A targeted radio survey of infrared-selected bow shock candidates
Bow shocks around massive stars have primarily been detected in IR emission, but radio detections are becoming more frequent with the commissioning of sensitive and large field-of-view interferometers. Radio data probes both thermal and non-thermal emission, thereby constraining the relativistic electron population. We undertook a radio survey for bow shocks based on IR catalogues of candidates, using the VLA and the 100-m Effelsberg Telescope, aiming for new detections and to better characterise the multi-wavelength emission. We used Gaia DR3 to re-calculate spatial motion of the driving stars with respect to the surrounding stellar population. We studied the radio emission from bow shocks using emission maps and spectral-index measurements, and compared our results with data from catalogues and multi-wavelength emission. Of the 24 targets observed with the VLA in the 4-12 GHz band, six were clearly detected (including two previously reported) and 5 possibly detected. A subset of these were also observed and detected with Effelsberg at 4-8 GHz. The VLA-derived spectral index maps indicate non-thermal emission for most sources, but the statistical uncertainties are large for most sources and all Effelsberg observations indicate thermal emission. Assuming thermal emission, we obtain upper limits on the electron density within the shocked layer. We obtained upper limits on radio emission from the bow shock of Zeta Oph at a similar flux level to predictions from MHD simulations. Our survey marks a significant addition to the ca. 10 previously known radio-emitting bow shocks in the literature, and demonstrates that deep, targeted radio surveys can effectively detect IR-selected bow shocks. Follow-up observations of these targets at lower and higher frequencies are encouraged to determine whether any are non-thermal emitters like the bow shocks of BD+43, BD+60 and LS2355. (abridged)