Characterizing TESS-Identified Quadruple and Higher Order Eclipsing Binaries: I. Speckle Imaging with DSSI and HRCam

TL;DR

The study presents a diffraction-limited speckle-imaging follow-up of 57 TESS-identified quadruple and higher-order eclipsing binaries (Q$^+$EBs) using DSSI on the ARC 3.5 m and HRCam on SOAR, finding that about $60\%$ are resolved into two subsystems with separations down to $\geq 0.03''$. It provides precise astrometric and photometric measurements, assesses repeatability, and compares speckle results to Gaia DR3, highlighting Gaia’s limited resolving power for subarcsecond, high-contrast systems and the resulting unreliability of Gaia-based distances in many cases. The paper derives distance- and size-related constraints, identifies particularly compact systems, and explores special cases including a sextuple TIC 168789840 and ETV-bearing sources, while introducing Speckle Imaging During Eclipse (SIDE) to connect eclipses with resolved components. These results refine the architectural census of compact Q$^+$EBs, inform dynamical evolution studies, and establish SIDE as a valuable technique for future multiplicity investigations with large telescopes. Overall, the work emphasizes the necessity of high-resolution imaging to complement space-based astrometry in studying hierarchical stellar systems identified by TESS and sets the stage for broader 4–8 m and SIDE-enabled follow-up campaigns.

Abstract

NASA's TESS mission has unveiled a plethora of eclipsing binaries (EBs), among them hundreds of triples and higher order, hierarchical systems. These complex targets require follow-up observations to enable full characterization of system architectures and identify the most compact multiples expected to undergo the most dramatic dynamical evolution. We report first results from a long-term effort to perform such follow-up, focusing here on multi-band speckle imaging of a majority, 57, of the sample of 97 quadruple and higher order eclipsing binaries (Q+EBs) identified via TESS light curves by V. B. Kostov et al. (2022). Diffraction-limited imaging with the Differential Speckle Survey Instrument (DSSI) on the ARC 3.5-meter telescope and HRCam on the SOAR 4.1-m telescope reveals nearly 60% of the 57 to resolve into two sources separated by $\geq$ 0.03 arcseconds. For these partly resolved systems, we report derived characteristics (e.g., relative position angle, angular separation, and magnitude differences in multiple passbands) from the speckle imaging. We find those Q+EBs partly resolved with 4-m class telescopes to have significantly inflated Gaia parallax errors and large Gaia RUWE, particularly for systems with separations comparable to Gaia's resolution limit (~0.6 arcseconds). For unresolved systems we report upper limits on angular and linear projected separations. We find two partly resolved Q+EBs with wide linear separations having eclipse timing variations that are therefore candidates of higher than quadruple multiplicity. Finally, we demonstrate how speckle imaging of resolved Q+EBs during an eclipse can clarify which speckle-resolved Q+EB subsystem is associated with a particular set of TESS eclipses.

Figures (5)

• Figure 1: A representative sample of DSSI reconstructed images from the resolved systems in Table 1. Sources are organized left to right from brightest (TIC 344541836; $G=7.88$) to dimmest (TIC 375325607; $G=13.14$). The panels show the reconstructed image in the 692 nm channel ( top, blue-tinted panels) and in the 880 nm channel ( bottom, red-tinted panels). For each panel, North is up and East is to the left. The white arrow points to the real secondary source to distinguish it from the ghost image mirrored 180$^\circ$ in position angle. The flux scaling is logarithmic to highlight the detected sources. The $\Delta m$ contrasts between primary and secondary are given in Table \ref{['tab:speckle measurements']}.
• Figure 2: The average detection limits for the targets in our sample from DSSI in the 692 nm channel (top) and in the 880 nm channel (bottom). The gray lines show the detection limits from each of the targets with a high-quality non-detection, and the colored dashed line is the mean of all the detection curves. Targets with speckle-resolved companions are denoted with open circles (sometimes accompanied by right arrows, the latter to indicate those lying beyond the range of plotted separations).
• Figure 3: The Gaia DR3 RUWE as a function of the separation of Q$^+$EBs. The orange circles represent the eight Q$^+$EBs that are resolved both by speckle imaging and Gaia, while the blue circles represent those systems resolved only by speckle imaging. The stack of crosses shown at a separation of 0.05 arcsec represents the high quality non-detections in Table \ref{['tab:nondetections']}, assigned a separation corresponding to the approximate resolution limit for our speckle imaging. For the resolved systems, the size of the symbol represents the relative magnitude difference of the two components. The Gaia-resolved systems show a clear trend where those with smaller separations have smaller brightness contrasts, but larger RUWE. The shaded region to the left indicates the area below the approximate diffraction limit of the speckle imaging, which is 50 mas for both telescopes used, accounting for the wavelengths adopted for observations at each site.
• Figure 4: Relative parallax error, $\epsilon_{\pi}/{\pi}$, as a function of Q$^+$EB system angular separation. Similar to Figure \ref{['fig:Sep_RUWE']}, the vertical stack of crosses at separation of 0.05 arcsec represents the HQNDs, assigned a separation corresponding to the approximate resolution limit for our speckle imaging; the shaded region to the left indicates the area below this formal diffraction limit. Q$^+$EBs resolved by both our speckle imaging and Gaia are denoted with a black open circle (albeit sometimes shaded with interior coloring). As in Fig. \ref{['fig:Sep_RUWE']}, the size of the points corresponds to the relative magnitude difference of the two components as determined from our speckle imaging. Points with a purplish hue correspond to objects with Gaia RUWE $\lesssim$ 1.4, which is where Gaia is able to well-fit a single-star astrometric solution to describe its observations, and is often adopted as a delimiter above which sources are often suspected of multiplicity Fabricius2021. The dotted and dot-dashed lines at separations of 0.4 and 0.7 arcsec correspond to the approximate angular resolution achieved by Gaia DR2 and the visual binary completeness limit of Gaia DR3, respectively GaiaCollab2018Fabricius2021
• Figure 5: Light curves of TIC470710327 taken on the nights of 19 October (top) and 20 October (bottom) 2023 UT from the ARCSAT telescope at APO. Superposed on the lightcurves are the averaged $\Delta m$ values from DSSI speckle observations using the ARC 3.5-m telescope, with those made in the 692nm filter shown in blue and those made in the 880nm filter shown in red. The averaged $\Delta m$ values on the 19th (top panel) have seeing$\times$separation greater than 0.6, and are therefore considered upper limits. The error bars shown on all $\Delta m$ points are the standard deviations from Table \ref{['tab:speckle eclipse']}.