First order phase transitions are well-motivated and extensively studied
sources of gravitational waves (GWs) from the early Universe. The vacuum energy
released during such transitions is assumed to be transferred primarily either
to the expanding walls of bubbles of true vacuum, whose collisions source GWs,
or to the surrounding plasma, producing sound waves and turbulence, which act
as GW sources. In this Letter, we study an alternative possibility that has so
far not been considered: the released energy gets transferred primarily to
feebly interacting particles that do not admit a fluid description but simply
free-stream individually. We develop the formalism to study the production of
GWs from such configurations, and demonstrate that such GW signals have
qualitatively distinct characteristics compared to conventional sources and are
potentially observable with near-future GW detectors.
