OPEN FAU Online-Publikationssystem der Friedrich-Alexander-Universität Erlangen-Nürnberg
Not a member yet
23478 research outputs found
Sort by
Classification of reflective writing: A comparative analysis with shallow machine learning and pre-trained language models
Get PDFReflective practice holds critical importance, for example, in higher education and teacher education, yet promoting students’ reflective skills has been a persistent challenge. The emergence of revolutionary artificial intelligence technologies, notably in machine learning and large language models, heralds potential breakthroughs in this domain. The current research on analyzing reflective writing hinges on sentence-level classification. Such an approach, however, may fall short of providing a holistic grasp of written reflection. Therefore, this study employs shallow machine learning algorithms and pre-trained language models, namely BERT, RoBERTa, BigBird, and Longformer, with the intention of enhancing the document-level classification accuracy of reflective writings. A dataset of 1,043 reflective writings was collected in a teacher education program at a German university ( M = 251.38 words, SD = 143.08 words). Our findings indicated that BigBird and Longformer models significantly outperformed BERT and RoBERTa, achieving classification accuracies of 76.26% and 77.22%, respectively, with less than 60% accuracy observed in shallow machine learning models. The outcomes of this study contribute to refining document-level classification of reflective writings and have implications for augmenting automated feedback mechanisms in teacher education.Open Access funding enabled and organized by Projekt DEAL.Friedrich-Alexander-Universität Erlangen-Nürnberg (1041
Expression und Funktion von α4β7-Integrin als Prädiktoren für den Therapieerfolg von Vedolizumab bei chronisch-entzündlichen Darmerkrankungen
Get PDFChronisch-entzündliche Darmerkrankungen sind in vielen Fällen schwer zu behandeln und erfordern eine kontinuierliche Therapie. Obgleich eine zunehmende Anzahl an Biologika und small molecules zur Verfügung stehen, sprechen viele Patienten primär oder sekundär nicht ausreichend auf diese Optionen an. Darüber hinaus mangelt es an direkten Vergleichsstudien zur Therapieauswahl. Somit besteht ein großer Bedarf an Ansätzen, die die Vorhersage des Therapieerfolgs für den einzelnen Patienten ermöglichen.
Diese Publikationsdissertation thematisiert basierend auf einer prospektiven multizentrischen Studie, inwiefern die Expression und Funktion von α4β7-Integrin auf T-Zellen im peripheren Blut mit dem Therapieerfolg des anti-α4β7-Integrin-Antikörpers Vedolizumab korrelieren und als prädiktive Marker eingesetzt werden können
An optoacoustic field-programmable perceptron for recurrent neural networks
Get PDFRecurrent neural networks (RNNs) can process contextual information such as time series signals and language. But their tracking of internal states is a limiting factor, motivating research on analog implementations in photonics. While photonic unidirectional feedforward neural networks (NNs) have demonstrated big leaps, bi-directional optical RNNs present a challenge: the need for a short-term memory that (i) programmable and coherently computes optical inputs, (ii) minimizes added noise, and (iii) allows scalability. Here, we experimentally demonstrate an optoacoustic recurrent operator (OREO) which meets (i, ii, iii). OREO contextualizes the information of an optical pulse sequence via acoustic waves. The acoustic waves link different optical pulses, capturing their information and using it to manipulate subsequent operations. OREO’s all-optical control on a pulse-by-pulse basis offers simple reconfigurability and is used to implement a recurrent drop-out and pattern recognition of 27 optical pulse patterns. Finally, we introduce OREO as bi-directional perceptron for new classes of optical NNs.Optical recurrent neural networks present a unique challenge for photonic machine learning. Here, the authors experimentally show the first optoacoustic recurrent operator based on stimulated Brillouin scattering which may unlock a new class of optical neural networks with recurrent functionality.Max-Planck-Gesellschaft (Max Planck Society)501100004189Deutsche Forschungsgemeinschaft (German Research Foundation)501100001659Studienstiftung des Deutschen Volkes (German National Academic Foundation)50110000435
High‐Throughput Electrospinning of Unmodified and Aminated Poly(Pentafluorostyrene) for Fiber‐Reinforced Proton Exchange Membranes
Get PDFThis study demonstrates a high‐throughput fabrication of fiber interlayers for proton exchange membranes based on poly(pentafluorostyrene) (PPFSt) and its aminated derivatives. The fibers are produced by electrospinning, where the parameters are carefully screened. The controlled parameters are solvent composition, weight percentage, voltage, flow rate, and temperature, controlled with a self‐designed heating jacket. The parameters are iterated toward optimized fiber structure and maximum output. The yielded fibers are infiltrated with Nafion and sulfonated polymer from bisphenol AF and decafluorobiphenyl (SFS001) by spray‐coating and doctor blading to obtain the fiber‐reinforced proton exchange membranes. Tensile tests reveal a higher Young's modulus and yield stress than pure Nafion. Here, the basicity of the aminated PPFSt fibers correlates with the Young's modulus due to improved acid‐base interactions between amine groups and sulfonic acid. The acid‐base interactions influence the composite membrane's proton conductivity, varying from 23 mS cm −1 for strongly alkaline fibers to 69 mS cm −1 for non‐basic fibers. These findings can be transferred to fabricating fiber reinforcements beyond routinely used poly(benzimidazoles).This work presents a high throughput fabrication of modifiable poly(pentafluorostyrene) fibers as reinforcement for Nafion and SFS001 in proton exchange membranes. The mechanical properties and proton conductivities correlate with the basicity of the fiber material by interactions between amines (fiber) and sulfonic acid (matrix). These findings serve as a blueprint for the design of novel fiber materials with high throughput. imageDeutsche Forschungsgemeinschaft (DFG)Indonesia Endowment Fund for Education (LPDP
Influence of basic knowledge about female health, physiology, and contraception on unintended pregnancies: an international multicenter survey among women in Austria, Germany and Switzerland
Get PDFOpen access funding provided by Johannes Kepler University Linz.Johannes Kepler University Lin
Electrospun Thermoplastic Polyurethane Fabric for Flexible Strain Sensors
Get PDFIn recent years, there has been a significant surge of interest in wearable high-performance strain sensors, capable of detecting physiological activities and monitoring changes in physical quantities, from both academia and industry. This trend is mainly attributed to the soaring demand for high-performance sensors in various fields, including smart electronics, motion signal analysis, and human monitoring, which are witnessing a rapid expansion of applications and potentials.
However, conventional strain sensors have significant limitations in terms of their stretchability. Electrospun strain sensors offer advantages such as lightweight, good elasticity, high stretchability and ease of processing, and have great potential for use in the field of wearable strain sensors. The aim of this work is to develop electrospun TPU fabric strain sensors with good wearability, mechanical flexibility, and electrical conductivity for use in new flexible electronics. The influence of materials, processing methods, and microstructures of TPU electrospun strain sensors on wearable strain sensors is also investigated, and future directions are predicted.
In this thesis, a simple and efficient fabrication method for high-performance flexible conductive polymer composite strain sensors are developed via using electrospun thermoplastic polyurethane fabric films as the matrix and carbon black (CB), carbon nanotubes (CNTs) and silver nanowires (AgNWs) as conductive fillers, respectively. TPU fibrous films with adjustable 3-D scaffold network was fabricated via electrospinning processing. Study on the preparation method and performance of strain sensors indicated the influence of scaffold network structure built under various collection rotating speed of electrospinning process on the conductive sensitivity of strain sensor. Scaffold network makes the sensor exhibited high sensitivity and good stretchability. TPU/CB strain sensor have a higher gauge factor (GF) of 8962.7 at 155% strain; TPU/CNTs strain sensor exhibits an excellent workable stretching range (0 ~ 400%) with GF > 1571; TPU/AgNWs strain sensor shows a superior sensitivity (GF > 16,000) with strain of 360%.
In addition to the aforementioned, the present work employs a theoretical approach to investigate the mechanical properties of the sensor. Two equations for a model based on tunneling theory are introduced for the first time, which offer an efficient and straightforward means of analyzing alterations in the conductive pathways and inter-particle distance during tensile deformation of the sensor. As a result, not only can changes in electrical sensitivity during sensor operation be more comprehensively elucidated, but also the conductive sensitivities of strain sensors can be forecasted and examined. This theoretical approach presents a promising avenue for the industrial manufacture of flexible strain sensors
The Influence of Confinement on the Adsorption Behavior of Supercritical Fluids in Nanoporous Materials for Gas Storage Applications
Get PDFDuring the last decades, major progress was made concerning the understanding of subcritical, low
pressure adsorption of fluids in nanoporous materials. Within this context, it was possible to understand
how structural properties affect the shape of the adsorption isotherms, allowing the development
of advanced methodologies now commonly used for textural characterization. However, for
gas storage applications, supercritical gas adsorption at high pressures plays an important role. Highpressure
adsorption experiments introduce several complexities, both in terms of collecting isotherm
data and interpreting the results. A key feature here is that the experimentally determined surface
excess adsorption isotherm may exhibit a characteristic maximum at a certain pressure. For a given
temperature and adsorptive/adsorbent system, the surface excess maximum (and the corresponding
(absolute) adsorbed amount) is related to the storage capacity of the adsorbent. However, there is still
a lack of understanding of how key textural properties such as surface area, pore size and pore volume
affect the shape of supercritical high pressure adsorption isotherms (and here in particular the position
of the surface excess maximum).
In order to address this open questions, first an experimental setup was modified and validated, allowing
for accurate measurement results even for demanding measurement conditions. Based on that, a
systematic experimental study was performed to assess the effect of pore size/structure on the supercritical
adsorption isotherms of pure fluids such as C2H4, CO2, SF6 and CH4 over a wide range of
temperatures and pressures (from well above the bulk critical point to the near-critical region) on a
series of model materials exhibiting well defined pore sizes/geometries, i.e. ordered micro- and mesoporous
materials such as zeolites and mesoporous molecular sieves (e.g., NaY zeolite, KIT-6 silica,
MCM-48 silica). A fundamental result of the experiments is a unique fluid-independent correlation
between the pressure of the surface excess maximum pmax (at a given temperature) and the pore size (by
taking into account the kinetic diameter of the fluid and the underlying effective attractive fluid-wall
interactions). To further investigate the adsorption mechanism, a series of complementary simulations
was performed. With regard to the shape of the surface excess isotherms, the pore-size effect and
the temperature effect, the simulations are in qualitative agreement with experimental data. The results
suggest important structure-property relationships and allowed to derive a rigorous approach for
predicting gas storage properties of nanoporous materials at given thermodynamic conditions based
on their textural properties. The fundamentals insights may also serve as basis for a new generation of
process specific tailored materials
Biochar—just a black matter is not enough
Get PDFWhat differs biochar from charcoal? The simple answer is that biochar is a carbon-rich product obtained from the thermal decomposition of organic material, at the presence of no or only a bit of oxygen. In principle, the production of biochar is comparable to the production of charcoal, one of the oldest and most established processes developed by mankind. While charcoal is made traditionally from wood, biochar can be based on a wide range of biomass and biomass residues. However, a variety of technologies for the production of biochar has been developed in recent years. The technologies are based on pyrolysis, gasification, or hydrothermal carbonization and are ranging from simple units, like heated steel drums to full automated and controlled processes. Therefore, the obtained products have tremendous differences in its properties and resulting qualities. The quality defines the field of application. To obtain the required quality for each application, the right process must be applied. Consequently, it is not enough only to enrich the carbon content by thermal decomposition of organic material. The production of tailor-made biochar for specific high added–value application is much more complex. In addition, side products like liquid biofuels make business cases stronger. If it is done in the right way, biochar production combined with advanced biofuels can be an economic solution to overcome the problems of climate change. “So for the future of mankind, this black matter might give the light at the end of the tunnel.”Open Access funding enabled and organized by Projekt DEAL.Projekt DEA
Supramolecular Gas-Phase Host-Guest Chemistry of [n]Cycloparaphenylenes based on π-π Interactions
Get PDFIn recent years, [n]Cycloparaphenylenes ([n]CPPs) have gained great research interest in the field
of supramolecular chemistry due to their special structure enabling the formation of novel host-guest complexes with unique properties. The present thesis focuses on the gas-phase host-guest
chemistry of [n]CPPs with a particular emphasis on π-π interactions and thus aims for a
fundamental understanding of this type of interaction on the molecular level.
All CPP-based host-guest complexes were studied by means of (tandem) mass spectrometry. To
this end, the modern soft ionization techniques MALDI and API (ESI, APCI and APPI) were used
to transfer the host-guest complexes of interest from solution into the gas-phase. Furthermore,
tandem mass spectrometry using either collision-induced dissociation or metastable decay was
applied to study the conformation and charge distribution of these complexes. Further insights into
the relative stabilities of the investigated host-guest complexes were gained by energy-resolved
collision-induced dissociation experiments. A detailed description of the techniques applied and
instruments used in the scope of this thesis is provided in Chapters 3 – 9. Chapter 10 is dedicated
to the synthesis and properties of the investigated compounds.
The scientific results obtained during these doctoral studies are summarized in five already
published research articles. Two studies focus on the formation of host-guest complexes between
[n]CPPs and fullerenes. In this context, several novel complexes featuring the fullerene guests C60,
C70 and Li+@C60 were identified. Furthermore, the complexation behavior of the CPP derivatives
Aza[10]CPP, N-methylaza[10]CPP and benzyl Lasso-CPP was thoroughly investigated. Their
complexes with the fullerene guests C60 and C70 were studied with an emphasize on the
fragmentation behavior and the relative complex stabilities. Besides fullerene-based complexes,
ring-in-ring complexes, in which the CPPs serve as host as well as guest molecules, were also
examined, leading to the successful formation of several novel complexes with so far unreported
size differences. In Chapter 11, a more detailed summary of these five research articles is given.
The full research articles can be found in the appendix (open access articles only) or may be
downloaded from the respective journal website (corresponding links are provided in the
appendix)
Local electroneutrality breakdown for electrolytes within varying-section nanopores
Get PDFWe determine the local charge dynamics of a z-zelectrolyte embedded in a varying-section channel. By means of an expansion based on the length scale separation between the axial and transverse direction of the channel, we derive closed formulas for the local excess charge for both, dielectric and conducting walls, in 2 D (planar geometry) as well as in 3 D (cylindrical geometry). Our results show that, even at equilibrium, the local charge electroneutrality is broken whenever the section of the channel is not homogeneous for both dielectric and conducting walls as well as for 2 D and 3 D channels. Interestingly, even within our expansion, the local excess charge in the fluid can be comparable to the net charge on the walls. We critically discuss the onset of such local electroneutrality breakdown in particular with respect to the correction that it induces on the effective free energy profile experienced by tracer ions. Graphical abstractDFGMCIU/AEI/FEDEGeneralitat de Catalunyahttp://dx.doi.org/10.13039/50110000280