Superconducting micro-resonators are crucial in creating superposition states in quan tum computers and are gaining prominence as quantum technology evolves. Beyond their quantum applications, these resonators are instrumental as ultra-sensitive de tectors in astronomy, particularly Microwave Kinetic Inductance Detectors (MKIDs), which are key for both scientific exploration and industrial use. However, the is sue of electronic noise in these detectors remains a significant hurdle. To address this, our approach integrates machine learning strategies alongside conventional noise reduction methods to improve MKID signal fidelity. Our process involves gathering data and applying denoising models, with a strong emphasis on machine learning techniques. Our findings highlight the strengths of dif ferent denoising techniques, particularly deep learning architectures like Long Short Term Memory (LSTM) networks and Autoencoders, which demonstrate promising results in denoising MKID signals. These models exhibit adaptability to diverse noise sources, proficiency in identifying complex noise patterns, and continuous improve ment with additional data. This research represents a significant advancement in the field and provides es sential insights into improving MKID detectors, which in turn could lead to ground breaking developments in various scientific areas.
Superconducting micro-resonators are crucial in creating superposition states in quan tum computers and are gaining prominence as quantum technology evolves. Beyond their quantum applications, these resonators are instrumental as ultra-sensitive de tectors in astronomy, particularly Microwave Kinetic Inductance Detectors (MKIDs), which are key for both scientific exploration and industrial use. However, the is sue of electronic noise in these detectors remains a significant hurdle. To address this, our approach integrates machine learning strategies alongside conventional noise reduction methods to improve MKID signal fidelity. Our process involves gathering data and applying denoising models, with a strong emphasis on machine learning techniques. Our findings highlight the strengths of dif ferent denoising techniques, particularly deep learning architectures like Long Short Term Memory (LSTM) networks and Autoencoders, which demonstrate promising results in denoising MKID signals. These models exhibit adaptability to diverse noise sources, proficiency in identifying complex noise patterns, and continuous improve ment with additional data. This research represents a significant advancement in the field and provides es sential insights into improving MKID detectors, which in turn could lead to ground breaking developments in various scientific areas.
Superconducting microresonators are pivotal in the realm of quantum computing, essential for establishing superpositions. As quantum technology evolves, these components become even more integral, not just in quantum applications but also as precision instruments in astronomy—specifically as highly sensitive cameras. Microwave kinetic inductance detectors are notable in this field, recognized for their heightened sensitivity crucial for both scientific exploration and industrial use. Yet, electronic noise remains a formidable obstacle. Addressing this, our approach combines machine learning strategies with traditional noise reduction techniques to enhance the signal clarity of superconducting microresonators. Our process involves gathering data and applying various denoising models, placing a significant emphasis on machine learning methods. We have discovered that certain denoising approaches, especially deep learning frameworks such as long short-term memory networks and autoencoders, are highly effective in clarifying the signals from superconducting microresonators. These models demonstrate versatility in handling different noise types, skill in deciphering intricate noise structures, and they consistently improve their performance as more data is provided.
Microwave Kinetic Inductance Detectors (MKIDs) are superconducting sensors primarily used in astrophysics and quantum measurements for their energy- and time-resolution of single-photon detection. However, their sensitivity and precision are limited by various noise sources. Conventional denoising approaches enhance the signal-to-noise ratio (SNR) but often blur signal features and are ineffective with non-stationary noise. This work explores the use of machine learning (ML) denoising techniques for MKID signals in steady state resonators and in photon-induced transients. Different denoising techniques are evaluated on a dataset of frequency-domain S21 measurements and time-domain phase pulses with illumination of 660 nm, 780 nm, and 1310 nm wavelengths. Traditional filtering approaches (moving average, median, Wiener, Savitzky-Golay, wavelet, Kalman, total variation denoising) were evaluated in comparison with ML models including convolutional neural networks (CNNs), recurrent neural networks (RNNs), long short-term memory (LSTM) networks, and gated recurrent units (GRUs). Due to absence of clean reference signals in experimental settings, a Noise2Noise training approach was used for self-supervised learning from noisy data. The models were assessed using mean squared error (MSE), signal-to-noise ratio (SNR), variation in peak depth and optical resolution metrics. Findings of this thesis show that machine learning models achieve substantially better denoising performance compared to conventional methods across all metrics. GRU models outperformed other models when denoising steady-state resonator signals, as they captured local features of the spectral response, and they also were the most effective at denoising transient photon signals due to their capacity to capture temporal dynamics. Conventional filters, while being computationally efficient, were found to distort the signal and were not as effective in retaining low-amplitude photon-induced features. In summary, this study demonstrates that machine learning denoising is a robust and versatile method for improving MKID signals without prior knowledge of noise properties. The results also demonstrate the opportunity to enhance superconducting detector readouts by combining data-driven approaches, leading to improved sensitivity and resolution for applications such as astrophysical imaging and quantum sensing.
Microwave Kinetic Inductance Detectors (MKIDs) are extremely sensitive radiation detectors based on superconducting resonators that can be combined in large arrays on a single readout line within a limited frequency bandwidth. This makes MKIDs ideal detectors for the ultimate far-infrared observatory: a future space-based actively cooled telescope with its performance solely limited by the low universe background radiation. However, to reach these detector requirements, state-of-the-art MKIDs still need a order of magnitude improvement in device sensitivity. In this work, the MKID sensitivity is improved by reducing the aluminium volume that absorbs pair-breaking radiation into quasiparticle excitations, while making sure all radiation is still absorbed. Furthermore, a key requirement is sufficient reduction of excess noise as to keep the device intrinsically limited by thermally driven random fluctuations in the number of quasiparticles in absence of radiation, or Generation-Recombination (G-R) noise. To this end, a model is developed that describes the noise contributions as function of device geometry, readout power, material properties and radiation power. Subsequently, a realistic MKID design is presented and tested that reduces excess noise and maximises the sensitivity, expressed as Noise Equivalent Power (NEP). At high temperatures, good overall agreement is found between the measured noise spectra and the model. At low temperature T = 120 mK, the measurement results give an optical NEP similar to current state-of-the-art MKIDs. The NEP is not as low as expected due to short quasiparticle lifetimes, an unexpected decrease in the G-R noise level and a very high excess noise attributed to Two-Level Systems (TLS) noise that starts to dominate the already low G-R noise spectrum at low temperatures. Possibly, the quick quasiparticle lifetime saturation and noise level drop are caused by a strong readout power effect, as the readout power is known to create excess quasiparticles and to cause a strongly non-thermal electron energy distribution in the aluminium strip of the MKID. However, the exact microscopic details of these effects are unknown and not studied in this project. Based on the current chip design, a straightforward way to improve device performance and study the readout power effect in more detail is a reduction of the high TLS noise levels, which is possibly fabrication related. This would allow an unobstructed view of the G-R noise spectrum at low temperatures, thereby allowing both a study of the readout power effect on the quasiparticle system, and ultimately achieving the improvement in NEP needed reach the detector requirements for the ultimate space-based far-infrared observatory.
Microwave kinetic inductance detectors (MKID) are superconducting broadband direct detectors. If a superconducting thin film patterned into a resonator is illuminated with photons, Cooper pairs break up and the kinetic inductance of the film changes. This shifts the resonance frequency downwards and the dissipation of the resonator increases, which can be detected in the phase and amplitude of a transmitted signal. \n \nIn this thesis, a waveguide coupled MKID is designed and simulated. 350 GHz radiation is coupled into a copper waveguide via a horn antenna. A waveguide probe leads the signal onto a 9 µm silicon membrane where it is absorbed in a 40 nm superconducting aluminum film which is the inductive part of a superconducting resonator. The capacitor of the resonator is made of niobium. Two versions of the device are developed, a transmission line resonator and a lumped element resonator design. The new design will enable an accurate understanding of the optical sensitivity of MKIDs and allow comparison of different resonator materials. Since two different resonator types are used, the direct comparison of transmission line and lumped resonators will also be possible. \n \nA measurement setup is developed including an ADR cryostat to cool MKID devices down to below 40 mK. An analog readout system using either a vector network analyzer or an IQ mixer is designed and built. The necessary software to control the ADR, the temperature monitoring, the readout signal generation and detection and the signal acquisition of the setup is also developed. \n \nThe waveguide coupled MKID design is fabricated in our in-house microfabrication laboratory and measured in the new setup. In my measurements, the devices do not show a resonance feature. The VNA and DC measurements are analyzed to determine a plausible hypothesis of the failure. The devices are diagnosed for critical issues during fabrication which will be addressed in a future iteration. \n \nTo verify the measurement setup, a test device of the IRAM NIKA 2 MKID camera is measured and the complete analysis pipeline is exercised. The quasi-particle lifetime in the aluminum thin films is measured and a dark and optical noise equivalent power are determined. The results are within the uncertainty comparable to measurements of IRAM themselves. \n \nA test chip designed in my master thesis that contains eight NbTiN resonators is measured in the new setup and non-linear effects under strong readout power are studied. Parametric amplification and oscillation of a resonator are shown qualitatively.
Over the past two decades, microwave kinetic inductance detectors (MKIDs) have gained exceptional importance in millimeter and sub-millimeter astronomy. MKIDs consist of thin strip resonators capable of detecting changes in the surface impedance of superconductor strips, which result from variations in resonance circuit properties. The principal noise in MKIDs comprises excess frequency noise and two-level system noise. In this paper, we propose a technique to mitigate the effect of two-level system (TLS) noise in MKIDs using a parallel plate capacitor with three layers of high ε dielectrics. To achieve this, we employ three layers of Al2O3, HfO2, and TiO2 with equal thickness between the capacitor plates. The experimental results demonstrate a nearly 30% reduction in TLS power spectral density.
Astrophysical photons with a sub-millimetre wavelength are of great value to research star and galaxy formation. Superconductors have a small energy gap with which these sub-millimetre photons can be measured. Superconductors can absorb photons so that quasiparticles are created, which will in turn change the complex conductivity: incoming photons change the electrical response of the superconductor. This principle can be used to realise a detector. This is done with the Microwave Kinetic Inductance Detectors (MKIDs) that consist of a superconducting film embedded in a resonant circuit, so that photons falling in will change the resonant frequency. This can be measured with a readout signal. To enhance the sensitivity of the MKID, we want to increase the responsivity of an MKID as much as possible.A previous study has proposed a figure of merit regarding the material properties of a superconductor to achieve a high responsivity. However, the readout frequency used is not included in this figure. The goal of this work is to extend the figure of merit so that the readout frequency of the MKID can be optimised for a higher responsivity.