In recent years, deep neural networks have attracted the attention of both the academic community and the general public. An effort to theoretically understand the intricacies of these systems is...Show moreIn recent years, deep neural networks have attracted the attention of both the academic community and the general public. An effort to theoretically understand the intricacies of these systems is ongoing and physics-inspired approaches may have a part to play. In this thesis, we will discuss recent results in the theoretical study of deep linear neural networks. This class of neural networks has very limited real-world applications, but it could provide a good training ground for developing theoretical techniques that could prove useful beyond the simple linear case. We will also argue that Fisher information, and in particular “sloppy model” logic, can be a useful tool for future research on deep neural networks, in particular for network architecture optimization.Show less
Circuit partitioning is a powerful technique in the Noisy Intermediate Scale Quantum era. This method allows the evaluation of many-qubit algorithms with fewer qubits by writing the expectation...Show moreCircuit partitioning is a powerful technique in the Noisy Intermediate Scale Quantum era. This method allows the evaluation of many-qubit algorithms with fewer qubits by writing the expectation value of an observable as a weighted sum of products of inner products which are obtained by evaluating smaller circuits. However the number of terms in this sum generally scales exponentially with the depth of the circuit. The recently introduced Subset Partition Model only evaluates a subset of these terms. Here we explore the application of the Subset Partition Model for Variational Quantum Eigensolvers. Based on our investigation, we conjecture that this model may not be generally applicable to this Variational Quantum Algorithm due to the breaking of the Variational Principle. In addition, optimizing Partitioned Quantum Circuits is challenging due to the Barren Plateau problem that arises for even small circuit sizes. To address this problem, this thesis proposes a novel optimization algorithm that takes advantage of the additional structure on the cost function induced by circuit partitioning. By utilizing this new algorithm, we demonstrate that we can alleviate the barren plateau problem in the optimization of Partitioned Quantum Circuits.Show less
In our research project, we conducted an analysis of the impact of dark matter haloes on the motion of the Milky Way. Our study focused on dark matter haloes located within a radius of 200 Mega...Show moreIn our research project, we conducted an analysis of the impact of dark matter haloes on the motion of the Milky Way. Our study focused on dark matter haloes located within a radius of 200 Mega Parsecs (Mpc) from the Milky Way. The cosmological framework we implemented was the Lambda Cold Dark Matter model. Our primary objective was to deter- mine the peculiar acceleration of the Milky Way and derive insights about its motion. To achieve this, we compared the peculiar acceleration to the Hubble rate, a significant parameter in cosmology, as a reference point. By means of our study, we aimed to determine whether the Milky Way would eventually reach the Great Attractor or undergo a change in its direction of motion. Additionally, we generated an all-sky structure map for the haloes to explore the density distribution within each region. This anal- ysis allowed us to examine the concentration of dark matter throughout the universe inside the 200 Mpc from the Milky Way. In our research, we utilized data obtained from the SIBELIUS-DARK project, which provided a robust scientific basis for our study.Show less
In the quest for finding a room temperature stable single photon source (SPS), defects in hexgonal boron nitride (hBN) have gained more and more attention in the previous years. This thesis studies...Show moreIn the quest for finding a room temperature stable single photon source (SPS), defects in hexgonal boron nitride (hBN) have gained more and more attention in the previous years. This thesis studies the single photon emit- ting properties of defects in a multi-layer hBN flake with a home-build confocal microscopy setup. As a reference system, nitrogen vacancy (NV) centers in nanodiamonds (NDs) are measured first, as there luminescent properties are well documented in literature. In order to improve on pre- vious work in this group, polarization selective elements are included in the optical setup. This inclusion allowed measurements of the absorption and emission dipoles of the SPSs and improved our results by (1) reduc- ing the background and (2) allowing more efficient excitation. We found that defects in hBN showed promising results in terms of single photon purity (g2(0) =0.09 ± 0.06) and brightness (35 KHz), outperforming the best single NV center (g2(0) =0.29 ± 0.07, brightness: 4 KHz). We also found that the hBN defects showed long term (up to 3000 ns) bunching behaviour, possibly related to the excitation power during measurements. This, together with the measurement of different hBN flakes, constitute interesting topics for future investigation.Show less
Microcavities play a significant role in the study of cavity quantum electrodynamics (CQED), as they induce efficient coupling of light and matter. Confining matter within a cavity, increases the...Show moreMicrocavities play a significant role in the study of cavity quantum electrodynamics (CQED), as they induce efficient coupling of light and matter. Confining matter within a cavity, increases the probability of deterministic interactions between e.g. a (quantum) particle and the light cast into the cavity by a laserbeam. Cavity parameters such as cavity length and radius of curvature of the micromirror define the cavity regime and consequently the physical phenomena that can be observed, from the bad-cavity (Purcell) regime to the strong coupling regime. We are looking for a highly controllable and repeatable way of producing microcavities with a small mode volume and a high finesse. This translates into the creation of micromirrors with radii of curvature between 10 μm and 50 μm and a depth of up to 1 μm. We offer an overview of micromirror production methods; CO2 laser ablation, focused ion beam milling (FIB), direct laser writing and current controlled curvature. For CO2 laser ablation the key parameters to control the dimensions of the ablated structure, are the power of the laser and the beamwaist. We perform simulations of the ablation process and show that micromirrors with the desired dimensions can be obtained by using a CO2 laser in a range between 420-440 mW and a beamwaist of 40 μm. We find that for these parameters clipping loss is negligible. Therefore, CO2 laser ablation meets the criteria and proves to be a reliable way of producing micromirrors.Show less
Our research is related to testing dark energy/modified gravity theories. We determine the positivity bounds on effective field theories with spontaneously broken Lorentz invariance. We consider...Show moreOur research is related to testing dark energy/modified gravity theories. We determine the positivity bounds on effective field theories with spontaneously broken Lorentz invariance. We consider all the operators in a low-energy — effective field theory (EFT) approach and gain the conditions for EFT coefficients so that a theory is healthy (without instabilities). These conditions are called the positivity bounds, for which a theory works. These positivity bounds can give us constraints about the cosmological model. We mainly follow the paper Positivity bounds on effective field theories with spontaneously broken Lorentz invariance by Paolo Creminelli, Oliver Janssen, and Leonardo Senatore, where the positivity bounds are calculated from the two-point correlation functions of conserved quantities like the Noether current and stress-energy tensor. Then we show how this new mechanism of finding positivity bounds can be used for real cosmological models.Show less
Due to it’s chirality, the electron transport or electronic transmission through a molecule can have a spin preference. This is called the Chiral-Induced Spin Selectivity (CISS) effect....Show moreDue to it’s chirality, the electron transport or electronic transmission through a molecule can have a spin preference. This is called the Chiral-Induced Spin Selectivity (CISS) effect. Understanding CISS may contribute to the understanding of life but can also be useful for spintronic applications. Within this research, a setup for measuring the CISS-effect locally, using a Photo-Emission Electron Microscope (PEEM), is investigated. It is found that using a linear polarizer in combination with a lambda half plate can be related to the parallel and perpendicular polarization direction with respect to the sample. This is useful for tuning circular and elliptical polarization to the samples surface. Besides, it turned out that difference in illumination response for metallic structures for different handedness of the elliptically polarized light, is mostly related to the size (∼ 1µm) of the structures, but not significantly to its geometry. The setup is tested for a Self-Assembled Monolayer consisting of alkanethiol molecules, for which it shows equal response for left- and right-handed circularly polarized light. A Low-Energy Electron Microscopy (LEEM) measurement shows that the self-assembly process has led to an ordered layer on the gold on Mica substrate. In extension to this research, for future research on the CISS-effect, it would be interesting to fabricate a Self-Assembled Monolayer of chiral molecules with different handedness in order to measure whether or not opposite response to different elliptical light polarizations.Show less
In the process of metastasis, cancer cells may transmigrate through the endothelium barrier of the vascular walls and into the circulatory blood system. During this process, the cancer cells...Show moreIn the process of metastasis, cancer cells may transmigrate through the endothelium barrier of the vascular walls and into the circulatory blood system. During this process, the cancer cells interact with the endothelial cells, resulting in the alternation of their mechanical properties. Although this interaction has been broadly studied from the perspective of cancer cells, no thorough investigation of the endothelial mechanical properties has been performed. In this thesis, by using a micro-rheology AFM-based approach, we show that the properties of endothelial cells change when cultured in cancer cell conditioned medium, as well as when in contact with cancer cells. We found that the stiffness of endothelial cells increased when cultured in a low-dilution cancer cell medium, while it decreased in high-dilution medium. This change was also evident on the viscoelasticity of the cells, with the endothelial cells cultured in high dilution medium showing lower viscoelastic properties. Furthermore, endothelial cells in direct contact with a cancer cell showed an increased height, as a result of the cancer cell's transmigration through the monolayer. Our findings demonstrate that the properties of endothelial cells change indirectly due to cancer cell secreted substances, and directly during the cancer cells' physical transmigration. This indicates that the endothelium is actively responding to the presence of cancer cells, rather than being a passive barrier as once believed.Show less
Cyclically driving subharmonic units can result in emergent memory effects, and such systems have the potential to store information and perform computations. For this reason, we numerically...Show moreCyclically driving subharmonic units can result in emergent memory effects, and such systems have the potential to store information and perform computations. For this reason, we numerically investigated the response to cyclic driving of two linearly coupled toggling bits, which are an inherently subharmonic unit. We found that there are 380 topologically distinct two-t-bit systems. Cyclic driving sometimes resulted in subharmonic responses with emergent periods of T=3 and T=4, as well as nonzero transients (tau>0). A variety of different orbits is possible, but the behavior (tau, T)=(0, 2) inherent to the single toggling bit, remains common in systems of two t-bits. A period of T=4 already occurred for weakly coupled systems that can be manufactured in experiments. We also found the restriction that tau + T < 4. Lastly, we conclude that coupling strength strongly affects the likelihood of avalanches occurring. The findings demonstrate both the capabilities and limitations of coupled toggling bits, which will valuable for future research.Show less
This work explores axion-photon conversion in the TeV halo of the Geminga pulsar and in neutron star magnetospheres. Based on the observed synchrotron and inverse Compton scattering radiation, the...Show moreThis work explores axion-photon conversion in the TeV halo of the Geminga pulsar and in neutron star magnetospheres. Based on the observed synchrotron and inverse Compton scattering radiation, the magnetic field inside the Geminga pulsar TeV halo is constrained to be < 2 μG, similar to literature constraints. The photon-axion conversion probability is ≲ 10−8, requiring an extreme signal-to-noise ratio for detection. In contrast, significant flux transfer can take place in the strongly magnetised anisotropic plasmas of neutron star magnetospheres. Following up on recent literature, this work provides a three-dimensional calculation of axion-photon conversion in anistropic plasmas, including the Euler-Heisenberg photon-photon scattering contribution in the limit BNS ≪ Bc = 4.4 · 1013 G. This allows for resonant double lens conversion, through which axion-like particles of arbitrarily small mass can resonantly induce photons. For relativistic axions the resonance length scale can exceed other typical scales of change in the plasma, in which case the conversion region and probability is truncated. For very light, relativistic axions, non-resonant contributions are important. In the Goldreich-Julian model with relativistic plasma, double lens resonance can occur at observable radio frequencies in the 100 GHz−1 THz regime. NB: Erratum. The derivation in chapter 3 sets magnetic permeability mu=1. For double lens resonant conversion this is inconsistent. Upon including the correct expression for mu, the EH contribution to the resonance condition (Eq. 3.28) is slightly changed:14*eta*B^2 -> 22*eta*B^2. The error should, of course, also be incorporated in the whole derivation.Show less
Efficient single-photon sources based on semiconductor quantum dots typically rely on resonant excitation schemes with a high degree of control. In particular, having access to continuous-wave (CW)...Show moreEfficient single-photon sources based on semiconductor quantum dots typically rely on resonant excitation schemes with a high degree of control. In particular, having access to continuous-wave (CW) and pulsed excitation without changing the center frequency is highly desirable. CW excitation is useful for alignment and characterization, while pulsed excitation is essential for on-demand single-photon production. We present a technique based on ultra-fast electro-optic modulation to directly synthesize optical pulses from a narrow linewidth CW laser. With custom-built ultra-fast electronics, we demonstrate tunable pulse lengths down to 50 ps. Pulses longer than 100 ps achieve a typical extinction ratio of 300, and the 50 ps pulses still show an extinction ratio of 150. We then use these pulses to excite a single InAs quantum dot in a micropillar cavity and show the generation of true single photons. This technique allows for full control over the experiment in the temporal-spectral domain, and is significantly simpler compared to using conventional Ti:Sa mode-locked laser oscillators in combination with grating-based pulse shaping.Show less
This thesis aims to alleviate the final parsec problem by investigating the hypothetical intermediate-mass black hole environment lying at the cores of galaxies, a model first proposed by Ebisuzaki...Show moreThis thesis aims to alleviate the final parsec problem by investigating the hypothetical intermediate-mass black hole environment lying at the cores of galaxies, a model first proposed by Ebisuzaki et al. (2001) [1]. Although intermediate-mass black holes remain undetected, their nature could be the key to understanding supermassive black hole formation. If they are indeed present at the hearts of galaxies, their mutual interactions encourage supermassive black hole-intermediate-mass black hole merging events. Such merging events bypass theoretical constraints placed by binary dynamics and the Eddington limit, allowing for supermassive black holes to grow into their colossal sizes, and could potentially help explain their existence in the early stages of the Universe’s life. We investigate this model using both a Newtonian (Hermite) and post-Newtonian (HermiteGRX) algorithm. The post-Newtonian algorithm incorporates terms up to order 2.5, allowing it to model gravitational wave emission, which acts as an energy sink source and encourages merging events. In addition to comparing the results found using either algorithm, we forecast its corresponding gravitational wave events. More specifically, assuming a steady intermediate-mass black hole infall rate of one every 7 Myr, we predict a population of NIMBH = 15∼20 residing at the inner 0.4pc of the Milky Way galaxy. In turn, the future gravitational wave interferometer LISA and the proposed one µAres will be able to detect up to 926 supermassive black hole-intermediate-mass black hole merging events per year up to a redshift z ≤ 3. This value is three orders of magnitude larger than those found in various literature ([2]; [3]; [4]; [5]) due to the lack of observation of intermediate-mass black hole leaving a large parameter space in such analysis.Show less
We successfully produce Josephson junctions by stacking NbSe2 flakes, where the twist angle and strain are responsible for a weak link. Field sweep measurements are taken to characterize these...Show moreWe successfully produce Josephson junctions by stacking NbSe2 flakes, where the twist angle and strain are responsible for a weak link. Field sweep measurements are taken to characterize these devices. These measurements highlight the importance of the direction of the applied magnetic field. Prompted by this, simulations of the critical current given an arbitrarily shaped interface area are presented and compared to measurements.Show less
There is a demand for high bandwidth down links from space to earth. A cubesat in a GEO could function as a relatively cheap access point to a high bandwidth communication channel with earth. This...Show moreThere is a demand for high bandwidth down links from space to earth. A cubesat in a GEO could function as a relatively cheap access point to a high bandwidth communication channel with earth. This thesis explores new ways to increase the bandwidth by identifying bottlenecks in the GEO- Earth communication channel and how to circumvent them. The diffrac- tion limit causes large beam spreading at GEO distance, holding back ad- vanced modulation techniques due to the inability to capture the whole wave front. In this case, a modulation scheme using only a few bits should be chosen, allowing to modulate as fast as possible. The low signal inten- sity can be detected with more sensitivity by making use of a quantum enhanced receiver. From GEO to Earth, data rates around 50 Gbps are possible. Additionally, the atmosphere introduces spatial incoherence. To mitigate the effects of the atmosphere, a modulation scheme should be chosen that exploits modulation vectors which are orthogonal to the spa- tial dimension, such as polarization or wavelength. This gives a modula- tion scheme with many degrees of freedom. To deal with the complexity, a variational auto-encoder deep neural network is used to act as the modu- lator and demodulator. The variational distribution is chosen to match the noise introduced by an atmospheric channel. Using this scheme, we were able to find encodings that increase the density of symbols in phase space relative to the noise. This approach is especially promising in a bandwidth limited channel.Show less
The application of denoising machine learning to STM data has several advantages, such as improving data quality, aiding visual interpretation of data, and speeding up measurement time. With...Show moreThe application of denoising machine learning to STM data has several advantages, such as improving data quality, aiding visual interpretation of data, and speeding up measurement time. With experimental data, the absence of a ground truth poses a problem for traditional supervised learning techniques. In this work, state-of-the art self-supervised machine learning techniques are applied to reduce noise in quasiparticle interference data of overdoped cuprates, using only the noisy measurements. The machine learning methods are shown to outperform traditional denoising methods. Further ideas to improve and generalize the denoising of quasiparticle interference data are proposed.Show less
