Dissipation and Noise in Granular Aluminum Fluxonium Qubits

The dream of universal quantum computing promises to revolutionize many fields due to its potential to drastically speed up certain calculations compared to classical computers.
This has naturally attracted attention and investment.
For example, in 2024, Google unveiled their Willow superconducting quantum processor, achieving a significant milestone by demonstrating quantum error correction below the surface code threshold.
Despite these advances, the road to a fully fledged quantum computer is still uncharted, and current technology is not sufficient to get there.
Therefore, unconventional materials hold great potential for addressing challenges faced by traditional approaches.
This work focuses on fluxonium qubits made from the disordered superconductor granular aluminum (grAl), aiming to gain deeper insights into underlying loss mechanisms and noise sources.

One of the main results of this work is the identification of inductive loss as the dominant decoherence mechanism in grAl fluxonium qubits with frequencies lower than 300 MHz at the half flux bias.
The observed inductive loss tangent aligns with previously measured single-photon internal quality factors of grAl resonators, and the energy relaxation profiles are well described by a combination of inductive, dielectric, and Purcell loss.
... mehr
The study endeavors to deepen the understanding of decoherence by refining loss models via the quantum fluctuation-dissipation theorem and proposing an improved framework for Purcell loss.
A comparison with other qubit materials underscores the universal nature of inductive loss in grAl fluxoniums, in contrast to well established Josephson junction array (JJA) based fluxoniums which show no such limitation at even lower qubit frequencies.

The second outcome of this thesis is the validation of flexible striplines (`flexlines') for the use in future cryogenic microwave setups.
This enables at least an order of magnitude increase in the density of microwave input circuitry without thermally overloading the cryostat, paving the way for increasingly complex superconducting detectors and quantum devices.
The study found no significant differences in qubit performance between setups using flexlines and conventional coaxial cables.
Passive heat load measurements indicated comparable photon shot noise-induced dephasing for both setups.
The introduced heating pulse method demonstrated a faster thermalization time for flexlines compared to coaxial cables, and it can serve as a simple health check for other groups as well.
An extended thermal model was developed to better understand the contributions of various attenuators in the input chains to the measured heat loads and to propose improved input chains.

To further exploit the versatility of grAl for fluxonium qubits, the investigation of the microscopic origin of the limiting inductive loss is of interest to develop potential suppression methods.
JJA-based fluxoniums could play a crucial role in this context, as particularly long arrays with increased plasma frequencies can emulate the properties of grAl while leveraging their better-understood behavior.