A targeted radio survey of infrared-selected bow shock candidates

TL;DR

This study conducts the largest targeted radio survey of infrared-selected bow-shock candidates around massive stars, using the VLA and Effelsberg to detect and characterize radio emission. By combining multi-frequency radio data with Gaia DR3-based proper motions, the authors assess thermal versus non-thermal contributions and derive upper limits on the electron density in bow-shock shocked layers. They report six clear radio detections and several marginal cases, with spectral-index analyses suggesting predominant non-thermal emission in some sources, though uncertainties and H II-region contamination limit firm conclusions. The work demonstrates the effectiveness of deep, targeted radio surveys in expanding the sample of radio-emitting bow shocks and outlines follow-up strategies across frequencies to disentangle emission mechanisms and test MHD models of wind–ISM interactions.

Abstract

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)

Figures (6)

• Figure 1: Presentation of the detected targets listed in Table \ref{['tab:targets']}. Each row shows a single target, labelled in the title of the first column. From left to right, the panels show (1) the VLA intensity map in mJy beam$^{-1}$ with black contours at $[-10,-5,5,10,20,40,80,160]\,\sigma$; (2) the map of the power-law spectral index, $\alpha$, with $10\sigma$ intensity contour overlaid; (3) the map of the 1-$\sigma$ absolute uncertainty in $\alpha$, with a contour showing $\alpha=-0.5$; and (4) the Effelsberg intensity map at 7.639 GHz (colour scale in mJy beam$^{-1}$) with VLA intensity contours overlaid and the beamsize shown for cases where it is not too large (in all cases, about 3 pixels in diameter). Also shown in the first panel are a line showing ;1; scale at the bottom right, the synthetic beam FWHM at bottom left, blue dotted contours of WISE 22 emission, and a black arrow showing the direction of peculiar proper motion, where reliably estimated.
• Figure 2: Intensity plots of the targets listed in Table \ref{['tab:targets']} for which there is a hint of possible emission but not enough to claim a detection. The colour bar values are in mJy beam$^{-1}$, with negative values shown in blue. In the bottom right a line that corresponds to ;1; is shown, while in the bottom left the cross-section of the synthetic beam's FWHM can be seen. Contours of the most interesting features from the WISE 22 counterpart are overlaid with blue lines. Black contours are radio emission at levels $[-10,-5,5,10,20,40,80,160]\,\sigma$, with negative contours using a dashed line.
• Figure 3: Comparison of observations of the bow shock of $\zeta$ Oph with the VLA (left), Effelsberg (centre), and the mid-IR image from Spitzer Space Telescope MIPS 24 $\mu$m. All three images are plotted in units of megaJansky per steradian. The VLA and Effelsberg beams are plotted at the lower-left of the respective panels. For the VLA and Effelsberg plots, IR emission contours at levels 70 and 100 MJy sr$^{-1}$ are overplotted in dotted blue lines, and for Effelsberg and Spitzer plots the VLA intensity contours are overlaid in black, showing only the detected point sources.
• Figure 4: List of VLA targets and their driving star, sorted according to right ascension.
• Figure 5: Intensity plots of the non-detected targets listed in Table \ref{['tab:targets']}. We use a saturated colour scale extending to negative values to show the noise level in the maps. The colour bar values are in milliJansky per beam. In the bottom right, a line that corresponds to ;1; is shown, while in the bottom left, the cross-section of the synthetic beam's FWHM can be seen. Contours of the most interesting features from the WISE 22 counterpart are overlaid with blue lines.