In this research, we study the oxidation process of the Transition Metal Dichalcogenide (TMD) niobium diselenide, NbSe2. TMDs are a category of van der Waals materials, of which samples are...Show moreIn this research, we study the oxidation process of the Transition Metal Dichalcogenide (TMD) niobium diselenide, NbSe2. TMDs are a category of van der Waals materials, of which samples are obtained by exfoliation. The measurements are performed with the Low Energy Electron Microscope (LEEM), which measures reflectivity at different energy, resulting in the so-called IV curves. To analyze oxidation, we have developed a new method to obtain in situ exfoliated flakes in the LEEM set-up. First, we measure in situ cleaved NbSe2 flakes (bulk). The reflectivity upon adding (pure) oxygen is followed. Additionally, the reflectivity of ex situ cleaved NbSe2 flakes is assessed. Here, an intensity boundary between flake’s edge and center is recognized. The reflectivity measurements show that the electronic structure is different across the boundary: a V-shaped reflectivity minimum appears, which sharpens over time upon oxidation. The boundary is further researched with Atomic Force Microscopy and Energy Dispersive X-ray analysis. Additionally, we perform roughness analysis and Principal Component Analysis. The latter provides an alternative method to follow the change in electronic properties over time. We propose that already degraded NbSe2 flakes are more susceptible to oxidatin damage, compared to in situ exfoliated flakes. Upon further measurements, the in situ cleaved samples do not show any degradation signs, therefore we attribute the presence of an intensity contrast with the associated IV features, to oxidized NbSe2.Show less
This thesis explores the initial steps towards integrating high temperature scanning SQUID-on-tip (SOT) with quartz tuning fork atomic force microscopy (QTF-AFM). By combining these imaging...Show moreThis thesis explores the initial steps towards integrating high temperature scanning SQUID-on-tip (SOT) with quartz tuning fork atomic force microscopy (QTF-AFM). By combining these imaging techniques into one sensor, the local magnetic field variations and surface topography of a sample can be mapped simultaneously. This allows the SQUID to scan with ultra-sensitive flux sensitivity and nanometer spatial resolution, while its position on the surface remains identifiable. Specifically, this thesis addresses the low operating temperature of conventional SOT by developing a fabrication method that uses BSCCO, a high temperature superconductor, to create SQUIDs through gallium focused ion beam (FIB) milling. The electrical contacting procedure of BSCCO involves mechanical exfoliation and electron-beam lithography. The results yield contact resistances in the order of 100 ohm, which are sufficiently low to perform current transport experiments. The flakes are then structured into 1 micrometer SQUIDs. The Josephson junctions are created by introducing ion beam induced damage to the crystal lattice of BSCCO to suppress superconductivity. The transport measurements reveal no conclusive evidence of SQUID features. However, it is shown that milling sub-200 nm wide structures does not alter the electronic properties of BSCCO, indicating that this nanostructuring method can potentially be applied in fundamental research into high temperature superconductors. This thesis also focuses on depositing multiple SQUID electrodes along a QTF, while keeping the self-sensing and -actuating capabilities of the force sensor intact. The QTF is insulated with a 100 nm thick SiOx layer. It is then covered with a laser micro-machined hardmask, through which 50 nm of titanium is evaporated in the shape of SQUID electrodes. Through fabrication alterations, certain issues involving alignment and electrode interruption can be solved. However, the evaporation method inexplicably compromises the integrity of the insulating barrier, thereby forming electrical shorts. Overall, the findings indicate that while substantial progress has been made in developing fabrication methods for the different components, significant technical hurdles remain. These need to be addressed to realize the potential of BSCCO scanning SQUID-on-tip for atomic force microscopy.Show less
In this research, a recently proposed renormalization group approach for networks to the case of random directed graphs is being generalized: we present a scale-invariant description of directed...Show moreIn this research, a recently proposed renormalization group approach for networks to the case of random directed graphs is being generalized: we present a scale-invariant description of directed networks containing reciprocated edges. This allows us to neglect several strong assumptions that are currently necessary to renormalize directed networks such as financial transaction networks. As an application, a model of ING’s transaction data has been derived across multiple coarse-grained partitions. In this article we provide detailed information on how this particular model has been structured and how its parameters are obtained. We show how we can use this model to determine the expected cumulative degree and weight distributions of ING’s transaction network across multiple coarse-grained partitions of the network which we will compare to the empirical degree and weight distributions, respectively.Show less
This paper presents an initial approach to investigate the feasibility of fabricating FSF heterostructures, for potential applications in spintronics. These FSF stacks could help characterize long...Show moreThis paper presents an initial approach to investigate the feasibility of fabricating FSF heterostructures, for potential applications in spintronics. These FSF stacks could help characterize long-range triplet Cooper pairs that are created at its interfaces. Simulations inform the final design of a 3D FSF stack, composed of one low and one high aspect ratio rectangular shape. Subsequent efforts were directed on the focused ion beam milling of apertures in 1 µm thick Si3N4 membranes that are used for shadow evaporation of the heterostructures. Several stacks of cobalt and niobium were fabricated with these membranes through shadow evaporation. These heterostructures were contacted via electron beam lithography and measured in a vectormagnet cryostat. Results indicate that no long-range triplets were generated inside the FSF stacks. It can be concluded that either the niobium layers did not become superconducting or the magnetization of the two ferromagnets in the stacks were not perpendicular.Show less
At the moment, quantum computer development is in the NISQ (Noisy Intermediate Scale Quantum) stage. This means that quantum computers are relatively small and exhibit large amounts of noise. To...Show moreAt the moment, quantum computer development is in the NISQ (Noisy Intermediate Scale Quantum) stage. This means that quantum computers are relatively small and exhibit large amounts of noise. To run any mean- ingful computation, small noise-resistant circuits are necessary. This work proposes a new algorithm, QASNEAT, for finding small noise-resistant circuits. The performance is evaluated by ground-state energy estimation of three small molecules with shot noise and physical depolarizing noise. QASNEAT is able to find small accurate circuits both in noisy and noise- less casesShow less
One of the greatest remaining puzzles in physics is what particle dark matter consists of. For this project, the theory of dark pions is considered, a Hidden Valley model that extends the Standard...Show moreOne of the greatest remaining puzzles in physics is what particle dark matter consists of. For this project, the theory of dark pions is considered, a Hidden Valley model that extends the Standard Model with new, dark particles and a new force, dark QCD. A sensitivity study is performed to determine how many dark pions are expected to be in acceptance of the LHCb detector for Run 2 conditions; the LHCb is well-suited to search for particles in the considered O(1) GeV mass and O(1) - O(100) ps lifetime range. Additionally, a framework has been developed to study the dependence of the sensitivity on a number of theoretical parameters of the dark QCD model, namely the probability to form a dark vector meson instead of a dark pion, the number of colours in dark QCD, the dark QCD scale, and the Higgs mass. It is found that O(100) dark pions are in LHCb acceptance for different track categories, and that the considered the- oretical parameters do not drastically change the number of expected particles (with some small caveats), staying within a difference of about 20%. This is acceptable given the expected experimental uncertainty, showing theory inde- pendent searches for dark pions are possible.Show less
AdS/CFT gives a framework for using calculations from a weakly curved gravitational theory to describe phenomena in strongly correlated matter. In this work we study some holographic models this...Show moreAdS/CFT gives a framework for using calculations from a weakly curved gravitational theory to describe phenomena in strongly correlated matter. In this work we study some holographic models this has put forward, such as the Reissner-Nordstro ̈m metal and the holographic superconductor. Of these models we observe transport properties like the optical conductivity under both translational invariance and broken symmetry. Lastly, we merged the two systems into the two-charge holographic superconductor and managed to again see the phase transition, measure the optical conductivity and analyse the two- fluid model.Show less
This work evaluates the performance of both Neural Networks (NNs) and Parameterized Quantum Circuits (PQCs) in Reinforcement Learning environments. We present novel quantum gate based games that...Show moreThis work evaluates the performance of both Neural Networks (NNs) and Parameterized Quantum Circuits (PQCs) in Reinforcement Learning environments. We present novel quantum gate based games that can run on current NISQ hardware. Our results show that NNs and PQCs achieve very similar performance distributions across the different environments, showing the promise of PQCs for (Quantum) Machine Learning applications. It is also shown that the NNs tested are more sensitive to change in learning rate than our PQC models. NN performance is also more eratic with relation to the amount of parameters than PQC performance, showing hyperparameter tuning might be more predictable for PQCs. Lastly, the smallest PQC designs show strong performance, often outperforming NNs with more parameters.Show less
The past decades have shown a rise in skin cancer. This creates the need for prevention and efficient treatment. The most common skin cancer (melanoma) can only be treated when detected early. In...Show moreThe past decades have shown a rise in skin cancer. This creates the need for prevention and efficient treatment. The most common skin cancer (melanoma) can only be treated when detected early. In this thesis we propose a method of increasing awareness for people with a high risk of skin cancer as well as allowing for early detection. Skin cancer is hard to detect, even for experiences healthcare professionals. One of the signals of potential harm full lesions is change over time. We propose to develop an application with which changes in skin lesions can be identified early. By allowing patients to film their body with a mobile phone camera we aim to track the development of lesions. If a patient films their body regularly changes can be detected and the application can urge the patient to consult a dermatologist. In this thesis we explore the possibility of combining the frames of these films into an overview displaying the patients complete back or arm. Combining frames is called stitching. Different stitching techniques found in literature are explored and tested for effectiveness. The optimizations performed are reported and the final result is presented. The location of the different lesions on an overview of the body is needed to show the patient and the healthcare professional where potential harmful lesions are located on the body. This allows for further inspection at the dermatology department.Show less
In this thesis we present our experimental work towards realizing single photon spectroscopy. Based on theoretical predictions and experimental confirmation with laser absorption spectroscopy, we...Show moreIn this thesis we present our experimental work towards realizing single photon spectroscopy. Based on theoretical predictions and experimental confirmation with laser absorption spectroscopy, we identify a reltively strong transition of water vapor. This transition is in close proximity to the emission of our single photon source: a semiconductor GaAs embedded InGaAs quantum dot (QD) with emission close to 935 nm. We then show absorption spectroscopy with this single photon source where we scan the emission wavelength of the quantum dot by means of the Stark effect. In a 75 pm wide spectral range for which the source is capable of producing relatively pure single photons, which we confirm with Hanbury Brown Twiss interferometry, we measure the transmission of 261 cm humid air with QD light. We find that the observed peak wavelength and linewidth characterizing this transition compare well with the expected parameters obtained from the HITRAN database.Show less
In this thesis the relative spectral energy density of stochastic primordial gravitational waves is investigated. Decoupling of Standard Model particles and neutrino free-streaming affect the...Show moreIn this thesis the relative spectral energy density of stochastic primordial gravitational waves is investigated. Decoupling of Standard Model particles and neutrino free-streaming affect the expansion history of the universe and thus leave characteristic signatures on the amplitude of the gravitational wave spectrum. Adding extra light or heavy particles damps the spectrum at frequencies before the particle decouples. Including an extra neutrino species amplifies the spectrum at larger wave numbers, but damps it at shorter wave numbers. Measuring these primordial gravitational waves reveals the thermal history of the universe. One possible non-standard thermal history is early matter domination due to the inflaton. It is shown that, in this cosmology, the end of early matter domination and beginning of the radiation era depend linearly on the reheating temperature.Show less
In this thesis we present an experimental realisation of a double loop type Magnetic Paul Trap. We show that a microgram heavy NdFeB permanent magnet can stably be levitated for hours at room...Show moreIn this thesis we present an experimental realisation of a double loop type Magnetic Paul Trap. We show that a microgram heavy NdFeB permanent magnet can stably be levitated for hours at room temperature in this trap. Magnetic levitation of a magnetized particle is theoretically possible with this trap by generating opposed alternating magnetic fields. We show the fabrication of a printed circuit structure capable of producing these fields, as well as the engineering behind the realisation of the trap. Both by optical and magnetic readout we characterize the motion of the trapped magnet and show that its center of mass motion frequencies $\omega_z = 2\omega_{x,y} \approx 20Hz$. We characterize the damping on these modes and find that at low pressure the quality factor is strongly limited (to $Q \approx 90$) by coupling to the environment through generation of Eddy currents.Show less
The intergalactic medium (IGM) contains most of the baryonic matter of the Universe and serves as a suitable environment for probing the thermal history of the Universe. The crucial moment in IGM...Show moreThe intergalactic medium (IGM) contains most of the baryonic matter of the Universe and serves as a suitable environment for probing the thermal history of the Universe. The crucial moment in IGM evolution is the Epoch of Reionization, corresponding to the transition from neutral to ionized IGM. However, due to the observational limitations, this period is still not well understood. In this thesis, we focus on constraining IGM thermal history by using Lyman-alpha forests data. This method is applicable in a wide range of temperatures, densities, and ionization fractions of cosmic gas at z about 2 − 5. Observations show that the longitudinal flux power spectrum of the Lyman-α forest exhibits a cut-off at small scales. This phenomenon is caused by thermal Doppler broadening, peculiar velocities along the line of sight (LOS), Hydrogen pressure smoothing, and warm dark matter. The first two effects act only along LOS, while the last two affect all spatial directions. To separate the one-dimensional and three-dimensional effects, we used the method of close quasar pairs, which is based on studying the correlations between Lyman-alpha forests of close quasar pairs. We used the Kolmogorov-Smirnov test to analyze the differences between distributions of phase difference, which characterizes correlations between Lyman-alpha forests. The calculations were performed for various thermal histories, parameters characterizing IGM, LOS separations, and wavenumbers, and accounting for different effects (Doppler broadening and peculiar velocities). Our results indicate that this method can distinguish various thermal histories regardless of the IGM thermal state and one-dimensional effects. Moreover, at separations of the order of pressure broadening, there is a prominent feature caused by different influences of pressure smoothing at large and small scales. In addition, this simple and powerful approach has the potential to distinguish scenarios with warm dark matter.Show less
In this work, a near-zero stiffness mechanical filter is designed for use in STMs in a cryogenic environment. The filter is a Geometric anti-spring (GAS) filter which consists of a set of blades...Show moreIn this work, a near-zero stiffness mechanical filter is designed for use in STMs in a cryogenic environment. The filter is a Geometric anti-spring (GAS) filter which consists of a set of blades with a payload attached. This design allows for a low resonance frequency of 0.27 Hz and thus a low cutoff frequency in terms of filtering. First, a theoretical model is described in order to determine the relevant properties of the filter and its approximate workings. Second, the model was experimentally verified. From this, the resonance and damping of the filter were found. A limited amount of vibration measurements were also done to check if the filter is functioning as expected, however, due to instrumental limitations this was not conclusive.Show less
Quantum reinforcement learning (QRL) is a growing field of research that has gone through several advancements over the past decade. Nonetheless, one of the obstacles that this field still faces is...Show moreQuantum reinforcement learning (QRL) is a growing field of research that has gone through several advancements over the past decade. Nonetheless, one of the obstacles that this field still faces is that a quantum computer is needed for both the training of the models and their deployment. This same problem holds for general quantum machine learning and has been addressed in recent work by utilizing classical shadows, or classical approximations, of flipped quantum models which are linear models with parametrized quantum states and data-encoding observables. They show that there exists a supervised learning task in which shadow models, which were derived from flipped models, achieve a learning advantage over fully classical models, yet the question of whether this advantage holds in QRL is left unexplored. In this work, this question is addressed by investigating the performance of flipped models (particularly those that successfully admit to shadow models) on an RL benchmark and proving a learning advantage of shadow models over fully classical models within RL. The former is accomplished by performing numerical studies on flipped models before comparing their performance to that of an already established QRL model, whereas the latter is achieved by constructing a new discrete cube root RL environment for which an arbitrary learning separation between shadow models and fully classical models is proven under the widely believed discrete cube root assumption. The results show that flipped models can perform well in an RL benchmark environment, and that shadow models can achieve a learning advantage over fully classical models in RL.Show less
