Lord of LRDs: Insights into a "Little Red Dot" with a low-ionization spectrum at z = 0.1

TL;DR

This study presents SDSS J102530.29+140207.3 at z=0.1006 as the brightest, lowest-redshift local LRD, integrating SDSS/Gemini/GTC optical spectroscopy with Chandra/NuSTAR X-ray data and broad UV–IR photometry. The object exhibits the hallmark V-shaped UV–optical SED, broad Balmer emission with a two-component profile, strong Balmer and NaD absorption, and a dense, low-ionization Fe II–dominated gas environment, alongside extreme X-ray weakness suggesting Compton-thick obscuration or a soft EUV SED. Emission-line diagnostics place the narrow-line region in a low-metallicity, star-formation–like regime, while Fe II density indicators imply a stratified, high-density zone near the BLR. The results challenge simple stellar-dominated continuum interpretations and highlight a complex, gas-rich AGN–like environment that mirrors high-z LRDs, offering a local laboratory to test LRD formation scenarios and guiding future high-energy and UV follow-up observations.

Abstract

Recent observations by the James Webb Space Telescope (JWST) have revealed a puzzling population of optically red and compact galaxies with peculiar "V"-shaped spectra at high redshift, known as "Little Red Dots" (LRDs). Until now, most spectroscopically confirmed LRDs are found at $z>4$ and it has been speculated that LRDs are tracing the early stages of black hole evolution. We report an independent rediscovery of a broad-line active galactic nucleus (AGN), SDSS J102530.29+140207.3, at $z=0.1$, which shows spectral features matching those of LRDs seen in the early Universe, including the V-shaped spectrum, broad Balmer lines (with widths of 1000-2000 km/s), and deep Balmer absorption. We present a new GTC observation of this LRD, which reveals an optical continuum similar to those of G-to-K giant stars including an unambiguous G-band absorption originating from the CH molecule. In addition, this local LRD shows a series of absorption lines potentially related to low-ionization ions or atoms but are deeper than what is observed in empirical stellar templates. We further identify a series of [FeII] emission lines indicative of low-ionization gas, which we find also present in a JWST-selected LRD at $z=2.26$. We find small but statistically significant variability in the H$α$ of SDSS J102530.29+140207.3 consistent with previous findings. Finally, we report new observations with NuSTAR. We confirm the extreme X-ray weakness of this LRD, which might imply Compton-thick gas obscuration with $N_{\rm H}>10^{24}~{\rm cm^{-2}}$. All evidence suggests SDSS J102530.29+140207.3 has a complex gaseous environment and the strong ionic, atomic, and molecular absorptions are hard to explain with typical stellar and AGN models.

Figures (21)

• Figure 1: Comparison between the spectrophotometric SEDs of J1025+1402, the Rosetta Stone (one of the lowest-$z$ LRDs discovered by JWST at $z=2.26$, juodzbalis_rosetta_2024), the Cliff (one of the LRDs showing the strongest Balmer break at $z=3.55$, degraaff_lrd_2025), and the median stack of the color-selected LRDs at $z\sim6$ from the COSMOS field Scoville_cosmos_2007 provided by akins_lrd_2024 (where non detections are plotted as $5\sigma$-upper limits following akins_lrd_2024). The GTC spectrum (solid black) of J1025+1402 is convolved to the resolution of the PRISM spectrum ($R\sim 100$) of the Rosetta Stone (solid pink) for illustrative purposes. There is a close resemblance among the SEDs of J1025+1402, the Rosetta Stone, and the stacked LRDs, once normalized to similar flux density levels, with a common V-shaped turnover in the NUV-optical regime, a blue UV slope, a red optical slope, a peak in the NIR, and weak MIR emission. While the Cliff exhibits a much stronger break near the Balmer limit compared to all the other LRDs, its NIR SED is similar to that of J1025+1402. On the top, we show the SDSS york2000$gri$-composite image and the Legacy Survey Dey_2019$grz$-composite image of J1025+1402, where J1025+1402 is unresolved.
• Figure 2: Best-fit spectral models for part of the observed spectra of J1025+1402. Panel (a) SDSS spectrum around H$\beta$ and [O iii], where H$\beta$ has a narrow component, a broad component, and a redshifted absorption (modelled with Equation \ref{['eq:abs']}) shown in the zoomed-in panel. Panels (b) and (c) Gemini/GMOS spectrum around He i, NaD, and [O i], where NaD is modelled as an absorber with kinematics independent of narrow emission lines. Panel (d) Gemini/GMOS spectrum around H$\alpha$, [N ii], and He i, where H$\alpha$ has a narrow component, a broad component (modelled as a double-Gaussian function), and a blueshifted absorption shown in the zoomed-in panel. Panels (e), (f), and (g) Gemini/GMOS spectrum around [S ii], He i, [Ar iii], [Fe ii], [Ca ii], and [O ii]. [Fe ii] and [Ca ii] show different line profiles compared to other narrow lines, suggesting that they come from a different region.
• Figure 3: Best-fit spectral model for the GTC/OSIRIS spectrum of J1025+1402 based on MILES empirical stellar templates miles and additive polynomials with pPXFcappellari2004cappellari2017. We plot both the best-fit emission line + continuum model (magenta) and the best-fit continuum model (red) and compare them with the GTC spectrum (black with $1\sigma$ uncertainty in shaded green). Regions overlapping with the vertical shaded bands are excluded during the fit. We also plot the residual of the fit divided by the uncertainty (i.e., $\chi$). The best-fit models primarily consist of cool G-to-K type supergiant stars with $T_{\rm eff}\sim 5000$ K. The bottom panels show zoom-in views of the spectrum, where we mark tentative identifications of absorption lines and Fe ii emission lines that are rarely seen in galaxies. Despite adding the flexible polynomials, the strengths of many absorption lines are underfitted.
• Figure 4: JWST/NIRCam color-color diagram used for selecting high-$z$ LRDs. The mock NIRCam colors for J1025+1402 is generated by redshifting its SED to $z=5$. The V-shape zone in the diagram defined in greene2024 is indicated by solid demarcation lines, which also set a zone for excluding brown dwarfs that contaminate the selection of the high-$z$ sample but not in our case. For comparison, we show the locations of normal high-$z$ galaxies from the UNCOVER survey uncoveruncoverdr_2024 as well as LRDs selected by labbe_2023. Some of labbe_2023's sources are outside the V-shaped zone due to a slightly different selection based on F150W-F200W. The location of the stacked LRD of akins_lrd_2024 are also shown.
• Figure 5: Comparison between the UV-optical spectra of J1025+1402 and the high-$z$ LRD, the Rosetta Stone, at $z=2.26$juodzbalis_rosetta_2024. Left: the median-resolution (G140M, $R\sim 1000$) spectrum of the Rosetta Stone (red) is compared with the GTC spectrum of J1025+1402. Fluxes of both spectra are rescaled for illustrative purposes. The shaded regions correspond to $1\sigma$ flux density uncertainties. The vertical grey band masks the telluric absorption in the GTC spectrum. The vertical dash-dotted purple line marks the Balmer limit. The vertical dashed black line marks the turnover point seen in the GTC spectrum. The vertical dotted grey lines mark several narrow lines corresponding to the forbidden transitions of Fe ii seen in both spectra. The vertical dotted cyan line marks a potential absorption feature in both spectra. The two sources have overall similar spectral shapes, although J1025+1402 appears redder in the optical. Right: the low-resolution (PRISM, $R\sim 100$) spectrum of the Rosetta Stone is compared with the GTC spectrum convolved to the same LSF. At low resolutions, it becomes difficult to determine whether the turnover point is at the Balmer limit or at redder wavelengths due to the blended lines, although the spectrum of the Rosetta Stone seems to favor a turnover at $\rm H_\infty$.