Probing the non-thermal physics of stellar bow shocks using radio observations

We aim to interpret the radio emission of stellar bow shocks and assess under which conditions it could be either thermal or non-thermal, and how to use the observational data to infer their physical properties. We used an extended non-thermal emission model for stellar bow shocks for which we incorporated a consistent calculation of the thermal emission from the forward shock. We fitted this model to the available radio data, including largely unexplored data at low frequencies. In addition, we used a simplified one-zone model to estimate the gamma-ray emission from particles escaping the bow shocks. We can only explain the radio data from (BD+43{\deg}3654 and BD+60{\deg}2522) assuming a hard electron energy distribution below $\sim$1 GeV, a high efficiency of conversion of (shocked) wind kinetic power into relativistic electrons, and a relatively high magnetic-to-thermal pressure ratio. In the other systems, the interpretation of the observed flux density is ambiguous, although a non-thermal scenario is also favoured. We also show how complementary observations at other frequencies can allow us to place stronger constraints in the model. We also estimated the gamma-ray fluxes from the HII regions around the bow shocks of BD+43{\deg}3654 and BD+60{\deg}2522, and obtained luminosities at GeV energies of $\sim 10^{33}$~erg\,s$^{-1}$ and $10^{32}$~erg\,s$^{-1}$, respectively. Stellar bow shocks can potentially be very efficient particle accelerators. This work provides multi-wavelength predictions of their emission and demonstrates the key role of low-frequency radio observations in unveiling particle acceleration processes. The prospects of detections with next-generation observatories such as SKA and ngVLA are promising. Finally, BD+43{\deg}3654 may be detected in GeV in the near future, while bow shocks in general may turn out to be non-negligible sources of low-energy cosmic rays.