We explore a new paradigm to study dissipative dark matter models using gravitational-wave observations. We consider a dark atomic model which predicts the formation of binary black holes such as GW190425 while obeying constraints from large-scale structure, and improving on the missing-satellite problem. Using LIGO and Virgo gravitational-wave data from September 12, 2015 to October 1, 2019, we show that interpreting GW190425 as a dark matter black-hole binary limits the Chandrasekhar mass for dark matter to be below $1.4M_{\odot }$ at $>99.9\%$ confidence implying that the dark proton is heavier than 0.95 GeV, while also suggesting that the molecular energy-level spacing of dark molecules lies near ${10}^{-3}\mathrm{eV}$ and constraining the cooling rate of dark matter at low temperatures.

• Received 5 October 2020
• Accepted 20 May 2021

DOI:https://doi.org/10.1103/PhysRevD.104.044015