Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 78 of the 88 references it cites were found in a registry or an index; 75 of the 87 sources its claims cite were checked for support.

Hardware-efficient quantum error correction via concatenated bosonic qubits

Open access
Putterman, Harald, Noh, Kyungjoo, Hann, Connor T., MacCabe, Gregory S., Aghaeimeibodi, Shahriar, Patel, Rishi N., Lee, Menyoung, Jones, William M., Moradinejad, Hesam, Rodriguez, Roberto, Mahuli, Neha, Rose, Jefferson, Owens, John Clai, Levine, Harry, Rosenfeld, Emma, Reinhold, Philip, Moncelsi, Lorenzo, Alcid, Joshua Ari, Alidoust, Nasser, Arrangoiz-Arriola, Patricio, Barnett, James, Bienias, Przemyslaw, Carson, Hugh A., Chen, Cliff, Chen, Li, Chinkezian, Harutiun, Chisholm, Eric M., Chou, Ming-Han, Clerk, Aashish, Clifford, Andrew, Cosmic, R., Curiel, Ana Valdes, Davis, Erik, DeLorenzo, Laura, D’Ewart, J. Mitchell, Diky, Art, D’Souza, Nathan, Dumitrescu, Philipp T., Eisenmann, Shmuel, Elkhouly, Essam, Evenbly, Glen, Fang, Michael T., Fang, Yawen, Fling, Matthew J., Fon, Warren, Garcia, Gabriel, Gorshkov, Alexey V., Grant, Julia A., Gray, Mason J., Grimberg, Sebastian, Grimsmo, Arne L., Haim, Arbel, Hand, Justin, He, Yuan, Hernandez, Mike, Hover, David, Hung, Jimmy S. C., Hunt, Matthew, Iverson, Joe, Jarrige, Ignace, Jaskula, Jean-Christophe, Jiang, Liang, Kalaee, Mahmoud, Karabalin, Rassul, Karalekas, Peter J., Keller, Andrew J., Khalajhedayati, Amirhossein, Kubica, Aleksander, Lee, Hanho, Leroux, Catherine, Lieu, Simon, Ly, Victor, Madrigal, Keven Villegas, Marcaud, Guillaume, McCabe, Gavin, Miles, Cody, Milsted, Ashley, Minguzzi, Joaquin, Mishra, Anurag, Mukherjee, Biswaroop, Naghiloo, Mahdi, Oblepias, Eric, Ortuno, Gerson, Pagdilao, Jason, Pancotti, Nicola, Panduro, Ashley, Paquette, JP, Park, Minje, Peairs, Gregory A., Perello, David, Peterson, Eric C., Ponte, Sophia, Preskill, John, Qiao, Johnson, Refael, Gil, Resnick, Rachel, Retzker, Alex, Reyna, Omar A., Runyan, Marc, Ryan, Colm A., Sahmoud, Abdulrahman, Sanchez, Ernesto, Sanil, Rohan, Sankar, Krishanu, Sato, Yuki, Scaffidi, Thomas, Siavoshi, Salome, Sivarajah, Prasahnt, Skogland, Trenton, Su, Chun-Ju, Swenson, Loren J., Teo, Stephanie M., Tomada, Astrid, Torlai, Giacomo, Wollack, E. Alex, Ye, Yufeng, Zerrudo, Jessica A., Zhang, Kailing, Brandão, Fernando G. S. L., Matheny, Matthew H., Painter, Oskar · 2025 · Nature

Abstract

To solve problems of practical importance1,2, quantum computers probably need to incorporate quantum error correction, in which a logical qubit is redundantly encoded in many noisy physical qubits3–5. The large physical-qubit overhead associated with error correction motivates the search for more hardware-efficient approaches6–18. Here, using a superconducting quantum circuit19, we realize a logical qubit memory formed from the concatenation of encoded bosonic cat qubits with an outer repetition code of distance d = 5 (ref. 10). A stabilizing circuit passively protects cat qubits against bit flips20–24. The repetition code, using ancilla transmons for syndrome measurement, corrects cat qubit phase flips. We study the performance and scaling of the logical qubit memory, finding that the phase-flip correcting repetition code operates below the threshold. The logical bit-flip error is suppressed with increasing cat qubit mean photon number, enabled by our realization of a cat-transmon noise-biased CX gate. The minimum measured logical error per cycle is on average 1.75(2)% for the distance-3 code sections, and 1.65(3)% for the distance-5 code. Despite the increased number of fault locations of the distance-5 code, the high degree of noise bias preserved during error correction enables comparable performance. These results, where the intrinsic error suppression of the bosonic encodings enables us to use a hardware-efficient outer error-correcting code, indicate that concatenated bosonic codes can be a compelling model for reaching fault-tolerant quantum computation. Bosonic qubits can be engineered to feature intrinsic protection against certain kinds of errors, which makes quantum error correction across many bosonic qubits possible with less overhead.

Overview

CoChat Summary

TL;DR

This paper demonstrates a superconducting logical qubit memory built from concatenated bosonic cat qubits plus an outer distance-5 repetition code, using noise-biased CX gates for syndrome extraction. The best measured logical error per cycle is 1.65 ± 0.03% for the distance-5 section and 1.75 ± 0.02% on average for the distance-3 sections, showing hardware-efficient error correction with preserved noise bias.

Key findings

  • The device consists of five bosonic storage modes and four ancilla transmons; the storage modes have average \(T_1 > 60\,\mu s\) and average \(T_2 > 80\,\mu s\).
  • At \(|\alpha|^2 = 2\), the cat qubits achieve greater than 1 ms bit-flip times and 27–33 \(\mu s\) phase-flip times, corresponding to a noise bias greater than 30.
  • The best-measured overall logical error per cycle for the distance-5 section is \(1.65 \pm 0.03\%\) at \(|\alpha|^2 = 1.5\); the distance-3 sections reach \(1.83 \pm 0.03\%\) and \(1.67 \pm 0.04\%\) at \(|\alpha|^2 = 1\), with average \(1.75 \pm 0.02\%\).
  • With erasure information incorporated, the logical phase-flip rate is reduced by about 20% at \(|\alpha|^2 = 1.5\) for \(d=5\), and syndrome-measurement error probability is reduced by over a factor of two at \(|\alpha|^2 = 1\).

Methods

  • Built a microfabricated superconducting circuit implementing a distance-5 repetition cat code: five cat-qubit data modes stabilized by two-photon dissipation and four transmon ancillas for stabilizer measurement.
  • Used a noise-biased CX gate based on tunable dispersive coupling and \(\chi\)-matching, with ancilla states encoded as \(|0_a\rangle = |g\rangle\) and \(|1_a\rangle = |f\rangle\), to measure repetition-code syndromes.
  • Quantified performance by repeated syndrome-extraction cycles, decoding with minimum-weight perfect matching, and extracting logical bit-flip and phase-flip rates from exponential fits to logical-operator decay.

Limitations

  • The paper states that a repetition code of biased-noise qubits does not possess a proper asymptotic threshold because physical bit-flip errors are uncorrectable and set a lower bound on logical error.
  • The logical performance is currently dominated by intrinsic cat bit-flip and phase-flip errors, and ancilla-induced bit-flip errors remain a concern, though not the present limiting factor.
  • Evident from the text: the reported code is a logical memory demonstration rather than a full fault-tolerant computer, and several projected improvements are described as idealized or contingent on better coherence times and gate performance.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41586-025-08642-7

Cites

78 of 78 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Repetition code of bosonic qubits. a, Schematic of the repetition code device. Data qubits S1, …, S5 (blue) are encoded into the Hilbert space of a quantum harmonic oscillator. Each cat qubit is stabilized by a buffer mode B1, …, B5 (green). The ancilla qubits A1, …, A4 (orange) are transmon qubits that detect Z errors (see example with red arrows) on the data qubits by measuring repetition-code stabilizers. b, Circuit layout of the repetition code device using a flip-chip architecture (Supplementary Information). The five bosonic modes (Si) are coplanar waveguide resonators. Each resonator is connected to a buffer mode (Bi). Buffer modes are damped through a multi-pole filter. Ancilla transmons (Ai) are connected to the storage modes by tunable couplers (Ci,j). Scale bar, 1 mm. c,d, Magnified circuit sections showing a storage-buffer subsystem (c) and an ancilla transmon coupled to its neighbouring storage modes by tunable couplers (d). e, Cat qubit encoding in a bosonic mode with experimentally measured Wigner functions of the four basis states of a cat qubit and arrows representing X and Z errors. f, Bit-flip and phase-flip times of the five cat qubits in our device under simultaneous two-photon dissipation. Error bars (standard error) incorporate sampling and fit error.
Fig. 1 Repetition code of bosonic qubits. a, Schematic of the repetition code device. Data qubits S1, …, S5 (blue) are encoded into the Hilbert space of a quantum harmonic oscillator. Each cat qubit is stabilized by a buffer mode B1, …, B5 (green). The ancilla qubits A1, …, A4 (orange) are transmon qubits that detect Z errors (see example with red arrows) on the data qubits by measuring repetition-code stabilizers. b, Circuit layout of the repetition code device using a flip-chip architecture (Supplementary Information). The five bosonic modes (Si) are coplanar waveguide resonators. Each resonator is connected to a buffer mode (Bi). Buffer modes are damped through a multi-pole filter. Ancilla transmons (Ai) are connected to the storage modes by tunable couplers (Ci,j). Scale bar, 1 mm. c,d, Magnified circuit sections showing a storage-buffer subsystem (c) and an ancilla transmon coupled to its neighbouring storage modes by tunable couplers (d). e, Cat qubit encoding in a bosonic mode with experimentally measured Wigner functions of the four basis states of a cat qubit and arrows representing X and Z errors. f, Bit-flip and phase-flip times of the five cat qubits in our device under simultaneous two-photon dissipation. Error bars (standard error) incorporate sampling and fit error.CC BY-NC-ND · PubMed Central
Fig. 2 Noise-biased CX gate between a transmon and a cat qubit. a, CX gate sequence. At the start of the sequence, the ancilla is initialized and cat qubit stabilization (blue arrows) is on. The stabilization is then turned off and the CX gate is applied between the ancilla and cat qubit. After the CX gate, the stabilization is turned back on and the ancilla qubit is read out and reset to |g⟩. Through the experimentally measured Wigner functions, we show the evolution of the storage state during the sequence for each of the ancilla states |g⟩, |e⟩ and |f⟩, with an initial storage-mode state |α⟩. b, Storage-mode Wigner tomograms before and after 10 applications of the CX gate sequence with the ancilla in |g⟩ and storage initialized in |α⟩. The sequence is applied with and without stabilization. c, Characterization of the CX gate. We apply repeated CX2 cycles (see main text) with a cycle duration of 3 μs and plot the measured bit-flip time of the cat qubit as a function of cat qubit photon number, ∣α∣2, for different ancilla states. The inset shows the measured phase-flip times (points) and a fit (line) when the ancilla is initialized to |g⟩ + |f⟩. Error bars (standard error) incorporate sampling and fit errors. a and b use the same colour bar.
Fig. 2 Noise-biased CX gate between a transmon and a cat qubit. a, CX gate sequence. At the start of the sequence, the ancilla is initialized and cat qubit stabilization (blue arrows) is on. The stabilization is then turned off and the CX gate is applied between the ancilla and cat qubit. After the CX gate, the stabilization is turned back on and the ancilla qubit is read out and reset to |g⟩. Through the experimentally measured Wigner functions, we show the evolution of the storage state during the sequence for each of the ancilla states |g⟩, |e⟩ and |f⟩, with an initial storage-mode state |α⟩. b, Storage-mode Wigner tomograms before and after 10 applications of the CX gate sequence with the ancilla in |g⟩ and storage initialized in |α⟩. The sequence is applied with and without stabilization. c, Characterization of the CX gate. We apply repeated CX2 cycles (see main text) with a cycle duration of 3 μs and plot the measured bit-flip time of the cat qubit as a function of cat qubit photon number, ∣α∣2, for different ancilla states. The inset shows the measured phase-flip times (points) and a fit (line) when the ancilla is initialized to |g⟩ + |f⟩. Error bars (standard error) incorporate sampling and fit errors. a and b use the same colour bar.CC BY-NC-ND · PubMed Central
Fig. 3 Detecting and correcting phase-flip errors with the repetition code. a, Error correction circuit, showing repeated error correction cycles with a duration of 2.8 μs. b, Detection probabilities for the measured stabilizers versus error correction cycle for three different cat qubit photon numbers, ∣α∣2 = 1, 2 and 3. Bold traces correspond to the average over the individual stabilizer traces. Error bars represent the standard error of the mean. c, Depiction of erasures occurring in the repetition code experiment due to ancilla decay from |f⟩ to |e⟩. We show shots of the experiment from a representative stabilizer, in which |f⟩ is light grey, |g⟩ is dark grey and the erasure state |e⟩ is shown in red. d, Effective syndrome measurement error extracted from the QEC graph before and after accounting for the erasure for ∣α∣2 = 1. Bold traces correspond to averaging over the stabilizers. e, Example fits of the decay of the X logical operator for the distance-5 repetition code for different photon numbers. f, Error corrected logical phase-flip probability per cycle (ϵL,phase-flip) and logical X lifetime (TX) as a function of ∣α∣2 for the different repetition code sections. Data and fits are shown with (squares and solid curves) and without (circles and dashed curves) inclusion of erasure information. The fits are to the power law (|α|2)γ for ∣α∣2 ≥ 1.5. Faded points indicate fits binned by the number of even cat states |+⟩ in the initial state and serve to indicate the spread from asymmetric error rates at low photon numbers (Supplementary Information). The dotted purple curve shows the simulated logical phase-flip probability for the distance-5 section. Error bars (standard error) incorporate sampling and fit errors.
Fig. 3 Detecting and correcting phase-flip errors with the repetition code. a, Error correction circuit, showing repeated error correction cycles with a duration of 2.8 μs. b, Detection probabilities for the measured stabilizers versus error correction cycle for three different cat qubit photon numbers, ∣α∣2 = 1, 2 and 3. Bold traces correspond to the average over the individual stabilizer traces. Error bars represent the standard error of the mean. c, Depiction of erasures occurring in the repetition code experiment due to ancilla decay from |f⟩ to |e⟩. We show shots of the experiment from a representative stabilizer, in which |f⟩ is light grey, |g⟩ is dark grey and the erasure state |e⟩ is shown in red. d, Effective syndrome measurement error extracted from the QEC graph before and after accounting for the erasure for ∣α∣2 = 1. Bold traces correspond to averaging over the stabilizers. e, Example fits of the decay of the X logical operator for the distance-5 repetition code for different photon numbers. f, Error corrected logical phase-flip probability per cycle (ϵL,phase-flip) and logical X lifetime (TX) as a function of ∣α∣2 for the different repetition code sections. Data and fits are shown with (squares and solid curves) and without (circles and dashed curves) inclusion of erasure information. The fits are to the power law (|α|2)γ for ∣α∣2 ≥ 1.5. Faded points indicate fits binned by the number of even cat states |+⟩ in the initial state and serve to indicate the spread from asymmetric error rates at low photon numbers (Supplementary Information). The dotted purple curve shows the simulated logical phase-flip probability for the distance-5 section. Error bars (standard error) incorporate sampling and fit errors.CC BY-NC-ND · PubMed Central
Fig. 4 Characterizing logical bit-flip error rates. a, Fitted decay curve of the logical Z operator for several storage mean photon numbers for the distance-5 code. Error bars capture sampling error. b, Logical bit-flip probability per cycle (ϵL,bit-flip) and logical Z lifetime (TZ) as a function of ∣α∣2 for the two distance-3 repetition-code sections and the distance-5 section. Solid lines correspond to data and dotted lines correspond to a phenomenological model. Error bars (standard error), capturing sampling variance and fit uncertainty, are smaller than the markers.
Fig. 4 Characterizing logical bit-flip error rates. a, Fitted decay curve of the logical Z operator for several storage mean photon numbers for the distance-5 code. Error bars capture sampling error. b, Logical bit-flip probability per cycle (ϵL,bit-flip) and logical Z lifetime (TZ) as a function of ∣α∣2 for the two distance-3 repetition-code sections and the distance-5 section. Solid lines correspond to data and dotted lines correspond to a phenomenological model. Error bars (standard error), capturing sampling variance and fit uncertainty, are smaller than the markers.CC BY-NC-ND · PubMed Central
Fig. 5 Logical qubit memory performance. a, Overall logical error per cycle of the repetition cat code, ϵL = (ϵL,bit-flip + ϵL,phase-flip)/2, versus cat qubit mean photon number ∣α∣2. As in Fig. 3f, the faded points correspond to fits to groupings of the data by the number of even cat states in the initial state. The lines are guides to the eye, computed by interpolating both the logical phase-flip and bit-flip probabilities, ϵL,phase-flip and ϵL,bit-flip, respectively. b, Error budget for the distance-5 repetition code using erasure information. Different shades of colour correspond to the different per cat (or per CX gate) contribution to the error budget. Error bars (standard error) incorporate sampling and fit errors.
Fig. 5 Logical qubit memory performance. a, Overall logical error per cycle of the repetition cat code, ϵL = (ϵL,bit-flip + ϵL,phase-flip)/2, versus cat qubit mean photon number ∣α∣2. As in Fig. 3f, the faded points correspond to fits to groupings of the data by the number of even cat states in the initial state. The lines are guides to the eye, computed by interpolating both the logical phase-flip and bit-flip probabilities, ϵL,phase-flip and ϵL,bit-flip, respectively. b, Error budget for the distance-5 repetition code using erasure information. Different shades of colour correspond to the different per cat (or per CX gate) contribution to the error budget. Error bars (standard error) incorporate sampling and fit errors.CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

19 Valid · 3 Not valid · 1 Not checked · 69 Own results and interpretations

55 more claims · 11 with their cited sources checked

Create a free account to see all 92 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 9
  • It may state two separate claims as one. 25

Create a free account to see these claims.

Citations

Cites

Show all 88 references

Cited by

See all 108

Metadata

Authors
Putterman, H., Noh, K., Hann, C. T., MacCabe, G. S., Aghaeimeibodi, S., Patel, R. N., Lee, M., Jones, W. M., Moradinejad, H., Rodriguez, R., Mahuli, N., Rose, J., Owens, J. C., Levine, H., Rosenfeld, E., Reinhold, P., Moncelsi, L., Alcid, J. A., Alidoust, N., Arrangoiz-Arriola, P., Barnett, J., Bienias, P., Carson, H. A., Chen, C., Chen, L., Chinkezian, H., Chisholm, E. M., Chou, M.-H., Clerk, A., Clifford, A., Cosmic, R., Curiel, A. V., Davis, E., DeLorenzo, L., D’Ewart, J. M., Diky, A., D’Souza, N., Dumitrescu, P. T., Eisenmann, S., Elkhouly, E., Evenbly, G., Fang, M. T., Fang, Y., Fling, M. J., Fon, W., Garcia, G., Gorshkov, A. V., Grant, J. A., Gray, M. J., Grimberg, S., Grimsmo, A. L., Haim, A., Hand, J., He, Y., Hernandez, M., Hover, D., Hung, J. S. C., Hunt, M., Iverson, J., Jarrige, I., Jaskula, J.-C., Jiang, L., Kalaee, M., Karabalin, R., Karalekas, P. J., Keller, A. J., Khalajhedayati, A., Kubica, A., Lee, H., Leroux, C., Lieu, S., Ly, V., Madrigal, K. V., Marcaud, G., McCabe, G., Miles, C., Milsted, A., Minguzzi, J., Mishra, A., Mukherjee, B., Naghiloo, M., Oblepias, E., Ortuno, G., Pagdilao, J., Pancotti, N., Panduro, A., Paquette, J., Park, M., Peairs, G. A., Perello, D., Peterson, E. C., Ponte, S., Preskill, J., Qiao, J., Refael, G., Resnick, R., Retzker, A., Reyna, O. A., Runyan, M., Ryan, C. A., Sahmoud, A., Sanchez, E., Sanil, R., Sankar, K., Sato, Y., Scaffidi, T., Siavoshi, S., Sivarajah, P., Skogland, T., Su, C.-J., Swenson, L. J., Teo, S. M., Tomada, A., Torlai, G., Wollack, E. A., Ye, Y., Zerrudo, J. A., Zhang, K., Brandão, F. G. S. L., Matheny, M. H., & Painter, O.
Affiliations
  • Amazon (United States) (US)
  • Google (United States) (US)
  • University of Chicago (US)
  • Yale University (US)
  • California Institute of Technology (US)
  • Hebrew University of Jerusalem (IL)
  • University of California, Irvine (US)
Corresponding author
  • Putterman, Harald
  • Painter, Oskar
Venue
Nature
Published
2025-02-26
Volume
638
Issue
8052
Pages
927-934
ISSN
  • 0028-0836 (ISSN-L)
  • 1476-4687
OpenAlex
W4407957966
Cited by
108
References
88
Citations by year
2026: 51 · 2025: 57
FWCI
42.98
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
  • arXiv (Cornell University)
Funders
California Institute of Technology
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Quantum Computing Algorithms and Architecture
  • Physical Sciences › Physics and Astronomy › Atomic and Molecular Physics, and Optics › Quantum Information and Cryptography
  • Physical Sciences › Physics and Astronomy › Atomic and Molecular Physics, and Optics › Quantum Mechanics and Applications
Keywords
quantum error correction, cat qubits, repetition codes, superconducting qubits, transmon qubits, fault-tolerant quantum computing, hardware overhead

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-09-30, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-09-30 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Hardware-efficient quantum error correction via concatenated bosonic qubits. https://app.cochat.ai/papers/10.1038/s41586-025-08642-7
The paper
Putterman, H., Noh, K., Hann, C. T., MacCabe, G. S., Aghaeimeibodi, S., Patel, R. N., Lee, M., Jones, W. M., Moradinejad, H., Rodriguez, R., Mahuli, N., Rose, J., Owens, J. C., Levine, H., Rosenfeld, E., Reinhold, P., Moncelsi, L., Alcid, J. A., Alidoust, N., . . . Painter, O. (2025). Hardware-efficient quantum error correction via concatenated bosonic qubits. Nature. https://doi.org/10.1038/s41586-025-08642-7

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.