Cosmic phase-transition simulations reveal a two-scale gravitational-wave signal

Listening for a cosmic change of phase

A first-order phase transition resembles boiling in one important respect: bubbles of a new phase form, expand and merge. In the early Universe, such bubbles could have disturbed a hot plasma and generated gravitational waves, ripples in spacetime that might preserve information about otherwise inaccessible particle physics.

The question is how to predict that signal accurately. Many simulations neglect cosmic expansion, an approximation suited to sufficiently rapid transitions. But expansion cools the plasma, while the transition releases energy and can reheat it. Both effects change the material through which bubbles and sound waves travel.

In their revised arXiv preprint, Jintao Zou, Zizhuo Zhao, Yongtao Jia and Ligong Bian simulate these processes together. Their central finding is that plasma motion remains the dominant gravitational-wave source in the cases studied, while the resulting spectrum depends strongly on how bubble walls propagate.

A changing plasma on a three-dimensional grid

The researchers couple a scalar field, representing the state of the transition, to a relativistic fluid. A friction term transfers energy between the moving bubble walls and the plasma. Unlike a fixed-friction prescription, their coupling changes with the local field and temperature.

They evolve this system in an expanding cosmological background on a grid with 512 points along each spatial direction. The pressure–energy relationship and transition strength also evolve, rather than remaining fixed throughout the calculation.

The detonation runs use initial transition strengths of 0.15, 0.38 and 0.60. This parameter, called alpha, measures the strength of the transition relative to the thermal background. They also explore inverse transition rates, beta divided by the Hubble expansion rate, of 20, 25 and 30; larger values describe faster transitions.

Fast walls travel at roughly 81–95 percent of light speed and are classified as detonations. A slower branch, with walls moving at about 27 percent of light speed, produces deflagrations: walls moving below the plasma’s sound speed. These names describe propagation regimes, not chemical combustion.

Two scales leave different signatures

Across the simulations, the authors find that gravitational waves sourced by the fluid outweigh those sourced directly by the scalar-field bubble walls. Bubble collisions transfer energy into plasma motion, which continues generating gravitational waves after the walls merge.

For detonations, the fluid-generated spectrum has two peaks. The authors associate them with two characteristic lengths: the typical separation between bubbles and the thickness of the moving sound shells. These spatial scales translate into characteristic gravitational-wave frequencies.

For deflagrations, the two lengths are similar enough that the spectrum instead shows one peak. The high-frequency signal also declines much more steeply: the reported tails scale approximately as frequency to the minus fourth power for detonations and minus eighth power for deflagrations. These are measured features of the simulated cases, not universal laws for every possible transition.

The supplemental results reveal additional dynamics. Stronger detonation transitions briefly reheat the plasma during collisions before expansion resumes cooling it. Collisions also generate rotational fluid motion, whose magnitude generally grows with transition strength. That finding does not, by itself, establish a separate turbulence-dominated gravitational-wave signal.

Lifetime and amplitude need careful definitions

The authors report that slower transitions have a larger effective sound-wave source lifetime, while increasing transition strength generally reduces it. Here, “effective lifetime” is an expansion-weighted measure of gravitational-wave production, not simply the elapsed time during which sound exists. In their treatment it remains below one Hubble time, the characteristic expansion timescale.

The revised paper places the fully reduced spectral amplitude at roughly one hundredth, comparable in order of magnitude to earlier flat-spacetime simulations. This number has several physical scaling factors divided out; it does not mean that one percent of the Universe’s energy becomes gravitational waves. The revision therefore does not retain the original abstract’s claim of an approximately tenfold amplitude suppression.

What the simulations cannot yet settle

The numerical box covers only a finite range of wavelengths. Although the resolved low-frequency spectra rise approximately as frequency squared, the authors expect a different, cubic rise at still longer wavelengths that their simulations cannot directly resolve.

The results also come from a particular field potential, friction prescription and limited parameter sample. Some tabulated amplitudes and lifetime trends do not align cleanly with the surrounding discussion, so precise numerical comparisons warrant clarification.

For proposed observatories such as LISA, Taiji and TianQin, the work offers candidate spectral features to investigate—not a detection forecast or evidence that such a transition occurred. Its useful message is that expansion, evolving plasma properties and bubble-wall propagation deserve attention together when constructing gravitational-wave predictions.