Seconds matter when you zoom in far enough. At Delft University of Technology, researchers watched a single atomic nucleus switch its spin live, not as an average over minutes but in the moment, using a scanning tunneling microscope and clever electron whispering.
The team reports a stable nuclear spin that lingers for about five seconds, a small eternity in quantum land, and a practical doorway to single-shot readout, quantum sensing, and even atom-by-atom simulation.
The stunt turns an STM, which normally feels only electrons, into a kind of nuclear eavesdropper. The trick is the hyperfine interaction that ties an atom’s electron spin to its nuclear spin. Drive or probe the electron at just the right frequency and the nucleus gives itself away, visible as a two-level current that hops between states. Those hops, measured directly in time, tell a story of surprising calm. Five seconds before a flip. Five seconds to prepare, measure, and act.
“The general idea had been demonstrated a few years ago, making use of the so-called hyperfine interaction between electron and nuclear spins,” Otte explains. However, these early measurements were too slow to capture the motion of the nuclear spin over time.
That bottleneck is what TU Delft’s group set out to break. First authors Evert Stolte and Jinwon Lee targeted a single 49Ti atom on MgO/Ag(100), tuned the radiofrequency to a nuclear-state-dependent electron transition, then watched the tunneling current jump between two discrete levels on a computer screen. No averaging needed. Because their measurement is faster than the nucleus tends to flip, each snapshot stands on its own, the definition of single-shot readout.
Numbers anchor the promise. The nucleus sits in one state for roughly 5.3 ± 0.5 seconds in their pulsed scheme, compared with about 100 nanoseconds for the same atom’s electron spin. Seven orders of magnitude of breathing room. And when they turned on continuous driving, the lifetime shortened, a useful clue that electron–nuclear flip-flops are the main relaxation valve. It is an elegant, slightly perverse result. To listen too hard is to nudge the spin along.
Beyond the lab romance of catching a nucleus mid-flip, there is a practical beat. Single-shot readout is the baseline for control, and control is what lets you build. With STM, you can place atoms where you want them, one by one, then read and initialize their nuclear states with high fidelity. The commercial angle is still speculative, sure, but real enough. Quantum sensors that live on surfaces, exquisitely sensitive to fields or strain, and tunable by atomic neighborhood, can be made, iterated, and characterized without exotic fabrication.
“We were able to show that this switching corresponds to the nuclear spin flipping from one quantum state to another, and back again,” Stolte says.
The buried lede is customization. Because the relevant relaxation channel depends on binding site, magnetic field, and the small in-plane slice of the hyperfine coupling, lifetime becomes a design parameter. Shift the atom, change the neighborhood, adjust the field, and you can stretch or squeeze those precious seconds. That is catnip for anyone eyeing atomic-scale quantum simulators, where long-lived nuclear spins could serve as stable memory while electrons do the fast, messy work.
And yet, a caution. The same sensitivity that enables readout also invites noise, from Johnson currents in metals to spin-lattice couplings and the statistical tug of nearby nuclei. The real surprise came when that complexity simplified into a handle you can turn. Detune the drive. Pulse instead of blast. Let the nucleus rest between questions. Suddenly five seconds is not theoretical, it is measurable, and it is yours to spend.
Explainer: What Is Single-Shot Readout, And Why It Matters
Single-shot readout means you measure a quantum state in one go, fast enough that the system has not changed underneath you. Most experiments average many repeats to see a tiny signal. That hides the dynamics. Here, researchers used an STM to probe an atom’s electron at a frequency that depends on the nucleus. When the nucleus sits in one spin state, the electron responds and the current jumps. When it sits in another, it does not. Recording those jumps in real time lets you assign the nuclear state for each measurement. That unlocks initialization, feedback, and control, all crucial for quantum sensing and for building atomic-scale simulators where nuclear spins store information while electrons perform operations.
Journal: Nature Communications
DOI: 10.1038/s41467-025-63232-5