Viewing protected superconducting qubits through the lens of the cat qubit

Simon Lieu, Emma L. Rosenfeld, Kyungjoo Noh, Connor T. Hann · 2025

Read Viewing protected superconducting qubits through the lens of the cat qubit by Simon Lieu, Emma L. Rosenfeld, Kyungjoo Noh, Connor T. Hann (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-16425.txt.

Introduction

Simon Lieu, Emma Rosenfeld, Kyungjoo Noh, and Connor Hann view protected qubits through the cat qubit lens in two thousand twenty-five. Protected superconducting qubits resist noise by design. Bosonic cat qubits guard one error type at the cost of the other. The pair draw close parallels between the two families.

They study the fluxonium circuit in its Fock basis at its sweet spot. A phase map shows a zone with symmetry broken ground states. In the heavy limit those states look like squeezed coherent states. Noise rates split as in cat qubits: one falls off fast, the other does not worsen. Bias safe gates and a route to a passive memory follow.

Catalog record

CREATORSimon Lieu, Emma L. Rosenfeld, Kyungjoo Noh, Connor T. Hann
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE58 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-16425

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Viewing protected superconducting qubits through the lens of the cat qubit Simon Lieu, Emma L. Rosenfeld, Kyungjoo Noh, Connor T. Hann Abstract We draw analogies between protected superconducting qubits and bosonic qubits by studying the fluxonium Hamiltonian in its Fock basis. The mean-field phase diagram of fluxonium (at the sweet spot) is identified, with a region in parameter space that is characterized by ℤ 2 \mathbb{Z}_{2} -symmetry-broken ground states. In the heavy fluxonium limit, these ground states are well approximated by squeezed coherent states in a Fock basis (corresponding to persistent current states with definite flux but indefinite charge), and simple expressions are provided for them in terms of the circuit parameters. We study the noise bias in fluxonium via a universal Lindblad master equation and find that the bit-flip rate is exponentially small in E j / ( k B ​ T ) E_{j}/(k_{B}T) , while the phase-flip rate does not get worse with this ratio. Analogous behavior is found in cos ⁡ ( 2 ​ θ ) \cos(2 \theta) qubits. We describe cat-qubit-inspired bias-preserving X X and C ​ X CX gates for fluxonium. We discuss first steps towards generating an Ising interaction between protected superconducting qubits on a two-dimensional lattice, with the aim of achieving a passive quantum memory by coupling a static Hamiltonian to a generic thermal bath. I Introduction Quantum error correction relies on redundantly encoding quantum information into a large Hilbert space such that physical errors do not directly cause logical errors [ 1 , 2 , 3 , 4 ] . One popular approach is called “hardware efficient bosonic encoding” where information is encoded into the infinite-dimensional Hilbert space of a harmonic oscillator [ 5 ] , including qubits such as the GKP code [ 6 ] , binomial code [ 7 ] , and cat code [ 8 ] . For certain codes, such encodings are advantageous because there is notion of distance in “phase space”: codewords are well separated in phase space while errors only perturb locally. Thus a certain level of protection can be achieved at the level of the physical hardware before applying ideas from active error correction. Notably, certain bosonic qubits (e.g. the cat code) can exponentially suppress the bit-flip error rate by only linearly increasing separation in phase space. The cost is that such schemes typically require external drives and/or engineered dissipation to confine a harmonic oscillator into a bosonic code space, which can lead to complexities due to drive-induced resonances [ 9 , 10 ] and nonequilibrium dynamics that can be difficult to model. On the other hand, it is known that certain superconducting “protected qubits” can also exhibit hardware-level protection against noise, e.g. an exponentially-long T_{1} time (i.e. bit-flip time) as a function of circuit parameters [ 11 ] . Notably, this protection is achieved without an external drive or engineered dissipation. Canonical examples include: the fluxonium qubit [ 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 ] , zero-pi qubit [ 20 , 21 , 22 , 23 , 24 , 25 , 26 ] , and \cos(2\theta) qubit [ 27 , 28 , 29 , 30 , 31 , 32 , 33 ] . The standard explanation for T_{1} protection is that most errors are caused due to environmental noise that couples to the charge or flux operator of the circuit, and the relevant transition matrix elements between logical codewords (i.e. \langle\bar{0}|n|\bar{1}\rangle and \langle\bar{0}|\phi|\bar{1}\rangle ) are exponentially suppressed in a ratio of energy scales (e.g. the Josephson energy over capacitive energy: E_{j}/E_{c} in fluxonium). In this work we would like to explain the protection mechanism of a superconducting qubit using the language of bosonic encodings: The infinite-dimensional Hilbert space of a circuit hosts ground states that are nearly degenerate and far away from each other in phase space. For concreteness, we will describe fluxonium in a language that draws an analogy with the cat code, emphasizing similarities and differences. Figure 1: The limit of a robust classical bit for fluxonium (infinitesimally-thin insulating barrier): A superconducting ring (black) with half of a flux quantum threaded through it (green) will develop a persistent current (red) that either flows in the clockwise or counterclockwise direction since the total flux through a superconducting ring must be quantized in units of \Phi_{0} . This is an example of \mathbb{Z}_{2} spontaneous symmetry breaking that protects a classical bit in a “zero-dimensional” system. Note that fluxonium does not require an external drive or engineered dissipation, and has finite energy even in this classical limit (in contrast to the cat code). The cat code is the canonical example of a bosonic code that features passive bit-flip noise suppression: Information is encoded into two macroscopic coherent states: |\pm\alpha\rangle , which can be stabilized, e.g., via a two-photon drive ( \lambda ) and engineered two-photon loss ( \kappa_{2} ) [ 8 ] . In the limit of large |\alpha|^{2}\sim\lambda/\kappa_{2} , the codewords |\pm\alpha\rangle are well separated in phase space, while the typical error mechanisms of single-photon loss and dephasing ( L\sim a,a^{\dagger}a ) only perturb states locally [ 34 ] . Physically: A cavity that is strongly driven with pairs of photons (and loses pairs of photons) will stabilize to one of two classical states of light that are \pi out of phase; the classical coherent states break photon parity: \exp[i\pi a^{\dagger}a]|\alpha\rangle=|-\alpha\rangle . A robust classical bit is associated with a phase in parameter space characterized by symmetry-broken steady states [ 35 ] . In this work we show that fluxonium has nearly-degenerate ground states that spontaneously break flux parity, corresponding to squeezed states in the infinite-dimensional Hilbert space of its LC subsystem that have nearly-definite flux. Thermal processes will thus evolve the system towards a nontrivial bosonic state at low temperature. The separation of the two codewords (ground states) in phase space will be mostly controlled by the ratio of two circuit parameters: |\alpha^{\prime}|^{2}\sim\sqrt{E_{j}/E_{c}} . Physically these “classical” states in fluxonium correspond to persistent current states that circulate clockwise or counterclockwise in the circuit. Scaling |\alpha^{\prime}|^{2}\rightarrow\infty represents the limit where the physical Josephson barrier width (the insulator between two superconductors) goes to zero ( E_{j}\rightarrow\infty ). In the extreme limit of no physical barrier, the model represents a superconducting ring with half of a flux quantum ( \Phi_{0}/2 ) threaded through it. Since the total flux through the loop must be quantized in units of \Phi_{0} , the superconducting ring forms a current in the clockwise or counterclockwise direction to induce an additional flux of \pm\Phi_{0}/2 . [See Fig. 1 .] Currents in a superconducting ring have been observed to persist without decay on a timescale on the order of years [ 36 , 37 , 38 , 39 ] , suggesting that the upper bound on the classical-bit lifetime could be very long in practice. The current in this classical limit is finite , only costing a finite energy (unlike the cat code which requires an increasing number of photons, i.e. large electric field). Fluxonium thus provides a limit where a single bosonic mode can reliably store a classical bit at finite energy. Beyond protected superconducting qubits, the field of passive error correction more

Read the complete text