When a quantum environment changes the ground state

A quantum system’s surroundings can do more than make its motion fade. They can change which configuration has the lowest energy. In a theoretical preprint, Giuseppe Musolino and colleagues calculate how that distinction could produce a collective phase transition in an ensemble of two-level atoms or spins. Their central prediction is conditional: the transition depends on the mathematical structure of the coupling to the environment, not simply on stronger energy loss.

Two environments with different effects

The model contains identical, non-interacting two-level systems coupled collectively to two reservoirs. Each reservoir is represented by many harmonic oscillators—an idealized description of environmental modes that can exchange energy with the ensemble.

The reservoirs differ in one crucial respect. The first interaction includes a compensating quadratic contribution, called a counterterm. Combined with the coupling, it produces a displaced harmonic potential without introducing the same downward shift in the system’s restoring force. The second reservoir lacks that compensation.

For the second interaction, allowing the reservoir coordinates to adjust lowers the energy associated with the ensemble’s collective displacement. Increase this effect sufficiently, and the original ground-state configuration ceases to be the minimum. The spins’ finite two-level structure matters when finding the new configuration; the calculation does not simply continue an unstable harmonic approximation beyond its range.

The authors quantify this effect with an integrated coupling, χ. In their notation, the transition occurs at χ = ωa², where ωa is the uncoupled two-level transition frequency. Below that threshold, the collective excitation frequency decreases toward zero, a process called mode softening.

An important distinction follows: χ depends on coupling across the reservoir’s frequency spectrum, whereas energy loss near a particular resonance depends on the reservoir locally at that frequency. Far-detuned environmental modes can therefore alter the ground state even when resonant damping is negligible. The prediction is not that any sufficiently lossy environment must cause a transition.

Following motion when oscillations disappear

The calculation takes the thermodynamic limit: the number of spins grows without bound at fixed density. A Holstein–Primakoff mapping expresses their collective motion through a bosonic variable. The researchers then derive quantum Langevin equations, which describe the system’s motion together with environmental noise and memory.

They retain interaction terms commonly discarded in a rotating-wave approximation. This matters near the transition because the collective frequency becomes small enough that damping can dominate it. An approximation requiring losses to be much slower than oscillations is no longer reliable there.

The excitation frequencies are complex: their real parts describe oscillation, while their imaginary parts describe decay. Around the transition, the model develops an overdamped interval in which the relevant motion relaxes without oscillating. Merely seeing the oscillation frequency vanish would therefore not identify the transition uniquely. At the critical point, one branch has both zero oscillation frequency and zero decay rate.

Above the threshold, the authors minimize the energy of the entire system and both reservoirs together. They obtain a collective displacement whose occupation scales with the number of spins, accompanied by displacements in both reservoirs. These are ground-state properties of the coupled model, not a prediction of freely emitted radiation or a continuously powered output.

What a measurement might show

Using input-output theory, the paper derives reflection and transmission spectra. In the displayed model spectra, the reflection dip moves toward zero frequency as the transition approaches, and the line becomes increasingly asymmetric.

The calculation also distinguishes the dip’s position from the real part of the damped excitation frequency. Near criticality, those quantities need not coincide. For the response expressions presented, the dip follows the frequency of the statically renormalized system with dynamical losses removed. Equal coupling rates to the two ports give zero reflection at that resonance, with energy transmitted through the other channel rather than disappearing.

A prediction with explicit boundaries

The proposed experimental setting is a yttrium iron garnet magnetic film coupled to an open superconducting waveguide. The authors suggest that magnetic coupling could supply the counterterm-free interaction. They do not report observing this reservoir-induced transition.

The results also concern a zero-temperature ground-state analysis in an infinite-ensemble limit, focused on the maximally collective spin sector. They do not establish how sharply a finite, warmer device would reproduce the transition. Although the Langevin framework can accommodate structured reservoirs, the explicit treatment takes the decay functions as real and sets aside their dispersive contributions. Its broad claims about arbitrary reservoir spectra should be read with that qualification.

Published as an arXiv preprint, the work offers a theoretical mechanism and proposed measurements, not experimental confirmation. Its useful distinction is between an environment that damps motion and one whose coupling changes the energy landscape itself.