Plasma-Mode Engineering in Fluxonium Qubits

The fluxonium qubit [1] combines long coherence times with large anharmonicity by shunting a small Josephson junction with a high-inductance superinductance. Realizing large inductances typically requires long Josephson junction arrays [2], but these introduce stray ground capacitances C_g that lower plasma-mode frequencies [3, 4, 5, 6] and might limit coherence [7, 8, 9].

Our goal is to engineer fluxonium devices with reduced ground capacitance while quantitatively linking their measured spectra to a microscopic model that captures the full electromagnetic

environment of the qubit.

We developed a dry-etching technique that selectively removes silicon around/in between the fluxonium loop, reducing the ground capacitance C_g by 60–70% and shifting plasma modes to higher frequencies while preserving qubit coherence. In parallel, we implemented a microscopic framework following Nigg et al. [10] and Smith et al. [11], using an ABCD-matrix formulation of the Josephson junction array to compute the linear admittance, identify plasma-mode frequencies, impedances, and zero-point phase fluctuations. These quantities are incorporated into the fluxonium Hamiltonian, yielding quantitative agreement with measured spectra and establishes an approach that can be extended to analyze dissipation mechanisms.

This combined fabrication–modeling approach links microscopic circuit parameters to observable spectra and provides a robust, scalable approach for engineering the electromagnetic environment and optimizing coherence in next-generation superconducting qubits.