Investigating photometric and spectroscopic variability in the multiply-imaged Little Red Dot A2744-QSO1

TL;DR

This study uses the multiply-imaged, $z=7.045$ LRD A2744-QSO1 to probe variability with JWST data. Spectroscopic analysis reveals significant Hα and Hβ EW variability, consistent with an active galactic nucleus, while photometric variability remains inconclusive due to systematic uncertainties and limited sampling. A DRW-based variability framework aligns with the observations, underscoring the AGN nature and the promise of future reverberation mapping using lensing time delays to measure black-hole mass at high redshift. The work emphasizes the need for coordinated, repeat monitoring to robustly map the light curve and constrain accretion physics in the early universe.

Abstract

JWST observations have uncovered a new population of red, compact objects at high redshifts dubbed `Little Red Dots' (LRDs), which typically show broad emission lines and are thought to be dusty Active Galactic Nuclei (AGN). Some of their other features, however, challenge the AGN explanation, such as prominent Balmer breaks and extremely faint or even missing metal high-ionization lines, X-ray, or radio emission, including in deep stacks. Time variability is another, robust, test of AGN activity. Here, we exploit the $z=7.045$ multiply-imaged LRD A2744-QSO1, which offers a particularly unique test of variability due to lensing-induced time delays between the three images spanning 22 yr (2.7 yr in the rest-frame), to investigate its photometric and spectroscopic variability. We find the equivalent widths (EWs) of the broad H$α$ and H$β$ lines, which are independent of magnification and other systematics, to exhibit significant variations, up to $18\pm3$ % for H$α$ and up to $22\pm8$ % in H$β$, on a timescale of 875 d (2.4 yr) in the rest-frame. This suggests that A2744-QSO1 is indeed an AGN. We find no significant photometric variability beyond the limiting systematic uncertainties, so it currently cannot be determined whether the EW variations are due to line-flux or continuum variability. These results are consistent with a typical damped random walk (DRW) variability model for an AGN like A2744-QSO1 ($M_{\mathrm{BH}}=4\times10^7 \mathrm{M}_{\odot}$) given the sparse sampling of the light-curve with the available data. Our results therefore support the AGN interpretation of this LRD, and highlight the need for further photometric and spectroscopic monitoring in order to build a detailed and reliable light-curve.

Figures (6)

• Figure 1: JWST/NIRSpec prism spectroscopy from UNCOVER of all three images of A2744-QSO1 in both spectroscopic epochs (see Tab. \ref{['tab:epochs']}). Image A is shown in blue (dark for epoch 7, light for epoch 8), image B in red and image C in green. Left: Full NIRSpec-prism spectra, normalized to the total luminosity of image A in epoch 7, obtained by integrating the spectra in wavelengths $1.0\,\mu\mathrm{m}<\lambda_{\mathrm{obs}}<5.4,\mu\mathrm{m}$. Thanks to the new reductions (see section \ref{['sec:data']}), the broad H$\alpha$ line no longer falls off the detector as in previous work, but is seen entirely here for the first time. Right: Zoom-ins on the H$\alpha$ (bottom) and H$\beta$ (top) lines, scaled by their continua listed in Tab. \ref{['tab:line-fluxes']} such that they become independent of magnification, similar to the EW.
• Figure 2: Spectroscopic variability of A2744-QSO1 as a function of time, with image A shown in blue, image B in red, and image C in green as in Fig. \ref{['fig:multi-image_variability']}. --Top: H$\beta$ rest-frame EW, presenting a drop of $22\pm8$ % between images C and A. --Bottom: H$\alpha$ rest-frame EW, presenting a drop of $18\pm3$ % in the time between images C and A. We note that the variability is consistent in both broad Balmer lines, as can be expected since they both originate from the broad-line region of A2744-QSO1. The EWs represent excellent indicators of variability since they are insensitive to most systematics such as lensing, slit-losses, calibration, etc.
• Figure 3: De-magnified light-curve of A2744-QSO1 in each photometric band with at least one epoch in both 2022 and 2023, as a function of time, incorporating the gravitational time delays $\Delta t_{\mathrm{grav}}$ given in Tab. \ref{['tab:images']}. Image A is shown in blue, Image B in red, and Image C in green. Solid circles represent NIRCam (NC) data-points and solid triangles represent the synthetic photometry obtained by integrating the NIRSpec (NS) spectra in the broad-band filter band-passes. The dashed horizontal line represents the mean $\bar{m}$ from which variations $|\Delta m|$ are calculated in each band. Note that the magnification uncertainties are not propagated into the photometry here since for each image, the magnification remains constant and is only relevant when comparing to the other images. Instead, the colored crosses indicate the level of SL (magnification and time-delay) systematics for each image. We also note that image B lies close to a cluster galaxy which might contaminate its photometry.
• Figure 4: Histogram of maximum variation $|\Delta m|$ from 100 simulated DRW light-curves of an $M_{\mathrm{BH}}=4\times10^7\,\mathrm{M}_{\odot}$ black hole, sampled at the rest-frame effective times of our combined observations of all three images of A2744-QSO1 as shown in Fig. \ref{['fig:multi-image_variability']}. Rest-frame UV (F115W) variations are shown in light blue and optical variations (F356W) are shown in dark blue. We choose to show the F356W-band here because it has the most epochs (see Tab. \ref{['tab:flux']}) and therefore has the highest probability of detecting a significant event. The gray-shaded area represents variations that are below the typical errors of 0.3 magnitudes (including systematics) in our photometry measurements in both bands. Even with the sampling of our full light-curve, the probability to detect a significant variation event ($|\Delta m|>0.5$ magnitudes) is relatively low.
• Figure 5: Histogram of maximum variation $|\Delta m|$ from our simulated light-curves (see Fig. \ref{['fig:mock-lightcurve_prediction']}) in F115W (light) and F356W (dark), now sampled at the rest-frame observation times of each image individually. As in Fig. \ref{['fig:mock-lightcurve_prediction']}, we show the F356W-band to represent the rest-frame optical because it has the most epochs. Image A is shown in blue, image B in red and image C in green. The probabilities of observing a significant event are much lower for the individual images than when combining the images to a full light-curve.