For years, only massless light particles could be entangled. Scientists just did it with helium atoms — particles heavy enough to fall.
image from grok
ANU physicists observed Bell correlations in entangled metastable helium-4 atoms falling under gravity, demonstrating entanglement in massive particles' motional states for the first time. However, this does not constitute a loophole-free Bell test: the experiment uses equal phase settings on both arms, leaving the locality loophole open and allowing a hypothetical faster-than-light hidden signal to explain the correlations. The actual advance is extending Bell correlations to particles with mass that couple to gravitational fields, unlike prior momentum-based tests using massless photons.
Physicists have entangled helium atoms and watched them fall through a gravitational field. The press called it "atoms in two places at once." The paper is more careful.
A team at the Australian National University (ANU) reports in Nature Communications the first observation of Bell correlations in the motional states of massive particles — specifically, entangled pairs of metastable helium-4 (4He*) atoms falling under gravity. The result is genuine. The platform the team built is a real experimental advance. But the press-release framing got ahead of the physics in ways the paper itself does not endorse.
The core claim: Bell correlations — the statistical fingerprint of entanglement — detected in the external, motional degrees of freedom of atoms with mass. Previous momentum-based Bell tests used photons, which have no mass and do not couple to gravitational fields. Helium atoms do both. That is the actual news, as the APS Physics commentary on the result notes.
What the experiment does not do: violate a Bell inequality in a loophole-free way. The interferometer uses equal phase settings on both arms — a constraint the paper explicitly calls a limitation. A genuine loophole-free Bell test requires randomly chosen, space-like separated measurement settings, neither of which this experiment achieves. The locality loophole remains open: a hidden signal traveling between the detectors faster than light could, in principle, explain the correlations without any nonlocality at all. The paper says this plainly. The press release did not.
Bell correlations are a particular kind of quantum correlation between separated particles. When two particles are entangled, measuring one instantly fixes the state of the other — even across arbitrary distances. Bell showed that certain statistical combinations of measurement outcomes cannot be explained by any local hidden variable theory. Observing those correlations is evidence for entanglement. Demonstrating them in a way that definitively rules out all hidden-variable explanations requires closing the detection loophole (insufficient measurement efficiency) and the locality loophole (insufficient separation in space and time). This experiment addresses neither.
The lead author is Yogesh Sridhar, a PhD researcher at ANU's Research School of Physics. Senior authors are Andrea Truscott and Sean Hodgman, also at ANU. The collaboration includes teams from the University of Oklahoma and the University of Queensland. The group has been building toward this result since at least 2019, when they demonstrated spatially separated entangled helium atom pairs for the first time.
The experimental sequence: the team cools roughly 100,000 metastable helium atoms into a Bose-Einstein condensate — a state where the atoms lose their individual identities and behave as a single quantum object. Three clouds of BEC are held in a magnetic trap with frequencies of approximately 15 Hz vertically and 25 Hz horizontally. The trap is switched off. Two clouds collide in free fall. From that collision, pairs of atoms emerge with opposite momenta via spontaneous s-wave scattering — a quantum mechanical process that entangles their motional states.
Each entangled pair passes through a Rarity-Tapster matter-wave interferometer: a sequence of laser pulses that act as mirrors and a beam splitter for atomic de Broglie waves. A collision pulse at time t0 is followed 350 microseconds later by a mirror pulse, then another 350 microseconds later by a beamsplitter pulse. After a fall time of roughly 0.416 seconds, the atoms land on a micro-channel plate detector with delay-line readout, positioned 848 millimeters below the trap. The detector resolves individual atoms in three dimensions — a single-atom camera for quantum trajectories.
The average occupancy of the quantum state is very low: about 0.035 atoms per mode. The second-order correlation function at zero delay reaches roughly 30, indicating strong bunching — entangled pairs tend to arrive together. From the statistics of where entangled partners land, the team extracts Bell correlation parameters consistent with nonlocality, according to the paper's analysis.
The 90 percent Raman transfer efficiency to a magnetically insensitive atomic state is what makes the measurement clean enough to work. Without that, magnetic field noise would scramble the signal before the detector ever saw it.
The equal-phase constraint is the paper's own caveat, not a criticism being levied from outside. "We emphasize that our test does not close the locality loophole," the authors write, "as the measurement settings are not chosen randomly and are not space-like separated." They are clear about what they did and did not achieve. The press release from ANU's science office ran under the headline "Spooky Quantum Helium Atoms Give Hope for Theory of Everything" — which is a different document entirely.
What the experiment does open is a regime. Previous atom-based Bell tests used internal spin states — the atom's internal angular momentum as the quantum variable. This work uses external motional states — where the atom actually is in space. Mass enters the problem because atoms falling through a gravitational field are subject to gravity's pull, and that coupling is what the team wants to probe. A preprint posted to arXiv in late 2024 by some of the same authors outlines a concrete proposal: entangled atoms of different masses, coupled via their mutual gravitational attraction, as a direct test of whether gravity preserves quantum coherence or collapses it. That proposal is still theoretical. This result is the experimental prerequisite.
The distinction matters. Photons do not meaningfully couple to gravitational fields. Massive atoms do. If you can entangle the motional states of two massive particles and then let gravity act on them, you have a probe of whether gravity itself respects quantum mechanics — or whether it operates classically, collapsing quantum superpositions in the way some interpretations of quantum mechanics suggest it must. That question has been philosophical for decades. The ANU platform makes it experimental.
"Quantum mechanics has been proven to apply to a vast range of systems," said Sean Hodgman, in a statement on ANU's science page. "Bell tests prove that entanglement is actually how the world works. For two separated atoms that are entangled, if you change one of them, it will instantly affect the other." That much is established physics. What remains untested is whether gravity participates.
The honest version of this story is infrastructure, not a milestone. The ANU team has built a working platform for quantum-gravity experiments. The entangled atom pairs are real. The Bell correlations are there. The numbers — g²(0) ~ 30, n̄ ~ 0.035, fall time 0.416 seconds, trap frequencies (15, 25, 25) Hz — describe a delicate, low-density experiment operating at the edge of what current atom optics can resolve. What happens next is the more interesting question.
The follow-up the team is proposing — different-mass atoms entangled through their mutual gravitational interaction — would be the first direct experimental test of whether gravity preserves quantum superposition. It is not a quantum computer. It does not break encryption. It is a fundamental physics experiment addressing a question that has been open since the 1930s: does gravity behave quantum mechanically at the scale of everyday objects?
Whether that question gets answered — and how — is what to watch.
The paper is published in Nature Communications (volume 17, article 2357, 2026). The preprint is available on arXiv (2502.12392).
Story entered the newsroom
Research completed — 0 sources registered. First Bell correlations observed in motional states of massive particles (metastable 4He* atoms). Previous momentum-based Bell tests used photons only
Draft (804 words)
Reporter revised draft based on fact-check feedback (757 words)
Reporter revised draft based on fact-check feedback (1174 words)
Approved for publication
Published (1185 words)
Get the best frontier systems analysis delivered weekly. No spam, no fluff.
Quantum Computing · 16h 34m ago · 2 min read
Quantum Computing · 17h 31m ago · 4 min read
