Resolving the Nuclear Environments of Tidal Disruption Event Host Galaxies within 45 pc

TL;DR

This study resolves the innermost environments of four nearby TDE host galaxies using high-resolution HST/STIS spectroscopy to connect small-scale stellar ages and circumnuclear gas to TDE rates. By fitting star formation histories with BAGPIPES and isolating disk contributions, it finds a younger $\sim$340 Myr nuclear starburst within $\sim$44 pc in ASASSN-14li, with central densities approaching nuclear star cluster regimes. The data reveal high-ionization coronal lines confined to the nuclear regions of ASASSN-14li and PTF09ge, alongside suppression of [O III] at small radii, suggesting echoes of past accretion events or compact, dense gas possibly shaped by a minor merger. Together, the results support a scenario where merger-driven gas inflow creates steep central density profiles and variable ionization, contributing to the elevated TDE rates observed in post-starburst hosts and highlighting the importance of multi-scale nuclear diagnostics.

Abstract

Using HST/STIS observations, we present the highest-spatial-resolution spectroscopic study to date of four tidal disruption event (TDE) host galaxies, with the best observed being the post-starburst (PSB) host of ASASSN-14li. The stellar population of ASASSN-14li's host, within 44 pc of the nucleus, reveals a younger recent starburst ($\sim$340 Myr) compared to the population at an offset radius of 88 pc that excludes the nucleus ($\sim$550 Myr), a radial age gradient suggesting gas inflows from a minor merger. We estimate a stellar density of $\sim5900 \pm 800 \, M_\odot / \mathrm{pc}^3$ within 30 pc of the nucleus of ASASSN-14li's host, exceeding densities expected for nuclear star clusters. High-ionization ``coronal" emission lines, [Fe VI] $λ5677$, [Fe VII] $λ6087$, and [Fe X] $λ6375$, are also detected within the nuclear spectra of the hosts of ASASSN-14li and PTF09ge, importantly alongside the non-detection of [O III] at the same scale. We similarly do not detect [O III] in the nuclear region of ASASSN-14ae's host despite its presence in the SDSS spectrum. The different ionization radiation levels detected at various radii from TDE host nuclei may indicate echoes of earlier accretion episodes, including, potentially, a prior TDE. We posit that a minor merger driving gas inflow to the nucleus could drive the enhanced TDE rates in post-starburst galaxies, inducing variation in nuclear gas properties and star formation history on $<$150 pc scales in TDE hosts.

Figures (7)

• Figure 1: The nuclear spectrum of each TDE host as taken by HST (black, with errors ranges shown in gray) compared to the SDSS spectrum of each host (red). Each panel is a different host: ASASSN-14li (first/top), ASASSN-14ae (second), PTF09ge (third), iPTF15af (fourth/bottom). For ASASSN-14li (top panel) we also show the averaged spectrum from the HST extractions at $\pm$88 pc (purple, with error regions shown in light purple). Lines either expected based on ground-based spectra, or newly identified from these data, are marked with vertical lines and labeled in the third panel. Specific lines identified only for one source, as discussed in § \ref{['sec:results']}, are labeled with vertical lines in the corresponding panel (forbidden lines in ASASSN-14li, top panel; iron lines in PTF09ge, third panel). There are several differences between the archival SDSS spectra and the nuclear spectrum of each host from HST, most notably the identification of forbidden lines in both ASASSN-14li and PTF09ge, as well as the lack of nuclear oxygen emission lines in ASASSN-14ae in contrast with its SDSS spectrum.
• Figure 2: Top: A close-up of the nuclear spectrum of ASASSN-14li (with an extraction radius within 44 pc of the host center, in black with error ranges in gray) and the averaged spectrum from the extractions at $\pm$88 pc (purple, with error ranges in light purple), both smoothed to visually emphasize the similarities in absorption features. We label the key lines used to analyze the stellar ages: Ca K $\lambda$3933 and neutral metals Mg I $\lambda\lambda$3832,3838, Fe I $\lambda$4046, and Ca I $\lambda$4226 (blue vertical lines); and the Balmer series. Bottom: The late-time plateau of ASASSN-14li likely arises from an accretion disk (that was either obscured or only recently formed) as modeled by GuoloMummery2025 and reproduced here in blue. The black line shows our original nuclear spectrum as observed, while the red line shows how the spectrum of the host nucleus changes without the disk contribution. This does not significantly change the overall flux redward of 4000Å which defines the strength of the absorption lines used to determine star formation histories with BAGPIPES, however, the strong disk contribution on the bluest end does give the nucleus a steeper upturn blueward of the 4000Å break, which can still impact stellar population age and mass estimates. We use the resulting subtracted spectrum in red for our BAGPIPES fits.
• Figure 3: Fitting ranges and best-fit values of the fit paramters from BAGPIPES of the innermost nuclear region of ASASSN-14li's host. The first (old) starburst of the host is modeled as a delayed exponential power law, while the second (younger) starburst of the host is modeled as an exponential power law. The order of the rows are as follows: Age of the first starburst, mass of the first starburst, age of the second starburst, mass of the second starburst, the values of $\tau$ in the exponential power law, metallicity from the second starburst, ionization parameter, redshift, velocity dispersion and dust extinction.
• Figure 4: The nuclear spectrum of ASASSN-14li (with an extraction radius within 44 pc of the host center, in black) and the averaged spectrum from the extractions centered at $\pm$88 pc (purple), both smoothed to show the emission lines of ionized gas either detected in galaxy-integrated ground-based spectra or detected in this nuclear spectrum. We also include the MUSE spectrum of the same host at the smallest resolution radius of MUSE (0.2 arcsec) corresponding to light within 88pc of the SMBH, in blue. We label the locations of emission features (some not identified in any spectrum) with vertical lines, with $[$O3$]$$\lambda$$\lambda$4959,5007 in green, $[$Fe7$]$$\lambda$5276, $[$Fe14$]$$\lambda$5303, and $[$Fe10$]$$\lambda$6375 in brown, $[$N2$]$$\lambda$5755 and $[$N2$]$$\lambda\lambda$6548,6583 in dark pink, and $[$S2$]$$\lambda\lambda$ 6716, 6730 in red. Finally, we show the best-fit Gaussian to the $[$N2$]$$\lambda$5755 and $[$Fe10$]$$\lambda$6375 features, identified in all three spectra, in red overlays. These observations localize the high-ionization coronal lines to an emitting region closest to the host nucleus.
• Figure 5: Corner plot showing 2D projection of the fitted parameter space for the nuclear spectrum (light within a 44pc radius from the nucleus) of ASASSN-14li's host galaxy. The calibration values (N0, N1, and N2) corresponding to the Chebyshev polynomial coefficients used to calibrate the flux have been omitted from the plot for visual clarity of the fits to physical parameters, but their best fits were centered at 1.06$\pm$0.15, 0.17$\pm$0.04, and 0.02$\pm$0.01, respectively.