Bosonic cat qubits store information in superpositions of coherent states of a microwave cavity, and their error structure is radically lopsided: bit flips are suppressed exponentially in the mean photon number, while phase flips grow only linearly. At a mean photon number of eight the bit-flip suppression factor relative to the bare oscillator scale is of order 9e6, deep into territory where bit flips effectively vanish from the budget, leaving a nearly pure phase-flip channel that a simple repetition code, rather than a full two-dimensional surface code, can correct. This paper budgets the resulting overhead trade. The repetition code that remains is one-dimensional, its patches are ribbons rather than squares, and published analyses credit the architecture with order-of-magnitude savings in physical qubits per logical qubit at algorithmic quality. The paper prices the phase-flip side honestly: the linear growth of phase flips with photon number sets an optimum cat size, and the operations that preserve the bias are a restricted set whose restrictions reappear as compilation and factory costs elsewhere in the machine.
Abstract Realizing the potential of quantum computing requires sufficiently low logical error rates 1 . Many applications call for error rates as low as 10 −15 (refs. 2–9 ), but state-of-the-art quantum platforms typically have physical error rates near 10 −3 (refs. 10–14 ). Quantum error correction 15–17 promises to bridge this divide by distributing quantum logical information across many physical qubits in such a way that errors can be detected and corrected. Errors on the encoded logical qubit state can be exponentially suppressed as the number of physical qubits grows, provided that the physical error rates are below a certain threshold and stable over the course of a computation. Here we implement one-dimensional repetition codes embedded in a two-dimensional grid of superconducting qubits that demonstrate exponential suppression of bit-flip or phase-flip errors, reducing logical error per round more than 100-fold when increasing the number of qubits from 5 to 21. Crucially, this error suppression is stable over 50 rounds of error correction. We also introduce a method for analysing error correlations with high precision, allowing us to characterize error locality while performing quantum error correction. Finally, we perform error detection with a small logical qubit using the 2D surface code on the same device 18,19 and show that the results from both one- and two-dimensional codes agree with numerical simulations that use a simple depolarizing error model. These experimental demonstrations provide a foundation for building a scalable fault-tolerant quantum computer with superconducting qubits.
Noise-biased qubits have bit-flip errors that are exponentially suppressed relative to phase-flip errors, and offer a promising route toward fault-tolerant quantum computing. However, this bias can be compromised during gate operations with non-biased control qubits. To address this limitation, we propose a "0-n" Fock qubit architecture that maintains the noise bias by encoding information in the ground state and n-th excited state of a nonlinear multi-level system, such as a transmon. This encoding is achieved via a dissipative stabilization that acts as decay and gain for lower and upper intermediate levels, respectively. We first analytically demonstrate that bit-flip errors are exponentially suppressed with the number of levels. Then, we present a practical implementation using a multi-mode lossy filter to achieve the frequency-selective dissipation. Finally, we numerically demonstrate that bit-flip probabilities approaching
10−8 are achievable for controlled-X gates on cat qubits using the 0-n qubit as an ancilla with
n≥9 (i.e. ten or more levels), for realistic experimental parameters. Building on this, we simulate syndrome extraction in a repetition code, achieving logical error rates in the megaquop regime with only a distance of
d=9.