The Signal
Single-Ion Error Correction Extends Qubit Lifetime 12.9-Fold
A trapped-ion team reports autonomous quantum error correction that extended a logical qubit’s lifetime to 11.6 milliseconds under injected magnetic noise. The result exceeded the best physical-qubit lifetime without measurement or real-tim
Quantum.world Team · Written with AI · fact-checked ·
A trapped-ion experiment extended a logical qubit’s lifetime to 11.6 milliseconds under deliberately added magnetic noise—about 12.9 times the lifetime of the best physical qubit tested. The team reports the result in a paper published October 23 in Nature Physics, using autonomous error correction that does not rely on repeated measurements or real-time feedback.
One calcium ion holds the logical qubit
The experiment uses two trapped calcium-40 ions. Researchers encode the logical qubit in four internal Zeeman states of one ion, within its D5/2 energy level; the second ion serves as an ancilla for cooling. Rather than distribute the encoded information across several data qubits, the scheme uses the first ion’s multiple internal states to distinguish the protected logical states from states reached when errors occur.
The authors write: “Here we realize autonomous quantum error correction with a logical qubit encoded in multiple internal spin states of a single trapped ion and surpass the break-even point for qubit lifetime.” Yi Li et al., “Beating the break-even point with autonomous quantum error correction,” Nature Physics.
Cooling removes entropy without a measurement
The correction targets first-order dephasing, a phase error caused by magnetic-field fluctuations. Tailored Raman sideband drives move error-related entropy from the ion’s internal state into a shared motional mode. The ancilla ion then helps sympathetically cool that motion, carrying away the entropy and returning the system toward its logical state.
Each correction cycle combines this entropy transfer with cooling and reset. The experiment characterizes the process through quantum process tomography: under ambient conditions, the measured process fidelity is 0.92(2) when the initial state is in the error space and 0.98(1) when it begins in the logical space. Those measurements describe the complete correction cycle, not the lifetime comparison alone.
Injected noise makes the break-even test sharper
In the controlled test, the researchers add low-frequency magnetic noise with a 10-hertz cutoff. Repeated correction extends the logical-qubit lifetime to 11.6 ± 1.9 milliseconds, compared with 0.9 ± 0.1 milliseconds for the best physical-qubit encoding and 0.8 ± 0.1 milliseconds for an encoded qubit without correction. The reported enhancement factor is 12.9 ± 2.6.
The team also tests the system under ambient laboratory conditions. There, the fitted lifetime is 60.9 ± 33.0 milliseconds for the corrected logical qubit and 26.0 ± 5.4 milliseconds for the physical qubit, giving an enhancement factor of 2.3 ± 1.4. The larger uncertainty reflects fluctuating background noise; the deliberately injected-noise test provides the more controlled comparison.
A memory result, not yet a multi-qubit computer
The result establishes a lifetime advantage for a single-ion logical memory under the tested dephasing conditions. It does not demonstrate protection against every error channel: the authors identify higher-order phase errors as outside the code’s present correction capability. Extending the code to more internal states is one possible route to addressing additional error types.
Nor does the experiment show entangling operations between logical qubits. The paper identifies such gates as a next technical requirement for building larger systems from these memories. For now, its measured benchmark is specific: under injected low-frequency magnetic noise, the corrected logical qubit lasts 11.6 ± 1.9 milliseconds, versus 0.9 ± 0.1 milliseconds for the best physical-qubit reference.