Radio emission in star-forming galaxies: connection to restarted or relic AGN activity

TL;DR

This work investigates the origin of radio emission in star-forming galaxies lacking clear AGN signatures by constructing a matched sample of GHz-detected and non-detected SF galaxies using LOFAR 144 MHz and FIRST 1.4 GHz data, in combination with MaNGA IFU spectroscopy and WISE mid-IR photometry. The authors find that GHz-SFs, compared to their non-detected twins at fixed $L_{144MHz}$ and stellar mass, exhibit broader ionized-gas kinematics ($W_{80}$), higher central outflow fractions, redder colors, increased central obscuration, and emission-line ratios offset toward the AGN regime in off-nuclear regions, while their MHz morphologies are more compact. Although no active AGN is detected in these galaxies, the GHz-SF population resembles radio-detected AGN in several properties, suggesting that compact GHz radio emission may trace past AGN activity or low-power jets, possibly via fossil outflows or light echoes. The results imply a co-evolutionary link between AGN activity and star formation, with AGN history influencing gas dynamics and ionization states even in galaxies where current AGN signatures are absent, and they highlight the importance of spatially resolved spectroscopy for disentangling these effects.

Abstract

Increasing evidence shows that AGN with radio detections have more perturbed ionized gas kinematics and higher outflow detection rates, suggesting a link between radio emission and these processes. In galaxies with weak or ambiguous AGN signatures, some studies attribute the radio emission to star formation, while others propose AGN-driven winds or weak, unresolved jets as the dominant mechanism. To investigate this connection, we take a step back and analyze a sample of star-forming (SF) galaxies with no clear current AGN signatures. Using low-(LOFAR, 144MHz) and high-frequency (FIRST, 1.4GHz) surveys, combined with spatially resolved spectroscopy from the MaNGA survey, we compare SF galaxies with 144 MHz detections that either do or do not have GHz detections. Despite being matched in stellar mass, redshift, and radio (MHz) luminosity, GHz-detected SF galaxies systematically differ from their non-GHz-detected counterparts. The former display enhanced ionized gas-emission line widths, higher central outflow fractions, redder colors, increased central obscuration, and offset emission-line ratios that shift towards (or closer to) the AGN regime (in the [NII] BPT diagram). Furthermore, the non-GHz galaxies are likely undetected due to their extended radio morphologies, while the GHz-detected ones are significantly more radio compact. Most of the properties from the GHz-detected (compared to non-detected) remarkably resemble the behavior found in many studies of radio-detected AGNs. This suggests that the underlying physical mechanisms shaping GHz-detected SF galaxies properties are fundamentally similar. This raises intriguing questions about whether some compact SF galaxies represent a precursor phase of AGN evolution or a form of low-power AGN. The radio compact characteristic sizes of GHz-detected SF galaxies also suggest a connection between AGN and old starburst galaxies.

Figures (12)

• Figure 1: Comparison of the radio continuum luminosities between megahertz and gigahertz frequencies. Blue circles represent SF galaxies and red X-symbols represent AGN candidates chosen from multi-wavelength selection techniques (all the ones mentioned in Section \ref{['sec:All_AGN']}). The black solid line is a 1D-polynomial fitted to SF galaxies, and it is followed by a dashed-dotted line that extrapolates it to $L_{1.4GHz}>23$ [W Hz$^{-1}$]. The dotted line is the same fit for AGN candidates.
• Figure 2: Parameter space used to match pairs of SF galaxies with and without gigahertz detections. We compared the properties of each pair, with GHz-SF on the x-axis and nGHz-SF on the y-axis, where both axes represent the same parameter. The dashed line in each plot represents the 1-to-1 correspondence where the parameter pairs are equal.
• Figure 3: Uncontrolled global or integrated properties of SF galaxies with and without gigahertz detections.
• Figure 4: Comparison of modeled and empirically resolved properties between GHz-SFs and nGHz-SFs. The solid lines represent the median value at each annulus and the shaded region represents the 25th and 75th percentiles. The empirical properties are located in the top row, and the modeled, in the bottom row.
• Figure 5: Resolved emission line ratios binned by their $W_{80}$. The first column of plots shows the positions of the spaxels (from the DAP, see Section \ref{['sec:manga_DAP']}) binned by $W_{80}$ for GHz-SFs. The second column shows the same for nGHz-SFs. The distribution of the emission-line ratio spaxels of either GHz-SF or nGHz-SFs is shown in the opposite columns using black contours (see \ref{['sec:em_line_diagnostics']}). For example, in the left column, $W_{80}$ is binned for GHz-SF, but the contours show the spaxel distribution of nGHz-SF. In all plots, the Kewley_2001 and Kauffmann2003a lines for transition objects and extreme starbursts are shown in the red solid and black dashed lines, respectively. Bins where fewer than three scatter points were present have been excluded. On the bottom left of each plot, we display representative error bars (in black) for the individual scatter points that compose the shown binning.