In a cryogenic chamber at Delft, a single titanium atom's nucleus has, for the first time, been driven directly by radio waves, sidestepping the electron that usually serves as an intermediary. The achievement reflects a long human pursuit: to hold information in the quiet heart of matter, where it can survive the noise of the world. It is a demonstration of access rather than mastery, a proof that the nucleus can be reached without losing the isolation that makes it precious. Whether that reach can become a dependable, timed craft remains a question the next experiments will have to answer.
TU Delft Drives Nuclear Magnetic Resonance on a Single Titanium Atom Using a Scanning Tunneling Microscope
The method reached the nucleus, but at the cost of what made it worth reaching.
Mark
Start me off simply. What did the Delft team actually do?
Mimi
They drove the nuclear spin of a single titanium-47 atom on a surface, directly, using radio-frequency fields, inside a scanning tunneling microscope cooled to 0.4 kelvin.
Mark
And why is "directly" the important word?
Mimi
Earlier STM work reached nuclei through hyperfine hybridization at low fields, 100 millitesla or less. That mixes electron and nuclear states and undoes the nucleus's isolation.
Mark
So they changed the field.
Mimi
Up to 1.4 tesla. There the electron's Zeeman energy dominates, the states stay separate, and the nucleus responded in the 50 to 85 MHz range.
Luke
Let me slow us down. Is this a quantum memory?
Mimi
Not yet. The source says full coherent gate control needs pulsed sequences, and this used continuous driving.
Luke
Right. And the talk of qudits and memory registers is a prospect, not something shown. What's actually demonstrated is direct driving of one nucleus.
Mark
What's the electron doing, then?
Mimi
Its gigahertz channel gives spin-polarized readout, so it acts as the way of seeing the nucleus rather than the way of moving it.
Luke
One more gap: the seconds-long relaxation lifetimes are described as typical for nuclei on surfaces. The report doesn't give a measured figure for this atom, so I wouldn't read it as a result here.
Mark
Fair. So the next thing to watch is pulsed control.
Mimi
Yes. That would show whether the nucleus can be controlled coherently, not just addressed.
Der Puls
- A nucleus that refuses to be disturbed is now being addressed directly, at a temperature of 0.4 kelvin inside a scanning tunneling microscope.
- Earlier methods reached the nucleus only by entangling it with its electron, which erased the very isolation that made it valuable.
- The Delft team pushed the field to 1.4 tesla, where electron and nucleus stay distinct, and drove the 47Ti nucleus with radio signals between 50 and 85 MHz.
- The drive so far is continuous, not pulsed, so the work proves the nucleus can be reached but not yet commanded with the timed precision a processor needs.
- The next test is switching to pulsed sequences, which would decide whether atomic-scale nuclear qudits can become real memory in engineered surface lattices.
In a cryogenic chamber at Delft, a single titanium atom's nucleus has, for the first time, been driven directly by radio waves, sidestepping the electron that usually serves as an intermediary. The achievement reflects a long human pursuit: to hold information in the quiet heart of matter, where it can survive the noise of the world. It is a demonstration of access rather than mastery, a proof that the nucleus can be reached without losing the isolation that makes it precious. Whether that reach can become a dependable, timed craft remains a question the next experiments will have to answer.
At Delft University of Technology, physicists at the Kavli Institute of Nanoscience have done something that until now resisted them: they drove the nucleus of a single titanium atom directly with radio-frequency fields. The atom sits on a surface inside a scanning tunneling microscope cooled to 0.4 kelvin, and the work, published in Nature Communications, uses an approach called ENDOR that addresses the nucleus itself rather than reaching it through the atom's electron.
The reason this matters lies in the nucleus's quiet character. Its magnetic moment is roughly a thousand times weaker than an electron's, so it is shielded from environmental noise and can hold its state for seconds on a surface. That stillness is exactly what makes it attractive for storing quantum information. Earlier attempts to reach single nuclear spins relied on strong hybridization with the electron, which worked only at very low fields and blurred the isolation that gave the nucleus its value.
The Delft team took the opposite path, raising the field to 1.4 tesla so that the electron and nucleus remained distinct. The nucleus could still be driven, with resonances between 50 and 85 MHz, while the electron served as a readout channel. The experiment shows an isolated nuclear spin can be driven directly at high field.
The researchers are candid about what remains. Their drive was continuous, while a working quantum gate needs pulsed sequences with precise timing. The result opens a route toward atomic-scale memory registers and higher-dimensional qudits that could be built into engineered surface lattices, but the decisive step is still ahead: showing that this nucleus can be controlled coherently, not merely reached.
At a temperature of 0.4 kelvin, inside a scanning tunneling microscope at Delft University of Technology, a single titanium atom sitting on a surface has had its nucleus driven directly by radio-frequency fields. The atom is the 47Ti isotope. The work comes from physicists at the university's Kavli Institute of Nanoscience, and it appears in Nature Communications under the title "Nuclear magnetic resonance on a single atom with a local probe."
The technical heart of the study is an Electron-Nuclear Double Resonance scheme, known as ENDOR, built into the ultra-low-temperature microscope. In practice, that means the team addressed the nucleus itself rather than reaching it by way of the atom's electron. The researchers describe this as coherent nuclear addressing that does not depend on electron-nuclear spin hybridization.
That distinction matters because of how nuclei behave. Their magnetic moments are roughly three orders of magnitude smaller than those of electrons, which keeps them well shielded from environmental noise. On a surface, that shielding translates into relaxation lifetimes on the order of seconds. It is the quality that makes a nucleus attractive as a place to store quantum information.
The difficulty is that earlier efforts to address single nuclear spins with a scanning tunneling microscope leaned on strong hyperfine hybridization, and that approach worked only at low magnetic fields, at or below 100 millitesla. Hybridization mixes the electron and nuclear states, and in doing so it collapses the very isolation that makes the nucleus useful. The method reached the nucleus, but at the cost of what made it worth reaching.
The Delft team went the other way and raised the magnetic field to 1.4 tesla. At that strength the electronic Zeeman energy dominates, so the electron and nuclear states stay distinct. Under those conditions the 47Ti nucleus could still be driven, with resonances in the 50 to 85 MHz range. The electron, meanwhile, was put to a different job: its gigahertz-range channel provided spin-polarized readout of what the nucleus was doing. The demonstration verifies that an isolated nuclear spin can be driven directly by local radio-frequency fields at high magnetic fields.
The source is clear about the limits of the result. Full coherent quantum gate control will require pulsed-wave sequencing, whereas this work used continuous radio-frequency driving. So the experiment establishes that the nucleus can be addressed directly; it does not yet show the kind of timed operations a working quantum processor would need.
What the platform does suggest is a route toward atomic-scale quantum memory registers and high-dimensional qudits, which are quantum units that hold more than two levels. The researchers say atomic-scale nuclear qudits could be integrated into engineered surface lattices, the kind of atom-by-atom structures a scanning tunneling microscope is used to build. The next step to watch is the move from continuous driving to pulsed sequences, which would show whether this isolated nuclear spin can be controlled coherently as well as addressed.
Bemerkenswerte Zitate
Full coherent quantum gate control will require pulsed-wave sequencing rather than continuous RF driving.— Quantum Computing Report, summarizing the Delft study
Möchten Sie die ganze Geschichte?
Das Original lesen bei Quantum Computing Report ↗
Science & Space