Four cesium atoms hang in vacuum, held by laser tweezers like beads on an invisible thread. Around them is a ring resonator, a closed loop of mirrors where light circulates without obstacles. When a probe beam strikes this tiny chain, something strange happens: the atoms scatter light not uniformly in all directions, but predominantly in one direction. This is not a coincidence but a precise tuning, like a tuning fork that responds only to its own note.

Previously, collective light scattering was studied in free space or in standing waves of two-mirror resonators, where the mirrors set the boundary conditions. Here, the traveling waves of the ring resonator make the interference depend only on the relative positions of the atoms. It is as if musicians in an orchestra started playing in unison not because of a conductor's command, but simply because they see each other. The physicists varied the interatomic distance in steps smaller than a nanometer using an acousto-optic deflector, and observed the constructive interference peak narrowing as atoms were added, from two to four. For two atoms, the oscillation period in the clockwise mode was 783.1(4) nm, and in the counterclockwise mode 935.0(6) nm; the ratio of periods gave the probe beam's angle of incidence with an accuracy of hundredths of a degree.

But to make the atoms sing in harmony, they first had to be calmed down. Thermal jitter would smear the interference pattern, like a trembling hand ruining a fine brushstroke. Therefore, the atoms were cooled by the resonator method: by tuning the probe beam detuning to the red sideband of the oscillations, each scattered photon carried away one phonon, a quantum of vibrational motion. The radial motion of the atoms was cooled to a mean phonon number of 0.17(4), corresponding to a temperature of 2.1(3) μK and an 85% probability of being in the ground vibrational state. The axial motion reached 3.4(2.4) phonons. It is as if the tuning fork were frozen to near absolute zero, so it stopped trembling, leaving only a pure tone.

But the most surprising thing happened with a single atom. Prepared in a state with maximum spin projection, it scattered light in one direction 23 times more strongly than in the other. This is chiral scattering, a property of an object that does not coincide with its mirror image. In a sense, we are approaching the "optical lattices" from Greg Egan's novel "Permutation City," where ordered structures of matter serve as computational elements.

This work stands apart: the nearest neighbors are doing quantum squeezing of molecules, entanglement on chips, and ion cooling, but no one had combined tweezer arrays with a ring resonator to control the direction of scattering. The difference lies in using traveling waves instead of standing waves, which removes the dependence on mirror positions and makes the interference purely geometric. Previously, it was unclear whether one could control the direction of collective light scattering from a few atoms in a resonator without being tied to mirror positions. Now it is shown that an array of two to four atoms in a ring resonator gives a narrow Bragg peak, and a single atom can scatter light predominantly in one direction. This is the first realization of resonator cooling of single atoms in tweezers to the ground state in the resolved sideband regime.

The platform naturally extends to studying non-Hermitian and nonreciprocal physics, where the directionality of scattering plays a key role. Adding fast state readout and atom counting, already demonstrated earlier, will allow assembling molecules from atoms and tracking their collisions in real time. Resonator cooling does not depend on the internal structure of the particles, so it can be transferred to ions, molecules, and levitating nanoparticles. The sub-nanometer distance calibration method will be useful in quantum metrology, and deep cooling of atoms will improve the accuracy of quantum gates in Rydberg arrays, where residual temperature limits fidelity. The next step is to cool the axial motion to the level of the radial one and check how it improves the contrast of the Bragg peaks. Then one should study collective vibrational modes in arrays with identical trap frequencies. For quantum optics, an open question remains how to scale directional scattering to tens and hundreds of atoms without losing coherence. The problem of suppressing heating from probe beams during long experiments is also unsolved.