and dissolution, which is crucial for the realization of a successful Mg battery, the processes at the interface between the negative Mg electrode and the electrolyte need to be understood. To obtain deeper information on the structural evolution of possibly formed interphases, operando electrochemical quartz crystal microbalance with dissipation monitoring was applied. By systematically varying the electrolyte composition, conclusions on the influence of individual electrolyte components on interphase evolution were drawn. These conclusions were applied to develop novel electrolytes based on a solvent with beneficial properties compared to the commonly used ones.
Multimodal Recognition Architectures for microRNA Biosensing
(Universität Ulm, 2026-08-18) Abdelhamied Abdeltawab Abdelhamied, Muhammad; Kranz, Christine; Leopold, Kerstin
Reliable molecular diagnostics is essential for the early detection and therapeutic monitoring of major diseases, e.g., cancer and cardiovascular disorders. Within this context, microRNAs (miRNAs) have gained considerable attention, as their dysregulation has been documented in several pathological conditions, including cardiovascular disease, neurodegenerative disorders, and multiple cancer types. Their analysis has emerged as a promising strategy for cancer classification and prognosis. Yet, their small size, closely related sequences, susceptibility to degradation in biological matrices, and low physiological concentrations introduce significant challenges for their reliable detection and quantification. To date, most established detection strategies rely on complementary nucleic acid probes that recognize target sequences through base pairing. Although such systems can provide high sequence selectivity, their practical performance may be limited by probe degradation, desorption from sensor surfaces, and recognition mechanisms confined to base pairing alone, which restricts the diversity of molecular interactions available for target capture. Consequently, the development of alternative capture platforms with improved stability toward DNase/RNase degradation and broader interaction capabilities has become an important objective in biosensor research. Within this thesis, a hybrid biosensor based on molecularly imprinted polymers (MIPs) and peptide nucleic acids (PNAs) is introduced for sensitive and selective detection of miR 21 in complex samples such as artificial serum and RNA isolates from cancer cells (e.g., MCF-7 and HeLa). The polymer was synthesized by electropolymerization to embed PNA supported, pre-oriented miR 21 template molecules. PNAs are highly selective recognition elements for nucleic acids and play a dual role in this approach. Firstly, they serve as an organizing motif for the pre-arrangement of miR-21 during the polymerization process increasing the fraction of recognition sites with high template affinity. Secondly, PNA serves as an additional recognition element directly binding to miRNA-21 with high specificity as part of the binding moiety. In turn, MIPs complement this function by providing multifunctional recognition sites at the binding site for miR-21, while also serving as a protective polymeric jacket that stabilizes PNA. After optimization of the sensor architecture and experimental parameters, electrochemical impedance spectroscopy (EIS) was used as the readout. The measurements yielded a linear response from 0.5 to 5000 pM with a limit of detection of 0.11 ± 0.04 pM without any amplification step. The biosensor showed good selectivity for miR 21, even relative to a single mismatched sequence, and enabled quantification of miR 21 in artificial serum as well as in RNA isolates to discriminate between MCF 7 and HeLa cells. This cooperative PNA/MIP strategy represents the first application of molecular imprinting for miRNA sensing, addresses both the stability limitations of conventional DNA and PNA based sensors and the nonspecific interactions often associated with MIPs. In a second approach, natural peptides are investigated as non canonical capture probes for mature miR 21 in biosensing applications. A rationally designed library of 10 mer peptides was docked to mature miR 21 and evaluated by all atom molecular dynamics simulations, leading to the identification of N9 as a high affinity candidate with a stable groove associated binding mode, in contrast to the low affinity control peptide T1, which showed weak and unstable binding. N9 was synthesized by solid phase peptide synthesis, and the complex formation with miR 21 was confirmed by size exclusion chromatography coupled with mass spectrometry. The formation of a 1:1 N9-miR-21 complex was confirmed regardless of the mixing ratio used. Surface based fluorescence assays using N9 functionalized 96-well maleimide-activated plates yielded an apparent dissociation constant of 55 nM for 6 FAM labelled miR 21, while T1 produced a negligible signal, confirming that binding arises from specific recognition rather than nonspecific adsorption. Competitive assays with an equimolar single nucleotide mismatched analogue (S mis) retained 83% of the fluorescence response observed for miR 21 alone, indicating that N9 favors miR 21 by almost fourfold on a relative basis. Measurements in artificial serum preserved about 90% of the signal obtained in buffer, demonstrating good tolerance to matrix components. In a final step, the peptide-based platform was evaluated with RNA isolates from MCF 7 and HeLa cells in a competitive fluorescence plate assay. N9 modified wells showed a 27% decrease in the miR 21(6 FAM) signal in the presence of MCF 7 total RNA and a 16.7% decrease with HeLa RNA, indicating higher miR 21 levels in MCF 7 aligning with previous reports of higher miR 21 expression in breast cancer cell lines. These findings illustrate how the peptide based assay could, in principle, be applied to report relative miR 21 expression across different cancer cell types.
Long-term and sustained antisense oligonucleotides delivery in hiPSC derived motoneurons of spinal muscular atrophy desease
(Universität Ulm, 2026-08-18) Martinez Dominguez, Maria Victoria; Böckers, Tobias; Roselli, Francesco; Wagner, Karl
Antisense oligonucleotides (ASOs) are a promising therapeutic strategy for the treatment of neurological disorders such as spinal muscular atrophy (SMA). The FDA-approved ASO Nusinersen is able to restore exon 7 inclusion in SMN2 transcripts but requires the use of repeated intrathecal administration of the ASO via lumbar puncture in patients, highlighting the need for minimally invasive strategies that can have a controlled and sustained delivery of the ASO. We have developed and evaluated the use of a controlled and sustained delivery system using ethylene-vinyl acetate (EVA) or an EVA-silica composite. In our results in vitro, both formulations were biocompatible with hiPSC-derived motoneurons from SMA patients and healthy controls. We could demonstrate a sustained ASO release and an efficient cellular uptake, while polymer-released Nusinersen significantly increased exon 7 inclusion in SMN2 and showed modest improvements in motoneuron survival, showing preserved biological activity. These findings provide a proof of concept for a minimally invasive, controlled and sustained ASO delivery platform that could in the future reduce the need for repeated lumbar punctures in SMA and could be adapted for its use in other neurodegenerative diseases.
Impact of silicon on lithium distribution and plating in composite lithium‑ion battery electrodes : a simulative analysis
(Universität Ulm, 2026-08-18) Boveleth, Lioba; Latz, Arnulf; Waldmann, Thomas
Lithium-ion batteries, one of the most popular energy storage solutions for consumer electronics, electric vehicles, and grid-scale applications, have to be continuously improved to overcome performance limitations and safety issues. Consequently, research is constantly looking into novel negative electrode materials that offer higher energy density and are less prone to degradation. Silicon is a promising material, as it provides a ten times higher theoretical capacity than the well-established graphite. However, due to large volume changes during cycling, it is currently mixed with graphite. The incorporation of two active materials within a single electrode gives rise to complex electrochemical interactions, including competing lithiation dynamics and uneven current distributions. This affects the amount and distribution of deposited metallic lithium, a degradation mechanism termed plating, which is required to be detectable, predictable, and, eventually, preventable. Understanding and controlling the interactions of composite materials during battery operation is crucial for optimizing battery performance and predicting aging phenomena. This work contributes to the in-depth understanding featuring simulative and modeling methods as well as their comparison to experimental findings. Various experimental and simulative methods are validated during charging of full cells with and without silicon. Here, all methods observe the delayed lithiation of graphite in the silicon/graphite composite electrode compared to the graphite electrode. While initially silicon is lithiated preferentially, the preference switches to graphite over the course of charging. And this latter preference increases with increasing C-rate. The detailed experimental analysis of the subsequent relaxation indicates lithium redistribution from graphite to silicon during relaxation. The conducted three-dimensional microstructure-resolved simulations determine the lithium concentration within graphite as well as silicon throughout the electrode, supporting the experimental finding. Furthermore, the complementary methods should be utilized efficiently to detect lithium plating. Hence, preceding observations in graphite and silicon/graphite cells are discussed and recommendations for parallel or serial execution of the methods, stop criteria, and design of experiments planning given. To understand and mitigate plating in the first place, a parameter study is performed targeting the effect of electrochemical material parameters on plating in composite electrodes. Parameter values obtained from a thorough literature review are utilized to investigate critical parameters such as open-circuit potential, electronic conductivity, chemical diffusion coefficient, and exchange current density in simplified but sufficient model geometries. The three-dimensional microstructure-resolved simulations reveal not only the impact of the open-circuit potential along with the maximum capacity on the preferential lithiation and the resulting local currents, but also the importance of chemical diffusion and exchange current density, as well as their ratio, in plating onset and amount. Despite the observation of local structural inhomogeneities in a complex electrode geometry originating from commercial negative electrode material, the preferential plating on the electrode surface towards the separator dominates the previously mentioned findings. While significantly higher Si contents are desirable to achieve major improvements in the energy density, the displacement of electrolyte into the void regions of the cells caused by volume-changing active materials has to be analyzed. Past research revealed that the depletion of local lithium concentration in the electrolyte caused by the difference between inner and displaced electrolyte lithium concentration favors plating during prolonged cycling. To broaden the knowledge of the impact of active material volume changes and the resulting electrolyte motion on electrolyte lithium concentration, a homogenized electrochemical model of lithium-ion batteries including single-phase flow through the porous electrode media is developed. While the reservoir, meaning the void cell regions accommodating the displaced electrolyte, is not spatially resolved, the amount of displaced electrolyte and the lithium concentration therein are tracked in designated postprocessing steps. The model is utilized to investigate different silicon-to-graphite ratios and cell designs. Caused by the lithium concentration in the displaced electrolyte volume, the simulations reveal not only the depletion but also the accumulation of local electrolyte lithium concentration, which favors degradation during continuous cycling. As the displaced electrolyte's lithium concentration is controlled by the ratio of anode and cathode thickness and permeability, the concentration inhomogeneities can be mitigated by adjusting the cell design. By enhancing the understanding and detection of lithiation behavior and plating distribution, on the one hand, and giving detailed insights into the effects of the cell composition, on the other hand, this work contributes to improved design and utilization of lithium-ion batteries. In particular, application recommendations for the usage of highly abundant silicon and well-established graphite in silicon/graphite composite electrodes can be derived to increase the performance and safety of the resulting cells.
Sleep inertia in automated driving : how circadian and homeostatic factors influence post-sleep driver state, takeover, and driving behavior
(Universität Ulm, 2026-08-18) Tomzig, Markus; Baumann, Martin; Kunde, Wilfried
Napping during the trip is a promising use case for future automated driving systems. Surveys and case studies show that drivers want to sleep during automated driving and, in some cases, do so even illicitly. As long as autonomous driving is not fully realized, the system may prompt the person on the driver's seat to resume the driving task in the event of technical limitations. As soon as drivers are awakened from sleep by a takeover request, they may be confronted with sleep inertia. Sleep inertia refers to a phase after waking up that is characterized by pronounced sleepiness, disorientation, and cognitive slowing. Previous research demonstrated that sleep inertia can impair driving behavior in a safety critical way and may largely decrease the driver’s wellbeing and comfort. However, up to date, it is not sufficiently clarified which factors contribute to this impaired driver state. Knowledge about influencing factors would help to develop guidelines – similar to those in aviation – that indicate under which conditions napping during automated driving is safe from a human factors perspective and when sleep inertia is likely to impair post-sleep driving. The Two-Process Model of Sleep Regulation postulates that circadian and homeostatic processes considerably affect sleep and in this context alertness and performance. Derived from this model, this dissertation therefore examined how the circadian factors time of day and chronotype as well as the homeostatic factors sleep architecture and duration of the nap op-portunity affect the subjective state, driving behavior, and physiological processes immediately after sleep. The research questions were examined in two empirical participant studies. The methodology combined established methods from human factors research on automated driving and sleep science. Both studies were conducted in a high-fidelity driving simulator. Participants napped during an automated drive and were asked to resume manual driving upon awakening. Sleep depth was measured with an Electroencephalogram (EEG). The subjective state was assessed with different questionnaires. Driving behavior was assessed with various indicators such as takeover times, speed, lane keeping, and driving errors. Physiological processes were depicted by heart rate measures and EEG frequency analyses. Both studies resulted in two publications each (Study 1: Publications 1 and 2; Study 2: Publications 3 and 4). Publication 1 reports the influence of two circadian factors: the time of day and the chronotype. It was shown that the subjective state after a nap, as measured by arousal, wellbeing, and driv-ing motivation was significantly worse in the evening compared to a nap in the morning. In contrast, there were no significant differences in driving behavior as measured by speed, the standard deviation of lateral position and reaction to an acoustic vigilance task. The effects of the time of day interacted with the individual chronotype, resulting in so-called synchrony effects. Partic-ularly in morningness chronotypes, subjective state was impaired and driving speed was lower under sleep inertia when the time of day did not match the individual chronotype. Publication 1 also showed that the drivers experienced significantly more stage N3 deep sleep in the evening than in the morning. For Publication 2 it was therefore examined whether the pro-portion of N3 can predict post-sleep changes in the driver’s state and behavior. Publication 2 showed that N3 is significantly associated with a slowing down in takeover, driving and reaction behavior. The proportion of N3 is only one of several parameters used to describe the sleep architecture. In Publication 3, which primarily examined homeostatic influencing factors, the sleep architec-ture was therefore modeled more comprehensively using the last sleep stage and the propor-tions of the stages N1, N2, and N3 as parameters. The analyses showed that the sleep architec-ture can predict takeover times, the number of driving errors and subjective sleepiness after takeover. Impairments in driving behavior were particularly, but not exclusively associated with N2 and N3. In addition, different durations of the nap opportunity were experimentally manipulated in the study. It was found that sleep architecture is better suitable to predict takeover and driving behavior, as well as subjective sleepiness after sleeping than the duration of the sleep opportunity. While the preceding analyses had primarily examined the subjective state and driving behavior, Publication 4 focused on physiological changes during awakening. Cortical brain activity and heart rate during waking in a takeover situation were analyzed. It was shown that the last sleep stage had a significant influence on the post-sleep brain activity and heart rate level. A paradoxi-cal arousal pattern was observed, in which patterns of both low and high arousal were found in the brain and body. It is assumed that the cognitive resources needed to take over the driving task are limited under sleep inertia and that this limitation can cause stress, which in turn could be responsible for the paradoxical arousal pattern. The thesis concludes, that under conditions of moderate to high sleep pressure, sleep inertia may always occur immediately after awakening. Negative consequences are increased under the following conditions: • When naps are taken in the evening compared to morning. • When the individual chronotype does not match the time of day. This accounts particu-larly for morningness types sleeping in the evening. • When the sleeping driver experiences a high proportion of N3 deep sleep during the nap. • If the driver is awakened from N2 or N3. Avoiding awakenings from or high proportions of N2 and N3 may represent promising proactive measures to reduce sleep inertia related impairments. However, more research is needed to develop applicable technical solutions for an in-vehicle sleep stage detection. Further, it is con-cluded that proactive measures may fall short in reliably preventing sleep inertia, particularly in private vehicle usage. Therefore, before sleeping during automated driving can become a safe and comfortable use case, further issues such as the development of reactive countermeasures have to be solved.
Structural Analysis of Amyloid Fibril Seeding
(Universität Ulm, 2026-08-14) Pfeiffer, Peter Benedikt; Fändrich, Marcus; Kapurniotu, Aphrodite; Bieschke, Jan
Alzheimer’s disease is the most prevalent neurodegenerative disease and the main cause of dementia worldwide. The hallmarks of Alzheimer’s disease are the pathological misfolding and aggregation of Aβ and tau peptides into amyloid fibrils. The two main amyloidogenic Aβ isoforms, Aβ(1 40) and Aβ(1-42), consist of 40 or 42 amino acids respectively and are derived from the proteolytic cleavage of the amyloid precursor protein by β- and γ-secretases. While Aβ(1-42) fibril deposits accumulate in the hippocampus and cerebral cortex, Aβ(1-40) fibrils are deposited in leptomeningeal and cortical blood vessels. The accumulation of Aβ(1-40) fibrils is a central hallmark of Alzheimer’s disease and cerebral amyloid angiopathy, causing vascular lesions and haemorrhages. Recent advances in cryo-electron microscopy have enabled the structural determination of Aβ(1-40) amyloid fibril structures from Alzheimer’s disease. These findings revealed that previously reported seeded Aβ(1-40) fibrils differed structurally from the ex vivo fibril seeds and called their molecular origin into question. Therefore, this thesis aims to clarify the effect of seeding with brain-derived Aβ(1-40) fibrils by providing a detailed structural comparison of seeded and unseeded Aβ(1-40) amyloid fibrils. Initially, ex vivo Aβ(1-40) fibrils were extracted from the meningeal tissue of two patients with Alzheimer’s disease and cerebral amyloid angiopathy. The ex vivo fibrils were used as seeds to generate seeded Aβ(1-40) fibrils. An analysis with platinum side-shadowed specimen demonstrated that both ex vivo fibrils and seeded fibrils were right-hand twisted, whereas unseeded fibrils exhibited a left-handed twist. Moreover, lower seed concentrations and seeding over multiple generations led to the emergence of left-hand twisted fibril morphologies in seeded samples. High-resolution cryo-EM data enabled the reconstruction of 3D density maps for first- and third-generation seeded fibrils as well as for unseeded fibrils. The unseeded fibril sample contained two morphologies (unseeded i and ii), which shared a nearly identical fibril protein fold, but differed in their helical symmetry. In both unseeded morphologies, the stable fibril core extended from His13 to Val40, while the N terminus was structurally disordered and not resolved in the 3D maps. Additionally, both unseeded fibrils exhibited diffuse extra density, indicative of laterally bound Aβ(1-40) molecules. A subsequent comparison of the unseeded fibril structures with ex vivo Aβ(1-40) fibrils highlighted significant structural differences. In contrast, first- and third-generation seeded samples contained fibril morphologies (seeded I and I a) that closely resembled morphology I of brain-derived Aβ(1-40) fibrils. The stable core of the seeded I and I a fibrils extended from Asp1 to Gly37, while the C-terminus was structurally disordered. Other seed morphologies (II and III) were observed in the first generation, but essentially absent in later seeded generations. Interestingly, unseeded fibrils (unseeded i a and ii a) emerged in the third generation, demonstrating a competition between seeding and de novo nucleation. Following the structural analysis, computational and biochemical analyses consistently demonstrated that seeded fibrils were more stable than unseeded fibrils. In summary, the findings in this thesis demonstrate that brain-derived Aβ(1-40) fibrils are able to seed solutions of recombinant Aβ(1-40) protein and replicate the pathogenic seed structure in vitro. These findings challenge previous studies, which reported seeded fibril structures that resemble the unseeded fibrils observed in this study. Furthermore, seeding over multiple generations revealed a competition between seed extension and de novo nucleation. These observations offer insights into the prion-like propagation mechanism of seeding with amyloid fibrils and may provide an explanation for the presence of de novo nucleated fibril structures in previously reported seeded samples. Finally, the ex vivo-like stability of seeded fibrils supports the proteolytic selection hypothesis, which proposes that only stable fibril morphologies survive the proteolytic selection in vivo and subsequently propagate stable, pathogenic fibril morphologies.
Engineering of Color Centers in Isotopically Purified Diamond for Quantum Technologies
(Universität Ulm, 2026-08-13) Findler, Christoph; Jelezko, Fedor; Nesládek, Miloš
Silicon has turned from a niche laboratory-scale material into the backbone of modern electronics. Diamond functionalized with color centers could follow a similar path and grow from proof-of-concepts experiments to application-oriented quantum technologies. Quantum sensing, communication, and information processing with color centers has been proven to work in scientific experiments and enhanced sensitivities as well as novel functionalities have been reported. While the first generation of diamond quantum devices relied on color centers formed uncontrolled in natural diamond with a high statistical variation in composition and crystallinity, the situation has changed nowadays. With the emergence of high purity monocrystals and the option for isotopic purification of the carbon matrix, technological progress in materials science on the atomic-scale enabled tailoring the spin properties in diamond. But, since many color centers typically require vacancies to combine with impurities in the carbon lattice via diffusion, the generation of color centers underlies a statistical process providing only little control on the center's final position in the crystal lattice. The negatively charged nitrogen-vacancy (NV$^-$) center is the most studied optically active defect in diamond and exhibits remarkable spin properties and sensing capabilities, even on the nanoscale and at room temperature. The spin properties of NV$^-$ centers close to the surface (1\,nm - 30\,nm) have been thoroughly studied and the deeper the spins are located in diamond the less decoherence from surface defects is observed. Creating NV$^-$ centers at the surface by nitrogen ion implantation and burying them beneath a thin diamond layer via chemical vapor deposition (CVD) of diamond is reported to enhance the coherence times significantly. By separating the color center generation from controlling the resulting depth by an additional diamond layer, narrow depth profiles can be achieved by reducing the ion energy during implantation. Although the coherence properties improve, many of the implanted NV$^-$ center are reported to vanish in the hydrogen-rich plasma of growth CVD chambers. The mechanism responsible for the instability of NV$^-$ centers towards the plasma remains unclear, though. One hypothesis is the formation of nitrogen-vacancy-hydrogen centers from diffusing hydrogen and implanted NV$^-$ centers but is has not been proven, yet. Here, we demonstrate the fabrication of depth confined NV$^-$ centers by combining low energy (1\,keV - 5\,keV) ion implantation and CVD overgrowth. By increasing the thickness of the overgrown layer (6\,nm - 100\,nm) the influence of surface noise is shown to decrease leading to bulk-like coherence properties of up to $450\,\upmu$s (Hahn-echo) and $T_2^*$ dephasing times around $20\,\upmu$s. Monitoring the NV$^-$ density as a function of the overgrowth time, we analyze the kinetics of the passivation in case of implanting ion doses of 10$^{11}$ $^{15}$N $^{+}$/cm$^2$ and 10$^{12}$ $^{15}$N $^{+}$/cm$^2$. The passivation turns out to occur faster for lower doses offering insights into sub-surface hydrogen diffusion in diamond. Analyzing the spin bath before and after overgrowth reveals a drastic reduction in paramagnetic spins after the CVD process but no NVH$^-$ is found around the surviving NV$^-$ centers. Checking the chemical composition in overgrown implantation sites with secondary-ion-mass-specrometry (SIMS), however, confirms significant amounts of hydrogen in the $^{15}$N-implanted layers. Since the concentration found for hydrogen exceeds the one for nitrogen by far, we suggest that diffusing hydrogen interacts also with vacancies forming probably mostly diamagnetic defects. In the last part of this dissertation, NV$^-$ centers are generated with sub-nanometer precision below a thin $^{13}$C-enriched layer for quantum simulation. For fabricating the spin architecture starting from a $^{12}$C base layer, we overgrow implanted nitrogen with a $^{13}$C layer and bury the spins beneath a 30\,nm-thick $^{12}$C capping layer. By measuring the coupling strength between the electron and nuclear spins, we deduce a NV$^-$ placement accuracy reaching below one nanometer with respect to the $^{13}$C-layer. Regarding the spin properties of the electron spins, no reduction in $T_2$ is observed despite of the proximity to the dense $^{13}$C-bath. As a consequence, the overgrowth of NV$^-$ centers is demonstrated as a bottom-up approach to produce buried spin architectures with sharp isotopic transitions and efficient $^{13}$C-NV$^-$ coupling. Together with the potential for bulk-like coherence properties the synergy of implantation and overgrowth is a powerful tool to fabricate reproducible quantum systems with tailored properties and high depth confinement.

Browse

Recent Submissions