A research team from Quantum Elements, the University of Southern California's Center for Quantum Information Science & Technology, IBM, and RWTH Aachen University has made a significant breakthrough in quantum computing by achieving high-fidelity entangled logical qubits, providing crucial support for the development of fault-tolerant quantum computing. This achievement has been published in the journal *Nature Communications*. The research utilized IBM's 127-qubit superconducting processor to develop a hybrid protocol combining quantum error detection with normalizer dynamic decoupling techniques, effectively addressing challenges such as logical error accumulation in quantum scaling.

The core of the technical breakthrough lies in the application of normalizer dynamic decoupling, which repurposes the normalizer elements of a quantum code as dynamic decoupling pulses. Unlike traditional physical-level decoupling, this method operates at the logical level, suppressing both logical and physical error channels simultaneously through logical-level pulses, thereby enhancing system efficiency. This hybrid strategy utilizes a fixed logical pulse generator to improve code performance without increasing code distance or requiring additional qubits, demonstrating hardware efficiency.
Experimental results show that the fidelity of the logical Bell state significantly surpasses that of unencoded physical pairs, reaching a performance level described as "beyond breakeven." Specific metrics include: a peak post-selected fidelity of 98% for encoded Bell state preparation, representing a substantial improvement over previous transmon-based benchmarks; within a 55-microsecond time window, the average logical Bell state fidelity is maintained between 91% and 94%; the protocol also successfully suppresses logical Z errors caused by crosstalk, which were identified as a primary source in the 127-qubit transmon architecture. These results provide new pathways for quantum computing error control and advance the field further.









