Naturwissenschaftliche Fakultät

Permanent URI for this collection

Browse

Recent Submissions

Now showing 1 - 20 of 3992
  • Article
    Open Access
    Gain of Function Recyclable Photoswitches: Reversible Simultaneous Substitution and Photochromism Generation
    (Wiley-VCH, 2024-02-05) Zitzmann, Max; Fröhling, Matthias; Dube, Henry
    The use of molecular photoswitches has spread to virtually every field of pure and applied chemistry because of the extraordinary level of control they provide over the behavior of matter at the smallest scales. Photoswitches possess at least two different states with distinct structures and/or electronics and further functionalization of their core chromophore structures is needed to tailor them for a specific application. In this work we present a different concept for the generation and use of molecular photoswitches. It allows not only simultaneous establishment of photochromism and func- tionalization, but also full recyclability of a non-photo- chromic precursor material. Using a high-yielding and reversible ammonium salt formation, a functional group is introduced into a symmetric precursor while at the same time a strong electronic push-pull character is established in the structure. The resulting desymmetriza- tion leads to efficient photoswitching capacity and the functional group can be fully removed subsequently by a simple heating step recovering the precursor for another functionalization round. We finally demonstrate feasibil- ity of this concept over two consecutive closed loop functionalization/photoswitching/recovery steps. This concept offers great potential in any chemical research and application driven area but especially for the creation of responsive reprogrammable materials, no- background photoswitch labeling, and sustainable chemistry.
  • Doctoral thesis
    Open Access
    Anwendungen kovalenter Split-Isopeptidsysteme in vitro und in planta
    (2024) Lang, Martina; Sonnewald, Uwe
    Split-Isopeptidsysteme sind kovalente Proteininteraktionssysteme, die auf der natürlichen Ausbildung intramolekularer Isopeptidbindungen innerhalb von Domänen des CnaB-Typs basieren. Diese Art von Domänen ist besonders in Oberflächenproteinen Gram-positiver Bakterien zu finden (Pröschel et al., 2017). Die Spaltung einer solchen Domäne in zwei individuelle Teile, einem Tag-Peptid und einem größeren Catcher-Protein, markierte den Beginn der Split-Isopeptidsysteme. Daraus entwickelten sich zahlreiche weitere Systeme, wie das bekannte Spy-System, bestehend aus SpyTag und SpyCatcher. Treffen die Interaktionspartner Tag und Catcher aufeinander, rekonstituieren diese spontan unter Ausbildung einer kovalenten Isopeptidbindung, die zwischen den Reaktionspartnern ausgebildet wird und sie irreversibel miteinander verknüpft. Die kovalente Kopplung wird durch die räumliche Nähe der reaktiven Aminosäuren ermöglicht, welche bei der Rekonstitution eine katalytische Triade bilden. Somit erfolgt die Ausbildung der Isopeptidbindung spontan und ohne den Bedarf zusätzlicher Energie, wie beispielsweise ATP. Genutzt werden kann diese Interaktion z.B. durch die individuelle Fusion von Tag und Catcher an gewünschten Proteinen, um diese kovalent miteinander zu verknüpfen oder an Oberflächen zu immobilisieren. Seit der Erfindung des Spy-Systems 2012 von Zakeri et al., wurde das Prinzip der Domänen-Spaltung auf weitere Oberflächenproteine Gram-positiver Bakterien, die eine Faltung des CnaB-Typs aufweisen, übertragen. Das am besten charakterisierte Split-Isopeptidsystem stellt jedoch das Spy-System dar. Aufgrund der hohen Effizienz und Stabilität unter einer Bandbreite von Bedingungen (pH, Temperatur, Detergenzien) und der Funktionalität sowohl in vitro als auch in vivo, wurde es für eine Vielzahl an Anwendungen herangezogen. Dazu zählen unter anderem die Herstellung von Proteinkomplexen, Hydrogelen, thermostabiler Enzyme, bis hin zu Vakzinen. Bei vielen der publizierten Methoden wurde meist ausschließlich das Spy-System verwendet. Die Kombination von zwei oder mehr orthogonalen Interaktionssystemen ermöglicht den spezifischen, gerichteten Aufbau komplexer Strukturen verbunden mit einem breiteren Anwendungsgebiet. Im Rahmen dieser Arbeit wurden verschiedene Split-Isopeptidsysteme, sowohl in vitro als auch im in planta Bereich eingesetzt, um das Anwendungsspektrum besonders durch die Kombination orthogonaler Interaktionssysteme zu erweitern und neue Methoden in planta zu entwickeln. In dieser Arbeit wurden Multiproteinkomplexe mit Hilfe des Spy-, Snoop- und 4oq1-Systems konstruiert, die Bausteine rekombinant in E. coli hergestellt, in vitro zusammengebaut und der entstandene Komplex auf dessen Funktionalität überprüft. Dazu zählt die lineare Hintereinanderschaltung funktionaler Proteine innerhalb eines Komplexes, um diese in enge räumliche Nähe zu zwingen, die geometrische Lage der Proteine zu beeinflussen, sowie die Stöchiometrie der Komponenten zu bestimmen. Angewandt wurde dies in vitro bei der linearen Verknüpfung dreier Fluoreszenzproteine, mCherry, eGFP und mTagBFP, zur Herstellung weiß-emittierender Bio-LEDs (white hybrid light-emitting diodes; WHLED). Der fluoreszente Proteinkomplex wurde in eine stabilisierende Matrix eingebettet und als Beschichtung von UV-LEDs verwendet. Die kovalente Verknüpfung der Fluoreszenzproteine verbesserte den Energietransfer zwischen den Komponenten und ermöglichte die Emission weißen Lichtes. Das Prinzip der Kombination des Spy- und Snoop-Systems zur linearen Hintereinanderschaltung konnte auch bei der Herstellung eines Multienzymkomplexes Anwendung finden. Die Enzyme Invertase, Hexokinase und Phosphoglucoisomerase wurden mit den jeweiligen Interaktionsdomänen fusioniert und bildeten nach der Herstellung der modularen Bausteine und deren kovalenter Interaktion einen funktionalen Multienzymkomplex, der in der Lage war, Saccharose zu Glucose-6-Phosphat abzubauen. Über Anwendungen von Split-Isopeptidsystemen in planta wurde bisher nur wenig berichtet. Es ist lediglich bekannt, dass das Spy- und 4oq1-System erfolgreich und funktional in N. benthamiana hergestellt werden konnten, wohingegen das Snoop-System keine Funktionalität aufgrund fehlender Expression zeigte. Mit Hilfe der Kombination von Organell-spezifischen Lokalisationssequenzen und Komponenten des Spy- und 4oq1-Systems wurde in dieser Arbeit ein kovalenter Baukasten zur spezifischen Markierung pflanzlicher Organellen geschaffen, der sich individuell und einfach an verschiedene Anwendungen im in planta Bereich anpassen lässt. So führte die transiente Co-Expression von Tag-Lokalisationskonstrukten und fluoreszenten Catcher-Konstrukten zur spezifischen Markierung und Visualisierung von Chloroplasten, Mitochondrien und des Zellkerns in N. benthamiana. Die parallele Dekoration eines Organellentyps mit Catcher-Protein und eines anderen Organells mit Tag konnte für die Verknüpfung zweier unterschiedlicher Organellenarten verwendet werden. Dies konnte am Beispiel der Komplexbildung von Chloroplasten und Mitochondrien gezeigt werden (Organellenkleber). Die Kombination des Spy- und 4oq1-Systems konnte außerdem zur simultanen Kopplung von zwei Liganden an ein Organell herangezogen werden. Die Verknüpfung erfolgte entweder linear durch Hintereinanderschaltung oder durch Bindung an ein Gerüstprotein (Scaffold). Als Liganden wurden die Fluoreszenzproteine mCherry und eGFP verwendet, die spezifisch an Chloroplasten dirigiert wurden. Beide Ansätzen könnten als Plattform für die Kopplung funktionaler Proteine an Organellen verwendet werden und als kovalentes Werkzeug für das plant metabolic engineering dienen. Zusätzlich konnte eine Methode zur einfachen und spezifischen Isolation SpyTag-markierter Chloroplasten und Mitochondrien aus Blattrohextrakt mit Hilfe magnetischer SpyCatcher-beschichteter beads entwickelt werden. Die Isolation benötigte wenig Zeit und Ausgangsmaterial und lieferte Organellen mit hoher Reinheit und guter Ausbeute. Wie auch in Lang et al., 2020 beschrieben, ist die Besonderheit das breite Anwendungsspektrum von Split-Isopeptidsystemen in planta und die Möglichkeit, verschiedene Ziele und Strukturen spezifisch zu markieren, wodurch ein kovalenter Baukasten für die Pflanzenmanipulation entstand. Als langfristiges Ziel kann die Kombination verschiedener, orthogonaler Split-Isopeptidsysteme, in Verbindung mit zellspezifischen Promotoren, sowie die parallele Markierung und Isolierung verschiedener Organellen (wie Zellkerne, Chloroplasten und Mitochondrien), Polysomen und intrazellulärer Vesikel (aus dem ER, aus dem Golgi usw.) angesehen werden. Einzeln isolierte Organellen in Verbindung mit anschließender Analytik (Proteomik, Metabolomik) könnten einen tieferen Einblick in den zelltypspezifischen Stoffwechsel, die Signalübertragung und andere Mechanismen ermöglichen (M. Lang et al., 2020). Des Weiteren bildet die Möglichkeit der simultanen stabilen Kopplung mehrerer Liganden an Organellen ein großes Potential für die gezielte Manipulation des Pflanzenstoffwechsels, die genauer untersucht werden sollte. Die in dieser Arbeit entwickelten Methoden bilden somit die Grundlage für die Weiterentwicklung des kovalenten Baukastens basierend auf Split-Isopeptidsystemen im in vitro und in planta Bereich sowie ein nützliches Werkzeug zur Herstellung funktionaler, kovalenter Proteinkomplexe.
  • Article
    Open Access
    Non-projective two-weight codes
    (2024-02-26) Kurz, Sascha
    It has been known since the 1970's that the difference of the non-zero weights of a projective GF(q)-linear two-weight has to be a power of the characteristic of the underlying field. Here we study non-projective two-weight codes and e.g. show the same result under mild extra conditions. For small dimensions we give exhaustive enumerations of the feasible parameters in the binary case.
  • Doctoral thesis
    Open Access
    Sulfur- and Nitrogen-Substituted Hexa-peri-hexabenzocoronenes
    (2024-02-19) Mörsel, Sven; Hirsch, Andreas
    Within this thesis, various mono-, di-, tri-, tetra-, and hexasubstituted sulfur- and nitrogen-containing hexa-peri-hexabenzocoronenes (HBC) were prepared in an optimized microwave-assisted reaction from the corresponding fluorinated precursors. In that sense, this work describes the optimization of the reaction conditions for the reductive defluorination of different polycyclic perfluorocarbon precursors with different thiolates. By employing microwave irradiation, the reaction time was decreased from multiple days to only one hour compared to analogous procedures. The optimized conditions were then employed to produce a library of thioether-HBCs and N-heterocyclic HBCs with varying amounts of substituents in defined substitution patterns, to elucidate the effect of attached moieties on the electronic properties of the HBC core. Additionally, the generation of sulfone and sulfonium compounds from the thioether-HBCs is described as a facile tool to further tailor the electronic properties of the core, and the prepared sulfur- and nitrogen-substituted HBCs were investigated as precursors for heteroatom-doped nanographenes in on-surface or wet chemical syntheses. Finally, this thesis describes a cooperative project for the attempted preparation of thiolated 1,2-bis-(2-methyl-5-R-thien-3 yl)hexafluorocyclopentene (BMTFP) switches to be anchored to gold nanoparticles.
  • Doctoral thesis
    Open Access
    Fundamentals of imaging in interferometric scattering microscopy: theoretical analysis and data science techniques
    (2024) Gholami Mahmoodabadi, Reza; Sandoghdar, Vahid
    Studying cellular and subcellular processes in biology provides vital insights into our fundamental understanding of life. Labeling biological features with fluorescent labels and imaging them using a fluorescence microscope has been the method of choice in these investigations. However, this approach suffers from the limited photophysical properties of conventional fluorophores. In addition, labeling specimens with fluorophores is generally required, and this might affect the physiological state of the biological system under study. Therefore, noninvasive microscopy modalities that do not require labels are highly desirable. In the pursuit of achieving label-free imaging, a contrast mechanism native to specimens must be utilized. Scattering is one of the most attractive possibilities for attaining contrast without using labels. However, the scattering intensity of objects at nanoscales is very weak, resulting in significant technical challenges for their detection. As a means to circumvent these issues, the scattered electric field can be interfered with a reference beam to produce a pronounced interferometric contrast overlaid on a strong reference intensity. The readily accessible signal using this methodology relaxes several technological restrictions for detection. This technique, known as interferometric scattering (iSCAT) microscopy, allows the detection and tracking of matter at the nanoscale, even down to single, unlabeled proteins. Since the invention of iSCAT microscopy approximately twenty years ago, this technique has emerged as a powerful tool for label-free studies. However, the quantitative details for its image formation had not been examined prior to this doctoral dissertation. This restricted 3D tracking capabilities of iSCAT to an axial range of barely a few tens of nanometers. The current work presents a comprehensive computational study exploring and integrating various fundamental aspects of image formation in iSCAT and approaches from data science to tackle the sample’s complex scattering signal, as well as advance the detection sensitivity of iSCAT. The scope of this doctoral thesis ranges from the mathematical modeling and simulation of the interferometric point spread function (iPSF) in iSCAT to addressing challenges in iSCAT applications, i.e., algorithmic solutions for applications such as tracking proteins in cellular membranes and investigating membrane diffusion and fusion in model membranes. The findings and methodologies presented here contribute to advancing label-free detection and analysis of nanoscale matter in diverse research domains. The first aspect of this study addresses the formulation of the iPSF in iSCAT, for which a vectorial diffraction model is developed. Experimental measurements and rigorous simulations are employed to examine theiPSF in various imaging scenarios. In both the reflection imaging mode a transmission imaging mode of iSCAT, the effect of interfaces on the scattered light from a nanoparticle is examined in detail. Analyzing the information encoded in iPSFs demonstrates nanometric three-dimensional (3D) localization of nanoparticles over a two orders of magnitude longer axial range than was previously possible. These findings have direct positive implications for single particle tracking (SPT) against a more complex scattering background. The cellular background in iSCAT images is an example of one such environment. Another aspect of this work is to employ computational techniques to enable the study of protein motion in cell membranes and to investigate fundamental processes in lipid bilayers. Through the use of novel image processing techniques, we address the challenge of a speckle-like background and perform nanometer precision tracking of transmembrane epidermal growth factor receptors in all three dimensions. We showcase recordings of nanoscale motion and confinement in mechanisms such as diffusion in the plasma membrane, transport on filopodia, and rotational motion during endocytosis. In one study, we focus on pore-spanning membranes (PSMs). Utilizing iSCAT, we compare the diffusion of lipids in supported and suspended regions of PSMs. This makes it possible to directly examine the influence of the substrate on the diffusion of membrane components in supported regions. This adds to previous fluorescence studies that only looked at suspended regions. In another research project, we delve into the study of membrane fusion. By tracking the movement of single vesicles across giant unilamellar vesicles (GUVs), we can look at the intermediate states before membrane fusion. Lastly, we examine the complications associated with interferometric detection, specifically shot noise, speckle-like background, and technological difficulties in experimental setups, which adversely affect the highest achievable iSCAT sensitivity. We draw attention to key technical and conceptual requirements and deliver a set of modeling methods, together with algorithms from the fields of computer vision and machine learning, for iSCAT data processing. This paves the way for a completely novel set of applications in biophysical and medical research, where quantitative information extraction with unprecedented spatiotemporal details is in high demand.
  • Doctoral thesis
    Open Access
    The Continuous Stochastic Gradient Method
    (2024) Uihlein, Andrian; Stingl, Michael
    We consider a class of optimization problems in which the objective function requires the computation of an expected value of a hard to evaluate property. Such structures arise in probabilistic settings, e.g., expected values in optimization under uncertainty, as well as fully deterministic contexts, e.g., optimizing a design for a (possibly infinite) set of scenarios. Therefore, both stochastic and deterministic optimization schemes have been considered to solve such problems. The latter typically rely on a fixed quadrature rule to approximate all appearing integrals numerically. As a result, the associated computational cost may render such approaches numerically infeasible in complex settings. Moreover, discretization can lead to artificial local solutions emerging in the problem, if the corresponding quadrature rule is not chosen carefully. Contrary, stochastic sample based methods do not suffer from this problem and provide a computationally cheap alternative. However, these schemes are typically limited to objective functions of very simple structure, i.e., minimizing an expected value. Settings in which the objective consists of several integrals, possibly coupled by nonlinear functions, are beyond the scope of such techniques. In this work, we consider continuous stochastic optimization approaches, specifically, a continuous stochastic gradient (CSG) scheme. To keep the numerical effort per iteration low, these methods also rely on random gradient samples instead of full integrals. However, unlike traditional stochastic sample based strategies, the gradient information is stored over the course of iterations. Thus, by calculating suitable design dependent integration weights, a convex combination of old gradient samples is used to build an approximation to the full gradient. We show that the approximation error almost surely converges to zero during the optimization process. As a result, we prove that CSG admits the same convergence results as a full gradient scheme, i.e., convergence for constant step sizes and line search techniques based on the continuous stochastic approximations. Moreover, it means that CSG is not restricted to the minimization of simple expected values and can instead deal with arbitrary combinations of integrals and Lipschitz continuous functions in the objective. In this sense, CSG can be considered as a hybrid of stochastic gradient methods and a classical gradient descent scheme. Several methods to obtain the aforementioned integration weights in practice are presented and discussed. Furthermore, we provide a full theoretical convergence analysis of CSG and perform numerical experiments, covering all important aspects of continuous stochastic optimization methods. For several large scale applications from the field of topology optimization, the proposed schemes are compared to different optimization techniques known from literature.
  • Doctoral thesis
    Open Access
    Kinetic models for label switching and stochastic gradient descent
    (2024) Rossi, Alex; Burger, Martin
    In this thesis we develop a new model, which consists in a linear integro-differential equation. This is introduced considering the structure of various models of the kinetic theory, with different applications, like, for instance, in the analysis of populations, material science and the stochastic gradient descent (SGD), the most famous algorithm for the stochastic optimisation. After the analysis of these motivations and the introduction of essential theoretical concepts, we present the main model and its principal properties: existence and uniqueness of measure-valued solutions, existence of stationary states and some features of such states. Moreover, we deal with the limit equation, in case the interaction parameter tends to infinity (grazing collisions limit). If we take into account the SGD-interpretation and the interaction parameter depends on time, we consider the limit of the solution to the equation, as time tends to infinity. Then, we add some diffusion to the main equation in order to create a stochastic gradient Langevin dynamics-type model. We present the problem of propagation of chaos, which leads us to a nonlinear equation, the Vlasov equation, of which we analyse the well-posedness. To conclude, we expose some further unsolved problems and issues.
  • Doctoral thesis
    Open Access
    Harnessing mucosal immunity against respiratory viruses using adenoviral vectors encoding cytokines/chemokines as mucosal adjuvants
    (2024) Vieira Antão, Ana; Tenbusch, Matthias
    Influenza (IAV) and respiratory syncytial viruses (RSV) are two of the main causes of severe respiratory tract infections; however, developing effective vaccines against these pathogens remains a challenge. Despite the recent approval of two vaccines against RSV, the annual vaccination against IAV has demonstrated that inducing systemic immunity is not sufficient to prevent reinfections. The recent discovery of tissue resident memory (TRM) T cells in mice and human lungs has highlighted the importance of mucosal immunity against these pathogens. In IAV and RSV infected individuals, the induction of lung TRM cells and IgA antibodies was shown to correlate with an improved prognosis and rapid viral clearance in secondary infections. Therefore, vaccines able to improve mucosal humoral and T cell responses against these viruses are considered a promising approach to reduce disease burden. However, as mucosal tissues are the main entry site for different pathogens, these compartments have evolved to rapidly elicit immune responses to counteract replication and the subsequent inflammation. For this reason, antigen delivery by mucosal immunizations can be impaired, resulting in a weak vaccine immunogenicity. Nevertheless, these problems may be overcome with good vaccine platforms as well as the usage of adjuvants. Indeed, our work has elucidated the effect of an adenoviral vector (Ad) expressing IL1β as mucosal adjuvant together with Ad-based influenza and RSV vaccines. The intranasal administration of IL1β showed to improve vaccine immunogenicity and efficacy against these respiratory viruses. However, the pleiotropic effect of this cytokine may induce undesired immune responses. In line with these findings, we aimed to develop and analyse the effect of different chemokines/cytokines encoded in adenoviral vectors as alternative mucosal adjuvants. For this purpose, the innate molecules CCL17, CCL20, CCL21, CXCL9, CXCL13, TGFβ, IL28A and IL28B were chosen based on the observations made in the in vivo experiments with Ad-IL1β as well as on their function described during influenza and RSV infections. Therefore, these chemokines/cytokines are sought to be involved in the generation of the host mucosal immunity against these pathogens. In the first part of this study, adenoviral vectors expressing the different genes were produced and characterized in vitro. Afterwards, their impact on vaccine-specific immune responses were investigated in vivo. Our results showed that Ad-CCL17 and Ad-TGFβ were the two most promising candidates as mucosal adjuvants. Their expression in the lungs of mice resulted in the enhancement of antigen-specific TRM cells, whereas the other chemokine/cytokine did not improve vaccine-induced humoral or T cell responses. The second part of this project aimed to establish a mucosal combinatory vaccine against IAV and RSV, as an advantageous strategy to control and prevent both infections. In vivo studies elucidated the impact of a mucosal adenoviral vector-based vaccine encoding the influenza HA and NP protein as well as RSV-F, in combination with the adjuvants Ad-CCL17 and Ad-TGFβ. The combinatory vaccine induced robust mucosal immune responses and protected mice against heterologous IAV strains or RSV viruses. The additional use of the adjuvants strengthened the formation of antigen-specific TRM cells, but had less impact on the antibody responses against IAV and RSV. In addition, the mucosal administration of these adjuvants did not significantly improve vaccine efficacy. In summary, an effective mucosal vaccine was established against two of the most clinically relevant respiratory viruses. In addition, the co-expression of TGFβ revealed a supportive role in the formation of antiviral lung resident memory T cells, which is an interesting observation for additional basic or mechanistic studies.
  • Doctoral thesis
    Open Access
    Wissenschaftsgeschichte der quantitativ-theoretischen Wende. Erzählungen, Netzwerke und lokale Arrangements in der deutschsprachigen Geographie zwischen 1950 und 1979.
    (2024) Paulus, Katharina; Glasze, Georg
    Diese Promotion untersucht die sogenannte “quantitative Revolution” in der deutsch-sprachigen Geographie. Unter diesem Begriff wird eine Phase der akademischen Geographie gefasst, in der aus einer Kritik an der länderkundlichen Geographie die Forderung nach der Übernahme moderner Wissenschaftstheorien, quantitativer Ver-fahren und angewandter Forschungsfragen formuliert wurde. In der Regel wird diese Phase in Bezug auf die anglophone Geographie auf die 1950er und 1960er Jahre und in Bezug auf die deutschsprachige Geographie auf die späten 1960er Jahre da-tiert. Insbesondere in der deutschsprachigen Geographie wird die quantitative Revo-lution an einigen wenigen Ereignissen festgemacht. Vor diesem Hintergrund war der zentrale Ausgangspunkt des vorliegenden Forschungsprojekts die Annahme, dass diese Erzählung und die disziplinäre Erinnerung einer genaueren wissenschaftsge-schichtlichen Untersuchung nicht standhalten, sondern vielmehr als Mythen zur Selbstvergewisserung der Disziplin fungieren. Ziel der Arbeit war es daher, eine hete-rogenere und situierte Geschichte der Geographie jener Umbruchsphase zu schrei-ben, die ihren Fokus weniger stark auf einzelne Autor*innen und einige wenige ideengeschichtliche Momente legt. Im Gedächtnis der Disziplin, wie es sich in Überblickswerken, Einführungen, theoretischen Grundlegungsversuchen und auch in informellen Erzählungen aus-drückt, wird die Erinnerung an die quantitativ-theoretische Wende auf einige wenige Ereignisse fokussiert. Die Aufmerksamkeit gilt dabei in besonderer Weise dem deut-schen Geographentag in Kiel im Jahr 1969 und Dietrich Bartels‘ Habilitation „Zur wis-senschaftstheoretischen Grundlegung einer Geographie des Menschen“ (Bartels, 1968). Ausgehend von Ansätzen in der Wissenschaftsgeschichte (Shapin, 2010; Rheinberger, 2010; Daston, 2009) und Projekten zur anglophonen Geographie (Bar-nes, 2004; Barnes & Sheppard, 2019) wird die Geschichte der quantitativ-theoretischen Wende in der westdeutschen Geographie im Rahmen der Dissertation jedoch als komplexer, heterogener und widersprüchlicher Prozess verstanden. Damit soll nicht allein eine ideengeschichtliche Perspektive formuliert werden, sondern Wis-senschaften als soziale, materielle und lokalisierte Praxis begriffen werden. Die vorliegende Dissertation hinterfragt dabei insbesondere die Darstellung der quantitativ-theoretischen Wende als naiven Positivismus und erste Artikulation einer wissenschaftlichen Geographie. Das Forschungsvorhaben sieht die zahlreichen Versuche der Modernisierung und Anpassung an veränderte gesellschaftliche, wis-senschaftspolitische und wissenschaftstheoretische Anforderungen als Indikatoren dafür, dass es sinnvoll ist, nach vielfältigeren Momenten des Aufbrechens zu su-chen. Deutlich wird dieser Aufbruch unter anderem in den Untersuchungsgegenstän-den der Disziplin, so lässt sich für den Zeitraum zwischen 1950 und 1970 eine ver-änderte Wahrnehmung der Natur beobachten. Das in der Länderkunde zentrale Ver-ständnis der Natur als harmonisches Zusammenwirken wird in eine zunehmend von Kybernetik und systemtheoretischen Überlegungen geprägte Disziplin überführt und angepasst. Denn auch nach 1969 unterschied sich die geographische Forschung nicht radikal von dem, was davor geforscht wurde. In Bezug auf die quantitativ-theoretische Wende wird deutlich, dass die Neuerung als langsamer und wider-sprüchlicher Prozess zu verstehen ist. Außerdem deutet sich an, dass sich die quan-titativ-theoretische Wende und die Entwicklungen der Geographie in den 1970er Jah-ren an den verschiedenen universitären Standorten unterschiedlich vollzogen. Hierfür wählte ich einen komplementären methodischen Zugriff: Einerseits wurde sich dem Forschungsgegenstand mit Hilfe einer quantitativ ausgerichteten historischen Netz-werkanalyse genähert und andererseits mit einer qualitativ vorgehenden Perspektive, die sich auf die ungleichen Geographien und Zeitlichkeiten der untersuchten Gegen-stände fokussiert. In den Fallbeispielen werden drei exemplarische Dimensionen des Gegen-stands dargestellt. Einerseits wird eine lokale und ortsbezogene Perspektive deutlich gemacht, um aufzuzeigen, dass jenseits der dominanten eine Reihe lokaler quantita-tiver Revolutionen stattgefunden haben, die eigenständige Verständnisse produziert haben. Diese lokalen Realisierungen eines quantitativ-theoretischen Denkens und quantitativ-theoretischer Epistemologien in der Geographie haben eine divergentere vielfältigere Praxis hervorgebracht, als dies in den konzeptionellen Begründungstex-ten dargestellt wird. Andererseits wurde die veränderte Epistemologie des Gegen-standes Natur in den Blick genommen werden, da er bis heute zentral für das Selbstverständnis der Disziplin ist und trotzdem im Kontext der quantitativ-theoretischen Wende ein relativ wenig reflektierter Begriff ist. Darüber hinaus wurden die Veränderungen der geographischen Forschung in Form einer historischen Netz-werkanalyse untersucht, welche systematisch und auf die Breite gerichtet das Feld der deutschsprachigen Geographie der 1950er bis 1970er Jahre analysiert. Die his-torische Netzwerkanalyse bietet, als exploratives Verfahren zur Beschreibung von Relationen und Interaktionen in Netzwerken aus Personen, Institutionen und Wissen die Möglichkeit, ex post formulierte und produzierte Erzählungen über zentrale Ak-teure und Ereignisse zu hinterfragen. Die Wahl des methodischen Zugriffs über eine historische Netzwerkanalyse bringt mehrere Vorteile mit sich: So kann einerseits die Vielfältigkeit dieser Aufbrüche graphisch dargestellt und nachvollzogen werden und andererseits ein erster Einblick in die Praxis des wissenschaftlichen Arbeitens der Zeit genommen werden. Die histo-rische Netzwerkanalyse hat eine bibliographische Analyse der zentralen Zeitschriften der deutschsprachigen Geographie zwischen 1950 und 1979 unternommen. Ziel die-ses quantitativen Zugangs war es unter anderem, den kollektiven Charakter von Wissenschaft stärker zu fokussieren und zudem zu untersuchen, inwieweit die Ver-änderungen sich tatsächlich in der geographischen Forschungsarbeit niederschlu-gen. Die qualitativen Tiefenbohrungen haben hingegen darauf gezielt, den zeitlichen, räumlichen und inhaltlichen Blick zu weiten. Während der bisherige Blick in der Regel auf einigen wenigen wissenschaftstheoretischen Grundlagentexten und Diskussionen liegt, wurde anhand der Untersuchung des in der Geographie verwendeten Naturbe-griffs aufgezeigt, wie sich gerade auch das empirische Arbeiten im Vorgang zu die-sen wissenschaftstheoretischen Entwürfen verändert hat. Insgesamt zeigt die Arbeit, dass die hegemoniale Erzählung eines klaren Bruchs und die Vorstellung einer kom-pletten Neuartigkeit der Ansätze nicht haltbar ist.
  • Doctoral thesis
    Open Access
    Detecting fundamental structures of the interstellar medium with machine learning algorithms
    (2024-02-13) Bernreuther, Dominic; Sasaki, Manami
    The detection and classification of fundamental structures of the interstellar medium (ISM) is tedious but tremendously important work for astronomers. Structures like supernova remnants (SNRs) shape and drive the ISM and have a big impact on evolving galaxies. Those fundamental structures are of bubble-like shape. Finding these bubble-like structures is therefore crucial in the study of galaxy evolution and of the universe in general. Currently, the detection of those bubble-like structures is done mainly by eye, by manually examining astronomical images. In the era of machine learning algorithms, however, this process can be automized. In this thesis, we present an automated detection and classification pipeline for bubble-like structures based on multi-wavelengths images. The pipeline is called BScan and makes use of two neural networks: a convolutional neural network (CNN) is used for detecting the bubble-like structures in the images and a fully connected neural network (FCNN) is used to classify the detected structures. Both networks are used and combined by classical algorithms. The CNN is part of the nested boxes detection algorithm (NBDA), which scans the available images, and the FCNN then uses the output of the NBDA. The configuration is done via a configuration file where all hyperparameters can be set. In areas of isolated structures or less crowded regions, BScan performs relatively well by localizing already cataloged sources with an average distance of 36arcsec to the source center given by the catalog. The radius of the detected strucutre, however, is systematically overestimated. In highly crowded regions BScan may detect multiple but separated structures as one single structure. BScan also shows some inconsistencies in detecting similar bubble-like structures in images of similar or even the same resolution. The classification has a small estimated error of around 0.1% but is heavily dependent on the NBDA. There are different approaches using different techniques to address the idea of automated structure detection, but yet there has been no attempt to combine the detection and classification aspect in one process. With some further developments and improvements, the idea of BScan could set new standards in detecting fundamental structures of the ISM and could be a great tool for astronomy in the future.
  • Doctoral thesis
    Open Access
    From the smoking seafloor to metalliferous mountains: Tracing ore formation and metal fractionation by sulfide microanalysis
    (2024) Falkenberg, Jan J.; Haase, Karsten
    Magmatic-hydrothermal ore deposits occur at varying depths in the Earth´s crust, including deeper-seated porphyry- and shallow epithermal-style mineralization, as well as (active) hydrothermal systems on the seafloor. These ore-forming systems represent important metal anomalies in the Earth’s crust and provide various metallic raw materials, which are crucial for today´s and future´s economy. However, the magmatic and hydrothermal metal fractionation processes, which occur at different crustal levels and essentially control the trace metal content (e.g., Te, Au, Ag, Se, Bi, As and Sb) of ore deposits, are still poorly constrained on the crustal scale. Porphyry deposits form in subduction zone environments, where water- and volatile-rich magmas exsolve S- and metal-rich magmatic-hydrothermal fluids. Fluid-rock interaction, cooling, and phase separation of the fluids trigger the formation of disseminated and vein-style sulfide mineralization at high temperatures (>600 – 350°C) within deep (~1 – 6 km) porphyry stocks. Epithermal mineralization form in shallow (<1.5 km) areas above porphyries at low-temperature (<350°C) conditions from similar magmatic-hydrothermal fluids with possible varying meteoric water/seawater contribution. Active submarine hydrothermal vent systems in subduction zone settings are potential submarine analogues to epithermal systems. These vents form from seawater circulation through the oceanic crust, where fluid-rock interaction transforms seawater into a hot (<450°C), reducing, sulfur- and metal-rich hydrothermal fluid which precipitate “black smoker” sulfide-sulfate chimneys during quenching with cold seawater. This thesis investigates ore-forming processes and the trace metal fractionation during magma degassing as well as during fluid phase separation, which result in precious (e.g., Gold) and critical element (e.g., Rhenium) enrichment in deep to shallow ore-forming systems. Furthermore, it focuses on how these processes are recorded by sulfide trace element composition and their stable S- and radiogenic Pb isotope chemistry. Chapter 3 examines the active submarine hydrothermal systems at Niuatahi volcano in the north Tonga rear-arc systems and shows that the involvement of magmatic fluids and phase separation result in a distinct metal zonation (e.g., Au, Te, Se, Pb) within the caldera. Chapter 4 investigates the influence of magma degassing on the pyrite trace element variance from subduction-related hydrothermal systems. Here, it is revealed for the first time, that Te/As and Te/Sb ratios correlate systematically with the δ34S isotope composition and that these trace element ratios record the contribution of magmatic volatiles between different vent sites. In Chapter 5, it is shown that Co/As, As/Sb, Se/Te, and Se/Ge ratios in pyrite can be used to resolve variations in the physicochemical fluid parameters and ore-forming history in mineralized porphyry-epithermal veins during the porphyry-epithermal transition at the Maronia porphyry-epithermal system, NE Greece. Empirical observations further indicate that the Re enrichment at Maronia is linked to a distinct hydrothermal stage during which oxidized fluids were reduced and cooled to <400°C, inducing extreme precipitation of Re-rich molybdenite. In summary, this thesis provides significant and novel approaches for fingerprinting magmatic-hydrothermal processes through different crustal levels by microanalytical sulfide chemistry and reveals the control of these processes on the metal endowment in porphyry-epithermal and active submarine hydrothermal systems in subduction zone environments. Particularly, it highlights the usefulness of coupled microanalytical methods, and the use of trace element ratios compared to absolute trace element concentrations, as these ratios are mainly controlled by the underlying magmatic-hydrothermal fractionation processes such as phase separation and magmatic degassing.
  • Doctoral thesis
    Open Access
    A comprehensive analysis of annual and intra-annual growth data for reconstructing and modelling past and future climate impacts on tree growth: Case studies from the Himalayan region
    (2024-02-05) Aryal, Sugam; Bräuning, Achim
    Climate change impacts and corresponding tree growth responses are crucial for understanding ecological dynamics. Climate change significantly affects natural systems and human societies worldwide, and almost every climate model indicates a further and even more pronounced warming trend in the future. The impact of climate change in the Central Himalayan region, including the Tibetan plateau, is more pronounced compared to other regions at similar latitudes, resulting in higher rates of temperature increase compared to the global average. However, the observed warming rates show significant spatial variability, with a more pronounced warming trend at higher altitudes. The impact of climate change can already be monitored in significant changes (response) of vegetation, like in species distribution, population structures, ecotone shifts, vegetation composition, phenology, and prolonged growing seasons on a regional and global scale. The South Asian mountainous area has, however, a limited number of available meteorological stations, which hampers our ability to quantify long-term trends and recent impacts of climate change on the local level. This thesis presents four interconnected studies that contribute to the topics of climate reconstruction and tree growth analysis, both addressing the effects of climate change, defoliators, dendrometers, and phenological changes. The study entitled “Spring season in western Nepal Himalaya is not yet warming: A 400-year temperature reconstruction based on tree-ring widths of Himalayan Hemlock (Tsuga dumosa)” focuses on the central Hindu Kush Himalaya, where a 619-year tree-ring-width chronology was developed to reconstruct spring temperature conditions. The resulting reconstruction revealed distinct periods of cooler and warmer conditions during the past centuries without a consistent long-term trend. A spectral analysis identified supraregional effective climate modes influencing the study region, such as the Atlantic Multi-decadal Oscillation (AMO) and the El Niño-Southern Oscillation (ENSO). In some tree species, the effect of insect outbreak can result in non-climate-related narrow tree-ring formation, leading to a biased climate reconstruction. Such bias due to the periodic insect outbreaks needs to be corrected to get a realistic past climate reconstruction. The second study, “Insect infestations have an impact on the quality of climate reconstructions using Larix ring-width chronologies from the Tibetan plateau”, deciphers the impact of defoliators (insect outbreaks) on tree-ring-width climate reconstructions and suggests an approach to mitigate such impact. A 308-year-long statistically improved monsoon season (June-September) temperature reconstruction was achieved by nullifying the effects of larch budmoth (LBM) outbreaks in Southwest China. The final reconstruction exhibited fluctuating trends, cool and warm periods, and revealed a recent alarming warming trend of 0.4o C per decade. The reconstructed monsoon temperature was validated by the strong coherency with other regional and global temperature reconstructions. Similar to the pre-monsoon temperature reconstruction in Chapter 3, there was a strong direct and indirect influence of supraregional climate modes, such as AMO and Pacific Decadal Oscillation (PDO), in the reconstructed monsoon temperature. Tree growth is a dynamic process highly sensitive to short-term weather changes. The third study introduces a new package called 'dendRoAnalyst' programmed in the statistical environment R as an essential tool for studying intra-annual tree responses using dendrometers. The package provides algorithms for processing and analyzing dendrometer data using all the contemporary approaches, such as daily, stem-cycle and zero-growth approaches. The dendRoAnalyst package further assists researchers in cleaning dendrometer metadata, identifying data artefacts, handling data gaps, and calculating various statistics based on different approaches. The fourth study proposes a novel approach, INTRAGRO, combining dendrometer data and climate models to predict future tree growth at intra-annual resolution. INTRAGRO employs supervised and unsupervised machine learning algorithms and successfully predicts enhanced tree growth despite expected seasonal precipitation changes. This approach offers a global-scale understanding of cambial activity changes, which is crucial for assessing species' growth performance and adaptation under different climate scenarios. This PhD thesis highlights the interaction between climate change, tree growth responses, and ecological dynamics in the Central and Eastern Hindu Kush Himalayan region. This synthesis study sheds light on the complexities of climate reconstruction, considering factors such as insect outbreaks that can bias the modelled temperature. Furthermore, innovative approaches, including tree-ring analysis, dendrometer data processing, and the INTRAGRO model, offer an in-depth understanding of the multidimensional impacts of climate change on tree growth. As our understanding of such cascading impacts deepens, the significance of these findings resonates for broader efforts to address the challenges posed by a changing climate on global and regional ecosystems.
  • Doctoral thesis
    Open Access
    The role of plasmablasts in the development of anti-DNA autoantibodies
    (2024-02-06) Gleußner, Nina; Winkler, Thomas
    The hallmark of systemic lupus erythematosus (SLE) is the production of autoantibodies against double-stranded DNA. This B cell-mediated autoimmune disease is associated with multiple genetic risk factors, e.g., both Dnase1l3 (Dnase1like3) and Fcgr2b (Fc gamma receptor 2b) null-allele mutation have been linked to the emergence of autoreactive B cells and anti-dsDNA antibodies in SLE patients and genetic mouse models. We have recently shown that Dnase1l3 and Fcgr2b double-knockout mice show very early and massive anti-DNA autoantibody production. Whereas Dnase1l3 is exclusively expressed in the myeloid compartment, the inhibitory Fcgr2b is highly expressed also in B cells, and it has been suggested that the negative signaling functions of the Fcgr2b plays an important role in B cell tolerance particularly within the germinal center. To analyze the contribution of deficiency of the Fcgr2b in B cells in several aspects of the SLE-like disease we crossed Dnase1l3- deficient mice with Mb1-cre mice carrying a floxed Fcgr2b allele. Dnase1l3-deficient mice with a complete Fcgr2b deficiency were compared with mice with a B cell specific Fcgr2b deficiency. B cell populations were analyzed by flow cytometry and autoantibody production was measured by ELISA. Double-deficient Dnase1l3/Fcgr2b mice show strongly elevated anti-DNA antibody titers, elevated germinal center (GC) B cell numbers, strong activation of T follicular helper (TFH) cells and considerable expansions of plasmablasts in the spleen already at 3 months of age. At the same time point of analysis B cell specific Fcgr2b deficiency resulted in no difference as compared to Dnase1l3 single knockout mice, i.e., only slightly elevated anti-dsDNA titers with no significant expansion of GC B cells, plasmablasts or TFH cells. Elevated anti-dsDNA antibody titers, elevated GC B cell numbers, strong activation of T cells and expansions of plasmablasts progressed even further at six and nine months of age in mice with complete Fcgr2b deficiency. B cell specific Fcgr2b deficiency resulted in B cell and TFH phenotypes at six months of age. Our data clearly establish a protective role of the Fcgr2b function for early exaggerated anti-DNA production as well as deregulation of GCs, plasmablast expansion and T cell hyperactivation outside of the B cell compartment since an isolated B cell specific Fcgr2b deficiency does not lead to strong symptoms until much later. Afterwards we generated Fcgr2b specific KO mice for other cell groups i.e., dendritic cells (DCs) and follicular dendritic cells (FDCs). Overall, we could show that while the lack of Fcgr2b expression on B cells by itself does not determine the early onset of SLE-like disease in mice, B cells could be very well the main initiators of disease since the lack of Fcgr2b on other cell groups have led to even less manifestations of SLE in mice. Furthermore, we were also able to track and analyze autoreactive B cells in single RNA sequencing data.
  • Doctoral thesis
    Open Access
    Spatial coherence and Coulomb-induced correlations of ultrashort electron pulses
    (2024-02-06) Meier, Stefan; Hommelhoff, Peter
    Ultrashort electron pulses, triggered by few-femtosecond laser pulses, offer the possibility to study electronic processes on the nanometer and femtosecond scale. This is of great interest for fundamental physical processes, such as charge transfer in various systems with its associated time sales or ultrafast light-matter interaction, as well as for other scientific fields, such as the study of chemical reactions or biological processes on the smallest time and length scales. For the stated applications, precise knowledge of the spatial and temporal properties of such pulsed electron beams is required. In the present work, we investigate the spatial coherence of ultrashort electron pulses emitted by multiphoton photoemission from a tungsten needle tip. The spatial coherence is characterized by the so-called effective source size, which is smaller the higher the spatial coherence of the source. Based on the evaluation of spatial interference patterns of ultrashort electron pulses, we obtain an upper limit for their effective source size of $r_\text{eff,MPP}\le(0.65\pm0.06)$ nm. A measurement with field-emitted continuous electron beams results in the same limit. Therefore, within the limits studied, we can state that the spatial coherence does not depend on the nature of the electron emission process. Since the electron emission, in our case, is spatiotemporally confined to the few-femtosecond and nanometer scale, strong interactions between electrons are expected as soon as more than one electron is emitted per pulse. By successively increasing the laser intensity, we study the spatial coherence as a function of the average emitted number of electrons per pulse. We observe an exponential decrease of the spatial coherence as a function of the average emitted charge due to Coulomb interactions, corresponding to an enlarged effective source size. This effect already occurs below an average of one electron emitted per pulse due to the extreme localization of the emission and the Poissonian distribution of the number of emitted electrons. The consequence is a change of the key parameters of an electron beam, namely an increase in the emittance of the electron beam as well as a decrease in the beam brightness. Using a new experimental setup based on a delay-line detector, we are able to post-select events in which exactly two electrons are detected. When we magnify the electron beam with a quadrupole lens, we observe clear energy anticorrelations of two-electron events due to Coulomb interactions. In the histogram of the energy difference, a gap with a peak-to-peak extent of 3.2 eV is observed, which has a repulsion visibility of $\mathcal{V}_r=56$ %. Sub-Poissonian electron beams with a value of the second order correlation function as low as $g^{(2)}=0.34$ can thus be generated by subsequent energy filtering. Simulations show that when we apply an energy filter to a beam with an average of 5 electrons per pulse, we can obtain a small Fano factor of $\mathcal{F}\approx 0.5$, which is of great importance for quantum imaging with electrons. By carrying out a two-pulse measurement with adjustable time delay, we are able to determine the temporal range of the anticorrelation signal and get a correlation decay time of $\sigma_\tau=81.7$ fs. This value allows to calculate a theoretical current limit of such a source before Coulomb interactions affect the beam quality or energy width. We determine a limit of $I_\text{max}=e\cdot f_\text{rep,max}=e\cdot\frac{1}{\sigma_\tau}=1.96$ µA. This value could be reached with a deterministic single-electron source that is exactly one electron per pulse, driven at the repetition rate of $f_\text{rep,max}$. Furthermore, we investigate the dependence of the anticorrelation gap on the laser intensity in a one-pulse measurement. We show that the anticorrelation gap is reduced as soon as strong-field effects contribute to the photo-electron spectrum. In the last part of the work, we investigate the possibility of further improving the multi-hit capability of the delay-line detector by post-processing its analog signals. We apply a least-square-fit based method as well as a new approach based on machine learning. The application of the machine-learning based evaluation allows us to more than halve the dead radius of the detector for the case of two-electron events compared to the previous electronic evaluation, which promises great improvements for future correlation measurements.
  • Doctoral thesis
    Open Access
    Coupling Organic Molecules to Integrated Photonic Circuits
    (2024) Rattenbacher, Dominik; Sandoghdar, Vahid
    Connecting individual solid-state quantum emitters via integrated photonic circuits presents a promising avenue for the controlled assembly of quantum optical networks. Such networks are a pivotal prerequisite for quantum technologies and fundamental studies on collective quantum many-body states. Despite substantial advances in recent years, the realization of larger networks has been impeded by several factors, including emitter inhomogeneities, decoherence, the absence of single emitter positioning techniques, and material incompatibilities. To overcome these challenges, there is a growing trend towards hybrid solutions which leverage the benefits of different emitter and photonics platforms. In this thesis, we combine the excellent spectral properties, high achievable concentrations, and DC Stark frequency control of the polycyclic aromatic hydrocarbon quantum emitter dibenzoterrylene (DBT) with four different integrated photonic circuits. We begin by presenting a one-dimensional (1D) architecture that uses waveguides with subwavelength cross sections (nanoguides) covered by a dye-doped organic crystal. Single molecules near a nanoguide couple evanescently to its guided mode, with their coupling efficiencies determined by the transverse mode profile and refractive index of the nanoguide. We demonstrate nanophotonic circuits made from titanium dioxide (TiO₂) and the high-refractive-index material gallium phosphide (GaP), achieving coupling efficiencies of 5 % and 7 %, respectively. Using the DC Stark effect to tune the molecules' frequencies, we show an integrated two-emitter quantum network using TiO₂. With GaP nanoguides, we observe spectral instabilities due to photo-induced charge diffusion. Using single molecules as nanometer-sized probes, we investigate the temporal, spatial, and statistical properties of this diffusion, showcasing the potential of single molecules as electric field sensors. These findings deepen our understanding of GaP and advance its use in integrated photonics. Due to the geometrical and material constraints of the photonics platform, as well as complex transition pathways and dephasing of the emitters, deterministic coupling in a 1D geometry is elusive. To address this issue, we explore nanophotonic resonators which can enhance the coupling efficiency via the Purcell effect. Initially, we investigate TiO₂ microring resonators covered by an organic crystal. However, the birefringence and grain boundaries of the crystal negatively impact the finesse, and thus the attainable Purcell effect. By partially shielding the resonator and optimizing its design, we achieve a finesse of 18 and 23 % Purcell-enhanced coupling efficiency. To overcome the limitations imposed on the microresonators by the crystalline host, we employ polyethylene (PE) as a host material. PE is a widely used plastic that can coat photonic circuits with thin and transparent films, enabling exceptional performance of integrated microresonators. Furthermore, individual polymer chains exhibit short-range order and form a semicrystalline structure, resulting in surprisingly narrow optical transitions of DBT at 2 K. We couple DBT molecules in PE to a 3 µm-radius disc resonator with a finesse of 230, achieving a cooperativity of three and a coherent coupling efficiency of 75 %. This achievement sets a new milestone for integrated quantum optics with molecules. Using local DC Stark shifts, we tune two well-coupled molecules into resonance. Due to the specific distance between the molecules along the resonator perimeter, their backward scattering is suppressed similarly to chiral emitters, resulting in an enhanced extinction dip of more than 75 %. This value surpasses the maximum possible extinction of a single dipole coupled to a circular resonator. We conclude our work by proposing the fabrication of multiple independent microelectrodes to scale up the number of resonant emitters and further refinements to the microresonators in order to increase the cooperativity of single emitters. These advances will pave the way for the practical implementation of integrated photonic networks consisting of multiple well-coupled organic molecules. Such networks can be utilized to investigate quantum many-body states, thereby opening up new frontiers in quantum physics and technology.
  • Doctoral thesis
    Open Access
    Tracing the diversity of hot subdwarf evolution: Surface composition, magnetic fields, and populations
    (2024-01-26) Dorsch, Matti; Heber, Ulrich
    Hot subdwarf stars of spectral types O and B (sdO/B) represent late stages of stellar evolution. They are located close to the hot end of the horizontal branch, and most of them are core helium burning stars. At radii of only 20% of the solar radius, they lack the extensive hydrogen envelopes of cooler horizontal branch stars. The formation of these objects and their evolutionary links to other classes of stars are still not fully understood, but their majority seems to result from binary evolution: Roche lobe overflow, common envelope episodes, and several types of stellar mergers. It was the aim of this thesis to provide an observational overview of the properties of hot subdwarfs as an important step towards understanding their complex formation and evolution. Therefore, both the overall hot subdwarf population and several peculiar stars were studied in detail. In the first part, individual hot subdwarfs were studied in great detail to establish them as testbeds for stellar evolution, in particular by determining their chemical signatures from high-quality ultraviolet and optical spectra. This includes two typical He-poor sdOB stars as reference objects: as determined from archival far-ultraviolet spectra, CPD-56 464 and the Schweizer-Middleditch Star exhibit similar surface compositions, including a distinct CNO-cycle pattern. Several heavy elements were identified for the first time in He-poor sdOB stars. The resulting heavy metal abundances are high compared to the Sun but lower compared to peculiar stars like LS IV -14 116. The absence of silicon in the Schweizer-Middleditch star likely results from the combination of diffusion and weak stellar winds. In contrast, CPD-56 464 has a silicon abundance about a third of solar, among the highest observed in helium-poor sdOB stars. Three of the most chemically peculiar stars, the so-called heavy-metal helium-rich sdOBs, were analysed with unprecedented detail and precision. High-quality UVES spectra of the Zr-rich LS IV -14 116 and Feige 46 revealed many strong lines corresponding to transitions of various heavy metals, some of which have not previously been observed in any star. The surface abundance of 19 metals is nearly identical in both stars. These abundance patterns differ significantly from those of typical He-poor hot subdwarfs with similar temperatures. The observed extreme overabundance of heavy metals suggests the presence of strong atmospheric diffusion processes that affect both stars similarly, although a contribution by s-process fusion is not excluded. The comparable abundances of C, N, O, and Ne in both stars provide evidence of a shared evolutionary origin. Recently proposed evolutionary models involving the merging of a hybrid He/C/O white dwarf with a more massive helium-core white dwarf offer a promising formation scenario for stars similar to LS IV -14 116. This scenario explains the atmospheric parameters, single-star nature, and unique pulsations observed in LS IV -14 116. The next analysis showed the lead-rich He-sdOB EC22536-5304 to be in a binary system with an extremely metal-poor subdwarf F-type companion. As a result, the derived lead abundance in the hot subdwarf is even larger than originally thought: about a million times solar, making it the most lead-rich star known to date. Based on the metallicity and atmospheric parameters of the F-type companion, the EC22536-5304 system seems to be older than about 10 Gyr. The system has an orbital period of about 457 days and was therefore formed through stable Roche lobe overflow. The experience gained with EC22536-5304 was then applied to another peculiar binary system: the sdOB + K-type subgiant BD-7 5977. The system was extensively observed with high-resolution far-UV, optical, and infra-red spectrographs. This allowed us to test the Roche lobe overflow scenario by determining the companion's C12/C13 isotopic ratio, which may have been decreased by accreted material. The resulting ratio (28+/-5) is clearly lower than the solar value (89), likely due to a combination of both mass transfer and mixing caused by convective dredge-up in the K-type companion. Magnetic fields are considered a smoking gun for the formation of single hot subdwarfs in the merger scenario, but despite several dedicated searches they have long eluded detection. Here I report the discovery of four magnetic He-sdOs and carry out spectral analyses based on high-quality optical spectra by modelling their Zeeman-split hydrogen, helium, and metal lines. The first magnetic hot subdwarf to be analysed, J0809-2627, was discovered by a low-resolution spectrum and later confirmed by X-shooter and UVES spectra. This intermediately helium-rich sdO star exhibits strongly Zeeman-split lines, indicating an average field strength of about 350kG. Although the star has a low projected rotation velocity, its overall properties are best explained as the product of a double helium white dwarf merger. In the next step, three additional magnetic He-sdOs that were discovered in the SDSS survey were analysed. These stars are almost identical to J0809-2627 in terms of atmospheric parameters, rotation, and field strengths. The occurrence of these magnetic stars suggests a lower limit of about 2% for the magnetic fraction in the He-sdO population. It remains unclear why the majority of He-sdO stars do not exhibit detectable magnetic fields, even though they are also thought to be formed by mergers. The Gaia space mission has provided photometric and astrometric measurements for more than one billion stars, which have transformed Galactic astrophysics in general and research into hot subdwarf stars in particular. Combined with large ground-based spectroscopic surveys and other photometric surveys, this huge dataset finally allows the study of statistically significant samples of the comparatively rare hot subdwarf stars. In the second part of this thesis, two analyses based on these datasets were performed. The analysis of spectral energy distributions (SEDs), from the UV to the infra-red, combined with Gaia parallaxes provided fundamental stellar parameters and characterised the population F/G/K-type companions to hot subdwarfs. In addition, the computation of Galactic orbits based on radial velocities and Gaia astrometry provided age estimates for various sub-populations of hot subdwarfs. The SED fits performed here are the most extensive so far: they include all ~6600 spectroscopically identified hot subdwarfs listed in the latest version of the hot subdwarf catalogue. In this study, 27% of these known hot subdwarfs turned out to belong to composite-SED systems, meaning that an F/G/K-type companion star was detected. An interesting result is that F- and K-type companions seem to be more common than G-types, which is currently not explained by theory. Further investigations based on Gaia parallax measurements, angular diameters from the SED, and spectroscopic surface gravities allowed the determination of radii, luminosities, and masses. As expected, the He-poor sdB stars on the extreme horizontal branch can be divided into two groups: a cooler one and a hotter one. These groups evolve into two groups of He-poor sdO stars once helium fusion moves from the core to a shell around it. The extremely He-rich sdO stars, which lack F/G/K-type companions, exhibit higher masses compared to other hot subdwarf classes, indicating that they were formed by merger channels. The intermediately He-rich sdOB stars are divided into two subgroups: the more luminous stars show a similar lack of companions, which suggests that low-mass white dwarf mergers contribute to their formation. In contrast, more compact He-sdOB stars have a companion fraction that is more comparable to the He-poor sdOB stars. Both the newly observed binary fractions for each spectral type and the lack of G-type companions should be compared to updated binary population synthesis models. From the observational side, future spectral analyses should be performed in a homogeneous fashion, which would improve the precision in particular in the important mass determination. Gaia parallaxes, proper motions, and literature radial velocities were then used to study the current Galactic velocities and orbital properties of the known hot subdwarfs. This kinematic analysis showed that most of these stars are associated with the Galactic thin and thick disk, each contributing about 44%, while the metal weak-thick disk and halo population contribute about 3% and 10%, respectively. In particular the helium-poor and helium-rich populations of hot subdwarfs show different kinematic properties. Helium-rich stars at effective temperatures of more than 32000K have a larger contribution by the thick disk and halo populations - these stars must therefore be predominantly formed in old stellar populations. The opposite is true for cooler helium-rich sdB stars, which seem to be dominated by the thin disk. In addition, the difference in age between the helium-rich sdOB stars and their more common helium-poor sdB counterparts seems to exclude the proposition that He-sdOB stars may evolve to become sdB stars. In conclusion, this thesis provides several detailed analyses of hot subdwarf stars, encompassing their abundance patterns, magnetic properties, binary nature, fundamental stellar parameters, and kinematic age. These results form an observational basis that can be combined with binary evolution models to shed light on the diverse formation mechanisms of hot subdwarfs and their evolution. In that sense, the observations performed here also contribute to the broader context of binary evolution, in particular in old stellar populations - in our Galaxy and beyond.
  • Doctoral thesis
    Open Access
    Transcriptional Targeting of Dendritic Cells as a New Therapeutic Vaccine against HIV-1
    (2024-01) Strack, Astrid; Winkler, Thomas; Steinkasserer, Alexander
    According to the World Health Organisation in 2021, more than 38 million people were living with the Human Immunodeficiency Virus (HIV). While antiretroviral therapy (ART) was a breakthrough in preventing progression into the acquired immunodeficiency syndrome (AIDS), there is still no cure available and traditional vaccination attempts have shown limited effects. More recently, dendritic cell (DC)-based vaccination strategies have sparked interest in controlling HIV infection. Targeting DCs for vaccination is particularly promising as they are extraordinarily potent inducers of strong and broad T cell responses. Currently, most DC vaccination attempts focus on cumbersome and expensive in vitro transduction of cells. Vectors targeting DCs in vivo would facilitate up-scaling and vaccine availability. Achieving these benefits without compromising on safety or risk pathogen tolerance requires new tactics to ensure targeting specifically only mature T cell stimulating DCs (mDCs). To address this challenge, this project aimed to investigate transcriptional targeting utilising the status-specific activity of the CD83 promoter to limit expression of HIV-1 antigens to mature DCs. We strived to create vectors based on the adenovirus serotype 5 (HAdV-5) that encode for HIV gag, a HIV gene as an antigen source, and a constitutive active variant of the IκB kinase (caIKK), an immune-stimulatory molecule to boost the DC activity. Expression of both those genes was controlled by the CD83 promoter. Before, however, we had to tackle limited transduction efficiency of DCs with HAdV-5. Towards the latter, LentiBOOST®/Polybrene (LeB/PB) significantly improved the transduction rate of human DCs. Additionally, LeB/PB enabled adenoviral transduction of human monocytes and murine bone marrow-derived DCs. Pivotally, cell viability, morphology and function was maintained in all assessed cell types despite exposure to the transduction enhancer. Next, we used the CD83 promoter to co-express two proteins within the same vector. By cloning the promoter twice into the vector or alternatively introducing an internal ribosomal entry site (IRES) and using fluorescent genes as placeholders, we created two vectors with comparable protein expression levels that provided the basis for the generation of HAdV-5 vectors for potential future use as vaccines: HAdV-5-IRES-gag-caIKK and HAdV-5-2xP-gag-caIKK (containing the CD83 promoter twice). qPCR confirmed protein expression to be restricted to mature DCs. DCs transduced with both viral vectors stimulated autologous T cell activity independently of the expression of antigen-specific T cells, as was demonstrated by upregulation of CD69 and increased production of pro-inflammatory cytokines including IL-6, IL-8, IL-12p70, and in addition IL-10. These results provide the basis in order to use the CD83 promoter to develop DC-based gene therapy against HIV-1.
  • Doctoral thesis
    Open Access
    Scanning Probe Studies of Porphyrins and Ionic Liquids on Metal Surfaces
    (2024) Adhikari, Rajan; Steinrück, Hans-Peter
    In this thesis, STM and nc-AFM disclose a tremendous richness of detail and a deep degree of comprehension on the molecular level. Both techniques yield real space information, which was applied to investigate the adsorption and reactivity of complex porphyrin derivatives and ionic liquids on different metal surfaces prepared under UHV conditions at or below RT. The publication [P1] addresses the adsorption behavior of three related cyano-functionalized tetraphenylporphyrin derivatives, namely Cu-TCNPP, Cu-cisDCNPP, and 2H-cisDCNPP on Cu(111) by STM as a function of temperature combined with DFT calculations in order to identify the role of the cyano groups and the central Cu atom. Cu-TCNPP forms a hexagonal honeycomb-type structure at RT that coexists with 1D molecular chains and completely transforms to a long-range ordered hexagonal honeycomb-type structure at 400 K. Cu-cisDCNPP forms flakes at RT, including mono-, bi-, and multiflakes, and then transforms to a hexagonal honeycomb-type structure at 400 K. 2H-cisDCNPP shows no ordered structure at RT, but orders after self-metalation upon heating to 400 K. Thus, all three molecules form the same long-range triangular porous hexagonal honeycomb-type structure, in which molecules adopt “saddle-shape” conformation. The observed structures share the same structure-forming element, that is, porous porphyrin triangles (distance of 3.1 nm) fused together via CN-Cu-NC interactions with Cu adatoms. Three porphyrin molecules are rotated by 60° to each other, forming regular triangular pores with Cu adatoms located at the corners (not visible in STM). Complementing STM, DFT calculations offer more insight into various energetic contributions leading to the thermodynamically stable hexagonal honeycomb-type structure. In addition, the internal structure of the unit cell indicates that cyano-phenyl groups (n = 2) in “cis” position are the minimum prerequisite to form a highly ordered 2D porous molecular pattern. To expand the understanding of the influence of linker groups (cyano- and isoindole (benzopyrrole) groups), publication [P2] addresses the adsorption behavior and structure formation of the novel cyano-functionalized benzoporphyrin 2H-TCNPTBP on Cu(111) by STM as a function of temperature. 2H-TCNPTBP, with four cyano- and four isoindole groups, shows the coexistence of three types of network structures, namely, a Kagome lattice structure, a quadratic pattern, and a hexagonal structure at RT. All three structures have molecules in a “saddle-shape” conformation. The Kagome and quadratic structures are porous with different pore sizes (diameters: ∼3 nm and ∼1.5 nm respectively), are stabilized by CN-Cu-NC bonds with Cu adatoms, and both have a molecular density of 0.18 molecules/nm2. The close-packed hexagonal structure has a molecular density of 0.42 molecules/nm2 and is stabilized with much weaker intermolecular H-bonds and dipole-dipole interactions of oppositely oriented cyano end groups. While the two Cu-coordinated porous structures (Kagome and quadratic) with same molecular density are stabilized by the energy gain due to the network formation (CN-Cu-NC), the hexagonal structure compensates the weaker intermolecular interactions (H-bond) by a factor of 2.3 higher molecular density. Heating to 450 K yields an immobile species with a “clover-shape”, which is attributed to the dehydrogenation and the formation of intramolecular C-C bonds between the isoindole and the phenyl groups. This finding suggests that cyano functionalization of benzoporphyrins results in unusual 2D self-assembled lattice structures. After gaining deeper insight into the role of linker groups, publication [P4] addresses the adsorption and self-assembly of exceedingly complex mixed benzoporphyrin derivatives 2H-diTTBP(x)BPs on Ag(111), Cu(111), and Cu(110) by STM as a function of temperature to provide further information on fundamental aspects. The mixture contains six different 2H-diTTBP(x)BPs with x = 0, 1, 2 (cis, trans), 3, and 4. On Ag(111), a long-range ordered 2D square phase was observed, which is stable up to 400 K. On Cu(111), an identical square phase coexists with a stripe phase, which disappears at 400 K. In contrast, on Cu(110) 2H-diTTBP(x)BPs adsorb as isolated molecules or dispersed short 1D chains (two to three molecules) along the < 11F0 > substrate directions, which remain intact up to 450 K. Notably, high-resolution STM images on all surfaces allow for the identification of different 2H-diTTBP(x)BPs. A “crown-shape” quadratic conformation on Ag(111) and Cu(111), an additional “saddle-shape” on Cu(111), and an “inverted” structure and quadratic appearance on Cu(110) were deduced. The different conformations are attributed to the different degree of interactions between the iminic nitrogen atoms of the isoindole and pyrrole groups and the substrate atoms on the three surfaces. Further, stabilization of the ordered 2D structures on Ag(111) and Cu(111) plus 1D short chains on Cu(110) is attributed to van der Waals interactions between the tert-butyl and phenyl groups of neighboring molecules. In addition, the thermal stability of the “crown-shape” conformation on Ag(111) and Cu(111) plus intact isolated “inverted” molecules on Cu(110), indicate that no reaction occurred. However, the loss of stripe phase on Cu(111) at 400 K implies a reaction, the nature of which is difficult to determine. A simple self-metalation with Cu atoms is unlikely, as metalated porphyrins exhibit island formation. Further, the presence of an intact “inverted” structure on Cu(110) indicates the absence of self-metalation. Moreover, self-metalation cannot be ruled out for molecules with four almost quadratically arranged protrusions. Overall, these findings show that the long-range 2D order or 1D short chains are stabilized by the outer periphery of the molecules rather than the substrate or the number of isoindole groups. Following the detailed characterization of the adsorption behavior of highly complex porphyrin derivatives on different metal surfaces, publication [P3] addresses the adsorption and reaction behavior of the ionic liquid [C1C1Im][Tf2N] on Cu(111) by nc-AFM, STM, and complementary XPS in combination with DFT calculations. Understanding the IL-substrate interface is of utmost importance for further developing the SCILL approach. [C1C1Im][Tf2N] on Cu(111) was deposited using two different preparation routes: either at RT followed by very fast cooling to 110 K, or directly at low temperature (< 160 K). At 200 K, IL films self-assemble into highly ordered islands (stripe phase) with intact cations and anions arranged next to each other. DFT calculations reveal two identically oriented cations nearly perpendicular to the Cu surface and two anions with different orientations, in line with the distinct contrast of anions observed in nc-AFM at 200 K and an adsorption energy (from DFT) per ion pair of 3.5 eV. Extended heating to 300 K triggers the stripe phase to evolve first into a hexagonal phase, and then into a porous honeycomb structure that coexists with many small, disordered islands. The hexagonal structure is a transition structure from the stripe phase to the honeycomb phase. The honeycomb structure is proposed as an oppositely orientated stacked “sandwich structure” of intact anions and cations. DFT calculations support the proposed structure and show that it is stable on the surface with an adsorption energy of 3.1 eV. The chemical composition measured by complementary XPS reveals that intact anions and cations are adsorbed next to each other at 200 K, and that no IL desorption occurs until 300 K. A large fraction of the IL is transformed into a new dissociated species at around 275 K and increases with time at 300 K, as evidenced by XPS. The decomposition products appear as disordered islands in nc-AFM and STM. Upon heating to 350 K, only small, disordered islands are observed by nc-AFM, and XPS indicates a complete decomposition of the IL on the surface. In summary, this thesis advances the molecular level understanding yielding novel detailed insights into the adsorption behavior of complex porphyrin derivatives and ionic liquids on different single crystal metal surfaces, with a focus on the specific roles of molecule-molecule and molecule-substrate interactions. The fundamental understanding obtained from the investigated systems in publication [P1-P4] might grant protocols as nanoscale templates for developing molecular devices or catalytic concepts.
  • Doctoral thesis
    Open Access
    The hydration of calcium aluminate cement in mixtures with calcite: Investigations at different temperatures and CA2/CA ratios
    (2024-01-24) Goergens, Julian Wenzel; Goetz-Neunhoeffer, Friedlinde
    This thesis represents a further step towards understanding the containment of the phenomenon of "hydrate phase conversion", which is considered to be the main challenge in the use of calcium aluminate cements (CAC) today. In this thesis, two main variables and their influence on the nature and quantity of all crystalline phases in hydrated CAC were investigated: The influence of hydration temperature was investigated with respect to the early hydration (0 - 48 h) and the late hydration (2 - 365 d). The effect of varying clinker composition, i.e. varying CA2/CA ratios in the CAC was investigated as a second variable. CAC reacts without chemical participation of the calcium carbonate (calcite) at 10°C, since the precursor phases C2AHx, which are conditional for the massive precipitation of carbonate-AFm phases, do not precipitate. At this temperature, CAH10 is the by far most dominant hydrate, without a sign of conversion for up to one year. At 23°C, CAH10 together with C2AHx, at 40°C and 60°C only C2AHx are formed. At all temperatures at which C2AHx is precipitated first, it is converted into monocarbonate within hours to minutes, and this process is accelerated with increasing temperature. Thus, after the main reaction, monocarbonate is always (except at 10°C) the dominant hydrate phase besides AH3. CAH10 does not reveal a "precursor function" for monocarbonate, it remains at 23°C until several weeks after hydration. At 10°C, CAH10 is still the predominant and, in the presence of amorphous AH3, the stable hydrate phase even after one year. C3AH6, which at later hydration times is quantified at 60°C, is causally dependent on the amount of evaporated mixing water in the early hydration. Thus, a reduced w/CAC ratio due to evaporation appears to determine the ratio of monocarbonate to C3AH6 rather than a thermal stability limit of monocarbonate. C3AH6 partially converts to monocarbonate in carbonate-bearing solution. On the other hand, if the CA2/CA ratio in the clinker of the CAC is varied, the amount of quantifiable AH3 changes in direct proportion to this ratio. However, the quantity of AH3 is not directly proportional to the strength, nor is the degree of hydration or the absolute amount of CA but specific populations of macro-porosity. It is much likely that the paste viscosity while stirring plays the decisive role in the formation of different degrees of macro-porosity. With a novel combination of three methodologies, it is now possible to exactly determine which porosity sizes are responsible for the fracture failure of individual samples of CAC-Cc-mixes.
  • Doctoral thesis
    Open Access
    Mechanistic, Integrative Multiscale Modeling of the Turnover of Soil Microaggregates
    (2024-01-19) Zech, Simon; Ray, Nadja
    This work is concerned with the mechanistic mathematical modeling of the turnover of soil microaggregates in interaction with the microbial turnover of soil organic matter at different scales. Microaggregates - composite soil structures smaller than 250µm - are considered to be the fundamental structural building blocks of soils and are therefore intrinsically linked to soil functions such as water retention, microbial habitat, and carbon storage and turnover. Soil microaggregates are composed of various mineral, organic and biotic components and are formed by an interplay of physical, biological and chemical mechanisms. Mechanistic models operating at the microscale can help to unravel the complex interplay of factors affecting aggregate formation, complement experimental evidence and are a prerequisite for deriving macroscopic understanding. However, an explicit description of the dynamically evolving soil structure has been neglected in most modeling approaches of microbial turnover of soil organic matter (SOM). In the beginning of this thesis, we present a versatile discrete continuum microscale model that operates on a spatially explicit domain. The model combines partial and ordinary differential equations for the reaction and diffusive transport of different species with a cellular automaton method (CAM) to describe the evolution of the microstructure in terms of solid phase reorganization and biomass evolution. A key component of the model is the ability to account for the effect of soil structure on SOM degradation and vice versa. In addition, data from real soil samples in combination with different imaging techniques (CT, dynamic image analysis) and a morphological fluid distribution model are used to initialize the model in a realistic way. A major focus of this thesis is the application of the microscale model in realistic settings based on experimental data to gain insights that are difficult to obtain from experiments. In the first study, we investigated microbial population dynamics and particulate organic matter (POM) turnover in response to carbon supply, pore space heterogeneity, and water connectivity in soil microaggregates using CT scans. The simulations revealed that heterogeneities in substrate and bacterial distribution have a strong local effect, but are averaged out over the scale of a microaggregate, with water connectivity identified as the key factor at this scale. Second, we studied the formation of microaggregate building units influenced by the size, shape and surface charge of the model minerals goethite and illite under different driving forces. Numerical simulations were in good agreement with laboratory experiments and facilitated the understanding of the formation and stability of microaggregate units with quantitative data that are difficult to obtain through experimental approaches, such as the contact area between particles. In the third model application, we disentangled the interplay between changing soil structure, POM turnover, and the simultaneous modulation of soil surface properties that may promote aggregation, as affected by soil texture, POM input, and POM decomposition rate. Our results showed that POM storage increases with clay content, demonstrating that surface interactions can delay the turnover of OC to CO2. In addition, we observed a priming effect, in which increased POM input stimulated the mineralization of old POM and revealed its origin through an increased rearrangement of the soil structure. As a next major objective, we addressed the challenging task of upscaling microscale models to larger scales. For this purpose, we introduced a macroscale model concept describing the diffusive transport of CO2 over a soil profile, where the CO2 production is calculated by representative microscale models taking into account macroscopic environmental conditions varying with soil depth. Model applications based on the previously investigated microscale scenarios underlined the importance of explicitly taking into account the microscale structure at the macroscale. Finally, we deal with the numerical investigation of a macroscopic flow and transport problem that can be challenging in many porous media applications when structural changes of the pore space lead to the degeneration of hydrodynamic parameters such as porosity, and permeability and diffusion tensors. Applying scalings of the unknowns in terms of the porosity as a remedy, we discretize a combined degenerating flow and transport problem with a mixed finite element method and prove the stability of the scheme. Numerical simulations confirm that optimal convergence is obtained for the transformed variables, while the non-transformed variables may not converge.