The ytterbium-171 ion hyperfine qubit

Short answer: The ytterbium-171 ion has a hyperfine qubit that is insensitive to first-order magnetic field changes, which makes it a long-lived qubit. A 2007 Physical Review A paper, with an arXiv version, reported preparation, detection and a 2.5 second coherence time, and a 2021 Nature Communications paper reports a coherence time of around 5,500 seconds. These are single-ion laboratory experiments, and the long coherence time is a quantum memory result and not a computing result.

2007: ytterbium as a qubit

The Physical Review A abstract says the authors demonstrate the use of trapped ytterbium ions as quantum bits for quantum information processing. They implement fast, efficient state preparation and detection of the first-order magnetic field-insensitive hyperfine levels of ytterbium-171, with a measured coherence time of 2.5 seconds. They say the high efficiency and fidelity result from stabilizing and frequency-modulating the relevant laser sources. The same abstract appears in an arXiv version posted in August 2007, which is not independent evidence.

2021: from 660 seconds to about 5,500 seconds

The Nature Communications abstract describes realizing a long-coherence quantum memory as a major challenge. It says the longest coherence time of a single qubit had been reported as 660 seconds, in a single ytterbium-171 ion, through sympathetic cooling and dynamical decoupling pulses that addressed heating-induced detection inefficiency and magnetic field fluctuations. It was not clear what prevented further gains. The authors identify and suppress the limiting factors, which they list as remaining magnetic field fluctuations, frequency instability and leakage of the microwave reference oscillator. They observe a coherence time of around 5,500 seconds, which they define as the time constant of an exponential decay fit to measurements up to 960 seconds. That means the figure is extrapolated from a fit and was not measured over the full 5,500 seconds.

They also study the decoherence process of the memory with quantum process tomography and analyze it using recently developed resource theories of quantum memory and coherence. They say their demonstration will accelerate practical applications of quantum memories, particularly in the noisy intermediate-scale regime. This is their expectation.

How to read the two numbers

The jump from 2.5 seconds to around 5,500 seconds spans about fourteen years and a change of experimental method. The 2007 work concentrated on state preparation, detection and laser stabilization, while the 2021 work concentrated on finding and suppressing noise sources that limit memory time. The 2021 abstract also notes that the previous longest single-qubit coherence time was 660 seconds, so the new figure is a large step beyond that earlier record as the authors present it.

Limits

The two experiments differ in what they measure and by how much. The 2007 value of 2.5 seconds was a first demonstration, while the 2021 value is a fit-based time constant from a long, carefully shielded setup. Coherence time is one property of a qubit, and neither abstract reports a two-qubit gate rate or error correction. The coherence of a memory in a quiet apparatus does not describe the same ion during computation.

FAQ

Why is ytterbium-171 a good qubit?

Its hyperfine clock levels are first-order insensitive to magnetic field changes, per the 2007 abstract.

How long did the qubit stay coherent?

2.5 seconds in 2007, and around 5,500 seconds in 2021, as reported by the authors.

Was 5,500 seconds measured directly?

No. It is the time constant of an exponential fit to measurements up to 960 seconds.

What limited earlier work?

Magnetic field fluctuations, frequency instability and microwave oscillator leakage, per the 2021 authors.

Is this useful for computing?

It shows a very long memory. Computing also needs gates, which the abstracts do not report.