Department Physik
Refine
Year of publication
Document Type
- Doctoral Thesis (135)
- Article (32)
- Part of a Book (3)
- Preprint (2)
- Book (1)
- Master's Thesis (1)
- Moving Images (1)
Has Fulltext
- yes (175)
Keywords
- - (32)
- Graphen (5)
- Nichtlineare Optik (4)
- nonlinear optics (4)
- photonic crystal fiber (4)
- quantum optics (4)
- Biophysik (3)
- CTA (3)
- Density functional theory (3)
- Diffusion (3)
Die Struktur von adsorbierten dünnen Filmen an Fest-Flüssig-Grenzflächen ist für viele wissenschaftliche Fragestellungen von großer Bedeutung. Unter anderem spielt die Grenzflächenstruktur eine entscheidende Rolle für die Bioaktivität von bioaktiven Gläsern wie dem Glas 45S5, für das initiale Schichtwachstum bei der Atomlagenabscheidung aus der Lösung bei der Herstellung von Solarzellmaterialien oder für die (hydrophobe) Funktionalisierung von Oberflächen.
Ziel dieser Dissertation war es, ein umfassenderes Verständnis der Struktur dünner adsorbierter Filme an der Fest-Flüssig-Grenzfläche zu erlangen. Der Fokus lag dabei auf selbstorganisierenden Monolagen (SAMs) und der Proteinadsorption auf bioaktiven Gläsern. Die Untersuchungen umfassten hauptsächlich Röntgenreflektometrie (XRR), Kontaktwinkel- und AFM-Messungen.
As new machine learning technologies are conceived, endless possibilities for breakthrough applications in science open up. We wonder how far we are from developing an artificial scientist that acts like a human one. It would be capable of making observations, analyzing data, building a model, deriving hypotheses, and designing experiments to falsify them. After an overview of common machine learning techniques, we review recent work that attempts to achieve some of these goals. We then present the main ideas that we believe are necessary for future artificial scientific discovery. In particular, we take the perspective of a physicist, and show examples from physical domains that could benefit from the presented approaches. We develop these ideas in various works. First, we focus on learning to make a high-level description of a physical system, e.g. by extracting collective variables. Therefore, we present the idea of using information theory to describe them and introduce the concept of “Renormalized Mutual Information”. Second, we study the problem of experimental design and investigate how an artificial scientist should choose which experiments to perform. We show applications of a deep learning approximation technique to quantum many-body systems. We then generalize our discussion to the problem of scientific exploration, diving into reinforcement learning techniques for intrinsic motivation as a framework for creating “curious” artificial scientists. Furthermore, we imagine ways in which an algorithm could learn concepts from experience and reuse them in different settings. We adopt the program synthesis approach and explore applications to the synthesis of quantum circuits given the associated unitary matrix. Finally, we discuss additional proof-of-concept examples and future developments in the field.
Hybrid quantum systems that couple different physical degrees of freedom for optimal functionality are essential for developing new quantum information platforms at the nanoscale.
A new class of hybrid systems based on magnonics has recently emerged, such as optomagnonic systems where light couples coherently to collective excitations in magnetically ordered systems. In this thesis, we focus on numerical methods to model these systems.
Most efforts so far, both on the theoretical and experimental level, focused on optomagnonic systems where a non-optimal overlap between the optical and magnetic modes is achieved, resulting in a relatively low coupling.
In the first part of this thesis to tackle the issue of improving the optomagnonic coupling strength, we put forward the concept of an optomagnonic crystal: a periodically patterned photonic structure at the microscale based on a magnetic dielectric, which can co-localize photon and magnon modes simultaneously.
This co-localization in a small volume can lead to considerable values of the optomagnonic coupling at the single quanta level, making these systems a suitable candidate for quantum information processing and quantum conversion schemes.
We analytically and numerically study a simple geometry consisting of a 1D array of holes with an abrupt defect, considering the ferrimagnet Yttrium Iron Garnet (YIG) as the basis magnetic material.
We show that both magnon and photon modes can be co-localized at the defect, and choosing an optimal pair of modes using symmetry arguments leads to maximized coupling values in the kHz-range.
We also discuss possible optimization routes to improve the coupling strengths and optical losses.
Since the developed simulation framework disregards effects originating from the interaction between light and the magnetization, we propose performing electromagnetic simulations incorporating the material's magnetic properties.
We therefore extend the Finite-Difference Time-Domain method to treat dispersive magnetic media by incorporating magneto-optical effects through a frequency-dependent permittivity tensor. For benchmarking the extended method, we consider the light scattering on a magnetic sphere in the Mie regime.
We first derive the analytical scattering expressions, which predict a peak broadening in the scattering efficiency due to the atomic energy level splitting in the presence of a magnetic field, together with an additional rotated part in the scattered field profile due to the Faraday rotation.
We show that our numerical method can capture the main scattering features to a remarkable extent and discuss its limitations and possible improvements in accuracy.
To reduce computational costs, especially in micromagnetic simulations for magnetic ground state prediction, we explore an artificial neural network approach to accelerate the ground state prediction of nonlinear time evolutions.
As a proof of principle, we consider a nonlinear wave evolution governed by the nonlinear Schrödinger equation in a model-free framework solved by the imaginary time evolution method.
We develop a suitable training strategy and network model to train a residual neural network based on convolutional layers to predict the evolved wave function at a desired time given the initial wave function and a random parameter set as input.
We show that the neural network approach can predict the desired evolved states for long evolution times in an accelerated fashion and discuss possible optimization routes for a better network performance.
This thesis with the title ‘Assembly of photonic nanostructures by AFM-based pick-and-place handling of individual nanoparticles inside an SEM’ was concerned with the development, setting-up, and application of a nanomanipulation system for the fabrication of particle patterns composed from individually selected and arranged nanoparticles.
This nanomanipulation system permits to collect single nanoparticles from samples, here referred to as ‘reservoir’ samples, and to transfer them individually to a new target location and even to a different sample if required. The exchange of the samples is performed in an automated fashion whereby previously stored sample positions can be changed within seconds across centimeter distances and with a repositioning accuracy in the low single digit micrometer to sub-micrometer range. The AFM-based nanomanipulations are performed by a human operator who can interactively control the movements of the AFM tips with a 3D-controller input device with nanometer precision and under the visual observation from the SEM. Additionally, the nanomanipulation setup permits to make use of two independently movable AFM tips in a shared volume of operation in a cooperative fashion to perform nanomanipulations. This is particularly helpful for manipulating higher-dimensional nano-objects than particles, such as wires and flakes, as was also demonstrated during the work on this thesis.
The AFM-based nanomanipulation setup utilizes different commercial positioning components and was designed to integrate these components into a commercial scanning electron microscope (SEM) to form a coherent and operable nanomanipulation setup. Electronic signal amplifiers to measure the mechanical bending motions of piezoresistive AFM cantilevers during these nanomanipulations were realized with the support of the electronics workshop at the Max Planck Institute for the Science of Light (MPL) in Erlangen. The required software control architectures for inter-operating the constituting components of the nanomanipulation system were developed and integrated. The procedures how to align the system for nanomanipulations with automatic sample switching and how to perform pick-and-place handling of nanoparticles are described in detail.
Besides working with gold and silica nanoparticles for test purposes, nanoparticle patterns from gold and silica nanoparticles were prepared for investigating directional Kerker scattering, optical chirality in heterogeneous particle assemblies, and mode sorting in inhomogeneously sized dimers in cooperation with colleagues at the MPL. Similarly, an optical displacement sensor for nanometer-precise detection of position shifts was realized by functionalizing a photonic crystal waveguide architecture with a precisely placed silicon nanoparticle.
Different problems encountered during the development and work with this nanomanipulation approach such as charging and carbon contaminations in the SEM environment and the local modification of optical properties in thin films of transparent conductive oxides (TCOs) are described. Explorative steps that were undertaken to develop pre-characterization strategies for selecting nanoparticles based on their optical properties are presented. The thesis concludes by considering potential mitigation strategies for relieving the encountered problems and for improving the nanomanipulation system itself, and by discussing prospective applications for what nanoparticle-based structures could be addressed with good prospects for future research.
We present a Raman study of MoS2 powders and MoS2 individual layers covalently functionalized with organic molecules. In MoS2 powders, the defect-induced “LA” Raman mode shows evidence for successful functionalization. Increasing temperature induces oxidation of both functionalized and nonfunctionalized MoS2 into MoO3. In contrast, mechanically exfoliated individual MoS2 layers do not transfer into MoO3 under the same conditions. Instead, the Raman spectra show that the procedure of covalent functionalization leads to a partial transition from the 2H into the 1T' crystallographic phase in few-layer MoS2. We support the identification of the 1T' phase by DFT calculations of the corresponding vibrational modes in the mono- and bilayer 1T'-MoS2.
The main object of investigation of this thesis was the cosmological and semiclassical realm of loop quantum gravity, which was addressed from two different directions. On the one hand, we analysed a cosmological toy model, and on the other hand a new procedure for conducting the computation of semiclassical expectation values was introduced. With the help of the latter, we revisited possible singularity avoidance in loop quantum gravity — which was so far only either analysed in so-called loop quantum cosmology, was limited to special configurations like cubic graphs, or was only possible by utilising estimations.
The cosmological toy model considered in this thesis, a so-called Gowdy model, features a T^3 symmetry and is of special interest when the cosmological realm of (loop) quantum gravity shall be investigated as it still, despite its simplifications, yields a field theory after quantisation. Loop quantisations of such models relying on Dirac quantisation already exist in the literature. We extend those results by applying a reduced quantisation via coupling Gaussian dust to gravity as a dynamical reference frame. The quantisation is performed for two different frameworks: reduced loop quantum gravity and algebraic quantum gravity, where for both approaches a graph preserving prescription is applied. Analysing a Schrödinger-like equation and finding special solutions thereof then constitute first applications of this model. We find zero volume states and states that experience a vanishing action of the Euclidean part of the physical Hamiltonian. When it comes to the corresponding Lorentzian part, in turn, we analyse degeneracies caused by its action. Overall, these are first steps for gaining an overview over the different aspects of the action of the physical Hamiltonian of such T^3 Gowdy models.
Addressing the question of singularity avoidance in full loop quantum gravity, we introduce a technique relying on Kummer’s confluent hypergeometric functions. It turns out that they feature a lot of very handy properties like an asymptotic expansion for large arguments that allow for an exact calculation of certain semiclassical expectation values by means of a power series in the semiclassicality parameter. These are taken of a specific class of operators that play a pivotal role in the dynamics of the theory and with respect to complexifier coherent states — the state-of-the-art coherent states used in loop quantum gravity. Seminal results of the literature that addressed singularity avoidance are generalised and extended with the help of this new method. Specifically, these improvements are that it is not always necessary to use estimates and that it is in fact possible to also conserve the correct powers of the momentum, e.g. The latter is also exemplified by applying the new procedure to standard quantum mechanics, where expectation values of fractional powers of the momentum operator can be computed analytically, resulting in a power series in hbar.
On a more fundamental level, we use the Zak transformation to link coherent states on the circle to those of the harmonic oscillator. This further allows for a more efficient computation of (the zeroth order of) semiclassical matrix elements as we provide a link between semiclassical matrix elements in L_2(R) and L_2(S_1). What is more, also Kummer’s confluent hypergeometric functions offer new insight on the fundamental level: As Kummer’s differential equation can be linked to the heat equation, we can associate Kummer’s confluent hypergeometric functions with solutions to the heat equation.
Monte Carlo simulations of extensive air showers, their Cherenkov emission, and the
telescope response to this emission are essential to many modern Imaging Air Cherenkov
Telescope (IACT) experiments and their event reconstruction techniques. Therefore, the
validity of these simulations is of critical importance for all scientific analyses performed
with the reconstructed data. One such Imaging Air Cherenkov Telescope (IACT) array
is the High Energy Stereoscopic System (H.E.S.S.), featuring four small (CT1-4) and
one central large-mirror telescope (CT5). A recent analysis of CT5 Crab data revealed
a mismatch of ≈46 % in the reconstructed spectral norm compared to results using
CT1-4, found to be caused by an incorrect telescope response simulation. CORSIKA is
used for the air shower and Cherenkov emission simulation and sim telarray to simulate
the telescope response. This work presents the author’s significant contributions to
this Monte Carlo validation effort. Multiple discrepancies between observations and
simulations are identified and include the size of the optical Point Spread Function
(PSF), the atmospheric transmission profile, the telescope trigger thresholds, the mirror
reflectivities, and the Night Sky Background (NSB) settings. The observed deviations
are reduced to satisfactory levels, resulting in a better match between the simulated
and observed telescope trigger rates. Post-validation, the deviation of the reconstructed
spectral norm is reduced to 11 % and now matches previous results within systematic
errors.
A 3D analysis of Crab data using the ABRIR method for improved background rejection
is presented and compared to the results obtained with the existing BDT approach. No
statistically significant difference is found for the derived spectral and spatial parameters.
Whilst the significance of calculated flux points below 10 TeV is reduced, it rises again
for higher energies when the ABRIR method is used.
Background events can also be rejected by an array of telescopes that issue an event-wise
veto. A candidate for such a telescope is the IceACT telescope. It is a low-cost (<10 ke)
and compact IACT conceived initially as part of the surface detector of IceCube. This
work presents a simulation of a modified version of this telescope with TARGET, based
on CORSIKA and sim telarray. The simulation implementation is discussed extensively
and validated on data acquired during a joint measurements campaign with the H.E.S.S.
array in early 2019. The determined proton energy threshold Ep
th = (11.3 ± 0.5) TeV is
in agreement with previous estimates.
Moderne kardiologische Implantate wie Stents und Herzklappenersatzimplantate werden oft kathetergestützt in den menschlichen Körper eingebracht. Der Durchmesser des Katheters mit dem darauf befindlichen Implantat limitiert aufgrund der Passierbarkeit der Blutgefäße bis zum Implantationsort das erreichbare Patientenkollektiv. Bei Transkatheteraortenklappen (TAVI) und Covered Stents wird der Durchmesser des Gesamtsystems durch die Dicke des auf dem Katheter komprimierten Stents und der gefalteten Klappen- bzw. Hüllenkomponente bestimmt. In den letzten Jahren konnten zur Reduzierung des Durchmessers von TAVIs nur marginale Verbesserungen erreicht werden.
Die Kombination aus Materialdicke und mechanischer Stabilität ist für die Stents bereits optimiert, während noch Verbesserungen bei der Gewebekomponente erzielt werden können. Die Dicke des am häufigsten verwendeten Materials für die Klappe einer TAVI, Perikard verschiedenen Ursprungs, ist natürlich vorgegeben und beruht nicht auf konstruktiven Überlegungen. Hier liegt der Ansatz des Vorhabens dieser Dissertation.
Perikard besteht zu etwa einem Viertel aus einer Matrix aus Kollagenfasern, während der Rest des Volumens mit Wasser gefüllt ist. Da die mechanische Festigkeit fast ausschließlich auf der Fasermatrix beruht, wurde der Ansatz verfolgt, die Dicke des Gewebes über den Wassergehalt zu kontrollieren, was in dieser Arbeit erstmals für Perikard erfolgreich umgesetzt wurde. Man erhält dünneres Gewebe durch Dehydrierung und gleichzeitige Fixierung mit Glutaraldehyd. Entscheidend für die Eigenschaften des so erhaltenen Materials ist die Art des Wasserentzugs, dieser kann durch Lufttrocknung, mechanische Krafteinwirkung oder weitere Verfahren erfolgen.
In der vorliegenden Arbeit wurden Verfahren entwickelt, bei denen eine durchlässige Schicht in Form eines Polyestergewebes eingesetzt wird, um den ungehinderten Transport des Wassers zu gewährleisten. Bei der Dickenreduktion erhält man eine stufenlose Kontrolle der Dicke bis hin zur natürlichen unteren Grenze nach fast vollständiger Entfernung des Wassers. Der angelegte äußere Druck bestimmt hierbei die Enddicke des Gewebes, bis hinunter zu weniger als 50% der Ausgangsdicke. Für ein Klappenimplantat mit mehrfacher Faltung des Gewebes auf dem Katheter wird eine deutliche Verringerung des Durchmessers erreicht.
Als Folge einer sehr starken Kompaktierung wird das Perikardgewebe transparent, im Gegensatz zum milchigen Erscheinungsbild im Ausgangszustand. Außerdem führt die stärkere Wechselwirkung der Kollagenfasern untereinander zu einer erhöhten Biegesteifigkeit, was bei der Anwendung in einem Implantat berücksichtigt werden muss.
Das auf diese Art hergestellte Gewebe kann einer Trocknung mit hygroskopischen Austauschstoffen unterzogen werden, ohne Einbußen in seiner Eignung für Implantate. Um diese zu prüfen, wurden Prototypen eines minimalinvasiven Aortenklappenersatzes erfolgreich hergestellt und getestet.
Neben dem Herzklappenersatz ist das entwickelte neue Gewebematerial auch für Covered Stents interessant, erste Versuche dazu wurden bereits durchgeführt, die Feinausführung und Weiterentwicklung dieses Ansatzes steht jedoch noch aus.
Die Kontrolle und Steuerung von Strom ist für die moderne Elektronik unerlässlich. Je schneller Elektronenströme kontrolliert werden, desto schneller werden Informationen übertragen und verarbeitet; die Anforderungen an die Übertragungsgeschwindigkeit steigen dabei stetig mit der fortschreitenden Entwicklung der Technik. Im Rahmen der vorliegenden Dissertation werden verschiedene Nanostrukturen für ultraschnelle Licht-Materie-Wechselwirkung vorgestellt, welche als Grundlage für eine moderne Lichtwellenelektronik dienen können. Die Herstellung der Strukturen mittels Photolithographie ermöglicht die Realisierung maßgeschneiderter Geometrien und komplizierter Anordnungen auf kleinsten Längenskalen. Eine hohe Reproduzierbarkeit und Stabilität durch die Herstellung auf Substrat bilden eine hervorragende Grundlage für optische Experimente. Aufgrund ihrer Größe sind Nanostrukturen eine räumlich wohldefinierte Quelle von Elektronen. In Kombination mit ultraschnellen Laserpulsen weniger Zyklen sind die Wechselwirkungen zwischen Licht und Elektron nicht nur räumlich sondern auch zeitlich wohl definiert und bilden die ideale Grundlage für moderne Lichtwellenelektronik.
Der Fokus der Arbeit liegt auf der elektronenlithographischen Fabrikation von Gold-Nanospitzen, von denen zwei Typen unterschiedlicher Orientierung hergestellt wurden und im Folgenden kurz zusammengefasst werden:
Eine planare Spitze mit Orientierung entlang der Probenoberfläche bietet die Möglichkeit, die Photoemission von Elektronen sowie deren Richtung zu manipulieren und zu detektieren. Aufgrund von Spitzenlängen zwischen 120 und 250 nm können die Auswirkung von Nanoantennenresonanzen auf die Licht-Materie-Wechselwirkung und damit verbunden die elektrischen Nahfelder der Struktur direkt im Photostrom nachgewiesen werden.
Spitzen mit Orientierung senkrecht zur Probenoberfläche weisen hervorragende geometrische Eigenschaften auf, um die elektrischen Nahfelder zu manipulieren und die Photoemission von Elektronen zu kontrollieren. Typische Spitzen-Radii kleiner als 10 nm resultieren in hohen Feldüberhöhungsfaktoren des elektrischen Nahfeldes und ermöglichen vielversprechende Starkfeldeffekte. Zudem ist es möglich, ganze Anordnungen von Spitzen, mit Abständen bis zu 75 nm, herzustellen, was sie zu hervorragenden Kandidaten für Korrelationsexperimente macht.
Neben den Nanospitzen wurden noch zwei weitere Strukturen vorgestellt.
Ebenfalls im Bereich der Licht-Materie-Wechselwirkung anzusiedeln ist die Anregung gebundender Elektronen in Graphen. Zur Untersuchung des wellenformabhängigen Stromes werden einzelne Graphenstreifen mit einer Breite von bis zu 43 nm sowie 11 Graphenstreifen mit 120 nm Breite und 600 nm Abständen zueinander hergestellt und charakterisiert.
Die vierte Struktur dieser Arbeit dient der Vorbereitung auf die Kontrolle freier Elektronen mittels eines ponderomotorischen Potentials. Dafür werden elektrische Durchkontaktierungen benötigt, welche mittels Laserlithographie und plasmaassistiertem Ätzen hergestellt werden.
Tailoring the environments of 2D Materials: Towards structural, electrical and optical control
(2023)
The field of 2-dimensional materials started in 2004 with the discovery of graphene and has been growing ever since due to its promising applications, which could be part of transistors, sensors, photovoltaics and many more.
Inherently, 2D materials have a large surface compared to their volume. This is why the tailored manipulation of the material via their surface is particularly interesting both for fundamental science and for applications. In this thesis, a rather broad variety of surface manipulation techniques is presented, which in turn are applied to a variety of 2D materials. The latter includes the well-studied materials graphene and transistion metal dichalcogenides (TMDs), but also the less known 2D materials of the pnictogens.
The unifying element is the structured, tailored surface control. In some of these
experiments, surface manipulation was carried out by research partners using synthesis chemistry; here, my contributions lay in the measurement of the electrical properties on the one hand and in the lithography adapted to the different challenges on the other. Other projects aimed to optically characterize the influence of 2D materials on other 2D
materials using novel experimental approaches.
Contributions to transport measurements on 2D pnictogens:
• Few-layer black phosphorus, which was noncovalently functionalized with perylene
diimide to prevent degradation in air condition, was measured in a field effect
transistor (FET) configuration, proving the preservation of its suitable electrical
properties. The hysteresis of the gate-dependent 4-probe conductance measurements
could be explained within a developed model.
• Hexagonal few-layer antimony was electrically investigated. Therefore, electrical
contacts were applied on one small synthetically produced sample. The essential result
shows that the material is metallic under all measured circumstances. Thereby,
this thesis contributes to the rare experimental data of thin antimony.
• A highly interesting, novel hexagonal bismuth/bismuthene hybrid material was
electrically characterized using a precisely realized FET configuration. The results
are in perfect agreement with the density functional theory-based simulations, which
predict a metallic behaviour of the outer hexanethiol-functionalized bismuthene
layer on a semi-metallic bismuth core, and with the structural investigations of our partners.
Contributions to patterned functionalization (PF) of 2D materials:
• A new strategy was developed by the combination of e-beam lithography and
functionalization, namely the patterned functionalization of 2D materials. The
technique was applied on graphene and molybdenum disulphide.
Contributions to photoluminescence and environmental control of 2D semiconductors:
• For optical control, a heterostructure was created, consisting of a monolayer MoS2
flake and epitaxial grown graphene on SiC, prepared in a stripe pattern. The photoluminescence
intensity of the A exciton as well as the peak position of the trion
and the A exciton are periodically changing according to the graphene stripe design.
• These experiments shed light on the continuous transition from monolayers to
heterostructures, which was so far experimentally not accessible. However, using
the custom-built setup from our research group, the squeezable nanojunction, a
controlled approach in slow motion was executed and optically characterized. The
change in the optical properties of MoS2-MoS2 or MoS2-WS2 stacks can be observed
in dependence of the separation between the samples. The overlap is further proven by PL-intensity-xy-maps and analysed by multi-peak-fit procedure. Altogether, being well embedded in several vibrant physical-chemical teams, this work contributed to enlarge the scientific toolbox for targeted surface control in terms of electrical, optical and structural properties.
Strong near-infrared laser fields release electrons from a metal surface within a time
window much shorter than a single oscillation of the optical field. The emitted electrons
propagate in the field and can return to the metal surface, where they rescatter elastically and thereby gain additional kinetic energy. In this work, we study these processes by superimposing a fundamental driving field with its second harmonic and use their relative phase to selectively modify the emission and consecutive motion of electrons in the laser field.
Rescattering is negligible for low field strengths of the driving field such that the second harmonic exclusively modulates the emitted electron yield. In this so-called perturbative regime, we demonstrate the applicability of the two-color scheme to needle tips made from the plasmonic material gold. In contrast to tungsten, we observe a frequency component at four times the fundamental frequency in the delay-dependent yield, which an additional third quantum pathway can explain. The three-pathway model correctly
describes the scaling of the electron yield with the second harmonic admixture but
predicts relative pathway strengths differing from the experimental observations and
quantum simulations.
In our main experiment, we increase the field strength of the driving field and observe a
pronounced rescattering plateau in photoelectron spectra from tungsten tips. Now, the
second harmonic modulates the emission rate and the final energies of rescattered electrons, resulting in two phase-dependent signatures within the electron spectra. First, we find a modulation of the high-energy cut-off resulting from a gain or reduction of the electron energy, which reveals the second harmonic nearfield strength. Second, a characteristic optimal phase maximizing the yield at each energy is observed, which for known field strengths only depends on the distribution of instantaneous rates. By matching the experimental results with simulations based on the time-dependent Schrödinger equation, we can – for the first time at the surface of a solid – precisely confine the emission duration to 710 ± 30 attoseconds. An essential requirement for the presented analysis is the excellent match between the measurement and our simple quantum mechanical models. To assure the validity of our models, we compare the predicted rescattering spectra from a single driving field at various intensities and two wavelengths against corresponding experiments.
The spatial localization of the electron emission at the tip is not only crucial for the
clear phase-dependent signatures but conversely allows us to characterize the spatial
properties of the two-color field. By moving the tip within the focal volume, we obtain
the spatially varying relative phase. Two measurements with different second harmonic
focal sizes finally allow us to separate the relative phase distribution into the focal phases of the individual fields. As a first application, we map the focal phase distortions when the few-cycle two-color field is tightly focused through a glass window.
In the future, understanding the sub-cycle dynamics and quantifying the emission duration
may assist in developing light-driven devices and applications. Furthermore,
our results open a new route to identifying fundamental similarities and differences in
strong-field physics at metal surfaces compared to atomic systems and within solids.
Every cell in the human body, regardless of its type, is exposed to mechanical stimuli. These mechanical stimuli can take many forms: Muscle contractions cause stretch in the surrounding tissue, blood flow causes shear stress on cells at the surface of blood vessels, and the mechanical stiffness of tissue has fundamental effects on the proper functioning of cells within that tissue. Cells can directly sense mechanical stresses, deformations, and the stiffness of their environment. Cellular mechanosensors are for example integrins, which are transmembrane proteins that form physical links with the surrounding matrix, and stretch-activated ion channels in the cell membrane. Through complex intracellular mechanochemical signal transduction pathways, mechanical stimuli are transmitted and processed by the cell, leading to a large range of responses at the cellular or tissue level. These responses can be rapid, short-term, and physiological, such as in the form of altered cell shape or migration. However, in the case of longterm abnormal mechanical stimuli, cellular adaptation can also lead to pathological consequences. For example, increased fluid shear stress in blood vessels can lead to tissue inflammation and arteriosclerosis. Numerous seemingly unrelated diseases can be attributed to altered or impaired cellular mechanotransduction, including various muscular dystrophies, polycystic kidney disease, and some forms of hearing loss. Often, these diseases can be attributed to genetic mutations of individual structural components of mechanotransduction, such as intermediate filaments of the cellular cytoskeleton (for example, mutations of the protein desmin in muscle tissue), or in proteins that connect the cytoskeleton to the nucleus (for example, mutations in lamins in the nuclear lamina). It is therefore not surprising that mechanical properties and mechanotransduction of cells have increasingly become a focus in basic and clinical research.
The study of cellular mechanotransduction requires appropriate in vitro cell culture methods. Classically, two-dimensional cell cultures, e.g. in Petri dishes, are used for this purpose, which allow for controlled laboratory conditions with relatively high experimental throughput, but only reflect to a limited extent the complex geometric environment to which cells are exposed in vivo. A current research trend is therefore to use three-dimensional cell culture, whereby cells grow in a native or synthetic extracellular matrix. For this, methods and devices are needed to apply mechanical stimuli on cells. A simple approach here is to seed cells on substrates with controlled stiffness in order to mimic varied tissue mechanics. This can be easily achieved, for example, with polydimethylsiloxane, a biocompatible silicone-like gel with adaptable Young's modulus. Mechanical strain, on the other hand, can be applied by cell stretchers, which are devices that can deform flexible cell substrates in a controlled manner in one or more directions. Finally, methods and devices are needed to characterize the mechanical properties of cells. For this purpose, there are contact-based methods, such as indenters, which are used to investigate the relationship between applied stress and resulting cell strain (or the other way around), as well as contactless methods, such as optical or magnetic tweezers, in which the application of deformation (or force) is achieved by electromagnetic field gradients. In order to interpret the results obtained by such methods, models are needed that explain the mechanical, in particular the viscoelastic, properties of cells dominated by the cytoskeleton.
In the following cumulative dissertation, I will first provide an overview of the current state of research regarding mechanical stimuli on cells, the cellular mechanosensors and mechanisms of cellular mechanotransduction, their impairment in disease, as well as the mechanical properties of cells. I will then focus on the methods and devices that can be used to investigate the resulting questions in a laboratory context. During my review I will give concrete examples which are mostly taken from a series of reports that I published during the time of my doctorate. A selection of five of my first-author papers are attached in full to this dissertation.
In this thesis, the effects of quenched disorder on zero-temperature spectral
densities in gapped quantum antiferromagnets are studied. We calculate
effective models in the presence of quenched disorder for gapped systems perturbatively
with perturbative continuous unitary transformations (pCUT) and
solve the models afterwards. Apart from exact diagonalisations (ED), Green’s
function techniques are used to solve the models. The one- and two-triplon
dynamic structure factors (DSFs) are calculated for Heisenberg quantum spin
ladders and the differences between rung and leg disorder are investigated.
For sufficiently strong bimodal rung disorder it is found that the DSF decouples
into finite ladder segments. The material(C5H12N)2Cu(ClxBr1-x)4
(BPCBC) is fitted and comparisons with inelastic neutron scattering (INS)
measurements of single crystals and powder are shown. For potential Raman
scattering measurements on BPCBC predictions are given. The DSF of the
Heisenberg bilayer model (HBLM) on square, triangular and kagome lattices
is calculated. For that, the linked-cluster property of pCUT is utilized by
performing a graph expansion for effective Hamiltonian and observable. The
distribution of disorder only has a minor effect on the inter-dimer disorder
results. Consequently, the self-consistent Born approximation agrees very well
with ED results. For intra-dimer disorder, one can not neglect the form of the
probability distribution. In the case of bimodal intra-dimer disorder, in-band
gaps emerge at certain momenta, where the value of the momenta depends on
the probability of the bimodal disorder. The coherent potential approximation
(CPA) does not give a quantitative agreement with the ED results for this
kind of disorder but can deliver the right qualitative features. INS data on the
material NiCl2-2xBr2x*4SC(NH2)2 with x=0.06 is modelled by first using
pCUT to obtain the best possible fit for the undoped material. Then the effect
of one impurity is determined non-perturbatively on a small one-dimensional
cluster with a CUT using the quasi-particle generator. In this way, reasonably
well agreement with the INS data is found using a bimodal disorder of D' and
Jc'. The CPA and molecular coherent potential approximation (MCPA) is used
to derive order by order a series in the Ising and the inter-dimer couplings in the
spin-diluted transverse-field Ising model (TFIM) and the dimer-diluted HBLM.
To this end, the self-energy for a given momentum is determined at the energy
where the imaginary part of the self-energy becomes zero. -ImG(k,w)/pi is
calculated and compared with perturbative pCUT results of finite systems,
where ED was used to diagonalise them, showing good agreement. Using extrapolations,
the critical behaviour of the series results is examined. This way
the phase transition in both models is studied. Rather surprisingly, in the
TFIM extrapolations yield very good agreement with preliminary quantum
Monte Carlo (QMC) simulations up to dilution concentrations of 0.35. In the
HBLM the agreement with former QMC simulations is only good up to dilution
concentrations of 0.2.
Electron emitters in the form of nanometer-sharp tips, nanotubes, and nanowires are an essential building block in a variety of fields, such as X-ray computed tomography, flexible displays, chemical sensors, and electron microscopy. In time-resolved operation, tip-based emitters triggered with femtosecond laser pulses are well known for their excellent transverse coherence and strongly enhanced near-field. Highly efficient coherent control of multiphoton photoemission in a two-color laser field has been demonstrated recently.
In this thesis, a novel electron emitter based on a single-crystalline nanowire (NW) of LaB6 is investigated under illumination with femtosecond laser pulses. The LaB6 NW emitter has been previously demonstrated as an ultrabright and monochromatic DC field emitter where the achieved normalized
DC brightness is up to two orders of magnitude higher than that achievable
with W(310) tips. Multiphoton photoemission from the LaB6 NW emitter is
demonstrated, and the photon orders are determined from the power scaling
measurements. The low half divergence angle (82 mrad) of the photoelectron
beam shows an increase with decreasing wavelength when the wavelength
is varied from 256 nm to 2650 nm. The proposed model based on the comparison between the determined total absorption energy for different wavelengths and the density of states (DOS) of LaB6 explains the observed increase with good agreements. Furthermore, a biprism interference measurement is performed to determine the effective source size of the emitter under the illumination of a few-cycle femtosecond laser. An effective source size of (1.5 +/- 0.16) nm is obtained. The obtained beam normalized rms emittance (1.30 pm rad) and the normalized rms peak brightness (3.77*10^17 A/m^2 sr) show excellent beam quality. The results enable an ultrafast and ultrabright nanowire emitter that meets the requirements of dielectric laser accelerators (DLA) and other applications such as ultrafast electron microscopes.
Furthermore, quantum interference visibility spectroscopy of two-color photoemission from tungsten needle tips is demonstrated. During the excitation by an intense bichromatic laser field, electrons can be emitted through different excitation pathways involving different combinations of photons of each color. Quantum interference between these pathways creates a strong oscillation of the photoelectron current, which is experimentally characterized by its visibility (contrast). The visibility of the quantum-pathway interference is studied over a nearly octave-spanning wavelength range of the fundamental (ω) and the second harmonic (2ω) laser pulses. The results show high interference visibility (90% +/- 5%) with a remarkably constant distribution over the measured wavelength range. Moreover, it is observed that by varying the relative intensity ratio of the two colors, the visibility from 0% to nearly 100% can be varied. The proposed simple yet insightful theoretical model explains all observations with excellent quantitative agreements. It is envisioned that the presented quantum interference visibility spectroscopy will offer a powerful tool for gaining deep insights into multiphoton processes and photoemission dynamics in materials such as individual atoms, molecules, clusters, and nanomaterials, as well as extended surfaces.
Das aktuell hohe Datenaufkommen, welches aufgrund von Internet of Things (Internet der Dinge), Industrie 4.0 oder auch autonomes Fahren weiter ansteigt, stellt die Datenübertragung vor neue Herausforderungen. Insbesondere die Gerätekommunikation benötigt neue technologische Lösungen, welche von der Geschwindigkeit und Effizienz der optischen Datenübertragung profitieren können.\\
Ein Ansatz, welcher von der DFG-geförderten Forschergruppe OPTAVER erforscht wurde, beruht auf dreidimensional gedruckten Polymer Optischen Wellenleitern (POW). In dieser Arbeit wird der Druckvorgang dieser POWs, auch OPTAVER-Prozess genannt, vorgestellt. Die aus dem Prozess resultierenden POWs werden auf ihre optisch relevanten Parameter hin untersucht, um ihre Qualität zu beurteilen, aber vor allem auch, um ein mathematisches Abbild für Simulationen so detailliert wie möglich zu schaffen.\\
Mit dem Hauptziel dieser Arbeit, ein Simulationstool zu schaffen, wird zunächst die Generierung eines detaillierten mathematischen Modells von POWs aufgezeigt. Dabei wird zwischen einem Modell für strahlenoptische Betrachtungen (Raytrace) und wellenoptischen Betrachtungen (Wave Propagation Method) unterschieden. Aufgrund der besonderen geometrischen Eigenschaften der POWs wird ein Simulationsalgorithmus (der sogenannten BARc, Bisektions-Algorithmus für Raytracing) entwickelt, welcher beliebig verlegte Wellenleiter mit beliebigem Querschnitt simulieren kann.\\
Um die Funktionalität des BARc-Algorithmus zu testen, wird an Beispielen von echten Wellenleitern demonstriert, wie ein mathematisches Modell aus den Wellenleitern entsteht, um sie anschließend zu simulieren. Dabei ist es wichtig, alle möglichen Ursachen für Dämpfungen wie Rauigkeit der Oberflächen, Welligkeiten oder direkte Absorption (hier im Modell nicht enthalten) zu berücksichtigen. Der durchschnittliche Unterschied in der Dämpfung zwischen experimentellem Aufbau und Simulation beträgt lediglich 0,07dB/cm. Mit einem solch zuverlässigen Tool werden daher verschiedenste Einflüsse auf die Dämpfung in einem POW untersucht. Mit dem Ziel der Optimierung des Druckprozesses können Einflüsse wie Rauigkeit der Oberflächen oder Randwelligkeiten der POWs als Quelle von Verlusten mit Hilfe von Simulationen ausgeschlossen werden.\\
Simulationen von sogenannten Droplets und Einschlüssen zeigen jedoch schließlich das Optimierungspotenzial des OPTAVER-Prozesses auf: Diese makroskopischen Fehlstellen treten auf, wenn der Materialfluss in der Düse des verwendeten Aerosol Jet Druckers nicht konstant ist, und können zu Verlusten von mehr als 1 dB/cm führen. Zirka ein Viertel aller produzierten POWs besitzen jedoch bereits einen Wert von weniger als 0,5dB/cm und können in der Spitze Werte von 0,2dB/cm erreichen. Diese Arbeit kommt daher zu dem Schluss, dass POWs, unter der Annahme einer weiteren Optimierung des Prozesses, das Potential besitzen, eine reale Alternative zur Kupfer-basierten Datenübertragung darzustellen. Weiterhin liefert sie ein zuverlässiges Simulationstool, um Dämpfungswerte für dreidimensional verlegte Wellenleiter mit beliebigem Querschnitt zu berechnen.\\
Neben der strahlenoptischen Simulation werden Simulationen der POWs mittels der wellenoptischen Simulationsmethode Wave Propagation Method (WPM) vorgestellt. Da ihre Anwendung noch nicht zur Simulation von Wellenleitern gezeigt wurde, wendet diese Arbeit die WPM an und nicht die typischerweise verwendete Beam Propagation Method (BPM). Die Validierung der Verwendung der WPM ergibt jedoch ein nicht zufriedenstellendes Resultat: während grundlegende Phänomene, wie die in der vorliegenden Arbeit untersuchte \glqq Totalreflektion in einem Wellenleiter\grqq{} oder \glqq Brechung nach den Snell'schen Gesetzen\grqq{}, korrekt simuliert werden, haben die Wahl der Anzahl der Stützstellen sowie der Brechzahlindex-Unterschied der Materialien einen enormen Einfluss auf das korrekte Ergebnis der Simulation. Bis hier hin besteht noch kein komplettes Verständnis des Einflusses, sodass die hier berechneten WPM-Simulationen als erste Untersuchungen der Anwendung der WPM zur Simulation von Wellenleitern zu verstehen sind.
In the context of fracture simulations of polymers, the molecular mechanisms in the vicinity of the crack tip are of particular interest. Nevertheless, to keep the computational cost to a minimum, a coarser resolution must be used in the remaining regions of the numerical sample. For the specific case of amorphous polymers, the Capriccio method bridges the gap between the length and time scales involved at the different levels of resolution by concurrently coupling molecular dynamics (MD) with the finite element method (FEM). Within the scope of the Capriccio approach, the coupling to the molecular MD region introduces non-periodic, so-called stochastic boundary conditions (SBC). In similarity to typical simulations under periodic boundary conditions (PBC), the SBC MD simulations must reach an equilibrium state before mechanical loads are exerted on the coupled systems. In this contribution, we hence extensively study the equilibration properties of non-periodic MD samples using the Capriccio method. From a thorough parameter study, which in general reveals only minor effects of the considered quantities on the equilibrium state, a best set of parameters follows. The behavior of an exemplary system equilibrated with this parameter set is further studied under uniaxial tension and we observe some peculiarities in view of creep and relaxation phenomena. This raises important questions to be addressed in the further development of the Capriccio method.
Hot subdwarf stars of spectral types O and B (sdOs/sdBs) are located on and beyond the very hot end of the horizontal branch in the Hertzsprung-Russell diagram. They are highly evolved (post-)extreme horizontal branch ((post-)EHB) objects that are generally believed to be in the core helium-burning phase or beyond. Since the (post-)EHB covers a wide range of objects of different subtypes showing a variety of different properties, sdOs/sdBs can be considered a stellar zoo. These rather compact objects with radii of ~0.10-0.30 solar radii have stellar masses of about half a solar mass and exhibit hydrogen envelopes that typically make up less than 1% of the total mass. Such thin envelopes cannot be reached during a canonical stellar evolution on the red giant branch involving a core helium flash. In the past, several evolutionary links between the subtypes of sdOs/sdBs and other classes of stars have been proposed. However, various investigations are lacking to fully resolve the question on how these remarkable objects form.
To this end, the present work entitled "Fundamental Analysis of Hot Subdwarf Stars in the Gaia Era" presents an in-depth fundamental stellar analysis of a carefully chosen set of 63 known and candidate sdOs/sdBs that represents all relevant subtypes. The sample covers the full range of atmospheric parameters (effective temperature, surface gravity, and helium abundance) of single and binary stars, pulsating and non-pulsating objects, and stars with particularly peculiar abundance anomalies, including ³He enrichment.
Precise and accurate atmospheric parameters are prerequisites in order to understand the nature and the evolution of the various subtypes of hot subdwarf stars. The first part of this work therefore focuses on in-depth spectroscopic analyses, making use of observed spectra of excellent quality in terms of signal-to-noise, spectral resolution, and wavelength coverage as well as of sophisticated model atmospheres and analysis strategies. A sample of 17 stars with optimum data quality (spectra taken with the XSHOOTER spectrograph at the ESO VLT) serves as a reference. In terms of model atmospheres, three sets are used in order to study the impact of different effects in great detail, including departures from local thermodynamic equilibrium (LTE) and metal line-blanketing. Classical metal line-blanketed LTE models (Heber et al., 2000) are compared to non-LTE (NLTE) model atmospheres with limited metal line-blanketing as well as to hybrid models that allow to treat NLTE effects and extensive metal line-blanketing. In terms of analysis strategies, two different methods are compared: the classical approach of analyzing preselected suitable spectral lines versus a global approach where the whole spectrum is fitted at once and only obvious outliers are excluded.
For the very first time, precise parallax measurements of the Gaia satellite allow to reliably convert atmospheric parameters into fundamental ones (radius R, luminosity L, and mass M). In order to do so, however, knowledge of the stellar angular diameter is required, which can be retrieved from spectral energy distribution (SED) fitting to appropriate photometric data. Hence, the second part of this work deals with the construction and the analysis of SEDs as well as with the subsequent derivation of the fundamental stellar parameters. In this way, a consistent comparison to the predictions of the theoretical evolutionary models for hot subdwarf stars is ensured.
The results of the present work show that the optimum effective temperature regime for the LTE models is between ~25,000 K and ~32,000 K. For lower and in particular higher temperatures, NLTE effects prevail. On average, the hybrid models result in higher surface gravities compared to LTE (by ~0.05 to ~0.10 dex), but at the same time the corresponding helium abundances are lower. This is explained by the anti-correlation between the surface gravity and the helium abundance. Compared to LTE, the NLTE models yield significantly higher effective temperature values (up to ~1700 K) for the hotter stars, which can mainly be explained by the backwarming effect due to the limited metal line-blanketing of the NLTE models. The same applies to the comparison between the hybrid and the NLTE models for which differences of up to ~2000 K are measured for individual program stars. Interestingly, the hybrid models also result, on average, in higher surface gravities (by ~0.08 to ~0.15 dex) compared to the NLTE models. Therefore, it is highly likely that metal line-blanketing also plays an essential role here. For the effective temperature regime between ~20,000 K and ~40,000 K, the hybrid models can be considered the new standard because these models yield consistent results, regardless of whether the selective or the global analysis approach is used.
Additionally, regression curves are determined, allowing to update published atmospheric parameters without having to actually perform the respective spectroscopic analyses based on the new hybrid LTE/NLTE approach.
In literature, the hydrogen Paschen series has very rarely been used as a diagnostic tool for hot subdwarf stars. Instead, the Balmer series is often used. The tests performed in this work show that both series deliver consistent results. This is very promising because many of the current and future spectrographs are configured for the near-infrared wavelength regime with the Paschen series.
Building on these investigations, the hybrid models are primarily used to evaluate the spectra of the sample quantitatively. It turns out that two thirds (~67%) of all program stars belong to the group of hydrogen-rich hot subdwarfs, whereas their (intermediate) helium-rich siblings make up ~17%. The majority of the program stars (~62%) are indeed core helium-burning objects that are located on the EHB. Another ~17% are more evolved post-EHB objects for which core helium burning has already ended. The nature of two stars remains unclear. One object (GALEX J080510.9-105834) is confirmed to be a progenitor of an extremely low-mass helium-core white dwarf (pre-ELM), whereas for two other stars (Feige 36 and BD+42° 3250) a pre-ELM nature seems highly likely. Two more objects may be pre-ELMs, too, and one star could also be a post-asymptotic giant branch (post-AGB) object. Most of the observed SEDs are matched with single model atmospheres, but four targets show clear infrared excesses indicating the presence of a cool companion. Three of these spectroscopic binaries (SB 290, Feige 36, and EC 01541-1409) are new discoveries.
The isotopic abundance anomaly of helium is also investigated. For the analyzed ³He hydrogen-rich sdB program stars, it is found to be restricted to a narrow temperature strip between ~26,000 K and ~30,000 K. Strikingly, about half of the analyzed ³He program stars show anomalous helium line profiles. The unusually broad wings and shallow cores of the spectral lines of the relevant stars indicate that the total helium abundance decreases with decreasing depth of the atmosphere by up to a factor of ~9.0, leading to an abundance stratification.
The next analysis step comprises the determination of the metal abundances. For this, almost all lines detected in the optical and near-infrared wavelength regime of the analyzed high and medium-resolution spectra are used. The abundances of the chemical elements C, N, O, Ne, Mg, Al, Si, S, Ar, and Fe are analyzed in NLTE. In addition, P, Ca, Ti, Sr, and Zr lines are found for some program stars of the XSHOOTER reference sample and are analyzed in LTE. In consequence, the metal abundance study presented in this work represents a major improvement over published results. The analyses of the sharp metal line profiles also allow to accurately determine the projected rotational velocities. As expected, most of the program stars turn out to be slow rotators. However, eleven stars show significant rotation, including three new discoveries. The exceptionally high projected rotational velocity of 142.0^{+9.0}_{-11.0} km/s of GALEX J203913.4+201309 is truly remarkable given the fact that faster rotation has only been reported for hot subdwarf stars that have filled their Roche lobes and started mass transfer onto a white dwarf companion.
Last but not least, the radii, the luminosities, and the masses of the sample stars are determined. The radius distribution (0.10-0.30 solar radii) determined for the hydrogen-rich hot subdwarf program stars is consistent with predictions of canonical evolutionary models, but is of bimodal shape, showing two well-defined peaks at 0.138 solar radii and 0.205 solar radii. The same applies to the corresponding luminosity distribution (with abscissa log L/L⊙), which peaks at 1.291 and 2.090, respectively. The mass distribution is also largely consistent with the predictions of the canonical models and can be described by a single Gaussian peaking at 0.465 solar masses. Regarding suggested evolutionary scenarios, this implies that the close binaries in the sample may have resulted from envelope stripping by the companion at the end of the red giant branch. For the single hydrogen-rich hot subdwarf program stars, however, internal mixing seems to be more likely than the competing scenario of the merger of two helium-core white dwarfs, although the latter could in principle explain the observed high-mass outliers. Some outliers at the low-mass end of the mass distribution are rather difficult to reconcile with any of the formation channels of hot subdwarf stars known today.
This work represents a milestone for future studies of hot subdwarf stars by quantifying the systematic uncertainties of the three atmospheric parameters effective temperature, surface gravity, and helium abundance. Together with the more precise trigonometric parallaxes from future data releases of the Gaia mission, this will allow to reliably determine the fundamental stellar parameters of thousands of hot subdwarf stars. The corresponding radius, luminosity, and mass distribution will lead to stringent tests of the theoretical evolutionary scenarios.
In this thesis I use multi-wavelength data from radio up to γ-rays to study emission processes
in the jets of flaring blazars and blazar-like active galactic nuclei (AGN). Powered by accretion
onto a supermassive black hole, AGN can produce luminosities that outshine their host galaxy.
A fraction of AGN form large-scale, collimated outflows of particles moving at relativistic
speed, which are called jets. In case such a jet is orientated towards our line of sight at a
small angle, the object is called a blazar. Blazars exhibit a variable behaviour across the
entire electromagnetic spectrum, and, on occasion, show flares, which are a dramatic increase
in luminosity at one or several wavelengths. The processes partaking in flares are not fully
understood, and rapid variability detected at γ-ray energies is a challenge for current jet
models.
The blazar Mrk 421 is one of the brightest sources in the extragalactic sky, and regularly
shows bright flares. I analysed multi-wavelength data obtained as part of a monitoring program
with the γ-ray telescope FACT and the optical, UV and X-ray instruments on-board Swift,
and, in particular, during a γ-ray flare in June 2019. A combined fit of the optical, UV and
X-ray data, including hard X-ray data from INTEGRAL, with a log-parabola model revealed a
shift of the peak position of the low-energy hump in the typical two-hump broadband spectral
energy distribution (SED) towards higher energies. This behaviour is also covered by a long
(90 ks) XMM-Newton observation. I displayed the X-ray data in hardness-intensity diagrams
and a hysteresis curve, and found a steady increase of high energy photons in the 4–10 keV
band during a general increase of the flux, known as the ‘harder-when-brighter’ trend, which
has been observed in Mrk 421 and other blazars before. Searching for signals of variability
on very short time scales, I discover two distinct features in the hard (4–10 keV) X-ray band,
which resemble ‘mini-flares’ on top of the generally variable emission. The associated time
scales are on the order of four to five minutes, and as short as the rapid variability detected at
(very high energy) γ-rays in a few other blazars. A process that could explain rapid variability
in both energy regimes is magnetic reconnection, which has been predicted to produce certain
polarisation characteristics that might be detectable in the future with IXPE.
In this dissertation, I also present a multi-wavelength study of the narrow-line Seyfert 1
(NLSy 1) galaxy PKS 2004−447 in the context of its first γ-ray flare. Typically, NLSy 1s
are non-jetted AGN, but a fraction of them have been found to host relativistic jets, and a
small number of them was also detected by the γ-ray telescope LAT on-board the Fermi
satellite. PKS 2004−447 belongs to the latter ones, and showed moderate variability since
Fermi/LAT began surveying the sky, until it exhibited a blazar-like γ-ray flare in October 2019.
Target-of-opportunity observations were conducted with Swift, XMM-Newton, and NuSTAR to
study optical, UV, and X-ray emission of the source during and after the flare. Due to good
multi-wavelength coverage at different activity phases of the source, it was possible to build
time-resolved broadband SEDs with quasi-simultaneous data. A leptonic model describes
the flaring behaviour sufficiently well, and reveals a strong resemblance to blazar, and in
particular to flat-spectrum radio quasars. Furthermore, I analyse the combined X-ray data
from XMM-Newton and NuSTAR, and find that the X-ray spectrum is strongly dominated by
the jet, and lacks typical properties of NLSy 1 X-ray spectra. While PKS 2004−447 shows
longer-term variability at all wavelengths, the γ-ray light curve reveals short-term variability on the order of hours during its state of enhanced γ-ray flux. Based on these findings, and due
to its possibly low-mass black hole (< 10^8 M⊙ ) and compact radio emission, PKS 2004−447
might be a powerful blazar in the making.
X-ray microscopy and phase imaging towards time-resolved applications in laboratory astrophysics
(2022)
In this work, a characterization and optimization of phase-sensitive X-ray imaging techniques with a focus
on the field of laboratory astrophysics is given. Here, the advent of hard X-ray free electron lasers offers
novel opportunities as single-pulse imaging with sub-picosecond temporal resolution becomes possible. The
use of phase-sensitive techniques is often mandatory as micro- and nanoscopic samples show little or no attenuation
contrast. In order to fully benefit from the short pulse lengths at X-ray free electron lasers, these
methods should be reconcilable with single-exposure acquisition schemes. This task is complicated by zeroes
in the respective transfer functions of the imaging systems. Therefore, direct inversions are typically not
possible and sophisticated algorithms are required for the image reconstruction. Overall, this thesis mainly
focuses upon the grating-based X-ray imaging technique, also known as Talbot interferometry. For comparison,
propagation-based phase contrast imaging will also be considered. The investigations are divided into
analytical considerations, numerical simulations, and experimental implementations of the respective imaging
techniques.
An analytical examination of the image formation within a Talbot interferometer is presented. This process
can become complicated, especially for applications in X-ray microscopes. Here, transverse shifts of the
interference pattern in general depend nonlinearly on the phase differences across the X-ray wave field. Existing
reconstruction methods on the basis of deconvolutions then rely on idealized conditions, thus limiting
the experimental applicability of the method. In addition, the achievable spatial resolution of Talbot interferometry
in single-exposure applications is typically limited to the demagnified fringe period of the interference
pattern.
In order to resolve the limitations regarding the applicability, three novel reconstruction methods for Talbot
interferometry are conceptualized and implemented: the design of a beam-splitting diffraction grating featuring
only two diffraction orders, a two-stage deconvolution approach, and a statistical image reconstruction
method based on an analytical forward model of the imaging process and a regularized maximum likelihood
approach. The three schemes are validated on the basis of simulated data. They all prove advantageous when
the premises for standard deconvolution-based reconstructions are not met. The statistical image reconstruction
technique seems most promising as it achieves the best reconstruction quality at low photon numbers and
also circumvents the abovementioned limitations regarding the spatial resolution.
Building up on the simulative studies, two experimental realizations of Talbot interferometry at synchrotron
light sources are presented. In the first experiment, the single-exposure phase imaging capabilities
of Talbot interferometry in conjunction with the statistical image reconstruction method are investigated and
characterized on the basis of simple test samples. The broadened experimental applicability is demonstrated
through the retrieval of Fresnel diffraction images in an X-ray projection microscope. While a comparative
implementation of propagation-based phase contrast imaging at the same instrument still yields a superior
spatial resolution, the mitigation of limitations due to the fringe period is also verified experimentally. In the
second experiment, single-exposure phase imaging with both the grating-based and the propagation-based
approach is employed in order to monitor the moistening process of wood on the level of single wood cells.
While a hypothesized humidity-induced pore system could not be detected, moderate swelling of the wood
samples in temporal correlation with the ambient air moisture was observed.
The feasibility of a potential experiment from the field of laboratory astrophysics is studied via numerical
simulations. Here, it is investigated whether the formation of laser-induced collisionless shock waves by the
Weibel instability can be observed and examined with X-ray phase imaging. Simulations of the image formation
are carried out for two optimized setups and phase images are reconstructed for different noise levels of
the intensity images. This allows the specification of limits regarding the detectability of different sample features.
The results indicate that a sufficiently accurate characterization is possible with the propagation-based imaging technique, assuming both an optimized X-ray detection system as well as the absence of significant
noise sources other than photon noise.
In conclusion, the results obtained in this work point out a broader and more flexible applicability of X-ray
Talbot interferometry. As demonstrated by the presented experimental and simulative studies, a superior image
quality in single-exposure applications can be achieved with the propagation-based method at the current
state. Potential future improvements of the developed statistical image reconstruction technique can entail
less idealizations in the forward model as well as more advanced regularization schemes. Another important
question is how the grating-based and the propagation-based approach differ in the way that intrinsic
pulse-to-pulse fluctuations at X-ray free electron lasers can be handled. The concepts and methods developed
within this thesis provide a good starting point for such investigations and hence represent a first step towards
a broad practical application of X-ray phase imaging at X-ray free electron lasers in general and for laboratory
astrophysics in particular.
In dieser Arbeit wird die Machbarkeit der Wasserstoff (1H) Magnetresonanztomographie (MRT) mit ultrakurzen Echozeiten zur direkten Bildgebung von ultrakurzen T2*-Komponenten der weißen Hirnsubstanz bei 7 Tesla unter Verwendung der inversionsvorbereiteten Doppelechodifferenzbildgebung mit ultrakurzen Echozeiten (IR-Diff-UTE) demonstriert. Eigenschaften, Chancen und Limitationen dieser Methode werden bei dieser neuen Feldstärke untersucht: Auftretende Artefakte werden reduziert, MR Parameter in der weißen Hirnsubstanz quantifiziert und darauf basierend ein klinisch hochaufgelöstes Protokoll mit UTE Differenz- und Fraktionskontrast vorgestellt.
Die IR-Diff-UTE Technik unterdrückt lange T2*-Signale in der weißen Hirnsubstanz durch Verwendung einer adiabatischen Inversionsvorbereitung in Kombination mit Doppelechodifferenzbildgebung. Mittels Bloch-Simulationen wurde das entsprechende Signalverhalten aller relevanten Gewebekompartimente nachgebildet und analysiert. Artefakte, die bei 7T durch Einfaltungen und Verschmierungen langer T2*-Fettsignale aus der Kopfhaut entstehen, wurden reduziert indem die IR Zentralfrequenz so verschoben wurde, dass die betreffenden Frequenzbereiche ebenfalls invertiert wurden. An 8 gesunden Probanden wurden anschließend die T2*-Relaxationszeiten der weißen Hirnsubstanz in Kompartimenten quantifiziert. Darauf aufbauend wurden an 20 gesunden Probanden das Signal- und Kontrast-zu-Rausch Verhältnis (SNR/CNR), sowie die Artefaktunterdrückung als auch die Stabilität der IR-Diff-UTE Kontraste evaluiert. Schließlich wurde bei 6 an Multipler Sklerose (MS) erkrankten Patienten die Fähigkeit der Technik untersucht, krankheitsbedingte Signaländerungen anzuzeigen.
Das in dieser Arbeit verwendete 7 Tesla MRT System ermöglichte UTE Messungen im Gehirn mit einer Totzeit von 30 μs. Wirbelstromeffekte nullter Ordnung wurden als vernachlässigbar charakterisiert (Δ < 15°), lineare Beiträge traten besonders auf der x- und y-Gradientenachse auf (Δ < 6m−1) und mussten retrospektiv korrigiert werden. An Messphantomen wurden die vorgeschlagenen Studien zunächst als geeignet für die in vivo Anwendungen verifiziert. Eine optimale IR Unterdrückung der langen T2*-Komponenten in der weißen Hirnsubstanz wurde bei TR = 1500 ms für TI = 430 ms im vorgeschlagenen Quantifizierungsprotokoll und für TI = 465 ms im klinischen Protokoll gefunden. Durch Simulationen und Probandenmessungen wurde aufgezeigt, dass eine Frequenzverschiebung des IR Pulses um −1,2 ppm (d.h. in Richtung der Fettfrequenzen) zu einer guten Unterdrückung der Fettartefakte führt. Damit konnte in der Quantifizierungsstudie ein ultrakurzes Kompartiment von (68 ± 6) % mit einer T2*-Zeit von (147 ± 58) μs und einer chemischen Verschiebung von (−3,6 ± 0,5) ppm von Wasser quantifiziert werden. In der klinischen Kontraststudie wurde für die weiße Hirnsubstanz gesunder Probanden ein stabiler ultrakurzer T2*-Fraktionskontrast von 0,57 ± 0,01 mit einer durchschnittlichen Standardabweichung von 0,20 ± 0,01 berechnet. Für den Differenzkontrast wurden SNRDiff = 4,7±1,1 und CNRDiff = 8,7±2,4 bestimmt. Bei den MS-Patienten wurde eine signifikante Reduktion der gemessenen ultrakurzen Fraktionswerte beobachtet, sowohl in den identifizierten Läsionen (−0,09 ± 0,09) als auch in der normal aussehenden weißen Substanz (0,54 ± 0,05).
Die im Rahmen dieser Arbeit gefundenen Ergebnisse deuten darauf hin, dass die in der weißen Hirnsubstanz gemessenen ultrakurzen T2*-Komponenten in erster Linie direkt aus dem Myelingewebe stammen. Die direkte IR-Diff-UTE Bildgebung von ultrakurzen T2*-Komponenten der weißen Substanz ist somit bei 7 Tesla artefaktfrei, mit hoher quantitativer Stabilität und guter Erkennung von Signalverlusten bei MS möglich.