arXiv:2609.22572
Modular fault-tolerant quantum computing on a non-CSS code
Robert Freund, Friederike Butt, César Benito, Ivan Pogorelov, Marcel Meyer, Alex Steiner, Alejandro Bermudez, Markus Müller, Thomas Monz
Abstract
Modularization promises to break down the design and implementation complexity of large scale quantum processors into smaller manageable subtasks. In this approach, quantum channels, realized for instance through physical rerouting of qubits or quantum teleportation, connect multiple modules. Each of those modules hosts a subset of qubits, e.g. multiple logical qubits, and provides quantum operations on them. In this work, we implement for the first time all logical operations required for modular fault-tolerant universal quantum computing with a non-Calderbank-Shor-Steane (CSS) code, the perfect $[[5, 1, 3]]$ code, on a trapped-ion quantum computer. This code is the smallest quantum error-correcting (QEC) code capable of correcting any single-qubit error, making it a compact alternative to larger CSS codes. We demonstrate logical state teleportation and a full suite of fault-tolerant operations required for universal logical control, including logical state preparation, QEC with real-time feedback, logical measurements, magic-state preparation, logical entangling operations, and magic-state injection. Moreover, we characterize the logical spectator error picked up by idling logical qubits during quantum operations on distinct qubit registers and demonstrate a logical Pauli quantum process tomography that minimizes required sampling resources for logical tomography.
AI Summary
The paper addresses how to implement universal fault-tolerant logical control and teleportation with a non-CSS code, where transversal CNOT and destructive logical readout are not directly available. It implements a toolbox for the $[[5,1,3]]$ code on a 16-ion trapped-ion processor, using flagged measurements, mid-circuit readout and real-time feedback. Two encoded blocks occupy one ion chain; physically separated modules are not demonstrated. The experiments cover logical Pauli-state preparation, error correction, logical measurement, teleportation, verified magic-state preparation, a pieceably fault-tolerant entangling gate and magic-state injection. Preparation and readout have an average infidelity of $5.24\%$ at $80\%$ acceptance; adding one error-correction cycle gives $26.8\%$ infidelity. Operating an error-correction cycle on one block adds roughly 20 percentage points of spectator-state infidelity. With error-detecting postselection, teleportation reaches $21^{+7}_{-5}\%$ infidelity at $0.035\%$ acceptance. Magic-state preparation reaches $4^{+3}_{-2}\%$ infidelity at $2.3\%$ acceptance; injection has an estimated average gate fidelity of $85\pm8\%$ at $0.023\%$ acceptance. Results are characterized through logical measurements and tomography and compared with Monte Carlo simulations using experimentally informed noise parameters. Single-fault enumeration verifies the magic-state preparation circuit. Reduced-resource logical process tomography uses Pauli-noise assumptions and removes SPAM contributions for error-correcting, but not postselected, operations. Stated limitations include residual physical errors that complicate concatenated-operation characterization, substantial two-qubit-gate and mid-circuit-measurement noise, sequential-operation idling, and very low acceptance for deep protocols; teleportation and injection retain only tens of shots per measurement configuration.
Central claim may not hold: The abstract leaves unstated that the principal multi-block demonstrations are conditional error-detecting protocols, with acceptance rates of 0.035% for teleportation and 0.023% for magic-state injection, rather than implementations of the full error-correcting protocols. With the paper’s logical Pauli definitions, its stated transversal phase-gate prescription implements logical S† rather than S, and its example Y_2Y_3X_5 represents −X_L rather than X_L. The printed Pauli-transfer-matrix definition omits the factor 1/2^N required for the subsequent composition and fidelity formulas. The same reported 85 ± 8% is inconsistently identified as process fidelity and average gate fidelity, which are distinct quantities.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 24 pages, 16 figures