Radio Variability in Recently-Quenched Galaxies: The Impact of TDE or AGN Driven Outflows

TL;DR

The paper investigates radio variability in four recently quenched, post-starburst galaxies to study newly launched nuclear outflows. Using quasi-simultaneous VLA observations across 1–18 GHz, the authors measure peaked, self-absorbed synchrotron spectra and model them with an equipartition SSA framework, deriving small source sizes and moderate magnetic fields. The targets exhibit significant brightening in the L band and variable behavior in the S band, with two showing infrared flares; optical data show no clear flares. The authors discuss whether the energy output is AGN- or TDE-driven, finding that either scenario could explain the observations, but the host stellar masses and radio properties are more consistent with TDE-related activity in several cases. Regardless of the driver, the inferred outflows possess enough energy to power molecular gas outflows and ISM turbulence, implying a meaningful impact on quenching processes in these galaxies.

Abstract

Outflows and jets launched from the nuclei of galaxies emit radio synchrotron emission that can be used to study the impact of accretion energy on the host galaxy. The decades-long baseline now enabled by large radio surveys allows us to identify cases where new outflows or jets have been launched. Here, we present the results of a targeted VLA program observing four post-starburst galaxies that have brightened significantly in radio emission over the past ~20 years. We obtain quasi-simultaneous observations in five bands (1-18 GHz) for each source. We find peaked spectral energy distributions, indicative of self-absorbed synchrotron emission. While all four sources have risen significantly over the past ~20 years in the 1-2 GHz band, two also show clear recent flares in the 2-4 GHz band. These sources are less luminous than typical peaked spectrum radio AGN. It remains unclear whether these sources are low luminosity analogs of the peaked radio AGN from accreted gas, or driven by tidal disruption events with missed optical flares. Regardless of the source of the accreted material, these newly-launched outflows contain sufficient energy to drive the molecular gas outflows observed in post-starburst galaxies and to drive turbulence suppressing star formation.

Figures (9)

• Figure 1: Cutout images (1$\times$1 arcmin) from FIRST (left column) and VLASS epoch 1 (middle column) of our four VLA targets, used in sample selection. The right column shows the 3$\sigma$ FIRST upper limits and VLASS epoch 1 detections. The blue shaded region shows the expected spectral energy distribution given the range of power law slopes observed for the sample of post-starburst galaxies with both VLASS and FIRST detections (487 galaxies). The four sources chosen were likely to have risen significantly since the FIRST epochs, despite uncertainties in the spectral slope.
• Figure 2: Observed SEDs from our targeted VLA observations with best-fit models as discussed in §\ref{['sec:sed']}.
• Figure 3: Radio and IR lightcurves. (Top) Lightcurves of 3 GHz flux from VLASS (squares) and our targeted VLA observations (circles). The top two sources show clear increases since the first epoch of VLASS, while the bottom two are more constant or variable. (Middle) 1.4 GHz lightcurves from the FIRST upper limits and targeted VLA observations. All four sources have brightened significantly in the L band since the non-detections in FIRST. The four sources have risen by at least $25, 6.6, 2.5, 2.5 \times$ (in order listed in each table, respectively), relative to the FIRST $3\sigma$ upper limits on non-detections. (Bottom) WISE W1 (3.4 $\mu$m) and W2 (4.6 $\mu$m) lightcurves for each source. No significant flares or trends are seen for two of our targets (PSB0800 and SPOG1033). A clear flare is seen for PCA1009, with a strong decline seen over several years. An IR increase is also seen for PCA1424, though in this case we see an initial increase and plateau, dissimilar to the light echoes seen after transient events.
• Figure 4: Left: Peak frequency vs. luminosity at peak frequency, Right: size vs. rest-frame peak frequency (legend shared). We compare the post-starburst targets to samples of AGN (copper colors; Patil2022Kunert2025 and compilations by An2012, Jeyakumar2016; peak--linear size relations from Odea1997 and Orienti2014) and TDEs (cool colors; relativistic jetted TDEs from Swift J1644 Eftekhari2018 and AT2022cmc Pasham2023, optically-selected events from Cendes2024Goodwin2022 (including AT2019azh), the delayed jetted event AT2018hyz Cendes2022, radio-identified TDEs from Somalwar2025b), and X-ray identified TDEs from Goodwin2025). For the post-starburst targets, we plot both the limits on the largest size from the VLA Ku-band beam sizes, as well as the inferred sizes from equipartition modeling. For comparison samples, sources with size corresponding to a direct measurement of the largest linear size are shown as $\times$ or $+$, while sources with equipartition-inferred sizes are shown as filled circles or lines. For inferred equipartition sizes for AT2018hyz and AT2019azh, models assuming a spherical outflow are shown as solid lines, while those assuming a collimated jet are shown as dotted lines. The post-starburst sources have lower luminosities and smaller sizes than most of the AGN samples, though the post-starburst source luminosities overlap with the transient sample from Kunert2025. The post-starburst sources have properties similar to the non-relativistic TDE samples, particularly those of Somalwar2025bGoodwin2025.
• Figure 5: Luminosity at peak frequency vs. redshift for the post-starburst sources and comparison samples, with plot points as in Figure \ref{['fig:size']}. While there is a clear trend with luminosity and redshift, the post-starburst sources (black squares) are among the lower luminosity sources, even at comparable redshifts.